UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Subclinical menstrual disturbances : are they more common in premenopausal vegetarian women than nonvegetarian… Janelle, K. Christina 1993

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1994-0013.pdf [ 3.82MB ]
Metadata
JSON: 831-1.0087266.json
JSON-LD: 831-1.0087266-ld.json
RDF/XML (Pretty): 831-1.0087266-rdf.xml
RDF/JSON: 831-1.0087266-rdf.json
Turtle: 831-1.0087266-turtle.txt
N-Triples: 831-1.0087266-rdf-ntriples.txt
Original Record: 831-1.0087266-source.json
Full Text
831-1.0087266-fulltext.txt
Citation
831-1.0087266.ris

Full Text

SUBCLINICAL MENSTRUAL DISTURBANCES: ARE THEY MORE COMMON IN PREMENOPAUSAL VEGETARIAN WOMEN THAN NONVEGETARIAN WOMEN?  By K. CHRISTINA JANELLE BAA, Ryerson Polytechnical Institute, 1991  A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in HUMAN NUTRITION  We accept this thesis as conforming to the required standard  October, 1993 Kathryne Christina Janelle, 1993  ___________________________  In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission.  Department of  €3  tyTla-TIoL)  The University of British Columbia Vancouver, Canada  Date  DE.6 (2/88)  ‘t4-k  ABSTRA CT  ABSTRACT  Wfth the current trends toward adopting a healthier lifestyle, many people are increasing their carbohydrate intake and decreasing their fat intake in the hope of preventing certain chronic diseases. Furthermore, social demands placed on physical appearance have resulted in increased dietary restraint, particularly in women.  Data are beginning to  accumulate however, that suggest that these high fibre, low fat diets may contribute to menstrual disorders and bone loss. The purpose of this prospective observational study was to determine whether a relationship exists between vegetarianism, the menstrual cycle, dietary restraint and bone mineral density. Statistical comparisons showed thatthe nonvegetarian women had significantly greater body mass index values and percent body fat evaluations than their vegetarian counterparts, and that they displayed a significantly higher level of dietary restraint. The results from this research, over six months in 45 women, support the concept that dietary restraint is associated with an increased prevalence of subclinical menstrual disturbances. Furthermore, body mass index was shown to be positively correlated with dietary restraint. It is therefore feasible that dietary restraint, stress and anthropometric variables act synergistically in affecting menstrual function.  The exact relationship remains to be  determined. The vegetarian women consistently displayed lower spinal bone mineral density values than the nonvegetarian women, although the differences were not significant and were largely eliminated when body weight was entered as a covariate.  Measurement of bone density  ABSTRACT  change over a period of time would be ideal and better represent the role of the diet on bone health, as well as eliminate small individual differences. However, if vegetarian women were, in fact, at increased risk of bone loss and osteoporosis due to some component of their diet, it is imperative that intervention and education be arranged to inhibit any further bone loss. As dietary restraint is related to women’s perceptions of weight and physical appearance, as well as society’s demands on body image, broad public education appears to be the ideal intervention. The feasibility of this solution is questionable however, as social and cultural pressures are not easily altered. Overall, the present study’s results indicated that vegetarian women tend to be leaner and less restrained eating than their nonvegetarian peers. It also confirmed the previously documented association of dietary restraint and subclinical menstrual cycle disturbances. More research is required to elucidate the role of vegetarianism, or certain components of the vegetarian diet, on bone density.  III  TABLE OF CONTENTS  TABLE OF CONTENTS ABSTRACT LIST OF TABLES  II  .  •  •..  LIST OF FIGURES  ix xl,  ACKNOWLEDGEMENT  Xl”  I. INTRODUCTION 1.BACKGROUND 2. PURPOSE OF THE STUDY 3. HYPOTHESES II. REVIEW OF LITERATURE 1. THE MENSTRUAL CYCLE 2. DOCUMENTING THE MENSTRUAL CYCLE A) Self-reports of the Menstrual Cycle B) Hormonal Measurements C) Endometrial Biopsies D) Ultrasound E) Basal Body Temperature Measurements  •  .  ..  .  3. FACTORS ASSOCIATED WITH MENSTRUAL CYCLE DISTURBANCES A) Body Fat and Dieting B)Age C) Menstrual History D) Athletic Training E) Stress 4. VEGETARIANISM AND THE MENSTRUAL CYCLE A) Dietary differences Between Vegetarians and Nonvegetarians B) Cross-sectional Reports of Menstrual Irregularities in Vegetarians C) Effects of Low-calorie Vegetarian Diets on the Menstrual Cycle iv  10 11 11 13 13 14  •  .  .  •  .  .  •  •  14 15 15 16 17 17  .  •  .  .  •  .  .  18 18 20 22  TABLE OF CONTENTS  D) Effects of Dietary Fat on Sex Steroids and the Menstrual Cycle E) Effects of Dietary Fibre on Sex Steroids and the Menstrual Cycle  25 25  5. DIETARY RESTRAINT  29  6. MENSTRUAL DISTURBANCES AND BONE A) Subclinical Menstrual Disturbances and Bone B) Sex Steroids and Bone C) Weight and Bone D) Exercise and Bone  33 34 35 37 37  7. VEGETARIANISM AND BONE  41  8. SUMMARY  44  III. EXPERIMENTAL DESIGN AND METHODOLOGY  45  1. STUDY DESIGN  45  2. SUBJECTS  46  METHODS  48  1. ANTHROPOMETRICS A) Weight B) Height C) Body Mass Index D) Skinfold Measurements E) Mid Arm Fat Area F) Muscle Mass  48 48 49 49 50 53 53  2. BASAL BODYTEMPERATURE  55  3. DIETARY INTAKE  58  4. EATING RESTRAINT ASSESSMENT  60  5. DUAL ENERGY X-RAY ABSORPTIOMETRY  61  6. URINE ANALYSIS  62  7. DEMOGRAPHICS  63 V  TABLE OF CONTENTS  STATISTICAL ANALYSIS OF THE DATA  64  THE ROLE OF THE CANDIDATE  66  SIGNIFICANCE OF THE STUDY  67  HUMAN STUDIES  68  IV. RESULTS  69  1. SUBJECTS  69  2. MENSTRUAL CYCLE DISTURBANCES  78  3. DIETARY INTAKE  82  4. DIETARY RESTRAINT  91  5. SPINAL BONE MINERAL DENSITY  94  6. URINE ANALYSIS  98  7. FACTORS RELATING CHANGES IN THE MENSTRUAL CYCLE A) Vegetarian/Nonvegetarian Population B) Normal/Abnormal Menstrual Function Population  100 100 106  8. FACTORS RELATING TO BONE  110  9. FACTORS RELATING TO DIETARY RESTRAINT A) Vegetarian/Nonvegetarian Population B) High/Low Restraint Population by Percentiles C) High/Low Restraint Population by Tertiles  115 11 5 11 7 1 21  10. SUMMARY OF RESULTS WITH REFERENCE TO HYPOTHESES  124 127  V. DISCUSSION 1. INTRODUCTION  127  2. SUBJECTS A) Demographic and Anthropometric Data B) Dietary Intakes (i) Phase of the Menstrual Cycle (ii) Vegetarian and Nonvegetarian Intakes  127 1 27 129 1 29 1 30  vi  TABLE OF CONTENTS  3. SUBCLINICAL MENSTRUAL DISTURBANCES  132  4. DIETARY RESTRAINT A) Between Vegetarians and Nonvegetarians B) Between High and Low Restraint Groups  140 140 142  5. BONE DENSITY AND THE MENSTRUAL CYCLE  146  6. BONE DENSITIES OF VEGETARIAN AND NONVEGETARIAN WOMEN  147  7. NUTRIENT INTAKES AND BONE DENSITY  148  8. POSSIBLE RELATIONSHIPS ELUCIDATED FROM THIS RESEARCH A) Dietary Restraint and the Menstrual Cycle B) Vegetarianism and Bone  150 150 152  9. IMPLICATIONS FOR FUTURE RESEARCH  154  10. STUDY DRAWBACKS  155  CONCLUSION  156  REFERENCES  158  APPENDIX A. Recruitment Advertisement  167  APPENDIX B. Subject Consent Form  168  APPENDIX C. Skinfold Techniques Used  169  APPENDIX D. Basal Body Temperature Record Form  177  APPENDIX E. Basal Body Temperature Measurements  178  APPENDIX F. Dietary Intake Record Form  179  APPENDIX G. Dietary Intake Records  180  APPENDIX H. Three-Factor Eating Questionnaire  .  182  APPENDIX I. Two-hour Urine Collection  187  APPENDIX J. Interview Form  188  VII  TABLE OF CONTENTS  APPENDIX K. Vegetarian Additional Interview Form  189  APPENDIX L. Basis for Methodologies Chosen 1. Anthropometrics A) Weight B) Height C) Body Mass Index D) Skinfold Measurements E) Mid Arm Fat Area F) Muscle Mass 2. Basal Body Temperature 3. Dietary Intake 4. Eating Restraint Assessment 5. Dual Energy X-ray Absorptiometry 6. Urine Analysis  190 190 190 190 191 191 193 194 194 196 198 199 200  .  .  .  .  .  .  .  .  APPENDIX M. Ethics Approval Form  202  APPENDIX N. Table: Additional Information Regarding Subjects  203  APPENDIX 0. Table: Mean (±SD) Characteristics of 62 Premenopausal Women, Categorized by Remaining Volunteers and “Drop-out” subjects  205  APPENDIX P. Table: Reasons for Becoming a Vegetarian  206  APPENDIX Q. Table: Mean (±SD) Nutrient Intakes for the Three Stages of the Menstrual Cycle  207  APPENDIX R. Table: Mean (±SD) nutrient Intakes for Weekday vs. Weekend Days for 45 Premenopausal Women  208  VIII  LIST OF TABLES  LIST OF TABLES Table 1. Mean (± SD) characteristics of 45 premenopausal women, according to dietary group  72  Table 2. Mean (±SD) characteristics of 23 premenopausal vegetarian women, according to dietary subgroup  76  Table 3. Mean (±SD) characteristics of 23 premenopausal vegetarian women, according to dietary subgroup  77  Table 4. Mean (±SD) menstrual cycle characteristics over the six months of study, according to dietary group  79  Table 5. Menstrual cycle characteristics over six months of study, according to dietary group  80  Table 6. Mean (±SD) menstrual cycle characteristics over the six months of study, according to vegetarian subgroup  81  Table 7. Mean (±SD) menstrual cycle characteristics over the six months of study, according to vegetarian subgroup  82  Table 8. Mean (±SD) total nutrient intakes for nine days, according to dietary group  85  Table 9. Mean (±SD) total nutrient intakes for nine days, according to vegetarian subgroup  89  Table 10. Mean (±SD) nutrient intakes, according to vegetarian subgroup  90  Table 11. Mean (± SD) total eating inventory scores and scores on subscales for the Three-Factor Eating Questionnaire, according to dietary group  91  Table 1 2. Mean (± SD) total eating inventory scores and scores on subscales for the Three-Factor Eating Questionnaire, according to vegetarians subgroups .  .  Table 13. Mean (±SD) total eating inventory scores and scores on subscales for the Three-Factor Eating Questionnaire, according to vegetarian subgroups .  ix  .  .  .  .  .  93  93  LIST OF TABLES  Table 14. Mean (±SD) spinal bone mineral density characteristics as measured by dual energy x-ray absorptiometry in 45 premenopausal women, according to dietary group  95  Table 1 5. Mean (± SD) spinal bone mineral density characteristics as measured by dual energy x-ray absorptiometry in 23 premenopausal vegetarian women, according to dietary group  97  Table 1 6. Mean (± SD) spinal bone mineral density characteristics as measured by dual energy x-ray absorptiometry in 23 premenopausal vegetarian women, according to dietary subgroup  98  Table 17. Mean (±SD) urinary excretion of calcium, cortisol and creatinine levels from a two-hour fasting urine sample, according to dietary group  99  Table 18. Mean (±SD) urinary excretion of calcium, cortisol and creatinine levels from a two-hour fasting urine sample, according to vegetarian dietary subgroup  100  Table 1 9. Correlation coefficients for demographic, anthropometric, bone and restraint data, as related to menstrual function characteristics  102  Table 20. Correlation coefficients for urine values as related to menstrual function data, according to dietary group  103  Table 21. Correlation coefficients of total nutrient intakes to menstrual cycle data, in 45 premenopausal women  104  Table 22. Variables correlated to the number of ovulatory cycles in 45 premenopausal women  105  Table 23. Variables correlated to the average luteal phase length in 45 premenopausal women  106  Table 24. Variables correlated to the average luteal phase index in 45 premenopausal women  106  Table 25. Mean (±SD) characteristics of 45 premenopausal women, according to menstrual cycle group  107  Table 26. Mean (±SD) total eating inventory scores and scores on subscales for the Three-factor Eating Questionnaire, according to menstrual cycle group .  x  .  .  108  LIST OF TABLES  Table 27. Mean (±SD) spinal bone mineral density characteristics as measured by dual energy x-ray absorptiometry in 45 premenopausal women, according to menstrual cycle group  108  Table 28. Mean (±SD) total nutrient intakes for nine days, according to menstrual group  109  Table 29. Correlation coefficients for bone values to demographic, anthropometric, dietary restraint and menstrual cycle data, for the entire study population  .  11 2  Table 30. Correlation coefficients for bone values to nutrients, for 45 premenopausal women  11 3  Table 31. Correlation coefficients for bone values to urine data for the two subgroups of women  114  Table 32. Characteristics correlated to the bone mineral content, in 45 premenopausal women  11 5  L in 4 1 L Table 33. Characteristics correlated to the bone density evaluation , 45 premenopausal women  11 5  L in 4 2 L Table 34. Characteristics correlated to the bone density evaluation , 45 premenopausal women  11 5  Table 35. Characteristics correlated to dietary restraint, in 45 premenopausal women  117  Table 36. Mean (±SD) characteristics of 45 premenopausal women, according to dietary restraint group (divided by percentiles)  11 8  Table 37. Mean (±SD) total nutrient intakes for nine days, according to dietary restraint group (divided by percentiles)  1 20  Table 38. Mean (±SD) characteristics of 45 premenopausal women, according to dietary restraint group (divided by tertiles)  1 22  Table 39. Mean (±SD) total nutrient intakes for nine days, according to dietary restraint group (divided by tertiles)  1 23  xi  LIST OF FIGURES  LIST OF FIGURES Figure 1. The anthropometric variables that significantly differed between the nonvegetarian and vegetarian groups  73  Figure 2. Significantly different mean macronutrient intakes for nine days in vegetarian and nonvegetarian women  86  Figure 3. Scores on the Three-Factor Eating Questionnaire, and three subscales (dietary restraint, disinhibition and hunger) in vegetarian and nonvegetarian women  92  Figure 4. Mean (±SD) bone mineral contents of the vegetarian and nonvegetarian , and 2 women measured using dual energy x-ray absorptiometry in g/cm controlling for weight  96  Figure 5. Differences in BMI and cycle information between high and low dietary restraint groups, divided by percentiles  xl’  11 9  A CKNO WLEDGEMENT  ACKNOWLEDGEMENT I appreciate and thank my supervisor Dr. Susan Barr for her constructive advice and excellent guidance throughout the completion of this thesis. I would also like to express my most grateful appreciation to Dr. Jerilynn Prior and Dr. Linda McCargar for the invaluable input and continual interest they offered as research committee members. Special thanks to Yvette Vigna, for her helpful instructions on temperature analyses, to Cathy Langdon and Sonia Kwan for their continual hard work in translating and correcting dietary intake data, and to Con Rexworthy for her unfailing patience in measuring subjects’ bone. I gratefully acknowledge the British Columbia Medical Services Foundation for funding this project. Many thanks are extended to the vegetarian and nonvegetarian women whose cooperative participation made this research project possible. Finally, extra special thanks goes to my mother and father, and my brothers and sisters for their encouragement, guidance and loving support in so many ways throughout this challenging experience. Ett mycket speciellt tack till Mans, for att han försett mig med sá mycket kärlek och stöd, for att han státt vid mm sida under denna provande tid, och for att han gett mig anledning att ta sâ manga lângvariga semestrar.  xiii  INTRODUCTION  Chapter I  INTRODUCTION  1. BACKGROUND  Vegetarian diets, especially those higher in dietary fibre and lower in fat than conventional diets, have been associated with the prevention of chronic diseases such as 1 With these heart disease and certain common cancers in women, such as breast cancer. current trends toward adopting a healthier lifestyle and with social demands placed on physical appearance, an increasing number of young women are choosing to follow vegetarian 2 There has also been an increase in the number of women displaying diets in recent years. signs of restrained eating, which is a conscious control of dietary intake as a means of 34 Unfortunately there has not been a similar increase in the maintaining a slim body shape. amount of research focusing on the effects of vegetarianism or dietary restraint on women. Data are beginning to accumulate, however, to suggest that high fibre, low fat intakes 56 This change in sex may be associated with lower circulating levels of some sex hormones. hormone levels may contribute to menstrual disorders, such as amenorrhea, oligomenorrhea, anovulatory cycles and short luteal phase cycles. Some studies have suggested that there is an increased incidence of menstrual irregularities among vegetarian women, although these 78 cross-sectional findings are inconclusive.  Data also suggest that restrained eating may  possibly be associated with menstrual cycle disturbances due to the stress experienced in  1  INTRODUCTION  9 Most importantly, association with food intake and the latter’s effect on hormonal status. however, these inferences have potential significance for young women, as the integrity of 1011 the menstrual cycle prior to menopause is an important determinant of bone density. ° showed that less profound menstrual disturbances (anovulatory and 1 Furthermore, Prior et al. short luteal phase cycles) were strongly associated with loss of trabecular bone over a one year period. If this same rate of loss continued over a period of years, risk of osteoporotic fractures would be significantly increased. Confirmation of this observation is necessary.  2. PURPOSE OF THE STUDY  The rationale for this study stemmed mainly from discrepancies seen in the literature with regard to vegetarian dietary practices and their effects on menstrual regularity. Most of the literature focused on menstrual histories reported by the subjects from memory, which did not permit the detection of anovulatory and/or short luteal phase cycles. Also, on average, studies performed only short dietary assessments, most frequently three-day dietary records which did not provide sufficient information on the subjects’ actual intakes to make any plausible conclusions regarding the association of dietary factors and menstrual function. And finally, some studies may have been influenced by recruitment bias, leading to an increased likelihood of recruiting vegetarians with menstrual disturbances. Furthermore, very little literature examined restrained eaters and the effect this may have had on menstrual regularity.  This study was therefore able to provide additional  beneficial information pertaining to this growing area of interest. In using an appropriate sample size and in exploring the subjects’ diets in greater detail  2  INTRODUCTION  than had been done in previous investigations, this study provided a more thorough assessment of the possible associations between dietary factors (e.g., vegetarianism, components of the diet, dietary restraint) and subclinical menstrual disturbances.  By  restricting the study to subclinical menstrual disturbances, recruitment bias was minimized. The main objectives of this study were: 1. To determine if subclinical menstrual disturbances (anovulatory and short luteal phase cycles) occurred more frequently in premenopausal vegetarian women than in premenopausal nonvegetarian women. 2. To assess whether restrained eating was associated with subclinical menstrual disturbances. 3. To compare the bone mineral densities of women with and without subclinical menstrual disturbances. 4. To compare the bone mineral densities of vegetarian and nonvegetarian women. 5. To examine whether intakes of any specific nutrients were associated with bone mineral density results.  3. HYPOTHESES The null hypotheses for this study were: Hypothesis 1: There would be no difference in the prevalence of subclinical menstrual dysfunction between vegetarian and nonvegetarian women.  3  INTRODUCTION  Hypothesis 2: There would be no difference in the prevalence of subclinical menstrual disorders between restrained and unrestrained eaters.  Hypothesis 3: There would be no difference in spinal bone mineral densities between women with normal ovulatory and normal luteal phase length menstrual cycles and those with irregular cycles.  Hypothesis 4: There would be no difference in spinal bone mineral densities between vegetarian and nonvegetarian premenopausal women.  Hypothesis 5: There would be no association between nutrient intakes and spinal bone mineral density results.  4  REWEW OF LITERA TURE  Chapter II  REVIEW OF LITERATURE  Recent  evidence  suggests  that  menstrual  disorders,  such  as  amenorrhea,  oligomenorrhea, anovulatory cycles and shortened luteal phases in premenopausal women, 1011 Many research can lead to increased yearly bone loss, and eventually to osteoporosis. studies incorporating a treatment program also have found that the bone density of these women with menstrual disturbances generally remains well below the average for their age 10 recovery. 1 1 2 group, although sometimes returns to normal with complete ’  It appears,  however, that it may not be possible to completely replace vertebral bone lost as a result of 11 extended interruptions in the normal menstrual cycle.  Furthermore, some findings have  implied that vegetarians are more susceptible to these menstrual disturbances due to their altered diet. 78 The role of prevention consequently appears critical. Given the increasing number of women choosing to follow vegetarian diets as well as those displaying dietary restraint, it appears important to establish whether or not this dietary pattern is associated with a higher prevalence of menstrual cycle irregularities. 4 This review ’ 2 of literature will describe the background for such a study, by first describing the normal menstrual cycle and the different types of menstrual cycle disturbances.  Next, different  methods whereby the characteristics of the cycle can be documented will be described, followed by a brief discussion of previous research on factors that appear to be associated  5  REWEW OF LITERA TURE  with disturbances of the cycle. Then a review of the available literature on vegetarianism’s role in menstrual cycle disturbances will be covered, followed by the introduction of the possible role of dietary restraint in the latter condition. Subsequently, the proposed effects of menstrual disturbances on bone, and of vegetarianism on bone will be reviewed. Finally, a summary of the literature review’s highlights will be provided, focusing on the gaps in the current knowledge.  1. THE MENSTRUAL CYCLE  The control of the menstrual cycle is, without any doubt, one of the most complex 13 Its regularity is managed by neurotransmitters and levels physiological processes in humans. 14 Scientifically, menstruation is defined as a of biogenic amines in premenopausal women. periodic flow from the uterus, accompanied by desquamation of the endornetrium. 15 In a study, however, participants will define menstruation as a periodic vaginal bleeding which the women herself must diagnose as “menstruation’ based on her previous experience and on accompanying signs and symptoms. 15 The length of the menstrual cycle is the difference between two dates: the onset of the previous and the day before the onset of the recent menstrual flow. Thus, the day of the 15 This is divided commencement of flow is conventionally called the first day of the cycle. 1 into four phases: menses, follicular, ovulatory, and luteaL’ The relationship of various hormones to the occurrence of menarche and the 16 There are changes in hypothalamic maintenance of an ovulatory menstrual cycle is complex. hormone secretions which occur with puberty.  6  As puberty begins, bursts of luteinizing  REVIEW OF LITERA TURE  hormone occur at night and, finally rhythmic activity at about one- to two-hour intervals is present throughout the day, which is an adult pattern and necessary for ovulation to occur. Estrogen levels also affect the pituitary response to gonadotropin-releasing hormone. Estrogen is synthesized in the ovary, adrenal, and by peripheral conversion  -  primarily in fat but also in  muscle tissue. 16 : in the early phase of follicular 3 The menstrual system operates in a feedback loop development in the ovary, levels of estradiol are low. There is a general increase until the 17 Following this mid-cycle when a peak is reached, which is necessary for ovulation to occur. peak, there is a substantial, although temporary, drop in estradiol as the peak in gonadotropins 16 occurs.  The hypothalamus secretes gonadotropin-releasing hormone into the anterior  pituitary, which results in the production of follicle-stimulating hormone and luteinizing hormone.  Rupture of the ovarian follicle occurs approximately 24-36 hours after the  , causing it to be converted into a corpus luteum which secretes 16 luteinizing hormone peak estradiol and increases the blood level of this hormone once again.  At this point large  17 The latter two ovarian hormones regulate the amounts of progesterone are also released. endometrium as well as provide feedback control of the anterior pituitary and hypothalamus. There are small amounts of progesterone which are produced by the ovary prior to ovulation, however, following ovulation, progesterone secretion increases during the life span of the corpus luteum. 16 The endometrium responds to the progesterone by becoming thick and spongy, marking the beginning of the luteal phase. Toward the middle or end of this phase levels of luteinizing hormone begin to fail, the corpus luteum degenerates and its production of estradiol and progesterone decrease rapidly.  7  Finally, the endometrium deteriorates and  REVIEW OF LITERA TURE  17 menstruation begins. To further emphasize the menstrual system’s feedback design, it has been shown that each individual part can only perform its normal function when it is able to interact with other parts of the system. For example, the hypothalamus is the master controller for the menstrual system; it receives signals from other parts of the brain, and is influenced by estradiol and progesterone secretions from the ovaries. However, the hypothalamus appears to be able to operate without such input and thus is not dependent, in the way that the ovaries are dependent on the anterior pituitary and the anterior pituitary is dependent on the 17 This demonstrates the ease with which the menstrual cycle can breakdown, hypothalamus. as well as the difficulties involved in pinpointing the cause of a disruption. A positive feedback at the pituitary occurs due to ovarian hormone production. The increase in estradiol immediately prior to the luteinizing hormone peak is necessary for this spike to occur. The site at which this modulation occurs may be both at the hypothalamus as well as the pituitary level.  Progesterone may also potentiate the positive feedback of  16 estrogens. Estradiol plays several roles in the menstrual cycle: it causes rapid growth of the endometrium; it increases the sensitivity of the anterior pituitary to gonadotropin releasing hormone; it causes the anterior pituitary to emphasize luteinizing hormone production over follicle-stimulating hormone production; and, it inhibits the release of luteinizing hormone by 17 the anterior pituitary. The gonadotropins are necessary for ovarian hormone synthesis as well as ovulation. Follicle stimulating hormone stimulates aromatase activity and causes follicular development,  8  REWEW OF LITERA TURE  and luteinizing hormone stimulates the synthesis of androgen precursors which are used for 1617 Furthermore, estrogen synthesis, and also induces ovulation and corpus luteum formation. if the follicle does not mature in a normal way during the preovulatory period, the corpus luteum may be inadequate, which results in a shortened luteal phase (normally 14 days) and relative infertility.  This may occur when gonadotropin levels are low in the preovulatory  period. The hormones which are secreted by the corpus luteum also have an effect on the 16 next cycle, thus a short luteal phase advances ovulation in the next cycle. Menstrual status is defined in the following ways: (i) normal ovulatory, with a cycle length of 21 to 35 days and a luteal phase 10 days; (ii) primary amenorrhea, when an individual has never menstruated; (iii) secondary amenorrhea, which is the absence of a cycle for  1 80 days, but having had at least one episode of menstrual flow previously; (iv) 36 and <180 days; (v) anovulatory, where cycle length  oligomenorrhea, with a cycle length  may be normal, long or short, and there is a failure to ovulate; and, (vi) shortened luteal phase, when the cycle length may be normal, long or short but the luteal phase is <10 ° Primary and secondary amenorrhea and oligomenorrhea are clinically detectable, but 1 days. anovulatory cycles and shortened luteal phases can go unnoticed by the individual as they can 1118 occur in cycles of normal length. Superficially, an anovulatory cycle cannot be distinguished from an ovulatory cycle. It may occur because of an inadequate luteinizing hormone surge or because ovaries do not 17 respond properly to its release.  Women with chronic anovulation often have chronic,  unopposed estrogen, which also leads to an increased risk of endometrial hyperplasia and 19 adenocarcinoma and this may induce profuse bleeding at unexpected times.  9  However,  REVIEW OF LITERA TURE  women with euestrogenic anovulation, i.e., estrogen levels >40 pg/mL, are also considered to be at risk of endometrial hyperplasia and adenocarcinoma, although this association has not 1920 been proved. A shortened luteal phase can occur from insufficient progesterone effect or hyperprolactinemia. Theoretically, the inadequate amounts of progesterone found in luteal phase insufficiency may lead to inadequate endometrial protection and also increase the risk 19 Also, luteal phase defects are important of endometrial hyperplasia and adenocarcinoma. 21 causes of fertility disturbances.  2. DOCUMENTING THE MENSTRUAL CYCLE The exact definitions of the phases of the menstrual cycle reflect the historic development of the different tests of ovulation. The foHicular and luteal phases are based on the assessment of the maturing follicle and the formation of the corpus luteum by cytologic criteria. The proliferative and secretory phases of the endometrium are estimated by the histologic changes in the glandular, interstitial, and vascular structures of the endometrium. The degree of acidophilia, karyopyknosis, and of desquamation of the vagiral epithelium are used to define the preovulatory and postovulatory phases. And, with the measurement of the plasma or urine levels of pituitary gonadotropin and ovarian steroid hormones, the 15 periovulatory phase can be clarified. When attempting to relate various aspects or characteristics of the menstrual cycle to potential health concerns, such as bone density, it is important to characterize the cycle completely and accurately. A number of different techniques are available to do this, including  10  -  REVIEW OF LITERA TURE  self-reports, hormonal measurements, endometrial biopsies, ultrasound, and basal body temperature measurements. However, as will be discussed below, these methods vary in terms of their accuracy, sensitivity, expense and invasiveness, and thus in terms of their suitability for longitudinal studies. A) SELF-REPORTS OF THE MENSTRUAL CYCLE  Traditionally, many investigators have relied on women’s self-reports of cycle interval. These have usually resulted in the classification of menstrual cycles as “normal”, oligomenorrheic or amenorrheic. This method, however, does not account for shortened luteal phases and anovulatory cycles which are important in determining the health concerns related to menstrual cycle disturbances. B) HORMONAL MEASUREMENTS  Hormonal measurements may be used both to characterize the cycle and to investigate mechanisms which may cause disturbances in the cycle. Women who experience endometrial shedding more frequently than every 21 days or at intervals of 35 days to 4 months probably have chronic anovulation and produce estrogen but not progesterone. Anovulation, however, also occurs with cycles of normal length. ° To determine if an abnormality exists, a women 1 can have her blood progesterone concentration determined several days before the expected onset of menstruation. Blood progesterone concentrations less than 1 6 nmol/L (5 ng/mL), obtained two to five days prior to the onset of menses, indicate ovulation did not occur. 17 This, of course, presents the problem of predicting the onset of menses, which is often difficult in women with abnormal cycles. On the other hand, women who engage in endurance training and who experience  11  REVIEW OF LITERA TURE  uterine bleeding less often than every 4 months, as well as those who are totally amenorrheic, are probably experiencing reduced levels of both sex steroids  -  estrogen and progesterone. 19  Estradiol tends to suppress the release of both luteinizing hormone-releasing factor from the hypothalamus and follicle-stimulating hormone from the anterior pituitary.  In some cases  estradiol is absent, such as with menopause, resulting in increased production of 17 gonadotropins and abnormal cycles.  However, the most frequently experienced and  documented menstrual cycle disturbances, such as those often seen in athletes and women with anorexia nervosa, are usually associated with decreased levels of luteinizing hormone and 13 follicle-stimulating hormone. De Crée 22 indicated in his review of endogenous opioid peptides in the control of the normal menstrual cycle, that endogenous opiates appear to be involved in the reproductive physiology of women, and more specifically in the control of the menstrual cycle and its pathology or disturbances. He explains that prolactin plays an essential role in the menstrual cycle, similar to its key function in pregnancy. Hyperprolactinemia is associated with amenorrhea, however, most studies have failed to show substantial evidence of an important role for prolactin in the control of the normal menstrual cycle. Overall, he concluded that the most frequently-observed finding was that women athletes had lower serum concentrations of progesterone, as well as lower estradiol levels, during the luteal phase than in sedentary 22 This is an important inference since most of the studies in the area of menstrual subjects. dysfunction in female athletes have concentrated on self-reported menstrual histories rather than actually following the women’s menstrual cycle in detail. Consequently, much of the necessary data for conclusive evidence of the role of menstrual dysfunction, and more  12  REVIEW OF LITERA TURE  specifically sex steroids, in the development of premature osteoporosis or above normal bone loss were not examined. Overall, however, there are many problems in using hormone levels as a valid predictor of bone density loss. Hormonal measurements are invasive, and must be measured frequently to obtain a composite picture. Usually due to the expense of these types of tests, they can only be measured once which really only provides a kind of “snapshot” approach and would not furnish enough values on which reliable conclusions could be based. Also, Baker and 23 found that low serum estradiol concentrations in their oligomenorrheic athletes were Demers compensated for through significant increases at later times.  They concluded that even  though the differences in hormone levels were significant, they were small, and can be interpreted as reassuring that future fertility may not be impaired in those women, and that osteoporosis is not merely a result of hypoestrogenism but rather from a synergistic effect of 23 a variety of factors. C) ENDOMETRIAL BIOPSIES  The morphology of the endometrium is characterized by various secretory changes in the luteal phase. Endometrial biopsies are performed less often now to determine when the menstrual cycle has entered this later phase as they are also invasive and painful. Basal body temperature measurements and progesterone assays now are more commonly used. The diagnosis of ovulation or anovulation based on endometrial histology is, however, highly 24 correlated with that based on plasma progesterone assay in the mid-luteal phase. D) ULTRASOUND  Vaginal ultrasonographic measurement of endometrial thickness has shown that the  13  REVIEW OF LITERA TURE  endometrium gradually increases throughout the menstrual cycle, and its thickness is an indication of its stage. 24 Serial ultrasound can also be used to trace follicular growth and to determine the exact time the follicle ruptures. Although these method may be less expensive and non-invasive when compared to the fore-mentioned methods for documenting the menstrual cycle, it is more time consuming and no less expensive or less invasive than basal body temperature measurements. El BASAL BODY TEMPERA TURE MEA SUREMENTS  To determine if an abnormality exists, particularly within a cycle of normal length, a woman can record her basal body temperature daily and graph it to determine the day of ovulation as well as the luteal phase length. Analysis of daily temperature with a computer program can also be done, avoiding potential errors associated with graphing. 25 Vollman 15 stated that the probability of over-re porting and under-reporting a menstrual date is less than one in 2,000 reported cycles when using basal body temperature as a test measure.  A  biphasic basal body temperature pattern, similar to a surge in the luteinizing hormone level, suggests that ovulation has occurred. If the woman’s next menses begins less than 10 days after the increase in basal body temperature, luteal phase inadequacy is probable.  A  monophasic basal body temperature pattern indicates that ovulation has not occurred.  3. FACTORS ASSOCIA TED WITH MENSTRUAL CYCLE DISTURBANCES  A number of hypotheses have been put forward regarding the underlying mechanisms involved in menstrual irregularities although interpretation of causal factors is still subject to 1326 The parameters that have been suggested as possible mechanisms in the investigation.  14  REVIEW OF LITERA TURE  development of menstrual cycle disturbances that are pertinent to this study include: (i) low percentage body fat, (ii) age, (iii) history of previous menstrual irregularities, (iv) athletic 2627 ’ 20 training, and (v) stress. A) BODY FATAND DIETING  It has frequently been suggested that a critical percentage of body fat is required for both the onset of menarche (17% body fat) and its maintenance (22% body ’ 18 14 fat), 2 6 0 27 Noakes hence the high frequency of amenorrhea present in women with anorexia nervosa. and Van Gend’s ° study showed that runners were approximately six kilograms below average 2 body weight for height and their higher incidence of menstrual dysfunction correlated with their lower percentage of body fat. It is however, argued that the rapidity of weight loss may be more important than the amount lost, since evidence suggests that a rapid loss of 1 5% ° Data 2 body weight (equivalent to approximately 30% body fat loss) influences menstruation. from Pirke et aI.’s study indicated that in the presence of certain risk factors, intermittent dieting with alternating losses and gains of approximately 5% of body weight, might be able 21 Drinkwater et al. ’ 7 11 also to induce menstrual disturbances without leading to weight deficit. indicated that as menstrual irregularities increased in severity in their female subjects, weight became a more important factor. However, there is no conclusive evidence that percentage body fat plays an exclusive role in determining menstrual patterns, but it may perhaps act ° 2 ’ 3 synergistically with other factors) B) AGE  Age is an important determinant of menstrual function. For example, age at the onset 19 After of menarche has been linearly related to the incidence of menstrual dysfunction.  15  REVIEW OF LITERA TURE  menarche, cycles are initially unstable, frequently anovulatory, or the luteal phase is 21 A refinement to more closely approximate the course of the biological process disturbed. is introduced by studying the length of the menstrual cycles in gynecologic ages, i.e., in years after menarche. This technique eliminates artificial distortions and interferences created by individual differences in the rate of adolescent development, but it still retains the isochronal, annual time intervals. 15 At a gynecologic age of six years, a large majority of women have ovulatory cycles with adequate luteal phases.  Temporal stability of the menstrual cycle  continues to further increase and the frequency of short luteal phases to decrease from this 21 age up until the age of 30 to 35 years.  Delayed menarche results in prolongation of the  hypoestrogenic state, which may lead to bone loss and osteopenia at the age during which bone density ideally should be increasing, and thus may lead to lower peak bone mass and earlier osteoporosis. 26 ’ 1619 C) MENSTRUAL HISTORY Most studies indicate that women who have some history of menstrual dysfunction 29 respond to exercise training by developing amenorrhea.  Furthermore, those who have  amenorrhea lack the beneficial effects of estrogen on calcium metabolism, lipid metabolism, and urogenital epithelial maturation.  Estrogen promotes the maturation of the urogenital  epithelium and maintains a thick vaginal epithelium that is resistant to many disorders. Estrogen deficiency leads to thinning of the vaginal epithelium and increased susceptibility to 19 atrophic urethritis and vaginitis, as well as increases the risk of developing osteoporosis. Interestingly, Fardy ° noted that even those runners who had regular menstruation thought 3 that their menses had been affected by exercise  16  -  they reported diminished premenstrual  REVIEW OF LITERA TURE  symptoms, shorter flow times and decreased menstrual flow. Menstrual irregularities may have gone unnoticed here since only histories were used and no measures for anovulatory or ° 3 shortened luteal phase cycles were made. D) A THLETIC TRAINING Many studies have focused on the effects of athletic training on menstrual function but the findings are inconclusive.  Prospective studies have documented changes in ovulatory  function in athletic women, including decreases in the endogenous production of progesterone during cycles with insufficient and short luteal phases (<1 0 days) and anovulatory cycles, 31 indicated that ° Pirke et al. 1 however these findings were not exclusive to the athletes. athletic amenorrhea is generally assumed to be hypothalamic in origin, and therefore , studying bone mineral content 29 luteinizing hormone is said to play a crucial role. Fisher et al. and levels of gonadotropins and estrogens in amenorrheic running women, found that luteinizing hormone pulse amplitude was significantly lower in the amenorrheic women, than 31 studied the balance of cyclic in their normal counterparts. 29 However, when Pirke et al. hormone secretion and episodic luteinizing hormone secretion patterns during normal and disturbed menstrual cycles, their overall finding was that major individual differences exist between anovulatory athletes, as well as high intra-individual variability. This indicates that the luteinizing hormone secretion patterns are likely of little significance for the development or the prediction of subclinical cycle disturbances in athletes. El STRESS Some studies have indicated that emotional stress may affect menstrual function, though studies of the relationship between stress and menstrual problems are few.  17  A  REWEW OF LITERA TURE  longitudinal study with frequent blood sampling considered information possibly linking nutrition and stress with a quantitative assessment of endocrine menstrual function in 1 8 32 The authors found that endurance-trained athletes and 25 age-matched nonathletic controls. luteal function was more easily disrupted than estradiol production by inadequate nutrition, and that caloric intake and stress interact with exercise to bring about disturbances to luteal 32 phase hormone production.  Also, other studies have shown higher cortisol levels,  9 found 27 Conversely, Clarvitt presumably reflecting increased stress, in amenorrheic women. that oligomenorrhea and amenorrhea were not more frequently experienced by women who 9 However, the latter study failed to include the reported a high level of psychological stress. possibility of subclinical menstrual disturbances occurring, thus the difficulty in extrapolating these findings. There are, therefore, no conclusive data available on the influence of stress on subclinical menstrual disorders.  4. VEGETARIANISM AND THE MENSTRUAL CYCLE  The influence of nutrition on the menstrual cycle, especially as it relates to low-calorie dieting and body composition, has been referred to previously. Another aspect of nutrition relates to the overall dietary pattern; namely, whether the diet is vegetarian or omnivorous, since a higher prevalence of menstrual cycle disturbances in vegetarian women has been cited 78 in some studies. A) DIETARY DIFFERENCES BETWEEN VEGETARIANS AND NON VEGETARiANS  It is not surprising that vegetarians as a whole tend to eat diets that differ in food choices from those of their nonvegetarian counterparts. It has been shown that they tend to  18  REVIEW OF LITERA TURE  include increased consumptions of fruits, berries, vegetables and herbal teas. Lacto-ovo and lacto-vegetarians are also inclined to consume increased amounts of dairy products. On the other hand, nonvegetarians have been shown to have increased consumption of biscuits, buns, sweets, alcoholic beverages, and coffee and tea. Along with the differences in food consumption patterns, it has been shown that disparity also exists in food preparation 33 These food consumption pattern differences are related to divergent food related methods. behaviours. Vegetarians tend to adhere more strongly to food related value-orientations of 34 ethics, religion, and health than nonvegetarians. It is difficult to specify the exact nutrient intake differences that occur between the vegetarian and nonvegetarian diets due to high intra-individual variability in food consumption patterns as well as the difference in food choices between different types of vegetarians. Furthermore, much of the data dealing with differences in vegetarian and nonvegetarian diets are not entirely reliable. Many studies use omnivores that have been placed on vegetarian diets which does not ensure that the diet followed is indicative of a typical vegetarian diet. Some do not specify the precise vegetarian subgroup types within which nutrient intake differences may exist. Others do not ensure vegetarians have been following the same type of diet over an extended period that would ensure the elimination of diet adjustment  And,  often dietary records are not long enough, thereby not controlling for the effects of seasonal variability or inconsistencies that may exist over different stages of the menstrual cycle. However, the major macronutrient differences that appear to be consistently seen between vegetarian and nonvegetarian diets are in the increased consumption by vegetarians of carbohydrates, particularly complex carbohydrates and dietary fibre, and in decreased  19  REVIEW OF LITERA TURE  8 There also tends to be greater consumptions of simple sugars, protein, fat and cholesterol. intakes of vitamin C, carotene and calcium in the vegetarian diet. Nonvegetarians are inclined 12 and selenium than their B to consume greater amounts of vitamin D, retinol, vitamin , 33 vegetarian counterparts. B) CROSS-SECTIONAL REPORTS OFMENSTRUAL IRREGULARITIES IN VEGETARIANS Work in the field of diet and hormones has shown that diet has significant effects on sex hormone levels, particularly on their metabolic clearance rate and peripheral availability, Diet appears to influence the production,  35 and thus also on their biologic activity.  metabolism, and excretion of endogenous estrogens, and may play a pivotal role in estrogen 36 metabolism thereby influencing menstrual cycle regularity. Vegetarian women, in particular, may be susceptible to irregular menstrual periods 8 It has been implied which suggests that they may have decreased reproductive capabilities. that these women are more sensitive to menstrual irregularities due to altered hormone profiles, however, whether these alterations influence bone density remains to be established. 37 that a meatless vegetarian diet decreased nocturnal It was reported by Hill et al. release of prolactin, plasma prolactin during the menstrual cycle, and the length of the menstrual cycle. They found that 1 6 women fed this diet had decreased levels of luteinizing hormone and follicle-stimulating hormone, and determined that the dampened release of luteinizing hormone could result in anovulatory cycles because of the hormone’s role in follicular development. They also concluded that the lower episodic release of gonadotropins and the shorter duration of the follicular phase implied diet played a crucial role in the control 37 of ovulation through the central nervous system hypothalamic axis.  20  REVIEW OF LITERA TURE  Although studies have reported an increased incidence of menstrual dysfunction in the presence of vegetarianism, the exact links have yet to be determined. Pedersen et al. 8 found that vegetarians consumed significantly greater amounts of polyunsaturated fatty acids, carbohydrates, magnesium, vitamin B , and dietary fibre, and less caffeine and cholesterol 6 than nonvegetarian counterparts. They found that the most significant predictors of menstrual regularity were protein/kJ, cholesterol/kJ, magnesium/kJ, and dietary fibre/kJ.  Increased  intakes of magnesium/kJ or dietary fibre/kJ were associated with an increased probability of being irregular, whereas high-cholesterol/kJ or high-protein intakes/kJ were associated with increased likelihood of menstrual regularity. The positive association between cholesterol/kJ and menstrual regularity may be secondary to cholesterol’s role as a biological precursor for estrogen synthesis. The frequency of menstrual irregularity in this study, for nonvegetarians (n=41) was 4.9%, which falls within the normal range of 25%38, yet it was 26.5% among vegetarian subjects (n=34). 8  Their results, however, are questionable due to problems  encountered in their study design.  First of all, menstrual regularity was based solely on  subject description and was not documented in any way, thus the incidence of subclinical menstrual disorders was excluded.  Also, their subjects’ ages ranged from 17 years to 47  years which is possibly entering the “grey zones” approaching menopause and of menstrual development following menarche.  Parous women were also included in the study.  Furthermore, their dietary data relied on a single three-day diet record which did not take into account variability due to phase of the cycle nor did it provide a sufficient example of normal intakes from which conclusions could be extrapolated. Goldin et al. 1 indicated that plasma levels of estrone and estradiol were lower in  21  REVIEW OF LITERA TURE  vegetarians, and a positive correlation was found between the concentration of plasma estradiol and dietary consumption of linoleic acid. 1 16 indicated that large amounts of vitamin A ingested by amenorrheic women UIIrich may result in inhibited ovulation, which has previously been seen in rats. This has been 7 This, however, ascribed to the 1-carotene deposited in fat tissue and in the corpus luteum. remains controversial. C) EFFECTS OF LOW-CALORIE VEGETARIAN DIETS ON THE MENSTRUAL CYCLE  Schweiger et al. 21 studied diet-induced menstrual irregularities and the effects of age and weight loss in 22 slightly active (one hour slow jogging/week) women who were placed on weight-reducing vegetarian diets. Their results showed that follicular phase and midcycle hormonal profiles were essentially unaltered, but the luteal phase, luteinizing hormone and estradiol were diminished, particularly in the 1 9 to 24 year group.  Overall, Schweiger et  21 data confirm the findings that a 1 ,000-calorie, vegetarian, high-carbohydrate diet has al.’s the potential to induce menstrual cycle disturbances, and that mainly the luteal phase is affected. It was noted however, that the extent of weight loss during a certain time period seems to be equally critical to the induction of menstrual disturbance.  Age was also  significantly and negatively correlated with luteal phase disturbances, and weight loss showed a significant negative correlation with luteal phase length and estradiol. Thus, they concluded that the risk of developing a diet-induced cycle disturbance is highly age-dependent, and that increased vulnerability will persist for several years once normal adult function of the hypothalamic-pituitary axis is established. The alteration of luteal phase induced by the high carbohydrate, vegetarian, 1 ,000-calorie diet can be viewed as an ontogenetic regression to  22  REVIEW OF LITERA TURE  21 early postpubertal menstrual patterns.  Similarly, the alterations in menstrual cycle in  amenorrheic runners have been found to exhibit the same hormonal profile as prepubertal 19 girls. , the effects of a mixed, relatively high-protein 7 In a comparable study, by Pirke et al. 1 ,000-calorie diet (n  =  9) were compared to a vegetarian, relatively high-carbohydrate 1 ,000-  calorie diet (n=9) over a six-week period, and it was found that menstrual disturbances reached a significant level only in the vegetarian group (P<O.02). These data indicate that a 1 ,000-calorie, vegetarian, high-carbohydrate diet may have a potential role in inducing menstrual cycle disturbances. They found that the different diets used to achieve weight loss had different effects on hormonal regulations during the cycle. In seven of the nine women they had placed on a vegetarian weight reducing diet, a typical pattern of hormonal abnormalities developed: luteinizing hormone values during and after the midcycle peak were lower (P<O.O1), the period of increased luteinizing hormone values were shorter (P<O.02), and consequently there were decreased progesterone (P<O.05) and estradiol values  (P<O.02), and a shorter luteal phase. Seven of the nine nonvegetarian women, however, displayed no significant hormonal changes, and maintained unaltered ovulatory cycles. They concluded that the kind of diet applied to lose weight may determine the impairment of the 7 Mainly the luteal phase was affected which might be hypothalamic-pituitary gonadal axis. related to the fact that dieting started at the beginning of the follicular phase in this study. Also, because all subjects were eating more than 500 grams of meat or fish per week as a part of their normal diet prior to the study, and because they were placed on the vegetarian diet for a six-week period only, it is feasible that this study’s findings are merely a result of  23  REVIEW OF LITERA TURE  the subjects’ adjustment to their new and very different eating patterns. It should also be emphasized that this research has focused on women placed on vegetarian diets who were not vegetarians initially. The dietary shift from a mixed diet to a lacto-vegetarian diet was shown to have the greatest change in nutrient intake between the period before and at three months after the dietary switch in 20 healthy-weight omnivorous volunteers, demonstrating the presence of an adjustment period. Significant differences were observed in intakes of carbohydrate (P=O.OO1), fat (P<O.O1), protein (P<O.05), alcohol (P<O.005),sucrose(P<O.005),fibre(P=zO.OO1),vitaminC(P=O.Q19),carotene(P=Q.QO1),  vitamin D (P=O.OO1), retinol (P<O.05), calcium (P=O.006) and selenium (P=OOO1). 33 Furthermore, the former studies focused on the weight-reducing capabilities of the vegetarian diet, rather than the diet in a normal setting. Thus, this leads one to question: would similar results be established in vegetarian women who practised normal levels of activity, maintained their normal weight, and had employed vegetarian habits for a period of time prior to the study? Thus, both cross-sectional studies and short-term studies of low-calorie vegetarian diets suggest that the vegetarian diet may be associated with menstrual irregularities. Evidence of interactions between sex hormones and nutritional components is accumulating, and present indications point to dietary macronutrients that alter metabolic pathways for testosterone and estradiol. 39 The search for a mechanism has focused on nutrients thought to differ between vegetarian and nonvegetarian diets, particularly fat and dietary fibre. In the following paragraphs, research relating to these nutrients and their influence on sex steroids will briefly be highlighted.  24  REWEW OF LITERA TURE  D) EFFECTS OF DIETARY FA T ON SEX STEROIDS A ND THE MENSTRUAL CYCLE  5 have demonstrated a significant effect of low fat intake on serum estrogen. Rose et al. They placed 1 6 premenopausal, regularly menstruating women, with cystic breast disease, on a low fat diet (21% of energy as fat) as compared to their usual diet (35% of energy as fat). Hormone levels were assessed in single blood samples obtained between days 1 7 and 20 of each cycle. Their reported energy intakes decreased significantly but no significant change occurred in the intake of dietary fibre. After three months on the diet, during which time a small loss of weight occurred, total serum estrogens had decreased from 299± 100 pg/mL to 200  ±  63 pg/mL (P< 0.00 1). The authors also concluded that increasing dietary fibre  to amounts commonly consumed by vegetarians, approximately 25-30 g/day, may result in 5 further suppression of serum estrogen concentrations. Conversely, dietary habits, including vegetarianism, low red meat consumption, eating disorders, pathogenic weight control methods, and weight reduction diets, were not ° cross-sectional study of 4 associated. with menstrual disturbances in Watkin et al.’s ultramarathon runners. Although they did not look at subclinical menstrual disturbances per Se,  they did find that the percentage of fat intake tended to be lower in subjects with short  menstrual cycles. Their overall conclusion, however, was that the development of menstrual disturbances (amenorrhea, oligomenorrhea, and short menstrual cycles only) was unrelated 40 to current nutrient intake. E) EFFECTS OF DIETARY FIBRE ON SEX STEROIDS A ND THE MENSTRUAL CYCLE  Goldin et al. , studying estrogen levels and diets of Caucasian American (n=1O) and 1 Oriental immigrant women (n  =  1 2), found that premenopausal Oriental women had a much  25  REVIEW OF LITERA TURE  higher percentage of calories as carbohydrate, with a fibre consumption exceeding the Caucasians by approximately 20%.  None of the subjects was strict vegetarian but the  Caucasians ate diets that were higher in animal foods with a significantly higher fat-to-fibre ratio than the Oriental women. The premenopausal Oriental women excreted in their feces approximately two times more estrone, three times more estradiol, four times more estriol, and three times more total fecal-estrogen per 24 hours than the premenopausal Caucasians. Furthermore, the premenopausal Oriental women had significantly lower plasma levels of 1 estrone and estradiol, approximately 32% less, than did their Caucasian peers. The mechanisms responsible for increased fecal excretion of estrogens are not known, 25 although a variety of mechanisms have been proposed.  8 suggested that Pedersen et al.  lowered plasma concentrations of estrogens in vegetarians may be due to direct binding of steroid hormones to fibre species in the gut or to alteration of steroid resorption caused by 8 Unfortunately, conversion in gut flora resulting from different amounts and species of fibre. dietary fibre embraces a variety of complex chemical structures, including celluloses, hemicelluloses, pectins, gums, and lignans. It is possible that any biological impact of dietary fibre on circulating sex steroids is mediated by a specific fibre subspecies, rendering total 41 For example, it is possible that differences of dietary fibre measurements of limited value. circulating sex steroid concentrations in vegetarian and nonvegetarian women may be due in part to the effect of lignans and phytoestrogens from dietary sources. Lignans inhibit the tissue aromatase enzyme responsible for peripheral estrogen formation, thereby lowering circulating estrogen concentrations. Fecal values of lignans and phytoestrogens were found 8 to be negatively correlated with plasma estrone and estradiol concentrations in subjects.  26  REWEW OF LITERA TURE  Unfortunately, most data analysis programs offer limited information on fibre content of foods, 42 and even less on fibre subspecies. It has also been suggested that, since estrogens are excreted in the bile as glucuronide and sulfate conjugates, vegetarians might hydrolyse fewer estrogen conjugates resulting in lower intestinal reabsorption and greater fecal excretion. 1 These are consistent with a highfibre low-fat diet, which is similar to a vegetarian diet.  The former has been shown to  diminish the extent of estrogen absorption from the intestinal lumen, thus influencing its contribution to the total body pool of estrogens by causing a reduction in the activity of 6 bacterial enzymes, particularly l-glucuronidase. Overall, mechanisms that have been proposed for the impact of diet on estrogens include: (i) the modification of synthesis or clearance rates of estrogens; (ii) the alteration of the conversion of androgens to estrogens; and (iii) the modulation of the enterohepatic circulation of estrogens. 39 Goldin et al. 43 examined estrogen excretion patterns and plasma levels in 10 vegetarian and 10 omnivorous premenopausal women on four occasions, approximately four months apart. On each of these occasions three-day dietary records were kept, and estrogens were measured in plasma, urinary and fecal samples. Overall, the vegetarian women consumed less total fat than their omnivorous counterparts, although the difference was not significant (30% of total energy vs. 40% of total energy), and more dietary fibre (28 g/day vs. 1 2 g/day; P<O.OO1).  There was a positive correlation between fecal weight and fecal excretion of  estrogens in both groups (P<O.OO1) with vegetarians having higher fecal weight and increased fecal excretion of estrogens. Urinary excretion of estriol was lower in vegetarians  27  REVIEW OF LITERA TURE  (P’<O.05) and their plasma levels of estrone and estradiol were negatively correlated with fecal excretion of estrogen (P=O.005). Also, the 1-glucuronidase activity of fecal bacteria was significantly reduced (P=O.05). The authors’ conclusions, similar to what has been stated previously, were that the greater fecal bulk and non-absorbed fibre somehow shielded the estrogens excreted in the bile from deconjugation and reabsorption. 43 Rose et al. 6 studied the effect of dietary fibre in 62 premenopausal women with regular ovulatory menstrual cycles. The women were randomly assigned to supplement their diets with corn, wheat or oat bran, for two months, the goal being to increase total fibre intake to a minimum of 30 g/day, representing an increase of  100% above baseline. Changes in the  percentage of energy from fat, which could confound the results, did not occur during this study. Their results indicated that wheat bran had a significant lowering effect on serum estradiol and estrone, whereas corn and oat bran had none. No changes were seen in the concentrations of sex hormone binding globulin, suggesting that a decrease in free estrogen levels occurred. Serum progesterone was significantly lower only after one month on the wheat bran supplement, but returned to baseline values by the second month. No changes in the menstrual cycle were noted. 6 The fact that only wheat bran caused a reduction in serum estrogen concentrations seems to validate the theory that unconjugated estrogens bind 8 ’ 6 to the fibre components in the gut, thus impeding their reabsorption. Furthermore, Hagerty et al. , who studied six ovulatory lacto-ovo vegetarian women 44 in a cross-over design, found that increasing short-term dietary fat intake in the healthy premenopausal women had no effect on fasting plasma hormone levels, and did not stimulate estrogen secretion.  This implies that it was the dietary fibre, rather than the low-fat  28  REVIEW OF LITERA TURE  component of the vegetarian diet, that was causing the reduction of estrogen, and menstrual 44 Although the diets were well controlled, the women followed each diet for dysfunctions. only one month and the sample size (n  =  6) may have been inadequate to detect changes.  5. DIETARY RESTRAINT  Although low-calorie dieting and central stress have both been suggested as being associated with disturbances in the menstrual cycle, to date relatively little work has been done to evaluate whether the degree of stress experienced in association with eating, irrespective of energy intake per se, could also play a role.  This concept of experiencing  stress in association with eating can be operationalized as dietary restraint, or the conscious 46 control of food intake to achieve or maintain a slim body shape.  Women who are  characterized as restrained eaters feel that they are always limiting the amount of food that they eat, and consciously eating less than they desire. 4 It is conceivable that this food-related stress could act centrally to affect the menstrual cycle, and preliminary evidence for this will be discussed below. It has been suggested that a low-fat diet, as well as the low-energy diets seen in restrained eaters, reduces plasma levels of sex hormones and can inhibit cycling and 16 14 ovulation. 2 3 4 ’ 9 5  Herman and Mack 4 stated that there may be a large proportion of  normal weight eaters who are biologically “underweight”, according to the theory that individuals have a biologically determined set-point for weight. Under ordinary circumstances, these “underweight” individuals would be expected to “overeat” in response to their set-point demands. However, these individuals are also subject to the countervailing pressure exerted  29  REVIEW OF LITERA TURE  by cultural and social demands; namely, to restrain their eating so as to maintain an “ideal” weight which is fairly low in absolute terms and extremely low relative to their biological setpoint weight.  Such individuals, then, should be characterized by normal weight levels,  restraint in their eating habits, and a form of “latent” externality which would be manifested strongly in the event that chronic restraints could be eliminated or overcome. Low restraint subjects, however, since they are able to maintain normal weight levels without the support of restraint mechanisms, are presumably at or near their set-point.  Herman and Mack 4  demonstrated this theory in their study on restrained and unrestrained eating. They showed that the more concerned, more restrained subjects (n  =  23) ate more when attractive food  cues were prominent once chronic restraint was experimentally eliminated. The experimental manipulation of situational restraint had little effect on those subjects who were not chronically restrained (n=22). Schweiger et al. 45 also showed that the eating behaviour of the restrained women was characterized by a tendency to overeat under conditions antagonistic to self-control, such as during alcohol consumption or emotional arousal. Tuschl et al. 47 demonstrated that the alterations in the eating behaviour of restrained eaters can affect biological functions susceptible to caloric intake, since their subjects had impaired menstrual cycles.  It has been postulated that chronic dietary restriction is the  additional essential component necessary to convert the impaired pulsatile luteinizing hormone release, which occurs with exercise and its attendant stresses in many female athletes with normal cycles, into oligomenorrhea or amenorrhea. 19 Schweiger et al. 45 found that increasing levels of restraint resulted in shorter total cycle lengths as well as decreased luteal phase lengths.  30  They also suggested that a possible  REVIEW OF LITERA TURE  mechanism contributing to the reduced caloric intake in restrained eaters may be due to the association of luteal phase length on increased energy expenditure. The authors hypothesized that women with shortened luteal phase lengths would therefore have decreased energy expenditure and develop restrained eating as an adaptation to their predicament. 45 Despite the lower caloric intake of the restrained eaters, their body weight was relatively elevated, perhaps due to a reduced level of energy expenditure. 47 Although none of this information has specifically focused on vegetarian women, the role of restrained eating may be vital in relation to menstrual irregularities, and crucial in identifying the possible causes determining these dysfunctions.  Based on the association  , it might be speculated 10 between subclinical menstrual disturbances and trabecular bone loss that women with high scores for dietary restraint would have lower values for bone density, or may be more likely to lose bone over time. A possible mechanism for the association among dietary restraint, menstrual cycle disturbances and bone loss is related to the effects of stress.  Women with high levels of dietary restraint may experience greater stress in  association with food intake. Central stress responses have been shown to be associated with increased corticotropin-releasing hormone levels, which in turn can affect the steroid milieu, in a number of ways, including increasing levels of cortisol. 27 A study by Grindoff and Ferin 48 showed that the administration of corticotropin releasing factor exerts a central nonadrenal-mediated inhibitory influence, resulting in decreased pulsatile secretions of both luteinizing hormone and follicle-stimulating hormone. This is mediated by endogenous opioid-peptides.  l-endorphin itself, which is secreted in  response to corticotropin-releasing factor, has been shown to suppress gonadotropin releasing  31  REVIEW OF LITERA TURE  hormone pulse frequency and/or amplitude, consequently interfering with normal ovulatory menstrual patterns. 133548 This hormone essentially needs a specific steroid milieu to exert its potential activity, and disruption can result in amenorrhea. 13  The corticotropin-releasing  hormone administration also acts through adrenocorticotropic hormone to increase cortisol 27 Glucocorticoids have been implicated as possible mediators of stress-induced production. gonadotropin inhibition, but it is unlikely that cortisol is directly responsible for the observed decrease in gonadotropins following the administration of corticotropin-releasing hormone. 48 As Hotta et al. 27 indicated in their study on anorexia nervosa patients, there was a low to zero response to corticotropin-releasing hormone administration, likely due to the already present hypersecretion  of that hormone.  They found that their subjects’ plasma cortisol  concentrations were significantly higher than the mean value in normal women. This is likely linked to the decreased food intake seen in these women and thus likely plays a vital role in dietary restraint and the menstrual cycle as well. 27 Dietary restraint can result in altering menstrual cycle status, and possibly impact bone, as will be discussed in more detail later, but has been shown to also act directly on bone due to the increased cortisol levels found in association with the stress response to food intake in restrained women. Increased levels of cortisol affect bone in two ways: (i) by decreasing levels of vitamin D, a nutrient required for the absorption of calcium in the bones; and, (ii) by sitting on bone osteoblasts thus interfering with bone formation. 49  High levels of cortisol  secretion have been reported in amenorrheic athletes 13 and in women with anorexia nervosa , 27 but this does not appear to have been evaluated in women with high levels of dietary restraint, nor are data available on bone density of women with restrained and nonrestrained  32  REVIEW OF LITERA TURE  eating patterns. Along a similar line, however, Joyce et al. ° examined 33 patients with eating 5 disorders, eight of whom fell into the subgroup “Eating Disorder Not Otherwise Specified”, made up of individuals not meeting the criteria for anorexia nervosa or bulimia nervosa. Their results showed that this subgroup had significantly lower bone density in all five sites measured and no evidence of a significant correlation of this decreased bone density with estrogen deficiency. However, neither cortisol levels nor emotional stress were measured in this study. ° Albeit the observations were made only on a small number of women, the 5 results do nevertheless show some promise for the role of subclinical eating disorders and bone.  6. MENSTRUAL DISTURBANCES A ND BONE  Most of the research in the area of menstrual irregularities has focused on female athletes. There is insufficient material on menstrual disturbances and bone in normally active women as well as little information on subclinical menstrual disturbances in normally active women and the latter’s effect on bone. Although many studies have appeared dealing with the immediate endocrinological and physiological changes in athletes, the underlying mechanisms have not yet been elucidated. 13 The oligomenorrhea/amenorrhea experienced by many young female athletes had been considered a benign and reversible condition until several investigators reported a lower vertebral bone mineral density for these women with low levels of sex steroids compared with regular menstruating ’ 51 peers. 5 2 3 Experts hypothesize that the maximization of trabecular peak bone density is reached  33  REVIEW OF LITERA TURE  in the late teens or early twenties and is important in preventing osteoporosis.  ,l82939.54  However, in a longitudinal study of 1 56 women of college age in whom no conscious intervention in lifestyle was undertaken, it was found that there was a gain in bone mass over the entire skeleton. They concluded that a woman’s gain in bone mass would be complete by about the age of 30 years, rather than during the early twenties. 55 Lloyd et al. 18 report a mean spinal trabecular bone density of 197 mg/mL in normally menstruating collegiate female athletes.  Trabecular bone loss is said to average 1-2%  annually, and atraumatic spinal fractures are said to be highly likely when the trabecular bone density falls below 70 mg/mL. It has been suggested that women who have missed 50% of their expected menstrual periods, such as those with delayed menarche, are likely to reach the age of 20 with a peak bone density of less than 1 50 mg/mL, and these women will be at substantially greater risk for the development of osteoporosis. 18 Drinkwater et al. 11 found a significant linear relationship between the current vertebral density of their female subjects and their past and present menstrual patterns. There was a significant difference in the vertebral bone mineral density between each menstrual group, in as much as those who had always had regular periods had the highest value, whereas those who had experienced occasional irregularities or who had never been regular averaged six percent and 1 7% less, respectively, It is interesting to note that those women who had never experienced regular menses were younger, weighed less, and had a later menarcheJ 62129 A) SUBCLINICAL MENSTRUAL DISTURBANCES AND BONE  Prior et aL’° documented that short luteal phases and especially lack of ovulation in menstrual cycles of normal length may be potential risk factors for excess bone loss in  34  REVIEW OF LITERA TURE  women. They found, through multiple regression analysis, that the luteal phase index (luteal phase length/cycle length) was the strongest explanatory variable for the change in spinal bone density, with caloric intake contributing minimally and family history of osteoporosis contributing about four percent.° B) SEX STEROIDS AND BONE  The role of sex steroids is as complex and as important in the maintenance of bone integrity as in the maintenance of a normal menstrual cycle. There is now a great deal of evidence indicating that sex steroids have a direct metabolic effect on cartilage and bone cells, thus the integrity of the menstrual cycle can potentially influence bone due to the vital interaction of sex steroids in both instances, Indirectly, steroid secretion is associated with stimulation of growth hormone secretion, which in turn triggers insulin growth factor-i which is partly accountable for bone maturation. Directly, however, estradiol has been shown to 56 It has also been suggested that increase osteoblast activity and enhance replication in bone. estrogen decreases sensitivity of bone to parathyroid hormone. The latter is a regulator of calcium and phosphorus metabolism that stimulates the bone resorption process.  When  estrogen is absent, bone appears to become more sensitive to the parathyroid hormone, thus shifting the remodelling process toward resorption. 57 Although much of these data were observed in vitro on cultured cells, they underlie the effect of sex steroids on skeletal tissue seen in vivo. 56 Furthermore, much recent data indicates that progesterone is also active in bone metabolism.  Directly, progesterone seems to act by engaging a bone osteoblast  receptor, and indirectly, it competes for a glucocorticoid osteoblast receptor. Progesterone appears to promote bone formation and/or increase bone turnover, and it may even play a role  35  REVIEW OF LITERA TURE  in the combining of bone resorption with bone formation. Overall, however, this sex steroid’s link to bone appears to be as crucial as estrogen’s. 58 Accelerated bone loss occurs with the cessation of menstruation at the time of menopause in women whose sex steroid levels drop and, as mentioned earlier, it has been shown that similar losses occur in premenopausal hypoestrogenic women, such as those with amenorrhea.’° 9 ’ 16 2953 Regular menses require normal plasma estrogen concentrations during the menstrual cycle and menses becomes disturbed or absent as these sex hormone plasma values decrease toward postmenopausal concentrations. 85358 Moreover, the results of Lloyd al.’s’ study support the notion that circulating sex steroid levels during adolescence play et 8 a fundamental role in determining peak bone density in women.’ 8 They found that the mean bone density of their moderately oligomenorrheic group (n=1O) was 88% of their normally menstruating peers (n=6), and that of their severely oligomenorrheic group (n=4) was 69% of their counterparts, which agrees with the hypothesis that bone density loss increases with 8 the level of cycle disturbance.’ 59 studied the hormonal characteristics of 28 premenopausal women Sowers et al. whose femoral bone mass values were below the lowest fifth percentile of the distribution observed in their initial epidemiologic study.  Their resufts were compared to 25  premenopausal women’s hormonal characteristics whose femoral bone mass values were within one standard deviation of the mean distribution. The group with a lower bone mass were found to have significantly lower mean estradiol levels, though within the range considered normal for this assay (76±45 pg/mL vs. 106±72 pg/mL; P<O.05) and significantly higher luteinizing hormone levels (3.8 ±2.7 mlU/mL vs 3.1 ±3.2 mlU/mL;  36  REVIEW OF LITERA TURE  P<O.07) than the control group.  There was no significant difference found in mean  progesterone levels between the two groups. Although these data are preliminary in that they identify hormonal characteristics at a single point in time within the female cycle (between days 1 8-22), they do imply the importance of estradiol in premenopausal skeletal integrity. 59 Snead et al. 53 examined the relationships among reproductive hormone concentrations and bone mineral density in 43 women runners classified as eumenorrheic (n=24), oligomenorrheic (n  =  8), or amenorrheic (n  normal cycle non-running women.  =  11). Their results were compared with eleven  Their major findings were that the amenorrheic and  oligomenorrheic women had significantly lower estradiol and progesterone levels (P<O.05), which were measured on a daily basis for 21 days, than the eumenorrheic runners and sedentary counterparts, as well as a 1 2% lower bone mineral density value at the lumbar spine (P<O.05). They also found that daily integrated progesterone levels were significantly correlated with spinal and femoral bone mineral densities in the eumenorrheic runners. 53 C) WEIGHT AND BONE Vico et al. ° found that estrogen-related parameters had more effect on the spinal bone 6 mineral density 4 -L measurement, whereas body weight was more important for predicting 2 (L ) bone mineral density in other areas.  Drinkwater et al. 11 found that body weight was a  significant predictive variable for current bone density, although menstrual history added significantly to the prediction of vertebral and femur density. DI EXERCISE AND BONE Because much of the research in this area has focused on athletes, a considerable amount of research has focused on the effects of exercise on bone. Also, because of the  37  REVIEW OF LITERA TURE  increasing awareness of health and body image, and the fact that women are more susceptible to the social pressures surrounding these factors, it can be postulated that many individuals are choosing to increase their activity levels. Thus the effects of exercise on bone are pertinent. However, research has demonstrated, as will be discussed in the following paragraphs, that, although exercise promotes bone formation and retards bone loss, it cannot compensate entirely for sex steroid deficiency. Heinrich et al.’s° 1 study on bone mineral content of cyclically menstruating female resistance and endurance trained athletes found that body builders (n mineral content than runners (n  =  1 6), swimmers (n  =  =  11) had higher bone  1 3), and the inactive group (n  =  1 8), on  all sites measured on the skeleton. Their multiple regression analysis indicated that body composition was significantly related to bone mineral content (P< 0.05). They also found that fat-free body weight was significantly correlated with bone mineral content (P<O.05), and fat-free body weight remained a significant predictor even after the effects of athletic group and body weight were removed from the analysis. Since muscle mass makes up over 40%50% of the fat-free body mass, it is reasonable to assume that individuals with greater fatfree body mass will also have the larger muscles capable of exerting greater tension on the 61 skeleton, thereby increasing their bone mineral content density. 54 supported the concept that vertical weight-bearing A similar study by Risser et al. exercise can help to maximize bone density in early life. They measured bone density in the , they found significantly lower (P< 0. 05) spine 61 calcaneus. However, unlike Heinrich et al. and calcaneus bone densities in their ten swimmers than in their 1 2 volleyball, seven basketball players, and 1 3 non-athlete controls.  38  They accounted for this by the lack of  REVIEW OF LITERA TURE  vertical weight-bearing activity involved in swimming. The relative weightlessness during many hours of swimming may decrease bone density, as does zero gravity in astronauts. 54 Contrary to most cross-sectional findings, Prior et  81.10  found no difference between  either the initial density or the changes in spinal trabecular bone density after one year among marathon (n=21) and consistent runners (n=22), and normally active individuals (n=23), although the mean decrease in bone density for the entire study population was significant over their study period (-3.0 ±4.8 mg/cm /year; P<O.OO1). They did, however, find that the 3 group of women with normal cycles or with only one short luteal phase per year (n=25) had significantly different changes in bone density to the women who had more than one short luteal phase (n  =  28) or one or more anovulatory cycles (n  =  1 3), with the latter group losing  bone over the study period (0.42 vs. -5.0 mg/cm /year; P<O.0001). 3 ° 1 It is possible, however, that exercise may still have a positive effect on bone mass, even in the amenorrheic state. Drinkwater et al. , in a cross-sectional study on menstrual 11 history as a determinant of current bone density, found that their amenorrheic elite runners had a greater vertebral density than amenorrheic sedentary women, though their values were still lower than their normally menstruating counterparts. It has also been shown that eight months of progressive training by either muscular strengthening or running exercise can increase the lumbar spine mineral content and areal bone mineral density of young women. Although the increase in spine mineral achieved in this study was modest, and the subjects were young (mean age  =  19.9±0.7 years), it was significantly greater than that of the control  group who maintained their physical activity patterns at baseline levels. 62 Some studies have also focused on the capabilities of these women to increase their  39  REVIEW OF LITERA TURE  bone mineral density to a normal level for their age group. Cann et al. 63 concluded, from their study on menstrual history as a determinant of trabecular bone density in female runners, that prior untreated amenorrhea as short as three years in athletic women leads to irreversible trabecular bone loss, although this remains controversial. 63  Intervention and prevention  through education are crucial to prevent the development of premature osteoporosis, particularly for those with shortened luteal phases and anovulatory cycles which can go unnoticed for extended periods. Overall, osteoporosis, either pre- or postmenopausally, seems to result not just from low levels of sex steroids but also from a combination or interplay of a variety of other 23 Even though exercise may promote bone formation and retard bone loss, it cannot factors. compensate for sex hormone deficiency. Furthermore, the beneficial effects of exercise on bone density rely on the adequacy of dietary calcium consumption and unfortunately, many amenorrheic women consume inadequate amounts of that mineral. Even adequate calcium, however, does not compensate for hormone deficiency. For example, Shangold et al. 19 stated that bone loss has even been demonstrated in hypoestrogenic women who ingested 2,000 mg of calcium daily. This ties into findings by Riis et al. 64 indicating that calcium treatment had minor effects in retarding bone loss in the absence of sex steroids. Therefore the effect of regular load-bearing exercise on the bone density of premenopausal women appears paradoxical: in the presence of normal sex steroid levels, as exist during regular menstrual cycles, trabecular bone density increases, whereas amenorrheic athletes have reduced bone density, and reduced circulating estradiol and progesterone levels. 18 It has become more clear that the prevention and adequate therapy of menstrual dysfunction are essential to prevent  40  REVIEW OF LITERA TURE  these long-term complications, particularly premature osteoporosis. 202930.52 The differing results in the studies on athletic women, as well as the results ° analysis of spinal bone loss and ovulatory disturbances, seem 1 documented in Prior et al.’s to indicate that it is more important to focus on the menstrual irregularities of the women rather than on their level of activity, but activity level should be controlled within the study population to eliminate confounding variables.  Also, it has been shown that obtaining  menstrual cycle histories are not sufficient in gaining relevant insight into menstrual irregularities and bone loss, since shortened luteal phases and anovulatory cycles are not detected and cannot be accounted for.  7. VEGETARIANISM AND BONE  If, as described previously, vegetarian diets are associated with alterations in sex steroid levels and the menstrual cycle, it is possible that differences could be observed in bone mineral density. To date, there have been a few cross-sectional studies which have compared bone density in vegetarians and nonvegetarians but no longitudinal studies have evaluated the changes in bone mineral density over time in these two groups. The data available will be described in the following paragraphs. The relationship between vegetarianism and peak bone density remains unclear and few data are available on the prevalence of osteoporosis in vegetarian and nonvegetarian populations. It should be noted however, that large inter-individual variability in bone density among premenopausal women renders the detection of small differences unlikely. There is no published evidence relating one single nutritional deficiency to menstrual  41  REWEW OF LITERA TURE  dysfunction. Calcium, vitamin D, phosphate, protein, fluoride, magnesium, ascorbic acid, sodium, and various trace minerals have been identified as part of bone mass modulation, though their exact influences have yet to be established. 39 Of all nutrients however, dietary calcium is most frequently identified as being most directly related to bone health and may play a vital role in the rehabilitation of bone loss in hypoestrogenic women. 3957 Approximately 99% of total body calcium is present in the bones and teeth and there is a constant turnover of bone calcium, although some always remains as a stable reserve. Unfortunately, many nutrients and hormones can influence the body’s absorption, retention and use of calcium. 57 Lloyd et al. 41 state that the calcium needs of many female athletes is greater than the recommended dietary allowance of 800 mg, due to the need to build and maintain the greater bone mass found in athletes, and it is possible that the recommended dietary allowance should be closer to 1 500 mg for all premenopausal women. 41 However, Riis et  81.64  found  calcium to have no effect on bone mass when provided as a treatment for retarding bone loss in postmenopausal women.  Nonetheless, Reid et al. 65 studied the effect of calcium  supplementation on bone loss in 122 postmenopausal women and they found a sustained beneficial effect of calcium supplementation. And, on that same note, in a longitudinal study of 1 56 college aged women, it was found that bone gain was enhanced by self-selected calcium intake (adjusted for protein intake).  This influence was closely followed by the  estimated level of physical activity. 55 Dual photon absorptiometry was used by Lloyd et  81.66  to compare spinal bone  densities in 27 premenopausal lacto-ovo vegetarian women and 37 premenopausal nonvegetarian women.  Compared to the nonvegetarians, they found the diets of the  42  REVIEW OF LITERA TURE  vegetarian women to be higher in dietary fibre and carbohydrate, lower in protein, and similar in fat content over a three-day period.  Menstrual cycles of abnormal length were more  common in the vegetarian women, and, although the difference was not significant, their spinal bone density tended to be lower (1.02±0.02 g/mL vs. 1.06±0.02 g/mL). 66 Baker and Demers , in their study on the menstrual status of female athletes, 23 concluded that increased dietary fibre intake can be associated with menstrual dysfunction and may contribute to the decreased bone density seen in female athletes practising a highcarbohydrate diet, one that is similar in composition to a vegetarian diet. 41 Woods et al.’s ’ 23 36 results indicated that plasma estradiol was positively correlated with total and saturated fatty acid intake and negatively correlated with dietary fibre intake. 35 These findings indicate, as ’ 10 previously discussed, that dietary habits and food fibre levels may alter the enterohepatic circulation of estrogens, resulting in increased fecal excretion, and thereby affecting peak bone 4 31 ’ 36 density. 9 Although controversial, some studies have also proposed that dietary fibre affects 41 As previously discussed, bone density is influenced calcium absorption and calcium balance. by several factors: weight-bearing activity, sex steroids and calcium intake. Thus, it appears that dietary fibre may be hindering proper bone mass modulation by interfering with sex steroid as well as calcium absorption, placing these women consuming high fibre vegetarian diets at even greater risk of osteoporosis. 39 found that dietary carbohydrate, a large component of the Leuenberger et al. vegetarian diet, was also associated with lower bone density but independently of sex-steroid hormones. They deduced that, because dietary fibre is a component of carbohydrate, there  43  REVIEW OF LITERA TURE  would be interdependence of these two with regard to trabecular bone density. A portion of the starch component in carbohydrate increases fecal bulk and thus, like dietary fibre, could also augment fecal excretion of androgens, and act independently of sex-steroid hormones. 39 On the other hand, however, Prior et al. ° found that those women with higher caloric intakes, 1 which may indicate increased carbohydrate intake, displayed preferable bone density evaluations. This, however, still requires more prospective research for clarification.  8. SUMMARY  Overall, there has been much research in the area of menstrual dysfunction but most of it has focused particularly on female athletes, and few studies have actually looked at subclinical menstrual disorders because they relied primarily on self-reported menstrual status. Furthermore, the research that has examined the relationship between diet and sex hormones predominately considered weight loss diets and omnivores placed on temporary vegetarian diets. There are no convincing data on the prevalence of menstrual cycle disturbances in vegetarian women, nor how it would further affect bone. Moreover, the information on the link between dietary restraint and the incidence of menstrual dysfunction is limited. In conclusion it is evident that more information regarding the-role of diet and dietary restraint in the evolution of menstrual cycle disturbances is essential to further understand the possible association between vegetarianism and bone loss.  44  EXPERIMENTAL DESIGN  Chapter III  EXPERIMENTAL DESIGN  1. STUDY DESIGN The basic design of the study was that of a prospective, observational study extending over six menstrual cycles. Initially, the subjects were asked to come to the university for anthropometric measurements, including height, weight and four skinfold thickness measurements. Forearm girth, thigh circumference, and mid arm circumference were also measured.  Subjects were then asked to fill out the Three-Factor Eating Questionnaire to  measure levels of dietary restraint. 3 All subjects were requested to provide a chart of their daily basal body temperature for the following six menstrual cycles, and three 3-day dietary intake records staggered to obtain one during the follicular phase, one around the time of expected ovulation, and the third during the premenstrual phase. They were also asked to provide a two-hour fasting urine collection which was used to measure urinary cortisol, calcium and creatinine. A bone mineral density assessment was conducted using dual energy x-ray absorptiometry. The observations made were the same in both the self-chosen control and the experimental group, i.e., in the vegetarian and nonvegetarian women.  45  EXPERIMENTAL DESIGN  2. SUBJECTS  Sixty-two women were recruited to participate in the study. Subjects were recruited using posted notices describing the study’s objective and the benefits it would provide for the participants. An example of a recruitment advertisement can be seen in Appendix A. The total sample size recruited was 62 women, of which 32 were vegetarians, making up the experimental group, and 30 women were the control group.  This number was  achieved by using the following sample size equation:  Sample Size (n)  =  2 2(z) x a 2 2 Difference  67  where npopulation size per group; z=1.96 so that all tests were done at a significance level of P<O.05; difference =0.04.  The “Difference term is that which one judges to be biologically meaningful and which one would like to be able to detect. For this study a difference in luteal phase index of 0.04 ° in which the mean luteal 1 was selected as meaningful, based on the study of Prior et al. phase index was 0.36 ±0.07.  Sample Size (n)=2(1.96) 2 x 0.072 0.042  Therefore, n  24, and the total sample size  48.  As a precaution against falling below the required sample size due to subject drop-out, an extra 30% of the required study population was recruited. The 32 initial vegetarian recruits were defined as those excluding meat, fish and poultry from their diets, but did include those following a vegan, a lacto-vegetarian, or a lacto 46  EXPERIMENTAL DESIGN  ovo vegetarian diet. The other thirty nonvegetarian women consumed a minimum of three servings of meat, beef or pork, a week to ensure that no individuals qualifying as “semivegetarians” were included. All women had to have maintained their current diet habit for at least the previous two years, to ensure that they had adjusted to their new diet. All participants were between the ages of 20 and 40, nulliparous, and experiencing menstrual cycles of normal length (i.e., between 21 and 35 days). They all had a Body Mass . Alcohol consumption was not 2 2 and no more than 25 kg/rn Index (BMI) no less than 1 8 kg/rn to exceed 30 mL of distilled spirits daily, or a total of 210 rnL a week. All participants maintained stable and, at most, moderate exercise habits since training 53 None exercised more has been shown to be associated with menstrual cycle disturbances. than seven hours weekly, and compulsive exercisers were eliminated initially during the first telephone screening process. Obligatory exercisers were identified by the need to continue to exercise despite physical injury or clear personal contraindications. Potential subjects were asked during the screening process whether they would continue to exercise regardless of 68 injury or illness. They were also asked if they felt guilty when they did not work-out. Exclusion criteria for the study included use of oral contraceptive agents within the six months preceding entry into the study. Also, women using medications that affect bone, such as glucocorticoids or other bone active drugs (e.g., prednisone, thiazide diuretics) were excluded. Women suffering from serious acne at the time of recruitment, or who had above average amounts of facial and body hair, were also eliminated from the study as acne vulgaris, ° found 6 as well as excess hair, are often signs of androgen excess in women. Medda et at. that women acne sufferers had elevated levels of luteinizing hormone, testosterone, and  47  METHODS  hyperprolactinemia, and thus concluded that it is a condition that can play an important role in gynecological endocrinology. They would therefore not be representative of the general population. Finally, those individuals who smoked or who displayed unusual sleep or waking patterns, due to shift work for example, were omitted from the study due to the possible impact on their basal body temperature measurements. All subjects provided written informed consent prior to participating in the study. See Appendix B for a copy of the consent form.  METHODS  7. ANTHROPOMETRICS A) WEIGHT Some subjects (n  —  26) were measured on a Sunbeam electronic monitor scale  (Sunbeam Corp. Canada Ltd., Toronto ON) but most (n36) were weighed on a Health-O Meter beam scale (Continental Scale Corp., Bridgeview IL), on which the subjects stood. The electronic scale was in increments of and accurate to the nearest 0.5 kg and the beam scale was in increments of and accurate to the nearest 0.1 kg.  Both were calibrated to zero  ° 7 between subject weighings. The subjects were asked to bring t-shirt and shorts with them, and to remove shoes, 7071 to enable the measurement to be made with minimal clothing.  When subjects were  weighed, they were asked to stand in the centre of the platform, with their arms at their side.  48  METHODS  The preferred time of day was upon rising, if possible, prior to eating or drinking and after ° 7 emptying the bladder, although this was not always possible. The subjects’ weights were recorded immediately, and each subject was weighed in duplicate. If the weights were grossly inconsistent, a third weighing was executed, and the 70 two most similar weights were averaged. B) HEIGHT  To obtain an accurate measurement of height, a stadiometer was used. It was read ° 7 to the nearest 0.1 cm. The subjects stood in the centre of the platform, in stocking feet or barefoot, with feet parallel and heels together. Their heels, buttocks, shoulders and head were in contact with the vertical surface, and their back was kept as straight as possible. Their arms were left to hang naturally at their sides. Their head was kept comfortably erect, with the Frankfort plane horizontal (the line from the lower orbit to the upper external auditory meatus). The horizontal bar was lowered at a  900  angle, crushing the hair and making contact with the head. 7071  As with weight, the measurements were done twice, and recorded immediately. Again, if there was a meaningful discrepancy between the two measures, a third was 70 performed, and the two most similar results were averaged. C) BODY MASS INDEX  Body mass index (BMI) was calculated from the averaged weight and height measurements using the following formula:  BMI =Wt (kci) 2 (m) Ht  72  49  METHODS  D) SKINFOL D MEA SUREMENTS  To increase reliability and more accurately assess fat stores, skinf old measurements were taken at several sites. The specific locale of each skinf old thickness measurement was 7072 kept consistent to ensure dependable results. Skinf old thickness measurements were made using Lange precision skinf old thickness calipers (Cambridge Scientific Industries Inc., Cambridge MD).  All precision calipers are  designed to exert a defined and constant pressure of 10 grams per square millimetre throughout the range of measured skinfolds, and to have a standard contact surface area or “pinch’ area of 20 to 40 mm . Skinfolds were recorded to 0.5 mm on the Lange calipers, and 2 triplicate skinfold measurements were made, agreeing within 1 mm. 73 An average of the three values was used, and if one value varied by more than 10% of the others, it was eliminated ° 7 and a fourth measure taken. As previously stated, four skinf old measurements were completed for all participants: these included the triceps, abdominal, suprailiac and thigh skinf old sites. To measure the triceps skinfold, the subjects let their left arm hang loosely by their side with their elbow extended. The distance betweenthe lateral projection of the acromion process of the scapula (on the backside of the shoulder) and the inferior margin of the olecranon process of the ulna (the tip of the elbow) was measured with the elbow flexed to 90°, and the midpoint between these two locations marked. Approximately one centimetre above the midpoint mark, located on the posterior side of the arm, a vertical pinch of skin and subcutaneous fat was pulled away using the thumb and index finger directed inferiorly. It was ensured that no muscle was included in the skinf old by having subjects flex their triceps. The  50  METHODS  caliper was then applied about one centimetre below the ‘pinch”, and the measurement was read after three seconds. The calipers and the fingers were then released from the skin and the results, to the nearest 0.5 mm, were recorded immediately. 70737475 For the measurement of the abdominal skinf old thickness, the subjects were asked to relax their abdominal muscles as much as possible during the procedure, and to breathe normally. The selected site for measurement was located three centimetres laterally and one centimetre inferior to the midpoint of the umbilicus. To maintain consistency, all measures were made to the left of the umbilicus. A horizontal skinfold was then raised with the left hand and its thickness measured to the nearest 0.5 mm. Again, the calipers were applied for three seconds before the reading was made, followed by the release of the calipers and of the fingers, and the immediate recording of the values. 7075 The suprailiac skinf old involved the same methods as those employed with the triceps and abdominal skinfolds, and was measured in the midaxillary line of the left iliac crest. The subjects were asked to stand with their feet together, in an erect position, with their arms suspended by their side. The skinf old was picked up obliquely just posterior to the midaxillary line and parallel to the natural cleavage lines of the skin.  The caliper jaws were applied  inferiomedially at 450 to the horizontal, approximately one centimetre from the fingers holding the skinfold, and the thickness was recorded to the nearest 0.5 ’ 73 mm. 7 4 5 The site for measurement of the thigh skinfold was located at the midline of the anterior aspect of the left thigh, midway between the inguinal crease and proximal border of the patella. The subjects were asked to flex their hip in order to locate the inguinal crease, at the midpoint of the long axis of the thigh. The subjects were then asked to extend their  51  METHODS  knee so as to locate the distal reference point.  The thickness of the vertical fold was  measured while the subjects stood, with their weight shifted onto their right leg, so that the measurement site was relaxed with their left knee slightly flexed and their foot flat on the floor. If maintaining balance was a problem, the subjects were allowed to hold the counter top. The calipers were applied about one centimetre distal to the fingers holding the fold and 75 the thickness was measured and recorded to the nearest 0.5 mm. For exact techniques used in skinf old thickness measurements, please consult a copy of Lohman et al.’s 75 instructions provided in Appendix C. Calculation of body fat from skinf old measurements involves the calculation of body density, using an appropriate population specific regression equation, then a calculation of percentage body fat from the body density result, using an empirical equation, and finally a Ward showed that multiple correlations calculation of total body fat. Jackson, Pollock and 76 for body density equations for the sum of four skinfolds was 0.849 with a standard error of 3.8% body fat. The equations are as follows:  BD  =  1 .0960950-0.0006952(X ) 4 -0.00007 14(X ) 2 ) + 0.0000011 (X 2  76  ); X 3 =the sum of triceps, abdomen, suprailium and 2 where BD=body density (kg/rn =age. 4 X thigh skinf old measurements in millimetres; and, 76  Percent Body Fat (%BF)=[(4.95/BD)-4.50]100  Total Body Fat (kg)=Body Weight (kg) x %BF 100  52  76  METHODS  El MID ARM FATAREA  Before the mid arm fat area could be calculated, the mid arm circumference had to be measured. To evaluate the mid arm circumference a metric flexible, non-stretch tape, of steel or fibreglass, accurate to the nearest 0.1 cm, was used. The subjects’ arm hung freely at their side, and the midpoint between the acromion process of the scapula and the olecranon process of the ulna was marked, as with the triceps skinfold measurement.  The arm  circumference at the midway mark was measured by placing the tape firmly around the arm and the mid arm circumference was then -recorded immediately.  Again, to maintain  consistency throughout the study and subjects, all measurements were performed on the 70 Details of the circumference methodology subjects’ left arm, and completed in duplicate. used are also provided in Appendix C. The mid arm circumference was then incorporated into the formula for the mid arm fat area. The formula for calculating the mid arm fat area is as follows:  Mid Arm Fat Area (MAFA) =tsfMAC 2  —  u(tsf) 2 4  70  ; MAC=mid arm circumference (mm); and, 2 where mid arm fat area is in mm tsf = triceps skinfold thickness (mm). 7073 Arm fat area calculated from this equation has been reported to agree within 10% to values measured by computerized axial tomography in adults. 73 F) MUSCLE MASS  The task of estimating muscle mass from anthropometry is to select one or more 77 found that forearm anthropometric variables that reflect total muscle mass. Martin et al. girth and mid-thigh circumference were the best predictors of muscle mass and thus were 53  METHODS  77 used here. Forearm circumference was measured at the area of maximum girth, where the humerus, radius and ulna join. A metric flexible non-stretch tape, accurate to the nearest 0.1 cm, was placed firmly around this area of the subjects’ left arm as it hung downward, but 77 The tape was placed loosely slightly away from the trunk, with the palms facing anteriorly. around the proximal part of the forearm, perpendicular to its long axis, and moved up and down until the level of the maximum circumference was located.  At this level, the  measurement was made and recorded to the nearest 0.1 cm with the tape in contact with the 75 skin but not compressing the soft tissue. The thigh circumference was also measured in a similar way.  The site for  measurement was located at the midline of the anterior aspect of the left thigh, between the 77 inguinal crease and the midpoint of the patella of the subjects’ leg.  As with the thigh  skinf old measurement, the subjects were asked to flex their hip in order to locate the inguinal crease, at the midpoint of the long axis of the thigh. The subjects were then asked to extend their knee so as to locate the distal reference point which was the midpoint of the patella in this case. The circumference was measured as the subjects stood, with their weight shifted onto their right leg, so that the measurement site was relaxed with their left knee slightly flexed and their foot flat on the floor.  If maintaining balance was a problem, the subjects  were allowed to hold the counter top. 75 Both the forearm girth and the thigh circumference measurements were evaluated in duplicate and the average of the two values used. If there was a large discrepancy between 75 results, a third was assessed and the inconsistent result ignored.  54  METHODS  For dimensional consistency, the variables used in the regression analysis were the square of each selected circumference, multiplied by the stature. The equation is as follows:  +0.1 1 1FG 2 MM=Ht(O.0739CTG )-2930 2 where MM = total skeletal muscle mass (g); Ht = height (cm); CTG = thigh circumference (cm) corrected for the front thigh skinfold thickness (see equation below); and, FG=uncorrected forearm girth circumference (cm). CTG  =  Thigh Circumference-u(Thigh Skinf old/i 0)  “  It is important to note that inter-individual variability in measurement error was eliminated by having the candidate perform all anthropometric measurements, ensuring more consistent and reliable results. See Appendix C for descriptions and photographs of the proper anthropometric 75 technique to be used as described by Lohman et al.  2. BASAL BODY TEMPERA TURE  Basal body temperature was taken orally, using a Becton-Dickinson Model 4051 digital thermometer (Becton-Dickinson Canada Inc., Missassauga ON), immediately after awakening and before rising. It was recommended that subjects take their temperatures between 6:00 a.m. and 8:00 a.m., however because many of them were students, taking temperatures between a set two hour span in the morning was permitted. This provided a temperature curve with a minimum of accidental fluctuations because the physiologic diurnal fluctuation of body temperature describes a sinuous curve with a peak from 3:00 to 7:00 p.m. and a nadir between 3:00 and 6:00 a.m. 78 Rectal and oral temperature data have been shown to provide analogous results when using mean temperature methods, and when temperature is 55  METHODS  taken before eating, drinking or exercising. It was thus decided that temperatures measured 8 invasive. 7 ’ orally would be least 25 A low-reading Becton-Dickinson electronic digital Celsius thermometer was used and temperatures were recorded (as displayed) to two decimal places.  The subjects received  instructions on how to read and record basal body temperatures, which were entered on a form that listed the menstrual cycle day, the calendar day, and included a section for 8 See Appendix D for a copy illness. 7 ’ comments regarding late-morning rising, insomnia, or 25 of the basal body temperature record form. Subjects began recording on the first day of their menstrual cycle, i.e., the first day of menses, and transcribed the information consecutively for the following six cycles. A copy of instructions provided to subjects on temperature measurements is available in Appendix E. A quantitative least mean square method of evaluation of the daily temperature records, previously validated against the luteinizing hormone peak, was used to categorize the menstrual cycles as ovulatory or anovulatory, and to determine the luteal phase length of 25 ovulatory cycles.  The least mean square method for basal body temperature analysis  examines the temperature data for any menstrual cycle and ascertains whether or not a division of the cycle into postmenstrual and premenstrual phases is statistically possible. An optimal splitting point, if one exists, is selected by least squares criterion, consistent with the two portions of the temperature data. This maximizes the mean difference in temperatures between the phases while minimizing the variations within them. A biphasic or ovulatory cycle is indicated by a statistically significant difference in temperature.  The least mean  square method also calculates the mean and standard deviation of the follicular and luteal  56  METHODS  25 phase temperatures and computes the luteal phase length. When analysing cycle information, temperatures that varied widely from preceding and proceeding values, and by more than 0.3° were eliminated from the analysis. Furthermore, when the subject indicated that there were some unforeseeable circumstances that may have affected their temperature reading, such as a late rising or performing the reading after rising or showering, and the value appeared to be affected, it was also eliminated from the analysis. Very few entire cycles were unanalysable (n=4); however, on certain occasions subjects were ill around the time of expected ovulation resulting in increased temperatures, often greater than 38°, that would affect the program’s interpretation of the temperatures. Also, if more than 50% of the temperatures were missing, particularly around critical times, the cycle was not analysed. The luteal phase index was also calculated, the proportion of total menstrual cycle 58 found it to be the spent in luteal phase (luteal phase cycle/cycle length), because Prior et al. strongest predictor of one year rate of change in trabecular spine density. The most common problems associated with the basal temperature method of assessing menstrual cycle function were resolved here as the women were instructed in accurate reading of the thermometer, noted influencing factors such as late-morning rising, insomnia, or illness and all data were plotted by computer, thus eliminating errors in subject 25 To further ensure consistency, all cycle or researcher plotting of temperatures on a graph. evaluations were entered into the computer by the candidate and were reviewed by a second individual for possible errors in temperature entries.  57  METHODS  3. DIETARY INTAKE  To estimate subjects’ food intake, the respondents were asked to record, at the time of consumption, all foods and beverages (including snacks) eaten, as well as any supplement use, type and amount, on a dietary intake form, for the specified time period. The act of recording food intake might in itself influence the actual intake because a person might avoid inconvenient or complex entries; however, subjects were encouraged to maintain their normal 55 eating patterns as much as possible.  Each subject supplied three 3-day dietary intakes,  contributing nine days of dietary records in total. Subjects were asked to avoid weekend days whenever possible. See Appendix F for a copy of the dietary intake form. Food intake studies have indicated that women may consume extra energy in the luteal ° Energy intake has been 798 phase of the menstrual cycle as compared to the follicular phase. shown to be the lowest over the period coinciding with the midcycle surge in blood estrogen that occurs at normal ovulation. This is an indication that menstrual cycle phase is important ° Thus, completion of the 8 when considering food intake in women of reproductive age. dietary records was staggered so as to obtain one three-day record during the follicular phase of the menstrual cycle, another around the time of expected ovulation (identified by each subject by changes in cervical mucous consistency and amount), and the third dietary record was transcribed during the premenstrual phase of the cycle. Also, to eliminate the possible effects of seasonal variations on dietary intake, subjects were randomly assigned to start their food records at any one of the three stages. A copy of dietary record instructions provided to subjects is available in Appendix G. The subjects were required to give detailed descriptions of all foods and beverages  58  METHODS  7381 Use of weights or standard household measuring cups consumed, including brand names. and spoons was encouraged wherever possible, supplemented by measurements with a ruler 73 for such foods as meat and cake, and Counts for other foods like eggs and slices of bread. Wherever necessary the investigator converted portions into gram weight or other portion 76 These portion size measures were then analyzed for energy and nutrient values available. composition by the investigator, using the computer program Food Processor II (enhanced, version 3.14, ESHA Research, Salem OR). The database used was the Canadian Nutrient File given that the study took place in Canada, and thus the food compositions would more closely 82 When a particular food was not available in the Canadian represent the subjects’ intake. database, values from the ESHA database were used instead. However, the Canadian Nutrient File has been shown to be incomplete for certain nutrients, whereas the American (ESHA) database is considerably more extensive and contains 42 For example, approximately 60% of the imputed values and therefore is virtually complete. foods in the Canadian database were shown to have missing values for dietary fibre, and 10-  42 Because these nutrients play an important role in the distinction 15% for zinc (P<O.OO1). between the vegetarian and nonvegetarian diets, more accurate values were obtained by re entering daily intakes solely in the ESHA database. These results were used for reporting fibre and zinc intakes. The average values derived from the nine days of records from each subject were used in comparative analysis. With the exception of carotene, retinol, vitamin E and selenium, all of the dietary components (macronutrients and micronutrients) calculated for the foods consumed in this study were used. These included calories, protein, carbohydrate, dietary  59  METHODS  fibre, total fat, saturated fat, monounsaturated fat, polyunsaturated fat, cholesterol, total , vitamin , 6 12 folacin, pantothenic B vitamin A, thiamin (B ), vitamin B 3 ), niacin (B 2 ), riboflavin (B 1 acid, vitamin C, calcium, copper, iron, magnesium, phosphorus, potassium, sodium and zinc. Percentages of calories from protein, carbohydrate and fat were also used. Carotene, retinol, vitamin E and selenium values were omitted either because data were frequently missing from the tables, or because they were not felt to be pertinent to the 42 study. Unfortunately, errors may have arisen as a result of the inability of the respondents to adequately quantify portion sizes consumed, however all dietary intakes were analyzed by the candidate, decreasing the amount of error by eliminating inter-individual variability in translation of intakes. All dietary records were reviewed by a second individual to ensure no 73 errors had been made in recording.  4. EA TING RESTRAINT A SSESSMENT , and each 3 Every subject completed the 58-item Three-Factor Eating Questionnaire score was tabulated and assessed to determine if that subject displayed characteristics of eating restraint or not, and if so, at what level. See Appendix H for a copy of the Three Factor Eating Questionnaire with answers. One point was given for each item in Part I and for each item in Part II. The direction of the questions in Part II was determined by splitting the responses at the middle: if the item is labelled  +0,  items with a  those responses above the middle are given a zero; and, vice versa for those  .,n3,83  60  METHODS  The Three-Factor Eating Questionnaire is made up of three independent components. Factor #1 is “cognitive restraint”, factor #2 is “tendency towards disinhibition” or “emotional 83 Scores on the restraint subscale, factor #1, lability”, and factor #3 is “perceived hunger”. which may range from 0-21, were used in the analysis. The values for disinhibition and hunger, as well as totals on the questionnaire, however, were also used for statistical purposes.  5. DUAL ENERGY X-RAYABSORPTIOMETRY  ) evaluated using 2 Participants had their lumbar spine 4 -L bone mineral density (g/cm 1 (L ) dual energy x-ray absorptiometry (DPX, Lunar Corp., Madison WI).  The subjects were  positioned with appropriate support blocks according to the manufacturer’s recommendations for lumbar spine measurements. The starting point was the palpated apex of the processus 84 A self-contained x xiphoidus, and the end point was located at the level of the iliac crest. ray source was mounted beneath the subjects providing alternating pulses at 70 and 1 40 kVp. Because of the high resolution images offered by this method, the clear visualization of the intervertebral spaces ensured that the exact region of interest was measured. An examination time of approximately 5-6 minutes was required at the site and it involved only very low levels 85 of radiation exposure (<3 mrem/scan). The densitometer was calibrated according to the manufacturer’s recommendations and quality control procedures in place at the Department of Nuclear Medicine at Vancouver 84 All measurements were made during the follicular phase of the General Hospital were used. cycle to ensure no measurements were made on pregnant women, and all evaluations were  61  METHODS  performed by the same technician.  6. URINE ANAL YSIS All participants were asked to provide a two-hour fasting urine collection. The subjects were requested to abstain from eating food and drinking anything other than water after dinner, on the evening prior to the urine sample collection to ensure at least 1 2 hours of 86 In the morning, the subjects emptied their bladder and flushed fasting preceded analysis. away the urine. At this point, the test began. For the proceeding two hours the participants were requested to drink several glasses of water and collect all of their urine in a clean glass container. The test was completed after a final bladder emptying, exactly two hours after the initial urination. See Appendix I for a copy of the urine sample instructions given to subjects. 87 found that urine volume was significantly increased during the Fong and Kretsch luteal phase relative to the follicular and periovulatory phases, and remained significantly different even after statistical adjustment for between phase differences in food water and drinking water intakes. The participants were encouraged to provide their urine sample during the follicular phase of their menstrual cycle to maintain consistency.  This was easily  accomplished because subjects brought their urine sample in when they went for their bone mineral density measurement, which was always done in the follicular phase of the menstrual cycle to avoid any possible damage to a growing fetus. The urine samples were brought to the out-patient chemistry laboratory at the Vancouver General Hospital and were assessed for cortisol, creatinine and calcium values according to standard procedures. Cortisol was determined by radioimmunoassay method.  62  METHODS  Creatinine was evaluated by a Jaffé-type colourimetric assay. And, calcium was measured 88 through atomic absorption.  If the sample could not be brought to the laboratory  89 indicated in his immediately, subjects were instructed to store it in the refrigerator. Kathol analysis that simple factors such as freeze/thaw, shaking or temperature had no significant effect on levels of cortisol or creatinine found in urine samples. Cortisol/creatinine and calcium/creatinine ratios were calculated to adjust the values for lean body mass.  7. DEMOGRAPHICS  Demographic information, such as age, age at menarche, occupation, exercise amount and type, menstrual history and family history of osteoporosis, was collected during subject interviews, and recorded immediately. These were variables that could possibly play a role in the development of some etiological factors and would thus be important when analyzing ° found bone density similarities between healthy mothers 9 data. For example, Lutz and Tesar and their daughters in areas of the skeleton most susceptible to osteoporosis; consequently, it could be important to know the family history of osteoporosis.  Also, as discussed  previously, age at the onset of menarche has been linearly related to the incidence of 19 See Appendix J for a copy of the interview form. menstrual dysfunction. Vegetarian subjects also provided specific information regarding their diet, the number of years they had been a vegetarian, why they became a vegetarian, and whether any family members were also vegetarians. See Appendix K for a copy of the vegetarian form.  63  ANAL YSIS OF THE DA TA  N.B. Please see Appendix L for the basis of choosing methodologies discussed.  STATISTICAL ANALYSIS OF THE DATA  Data entry was conducted using programs available through the Statistical Package for the Social Sciences (SPSS) Data Entry 1191 and the statistical analysis was achieved using 92 All results were SPSS Personal Computer (SPSS/PC+) version 4.0 (SPSS Inc., Chicago). entered into a data file, which was completely verified against the original analyses and data errors were corrected prior to statistical analyses. For most comparisons of vegetarians and nonvegetarians, a student’s unpaired t-test was used. 93 Specifically, comparisons were made in demographic and anthropometric data, levels of dietary restraint, urine and bone values, menstrual function, and mean nutrient intakes. An ANOVA analysis, with weight as a covariate, was used to control for weight in the bone values between vegetarians and nonvegetarians. These exact comparisons, using the same statistical tests, were used when comparing the different types of vegetarians, i.e. vegan, lacto-ovo and lacto-vegetarians. Furthermore, the student’s unpaired t-test was used when comparing demographic and anthropometric data, levels of dietary restraint, urine and bone values when subjects were split into menstrual cycle groups.  Again these analyses were completed when participant  were categorized according to their levels of dietary restraint. Correlation analyses were used when looking at the relationship of urine values to  64  A NAL YSIS OF THE DA TA  menstrual function; dietary nutrients to menstrual data; dietary nutrients to bone data; and urine values to dietary restraint scores. Stepwise multiple regression identified those variables most significantly linked to changes in the menstrual cycle, as well as the factors related to bone. Comparisons were made at a significance level of P<O.05. Overall, testing of the hypotheses was completed using:  Hypothesis 1: There would be no difference in the prevalence of subclinical menstrual dysfunction between vegetarian and nonvegetarian women. (Student’s unpaired t-test.)  Hypothesis 2: There would be no difference in the prevalence of subclinical menstrual disorders between restrained and unrestrained eaters. (Student’s unpaired t-test.)  Hypothesis 3:  There would be no difference in spinal bone mineral densities between women with normal ovulatory and normal luteal phase length menstrual cycles and those with irregular cycles. (Student’s unpaired t-test and ANOVA with weight as a covariate.)  Hypothesis 4: There would be no difference in spinal bone mineral densities between vegetarian and  65  ANALYSIS OF THEDATA  nonvegetarian premenopausal women. (Student’s unpaired t-test and ANOVA with weight as a covariate.)  Hypothesis 5:  There would be no association between nutrient intakes and spinal bone mineral density results. (Correlation and stepwise multiple regression analyses.)  THE ROLE OF THE CANDIDATE  The candidate was responsible for all measurements described in the experimental design and methods, as follows: 1. She performed all anthropometric evaluations. 2. She provided instructions for the subjects on how to read and record their basal body temperature. 3. She provided training for the subjects on how to record daily dietary intakes. 4. She obtained demographic information on all subjects. 5. She developed the forms for basal body temperature and daily dietary intake records, and produced questionnaires examining the possibility of eating restraint. 6. She organized the assessment of bone mineral density of the subjects through the Vancouver General Hospital (VGH) by dual energy x-ray absorptiometry. 7. She supplied instructions for the two-hour fasting urine collection and arranged its analysis 66  ROLE OF THE CANDIDA TE  at VGH. 8. She analyzed all data received through the study.  In having the candidate responsible for almost all dimensions of the study, interexaminer errors were eliminated.  lntra-examiner measurement errors were small due to  repeated measures, equipment calibration and consistent site choice in anthropometric 3 The interviewer was trained in anthropometric measurements, in nutrient evaluations. 7 ’ 70 73 record analysis and subject training, which decreased bias and random errors.  Also,  inconsistencies were eliminated because all the subjects received the same instructions on how to evaluate and record their basal body temperature, as well as how to record their daily 94 dietary intake, and complete their two-hour fasting urine collection.  SIGNIFICANCE OF THE STUDY  With the current trends toward adopting a healthier lifestyle, there has been an 2 increase in the number of individuals adopting vegetarian practices.  Some studies have  indicated that there is an increased incidence of menstrual irregularities in vegetarian 78 Also, because of the social demands placed on physical appearance, particularly women. that of women, there are large numbers of women who display characteristics of restrained 4 This in turn has been associated with menstrual disturbances. ’ 3 eating. The literature has indicated that menstrual disturbances may play a pivotal role in the  67  SIGNIFICANCE  , a debilitating disease for which there is no known 10 premature development of osteoporosis cure. Up until now, not much emphasis has been placed on the importance of menstrual dysfunction in women, believing the consequences to be both benign and reversible. Recent evidence seems to disagree with this theory, and many women may be at risk of developing related complications later in life, and unaware of this risk. This study ascertained if there was a link between dietary practices and menstrual irregularities, and set up precedence for preventive education in these populations. The specific focus of this study was on the role that dietary habit (vegetarian or nonvegetarian) played in the development of menstrual irregularities. Also examined were associations between particular nutrients and menstrual function. This provided information on important nutritional implications for vegetarianism, dietary restraint and components of the diet on menstrual and bone status.  HUMAN STUDIES  Permission for this study was received from the University of British Columbia’s Clinical Screening Committee for Research and Other Studies Involving Human Subjects. Appendix M for a copy of the acceptance form.  68  See  RESUL TS  Chapter IV  RESULTS  1. SUBJECTS  By the conclusion of the study in August 1 993, a total of 45 women remained in the study and had provided sufficient data for analyses, representing an attrition of 27 percent. Twenty-two of the remaining subjects were nonvegetarian participants and 23 were vegetarian women. The ethnic profile of the study population was: 35 Caucasian women (78%); four of Asian descent consisting of Japanese, Chinese and Koreans (9%); three women classified themselves as Jewish (7%); and, the remaining two were of mixed heritage. Fifty-six percent (n=25) of the participants were university students and the rest of the subjects’ occupations ranged from waitress to chartered accountant. Seven percent (n =3) of the subjects were not employed. Eighty-seven percent of the participants (n  =  39) engaged  in regular activity, whereas only 1 3% (n = 6) said they never exercised. The type of exercise the women participated in consisted of walking (n=9), aerobics (n=12), weight training (n  =  14), running (n  =  1 3), biking (n  =  14), swimming (n  =  1 2), and some other form of activity  such as tennis or skiing (n=3). Twenty-eight women (72%) fit into more than one of those exercise categories.  Only seven of the 45 women indicated that there was a history  osteoporosis present in their family. The above stated information has been summarized in Appendix N.  69  RESUL TS  The 1 7 who did not complete the study were similar to the remaining participants in anthropometric and demographic characteristics. The women who did not complete the study were however significantly younger (24.3±4.1 years vs. 27.2±5.1 years; P<O.05) and exercised significantly more (4.8 ±2.1 hours/week vs. 3.5 ±2.1 hours/week; P<O.05) than the remaining study population. Please see Appendix 0 for specific details pertaining to this group. Of the 1 7, eight subjects were nonvegetarian and nine were vegetarians. Reasons for withdrawing from the study included unexpected pregnancies (n=2), an inability to commit to the study (n=6), a decision to begin using oral contraceptives (n=2), and one individual required hormone therapy. The remaining drop-out subjects (n  =  6) could not be contacted  through follow-up. Only one subject had provided basal body temperature data, and none had furnished any dietary intake material prior to withdrawing from the study. Of the 23 remaining vegetarian participants, eight were vegan vegetarians, 11 were lacto-ovo vegetarians, and four were lacto-vegetarians. Within these categories, there were no significant differences in the number of years each individual had been practising her vegetarian diet: vegans averaged 3.12±2.32 years, lacto-ovo vegetarians’ mean was 5.82±3.49 years, and 4.25  ±  1.50 years was the mean for the lacto-vegetarian group. It is  interesting to note that although the three subgroups of vegetarians did not differ significantly with respect to the number of years they had been practising vegetarianism prior to the study, the mean number of years decreased with increasing restrictions within the diet. Furthermore, the difference between the vegan and lacto-ovo subgroups was quite large (P=O.073). During their initial visit to the University, all of the vegetarian subjects were asked to complete the sentence “I became a vegetarian because...’. Their responses to this question  70  RESUL TS  as well as other relevant information regarding the vegetarian women, are summarized in Appendix P.  Overall, 70% of the subjects acknowledged that health reasons were the  motivating factor behind their choice of an alternative diet, although not exclusively. The two other most common responses were that participants had adopted vegetarian practices for moral (57%) and/or environmental reasons (30%). Most of the vegetarians (57%) stated a combination of factors as their basis for becoming vegetarian. The mean age of the 45 women who completed the study was 27.2±5.1 years (range, 20 to 38 years) at the beginning of the investigation.  Table 1 displays subject  characteristics (mean ±SD) of the vegetarian and nonvegetarian women upon entry to the study. The subjects’ weights were within the ideal range since they had to have a Body Mass Index (BMI) within the healthy limits of 1 8 to 25 kg/rn 2 to participate in the study.  The  demographic data showed that the premenopausal vegetarian and nonvegetarian women were well matched: the two study groups did not differ with respect to age, age at menarche, the amount of exercise performed per week, and the number of years their exercise regimen had been maintained.  71  RESUL TS  Table 1. Mean (± SD) characteristics of 45 premenopausal women, according to dietary group. Variable’  All  Nonvegetarians  Vegetarians  Number of women  45  22  23  Age (years)  27.2±5.1  27.9±5.9  26.6±4.3  0.400  Age at Menarche (yrs.)  1 3.3 ± 1.5  1 3.5 ± 1.6  1 3.2 ± 1.5  0.538  Gynecologic Age (yrs.)  13.9±4.9  14.4±5.8  13.4±3.9  0.489  Exercise (hrs/wk)  3.6±2.1  3.1 ±1.8  4.0±2.3  0.127  Exercise Regimen (yrs.)  4.5 ±5.1  4.0 ±4.0  5.0±6.0  0.496  Weight (kg)  60.3±7.9  61.9±7.8  58.7±7.8  0.176  Height (cm)  166.0±8.1  165.0±7.0  167.0±9.1  0.415  Body Mass Index (kg/rn ) 2  21.8±2.2  22.7± 1.9  21.1 ±2.3  0.013”  Triceps Skinfold (mm)  20.7 ±6.8  22.2±7.0  19.2 ±6.3  0.144  Abdomen Skinfold (mm)  23.5 ±7.6  25.7 ±6.9  21.4±7.7  0.051  Suprailiac Skinfold (mm)  13.0±6.3  15.0±6.9  1 1,1 ±5.0  0.037”  Thigh Skinfold (mm)  33.6±8.0  35.7 ±7.7  31.6 ±7.9  0.082  Sum of Skinfolds (mm)  90.7 ±24.9  98.6 ±24.2  83.2 ±23.7  0.037”  Body Density (kg/rn ) 3  1.04±0.01  1.04±0.01  1.05±0.01  0.033”  Body Fat 1%)  25.7 ±5.5  27.4 ±5.1  24.0±5.5  0.032”  Total Body Fat (kg)  15.7±4.8  17.2±4.7  14.3±4.5  0.040”  Mid Arm Circumference (cm)  28.1 ±2.5  28.4±2.6  27.8±2.5  0.397  Mid Arm Fat Area (mm ) 2  2594±967  2790±1013  2407±903  0.188  Forearm Girth (cm)  24.7± 1.3  24.9± 1.2  24.5± 1.1  0.313  Thigh Circumference (cm)  51.3±3.6  52.7 ±3.9  50.00±2.9  0.013”  Corrected Thigh Circ. (cm)  40.7 ±2.9  41 .4±2.9  40.1 ±2.7  0.115  Total Muscle Mass (kg)  28.8±4.3  29.5±4.5  28.2±4.0  0.286  P value -  Vegetarians excluded all meat, fish and poultry from their diet, but could be vegan, lacto or lacto-ovo vegetarians. Nonvegetarians had to consume at least three servings of meat per week. Gynecologic age = age age at menarche. Significant difference at a level of P<O.05. +Comparisons were done using the student’s unpaired t-test. -  72  c)  ,1  0) 3  -I  CD Co CD -4 CD  3CD 3 0 3  ‘-4  CD CD 3  -4  CD CD 0. C. CD  -4. -4..  0.  -4  C) 0) 3  0 CO  -4  0)  -4  0) 0 CD 0  0)  0  ‘-4  CD  3  Co  2. D)CD  -4  CD  CD  0  p  A  ‘-4  CD  CD CD D  -,,  -4.  a  C) 0) D  -4.  D  CD  (I)  *  CO  ‘-4  CD  -n  0 0.  *  ‘1 0)  -4  0 0.  w  *  cJ  N)  H  H  CD 3  3  0  ci) 3  -‘  CD CD CD -4 0)  D 0  ci) a  D  ci)  -‘  ‘-4 ci)  CD CD CD  CD  -4  CD CD D  C. CD ‘-4  C’,  I,,  2 C)  m m  ‘-I  C)  m —I  0 -v 0  -I I  z  RESUL TS  The anthropometric data showed however that the vegetarian women had significantly lower means for some variables than their nonvegetarian counterparts: BMI (21 .1 ± 2.3 kg/rn 2 vs. 22.7± 1.9 kg/rn ; P<0.05), mean abdomen skinfold thickness measurement (21.4±7.7 2 mm vs. 25.7±6.9 mm; P=0.051), mean suprailiac skinfold thickness measurement (11.1±5.0 mm vs. 15.0± 6.9 mm; P<0.05), mean sum of the four skinfold thickness measurements  (83.2±23.7  mm  vs.  98.6±24.2  mm;  P<0.05),  circumference (50.0±2.9 cm vs. 52.7±3.9 cm; P<0.05).  and,  mean thigh  Furthermore, significant  differences were found for percent body fat (24.0 ± 5.5% vs. 27.4 ± 5.1%; P< 0.05) and total body fat (14.3 ±4.5 kg vs. 17.2±4.7 kg; P<0.05) evaluations, with the vegetarian women having the lower mean in these instances as well.  The most important differences in  anthropometric data are displayed in Figure 1. Overall, the subjects within the vegetarian sample population were well matched. Subject mean characteristics by vegetarian subgroup are listed in Table 2.  The only  differences in either the demographic or anthropometric data within the vegetarian study group were in the lacto-vegetarians (n=4) mean age (24.0±2.2 years vs. 28.0±3.2 years;  P<0.05) and mean gynecologic age (age-age at menarche) (15.0± 1.9 years vs. 11.5±2.9 years; P=0.03), which were found to be significantly lower than the vegan vegetarians’ (n=8). However, due to the size of the lacto-vegetarian group (n=4), the result cannot be generalized. Therefore, the lacto-ovo vegetarians and the lacto-vegetarians were combined to form one group (n  =  1 5) because of their common use of dairy products and were compared  to the vegan subgroup. The results from this analysis can be found in Table 3. There were no significant differences found in any of the demographic or anthropometric data compared  74  RESUL TS  between these two groups. The difference in years as a vegetarian approached significance (5.4±3.1 years vs. 3.1 ±2.2 years; P=O.082), with the lacto vegetarian group mean being greater than that of their vegan counterparts.  75  RESULTS  Table 2. Mean (±SD) characteristics of 23 premenopausal vegetarian women, according to dietary subgroup. Variable  Vegan Vegetarians  Lacto-ovo Vegetarians  Lacto-vegetarians  Number of women  8  1 1  4  Years asa Vegetarian  3.1 ±2.2  5.8±3.5  4.3±1.5  Age (years)  28.0±3.2  26.5±5.3  24.0±2.2”  Age atMenarche (yrs.)  13.0±1.6  13.6±1.4  12.5±1.3  Gynecologic Age (yrs.)  15.0±1.9  12.9±5.0  11.5±2.9”  Exercise (hrs/wk)  4.1 ±2.2  3.9±2.7  12.50± 1.29  Exercise Regimen (yrs.)  5.4±4.5  4.0±6.2  7.08±8.94  Weight (kg)  58.6±7.1  59.2±7.8  57.6± 10.9  Height (cm)  168.2± 10.2  167.3±8.5  163.9± 10.3  ) 2 Body Mass Index (kg/rn  20.7±1.8  21.1 ±1.9  21.1 ±1.9  Triceps Skinfold (mm)  19.6±6.6  19.1 ±5.8  18.8±8.7  Abdomen Skinfold (mm)  21.5±8.8  21.6±7.7  20.3±7.6  Suprailiac Skinfold (mm)  11.2±0.2  11.0±5.5  11.0±6.1  Thigh Skinfold (mm)  29.7 ±7.9  33.7 ±7.1  29.3± 10.2  Sum of Skinfolds (mm)  82.0±26.4  85.5±22.7  79.3±26.9  Body Density (kg/rn ) 3  1.05±0.01  1.04±0.01  1.05±0.01  Body Fat (%)  23.7 ±6.1  24.5 ±5.2  23.0±6.3  Total Body Fat (kg)  14.1 ±4.8  14.6±4.2  13.7 ±6.1  Mid Arm Circumference (cm)  27.3±2.3  28.1 ±2.3  27.8±3.8  Mid Arm Fat Area (mm ) 2  2396±886  2419±824  2398± 1374  Forearm Girth (cm)  24.2± 1.2  24.8± 1.2  24.1 ±0.7  Thigh Circumference (cm)  49.8 ±2.1  50.3 ±3.0  49.5 ±4.6  Corrected Thigh Circ. (cm)  40.4±2.4  39.7 ±2.8  40.3 ±3.8  Total Muscle Mass (kg)  28.5 ±3.6  28.2 ±4.6  27.5 ±4.2  Vegan was defined as those omitting all animal and dairy products from their diet; lacto-ovos included dairy products and eggs; whereas lactos eliminated eggs from their diet. Gynecologic age = age age at menarche. *Significant difference at a level of P<O.05 between lacto and vegan. Comparisons were made using the student’s unpaired t-test. -  76  RESUL TS  Table 3. Mean (± SD) characteristics of 23 premenopausal vegetarian women, according to dietary subgroup. VariabIe  Vegan Vegetarians  Lacto Vegetarians  Number of Women  8  15  Years as a Vegetarian  3.1 ±2.2  5.4 ±3.1  0.082  Age (years)  28.0± 3.2  25.8 ±4.7  0.253  Age at Menarche (yrs.)  13.0±1.6  13.3±1.4  0.688  Gynecologic Age (yrs.)  15.0±1.9  12.5±4.4  0.152  Exercise (hrs./wk)  4.1 ±2.2  4.0± 2.3  0.902  Exercise Regimen (yrs.)  5.4 ±4.5  4.8 ±6.8  0.830  Weight (kg)  58.6±7.1  58.8 ±8.4  0.968  Height (cm)  168.2± 10.2  166.4±8.7  0.658  Body Mass Index (kg/rn ) 2  20.7±1.8  21.2±2.5  0.635  Triceps Skinfold (mm)  19.6±6.6  19.0±6.3  0.851  Abdomen Skinfold (mm)  21.5±8.8  21.3±7.4  0.947  Suprailiac Skinfold (mm)  1 1.2 ± 4.2  1 1.0 ± 5.4  0.936  Thigh Skinfold (mm)  29.7 ±7.9  32.5 ±7.9  0.428  Sum of Skinfolds (mm)  82.0± 26.4  83.9 ±23.1  0.863  Body Density (kg/m ) 3  1.05 ±0.01  1.04 ±0.01  0.866  Body Fat (%)  23.7 ±6.1  24.1 ±5.3  0.871  Total Body Fat (kg)  14.1 ±4.8  14.4±4.6  0.882  Mid Arm Circumference (cm)  27.3 ±2.3  28.0±2.6  0.559  Mid Arm Fat Area (mm ) 2  2396 ±886  2413 ±943  0.967  Forearm girth (cm)  24.2± 1.2  24.6± 1.1  0.445  Thigh Circumference (cm)  49.8 ±2.1  50.1 ±3.3  0.801  Corrected Thigh Circ. (cm)  40.4± 2.4  39.9±3.0  0.659  Total Muscle Mass (kg)  28.5±3.6  28.0 ±4.3  0.791  P value -  Vegan was defined as those omitting all animal and dairy products from their diet; lacto include both lacto vegetarians and lacto-ovo vegetarians, thereby including dairy products and eggs overall. +Comparisons were made using the student’s unpaired t-test.  77  RESUL TS  2. MENSTRUAL CYCLE DISTURBANCES The mean menstrual cycle characteristics of the 45 women over the six month study period are presented in Table 4.  The mean menstrual cycle length was 26.7 ±4.2 days  (range, 20 to 60 days) and the mean luteal phase length was 9.5 ± 3.4 days (range, 0 to 18 days). Only 1 5 women (33%), nine vegetarians and six nonvegetarians, had normal menstrual cycles. This was defined by the presence of ovulation and luteal phases of normal length ( 10 days), consistently throughout the investigation. Twenty-seven women (1 3 vegetarians and 1 4 nonvegetarians) had one or more short luteal phase cycles (luteal phase <10 days)  -  1 2 (six vegetarians and six nonvegetarians) having only one short luteal phase, and 1 5 (seven vegetarians and eight nonvegetarians) having more than one short luteal phase.  Sixteen  women, six vegetarians and ten nonvegetarians, had at least one anovulatory cycle (besides having one or more cycles with short luteal phases). Only three of the 1 6 women who had at least one anovulatory cycle did not have one or more short luteal phases as well. The vegetarian women had menstrual cycle patterns similar to their nonvegetarian counterparts.  From Table 4 it can be seen that the former group had longer total cycle  lengths (27.4±3.9 days vs. 26.0 ±4.4 days; P=O.269) and luteal phase lengths (10.4± 2.6 days vs. 8.5 ±4.0 days; P=O.058), and a higher average luteal phase index (0.40 ±0.06 vs. 0.37±0.08; P=O.188) than the nonvegetarians, although none of the differences were significant. Luteal phase index is defined as the ratio of the luteal phase length to the total cycle length and is considered a more appropriate indicator of menstrual status then luteal phase length or cycle length alone. Overall means showed that the vegetarian women had more ovulatory cycles (5.4±1.0 cyclesvs. 4.6±1.8 cycles; P=O.072), and lessshortluteal  78  RESUL TS  phases (1 .1 ± 1 .2 vs. 1 .3 ± 1 .3; P= 0.623) than their nonvegetarian counterparts. The number of cycles analyzed within each subgroup did not differ: 5.5 ± 1.1 cycles for the nonvegetarian women and 5.7 ±0.7 cycles for the vegetarian women (P=0.385).  Table 4. Mean (± SD) menstrual cycle characteristics over the six months of study, according to dietary group. P value+  Characteristics  All  Nonvegetarians  Vegetarians  Number of Women  45  22  23  Cycle Length (days)  26.74 ±4.18  26.03 ±4.42  27.42 ±3.90  0.269  Luteal Phase Length (days)  9.49± 3.44  8.49 ±4.00  10.44±2.56  0.058  Luteal Phase Index  0.39 ±0.07  0.37 ±0.08  0.40±0.06  0.188  Number  of  -  Women  Luteal Phase Analysis: Normal Cycles  15  6  9  <0.10  Short Luteal Phases  27  14  13  <0.50  1  12  6  6  <0.90  >1  15  8  7  <0.70  16  10  6  <0.10  Long (>35)  8  5  3  <0.30  Short (<21)  2  1  1  <0.90  Anovulatory Cycles Cycle Length (days):  The normal length of the menstrual cycle is 21 to 35 days, and that of the luteal phase is 10 to 16 days. The luteal phase index is the mean ratio of the length of the luteal phase to that of the menstrual cycle. +Analysis was executed using student’s unpaired t-test. Analysis was executed using chi square.  The total number of cycles analysed was 249, of which 1 30 (52%) were from vegetarian subjects and 119 (48%) were from nonvegetarian participants. Overall, there were 24 anovulatory cycles (10% of all cycles) and 53 cycles with a short luteal phase (21% of 79  RESCJL TS  all cycles). When this was broken down into subgroups, the vegetarians experienced only six anovulatory cycles (5% of the vegetarian women’s cycles) and 25 cycles with a short luteal phase (19% of all vegetarian women’s cycles). On the other hand, the nonvegetarian women had 15% of their cycles as anovulatory (n= 18) and 24% with short luteal phases (n=28). These data are summarized in Table 5.  Table 5. Menstrual cycle characteristics over six months of study, according to dietary group. Characteristics’  All  Nonvegetarians  Vegetarians  Number of Women  45  22  23  Number of Cycles  249  11 9  1 30  Number  of  Cycles  Luteal Phase Analysis: Normal Cycles  172  73  99  Short Luteal Phases  53  28  25  Anovulatory Cycles  24  18  6  Long (>35)  12  7  5  Short (<21)  2  1  1  Cycle Length (days):  The normal length of the menstrual cycle is 21 to 35 days, and that of the luteal phase is 10 to 16 days.  Further analysis was performed on the menstrual cycle characteristics within the vegetarian group. These results are presented in Table 6. The vegan vegetarian women were found to have significantly longer cycles than both their lacto-ovo (30.0±3.0 days vs. 26.4±4.0; P<O.05) and their lacto-vegetarian counterparts (30.0±3.0 days vs. 25.1 ±2.9 days; P<O.05).  Both the vegan (11.7±1.8 days; P<O.02) and the lacto-ovo (10.7±2.1  80  RESUL TS  days; P<O.05) subgroups had significantly longer luteal phases than their lacto-vegetarian colleagues (7.4 ± 3.2 days). There were no significant differences noted in luteal phase index between vegetarian groups. Table 6. Mean ( ± SD) menstrual cycle characteristics over the six months of study, according to vegetarian subgroup. Characteristics  Vegan Vegetarians  Lacto-ovo Vegetarians  Lacto-vegetarians  Number of Women  8  11  4  Cycle Length (days)  30.0±3.0”  26.4 ±4.0  25.1 ±2.9’”  Luteal Phase Length (days)  11.7±1.8”’  10.7 ±2.1k  7.4±3.2  Luteal Phase Index  0.39 ±0.06  0.42 ±0.07  0.36 ±0.05  The normal length of the menstrual cycle is 21 to 35 days, and that of the luteal phase is 10 to 16 days. The luteal phase index is the mean ratio of the length of the luteal phase to that of the menstrual cycle. “Significant difference at a level of P<O.05 between vegan vegetarians and lacto-ovo vegetarians. “Significant difference at a level of P<O.05 between vegan vegetarians and lacto-vegetarians. +Significant difference at a level of P<O.05 between lacto-ovo vegetarians and lacto-vegetarians. Analysis was executed using student’s unpaired t-test.  As can be seen from Table 7, when the lacto-ovo vegetarians were grouped with the lacto-vegetarians (n=15) and compared to the vegan vegetarian group (n=8) there only remained a significant difference in cycle length.  The vegan vegetarians continued to  demonstrate longer mean cycles (30.0 ± 3.0 days vs. 26.0 ± 3.7 days; P< 0.02) than the new lacto vegetarian group.  81  RESUL TS  Table 7. Mean ( ± SD) menstrual cycle characteristics over the six months of study, according to vegetarian subgroup. Characteristics  Vegan Vegetarians  Lacto Vegetarians  Number of Women  8  15  Cycle Length (days)  30.0±3.0  26.0± 3.7  0.016”  Luteal Phase Length (days)  1 1.7 ± 1.8  9.8 ± 2.7  0.098  Luteal Phase Index  0.39 ±0.06  0.41 ±0.07  0.690  P vaIue -  The normal length of the menstrual cycle is 21 to 35 days, and that of the luteal phase is 10 to 16 days. The luteal phase index is the mean ratio of the length of the luteal phase to that of the menstrual cycle. “Significant difference at a level of P<O.05. + Analysis was executed using student’s unpaired t-test.  3. DIETARY INTAKE  Initially, each set of three-day dietary records was analyzed separately to investigate the possibility of intakes being significantly different from various stages of the menstrual cycle. These comparisons were completed using a paired t-test. Overall, the analysis showed that the mean intakes during the follicular phase were significantly lower that those recorded during ovulation, for the following nutrients: energy, protein, fat, saturated fat, riboflavin, vitamin B , pantothenic acid, calcium, copper, phosphorus, and zinc. It is interesting to note 6 that although all nutrients means were lowest in the follicular phase, the proportion of energy from carbohydrates was greater in the follicular phase than in the ovulatory phase of the menstrual cycle, although the difference was not significant.  When the follicular and the  premenstrual phases were compared, similar results were found, with the premenstrual nutrient intake means significantly exceeding the follicular means in the same nutrients listed as different between the follicular and ovulatory stage intakes, with the addition of  82  RESUL TS  . 6 monounsaturated fat, cholesterol, iron and potassium, and with the exception of vitamin B There were no significant differences in mean nutrient intakes from the ovulatory and premenstrual phases. The mean nutrient intakes for the three stages of the menstrual cycle that were used in the analysis can be found in Appendix Q. Also, weekend day dietary intake records were compared to weekday dietary intakes to ensure that they did not significantly differ in levels of nutrient intake. This would ensure that the menstrual cycle phase comparisons were reliable. The analysis was completed using the student’s unpaired t-test, and no significant differences were found in the levels of nutrient intakes between weekend and weekdays, with the exception of copper. Mean copper intakes were significantly higher on weekend days than on weekdays (1 .78±1.06 mg vs. 1.57±0.77 mg; P<O.05).  However, since copper was not one of the key nutrients of  concern in this study, no further analyses were conducted for this. The mean comparisons for weekday versus weekend day intakes can be found in Appendix R. Table 8 shows the women’s intakes for selected nutrients for the total nine days of recorded intake as total means and separated by diet group.  As would be expected, the  nonvegetarians ate diets that were higher in animal foods and different in other food choices from those of the vegetarian women. There were no significant differences in the mean caloric intake, in total fat, in the saturated fat, monounsaturated fat or polyunsaturated fat subgroup means measured, or in total carbohydrate evaluation between the vegetarian and nonvegetarian women. However, the vegetarian subjects did have a significantly lower intake of protein (55.3 ± 11 .3 g vs. 77.1 ±19.7 g; P<O.000), and cholesterol (132±82mg vs. 231 ±19 mg; P<O.000), and a  83  RESUL TS  significantly greater intake of fibre (28.3±9.8 g vs. 22.4±7.2 g; P<0.05) than their nonvegetarian counterparts.  Furthermore, when the macronutrients were compared as a  percentage of the study group’s diet rather than as a true mean, it was found that the vegetarian group’s percentage of calories from protein was lower (11.1 ±2.0% vs. 14.8 ± 2.3%; P< 0.000) than their nonvegetarian counterparts. Moreover, the percentage of calories from carbohydrate was higher (57.6±5.2% vs. 54.0±4.6%; P<0.05) for the vegetarian women than the nonvegetarian group, with their intake of fibre differing by approximately 21 % (Figure 2). Consumption of relatively greater amounts of carbohydrates by the vegetarians is probably due to their high grain and legume intakes. The results of the unpaired t-test comparisons of the micronutrient intakes in the two groups showed that levels of riboflavin (1.41 ±0.29 mg vs. 1.72 ± .047 mg; P< 0.05), niacin 12 (1.15 ±0.78 pg vs. 3.79±1 .6Opg; (13.4±3.4mg vs. 18.7±5.1 mg; P<0.000), vitamin B P<0.000), sodium (2210±774mg vs. 2789±757 mg; P<0.05), and zinc (8.32±2.14mg vs. 11 .10 ± 3.91 mg; P= 0.005), were lower in the vegetarian group than in the nonvegetarian subjects. However, the vegetarian women were found to have significantly higher intakes of folacin (347±ll6pg vs. 269±77pg; P<0.05), vitamin C (156±62mg vs. 116±39 mg;  P<O.05), and copper (1.83±0.64mg vs. 1.42±0.40 mg; P<0.05). Although vitamin and mineral supplement use was recorded by subjects, it was not incorporated into the dietary analysis. No difference was found when the frequency of its use was compared between the two groups.  Overall, roughly 61% of the vegetarians used a  vitamin or mineral supplement at least once over the nine days whereas only 36% of the nonvegetarians used them (2.48 ±2.59 times vs. 1.64±2.52 times; P=0.275).  84  RESULTS  Table 8. Mean (± SD) total nutrient intakes for nine days, according to dietary group. Daily Dietary Intakes  All  Nonvegetanans  Vegetarians  Number of Women  45  22  23  Calories (kcal)  2036±496  2086±528  1989±469  0.516  Protein (g)  65.9 ± 19.3  77.1 ± 1 9.7  55.3 ± 1 1.3  0.000”  Carbohydrates (g)  288.2±71.8  284.0 ± 77.4  292.2 ± 67.6  0.706  Fibre (g)  25.4±9.0  22.4±7.2  28.3±9.8  0.026  Total Fat (g)  73.3±23.4  75.0±23.2  71.6±24.1  0.632  Saturated Fat (g)  23.0 ± 10.4  25.2 ± 11.3  20.8 ± 9.1  0.154  Monounsaturated Fat (g)  27.9±9.3  28.4±9.0  27.4±9.9  0.724  Polyunsaturated Fat (g)  15.6±5.2  14.6±3.6  16.5±6.3  0.230  Cholesterol (mg)  180±97  231 ±87  132±82  0.000”  Vitamin A (RE)  1701±1160  1638±1099  1763±1235  0.725  Thiamin (mgI  1.52±0.48  1.55±0.41  1.49±0.55  0.686  Riboflavin (mgI  1.56±0.41  1.72±0.47  1.41 ±0.29  0.012  Niacin 1mg)  16.0±5.0  18.7±5.1  13.4±3.3  0.000”  Vitamin B 6 (mg)  1.59±0.49  1.61 ±0.45  1.56±0.53  0.739  (jig)  2.44± 1.82  3.79± 1.60  1.15±0.78  0.000”  Folacin (jig)  309 ± 105  269 ±77  347 ± 116  0.01 2  Pantothenic Acid (mgI  4.67 ± 1.38  4.95 ± 1.47  4.40 ± 1 .25  0.176  Vitamin C(mgl  136±56  116±39  156±62  0.013  Calcium (mg)  859 ±369  950 ±437  771 ±270  0.104  Copper (mg)  1.63±0.57  1.42±0.40  1.83±0.64  0.013  Iron (mgI  15.2±4.8  15.3±4.9  15.1 ±4.8  0.892  Magnesium (mg)  331 ±101  303±91  358±106  0.067  Phosphorus (mgI  1294±403  1409 ±470  1186±295  0.061  Potassium (mgI  3086±835  3042±833  3128±854  0.735  Sodium(mg)  2493±811  2789±757  2210±771  0.015  Zinc 1mg)  9.68 ±3.40  11.10±3.91  8.32±2.14  0.005”  Protein(%)  12.9±2.8  14.8±2.3  11.1±2.0  0.000”  Carbohydrate (%)  55.8±5.2  54.0±4.6  57.6±5.2  0.017  Fat (%)  30.8±5.1  30.9±5.1  30.8±5.3  0.909  Vitamin  812  ‘Significant difference at a level of P<O.05.  P value -  “Significant difference at a level of P<O.O1. +Student’s unpaired t-test.  85  __  __  __  __  ___ __  ___  RESUL TS  DIETARY INTAKES (between vegetarian and nonvegetarian women>  100  60 77.1 60  40  I  20  148i  —-  0  1  *protejn (g)  *%  *Fjbre (g)  Protein  *%  CHO  nonvEg. *significantly different at P<O.05  Figure 2. Significantly different mean macronutrient intakes for nine days in vegetarian and nonvegetarian women.  86  RESUL TS  Of the 1 3 nutrients, vitamins and minerals analyzed for which Recommended Nutrient Intakes (RNI) were available, both groups exceeded the RNI for 11 of the nutrients: vitamin A, thiamin, riboflavin, vitamin B , vitamin , 6 12 folacin, vitamin C, calcium, magnesium, iron B and phosphorus. Niacin and zinc were the two exceptions where the nonvegetarians both met and exceeded the RNI f or these two nutrients, but the vegetarians were slightly below the recommended amount. The vegetarian group reported an intake of 89% of the RNI for niacin and 92% of the RNI for zinc. Comparisons of nutrient intakes were performed between different types of vegetarians within the vegetarian study group using a student’s unpaired t-test.  These results are  displayed in Table 9. When contrasting the vegan vegetarian women (n=8) with the lacto ovo vegetarian subjects (n  =  11) several significant differences were found, as would be  expected from the different food compositions in each type of vegetarian diet. These included mean total fibre intakes (35.0  ±  9.9 g vs. 23.7  ±  7.8 g; P< 0.02), vitamin A (2687 ± 1682 RE  vs. 1298 ±560 RE; P=0.02), thiamin (1.83±0.67mg vs. 1.30±0.42 mg; P<0.05), niacin (15.9±2.7mg vs. 12.2±3.2 mg; P<0.02), vitamin B 6 (1.88±0.63mg vs. 1.37±0.38 mg; P<0.05), pantothenic  acid (5.32±1.54 mg vs.  3.97±0.74 mg; P=0.02), copper  (2.17±0.51 mg vs. 1.55±0.46 mg; P<0.02), and the percentage of energy derived from carbohydrates (60.7±5.1% vs. 55.7 ±4.0%; P<0.05), where the vegans consumed more on average than the lacto-ovo vegetarians. However, analysis also showed that the vegan vegetarians consumed significantly less saturated fat (1 5.1  ±  7.5 g vs. 26.0 ± 8.8 g; P< 0.02),  cholesterol (94±83 mg vs. 180±67 mg; P<0.05), vitamin B 12 (0.51 ±0.41 pg vs. 1.78±0.55pg; P<0.000), and calcium (578±184mg vs. 958±225 mg; P=0.001) than  87  RESUL TS  their lacto-ovo counterparts. When comparing the vegan vegetarian subgroup to the lacto vegetarians (n =4), the former group was found to ingest significantly greater mean amounts of niacin (15.9±2.7 mg vs. 11.5±2.2 mg; P<O.02), but no other significant differences were found. Finally, the lacto-ovo vegetarian women were compared to their lacto-vegetarian counterparts and these results indicated that the lacto-ovo vegetarian subjects consumed significantly greater mean amounts of cholesterol (180±67 mg vs. 74±35 mg; P<O.02), riboflavin (1 .56±0.28 mg vs. 1.20±0.27 mg; P<O.05), vitamin B 12 (1 .78±0.55 pg vs. 0.69 ±0.5Opg; P<O.005), and calcium (958 ± 225mg vs. 646±196mg; P<O.05) than the lacto-vegetarians. Again, because of the small sample size of lacto-vegetarians (n=4), the lacto-ovo vegetarians were grouped with the lacto-vegetarians and this new arrangement was compared to the vegan vegetarians. The same analysis was performed and the results are presented in Table 10. These results showed that the vegan vegetarians consumed significantly greater amounts of fibre (35.0 ±9.9 g vs. 24.7±7.9 g; P<O.02), vitamin A (2687± 1682 RE vs. 1268 ±479 RE; P=O.005), thiamin (1.83±0.67 mg vs. 1.31 ±0.40 mg; P<O.05), niacin (15.9±2.7mg vs. 12.0±2.9 mg;P=O.005), vitamin 6 (1.88±0.63mg vs. 1.40±0.40mg; P<O.05), folacin (416 ± 1 42pg vs. 310 ±83pg; P’<O.05), pantothenic acid (5.32±1.54mg vs. 3.90±0.72 mg; P<O.O1), and the percentage of the diet’s energy as carbohydrate (60.7 ±5.1% vs. 55.9±4.5%; P<O.05) than the new lacto vegetarian group. However, the latter had significantly higher intakes of saturated fat (23.8 ± 8.6 g vs. 1 5.1 ± 7.5 g; P< 0.05), vitamin B 12 (1 .49±0.72 pg vs. 0.51 ±0.41 pg; P<0.005), and calcium (875±255mg vs. 578± 184 mg; P<0.01) than the vegan vegetarian subgroup.  88  RESUL TS  Table 9. Mean ( ± SD) total nutrient intakes for nine days, according to vegetarian subgroup. Daily Dietary Intakes  Vegan Vegetarians  Lacto-ovo Vegetarians  Lacto-vegetarians  Number of Women  8  11  4  Calories (kcal)  1923±290  2125 ± 584  1748±354  Protein (g)  51.9 ± 12.1  59.5±9.4  50.7 ± 13.0  Carbohydrates (gi  299.5 ± 38.8  302.0±87.9  250.7 ±37.4  Fibre (g)  35.0±9.9’  23.7 ±7.8  27.3±8.8  Total Fat (g)  64.3 ± 22.4  79.0 ± 26.2  65.9 ±20.0  Saturated Fat (gi  15.1 ±7.5’  26.0±8.8  17.9±4.9  Monounsaturated Fat (g)  25.9 ± 11.0  28.9 ±9.8  26.6±9.7  Polyunsaturated Fat (g)  17.3±6.9  16.5 ± 6.4  14.8±5.9  Cholesterol (mgi  94±83’  180 ± 67”  74±35  Vitamin A (RE)  2687 ± 1682’  1299 ± 560  1184±114  Thiamin (mgi  1.83±0.67’  1.30±0.42  1.36±0.40  Riboflavin (mg)  1.32±0.22  1.56±0.28”  1.20±0.27  Niacin (mgi  15.9 ± 2.7’  12.2±3.2  11.5±2.2”  Vitamin B 6 (mgi  1.88±0.63’  1.37±0.38  1.49±0.49  Vitamin B 12 (jig)  0.51 ±0.41’  1.78±0.55”  0.69 ± 0.50  Folacin (jig)  416±142  310±85  308±88  Pantothenic Acid (mgi  5.32 ± 1.54’  3.97 ±0.74  3.72±0.76  Vitamin C (mgi  186 ±46  142 ± 65  136±78  Calcium (mgi  578 ± 184’  958 ± 226’  646±196  Copper (mgi  2.17±0.51’  0.55 ± 0.46  1.93 ± 1.06  Iron (mgi  17.7±3.8  13.8±4.6  13.6±6.1  Magnesium (mg)  396±92  326±87  370 ± 170  Phosphorus (mg)  1125±317  1256 ± 263  1107±372  Potassium (mgi  3587 ±908  2850±717  2976 ±914  Sodium (mgi  2275±770  2373±793  1632±598  Zinc (mgi  8.46 ±2.20  8.32 ± 2.26  8.03 ± 2.25  Protein (%i  10.4±1.6  11.5±2.4  11.5 ± 1.4  Carbohydrate (%i  60.9 ±5.1’  55.7 ±4.0  56.4 ± 6.4  Fat (%)  28.3 ± 6.8  32.0 ± 3.8  32.2 ± 5.3  *p<o.o5 between vegans and lacto-ovos.  P< 0.05 between (acto-ovos and lactos.  89  P<0.O5 between vegans and lactos.  RESUL TS  Table 10. Mean (±SD) nutrient intakes, according to vegetarian subgroup. Daily Dietary Intakes  Vegan Vegetarians  Lacto Vegetarians  Number of Women  8  15  Calories (kcall  1923±290  2024±548  0.631  Protein (g)  51.9±12.1  57.1 ±10.8  0.297  Carbohydrates (g)  299.5± 38.8  288.3±79.8  0.714  Fibre (g)  35.0±9.9  24.7 ±7.9  0.012  Total Fat (g)  64.3 ±22.4  75.5 ±24.7  0.298  Saturated Fat(gl  15.1 ±7.5  23.8±8.6  0.025  Monounsaturated Fat (g>  25.9 ± 11.0  28.3 ±9.5  0.592  Polyunsaturated Fat (g)  17.3±6.9  16.1 ±6.1  0.664  Cholesterol (mgI  94±83  152±77  0.106  Vitamin A (RE>  2687± 1682  1268 ±479  0.005”  Thiamin 1mg)  1.83 ±0.67  1.31 ±0.40  0.029  Riboflavin (mgI  1.32±0.22  1.47 ±0.32  0.258  Niacin 1mg)  15.9 ± 2.7  12.0 ± 2.9  0.005”  Vitamin B 6 (mgI  1.88 ±0.63  1.40 ±0.40  0.03T  Vitamin B 12 (pg)  0.51 ±0.41  1.49 ±0.72  0.002”  Folacin (pg>  416± 142  310±83  0.033  Pantothenic Acid (mgI  5.32 ± 1.54  3.90 ±0.72  0.006”  Vitamin C (mgI  186±45  141 ±66  0.098  Calcium 1mg)  578± 184  875 ±255  0.008”  Copper (mg)  2.17±0.51  1.65±0.65  0.067  Iron (mg)  17.7 ±3.8  13.7 ±4.8  0.057  Magnesium (mg)  396 ±92  337±109  0.210  Phosphorus (mgI  1125±317  1217±290  0.491  Potassium (mgI  3587 ± 908  2884 ± 742  0.058  Sodium (mgI  2275±770  2175±801  0.776  Zinc 1mg)  8.46±2.20  8.24±2.18  0.819  Protein (%)  10.4 ± 1.6  11.5 ± 2.1  0.244  Carbohydrate (%I  60.9 ±5.1  55.9 ±4.5  0.025  Fat (%I  28.3 ±6.8  32.1 ±4.1  0.108  P<O.O5  P<O.O1  +Comparisons using student’s unpaired t-test.  90  P vaIue -  RESUL TS  4. DIETARY RESTRAINT The vegetarian women were compared to the nonvegetarian subjects for total eating inventory scores on the Three-Factor Eating Questionnaire, as well as subscales for dietary restraint, disinhibition and hunger. The results are shown in Table 11.  Table 11. Mean (± SD) total eating inventory scores and scores on subscales for the Three-Factor Eating Questionnaire, according to dietary group. Scale  All  Nonvegetarians  Vegetarians  Number of Women  45  22  23  Total Earing Inventory (56)  19.3±8.0  21.9±8.0  16.8±7.3  0.033  Restraint Subscale (21)  7.9±4.3  9.5±3.7  6.4±4.4  0.015’  Disinhibition Subscale (20)  5.9±3.6  6.1 ±4.1  5.7 ±3.3  0.721  Hunger Subscale (15)  5.0±2.4  5.6±3.0  4.4±1.6  0.112  P vaIue -  ‘Statistically significant differences at a level of P<O.05. + comparisons were made using the students unpaired t-test.  The results of the unpaired t-test comparison between the two study groups indicated that the vegetarian women had significantly lower total eating inventory scores (P< 0.05) and levels of dietary restraint (P<0.02) than the nonvegetarian study group. These differences are displayed in Figure 3. When comparisons were made within the vegetarian group, a significant difference in the restraint scores between veganvegetarians and lacto-vegetarians emerged (5.3 ±3.2 vs. 11.0±4.7; P=O.03). Furthermore, the discrepancy in the restraint scores between the lacto ovo vegetarian women and the lacto-vegetarians was meaningful (5.6 ±4.2 vs. 11.0±4.7; P=0.054).  In both cases the lacto-vegetarians scored higher than the other vegetarian  subgroups. The results are displayed in Table 1 2.  91  ____  ___  ___  _  ___  __ ___  ___  RESUL TS  DIETARY RESTRAINT (between vegetarian and nonvegetarian women)  30  --  “—z 15 10  TT 61  5  —  —  *Restraint  I  Disinhibition  Hunger  non\.. *significantly different at P<O.05  Figure 3. Scores on the Three-Factor Eating Questionnaire, and three subscales (dietary restraint, disinhibition and hunger) in the vegetarian and nonvegetarian women.  92  RESUL TS  Table 12. Mean (±SD) total eating inventory scores and scores on subscales for the ThreeFactor Eating Questionnaire, according to vegetarian subgroups. Scale  Vegan Vegetarians  Lacto-ovo Vegetarians  Lacto-vegetarians  Number of Women  8  1 1  4  Total Eating Inventory (56)  14.4±6.5  16.6±6.1  22.3±10.6  Restraint Subscale (21)  5.3±3.2W  5.6±4.2”  11.0±4.7  Disinhibition Subscale (20)  4.3±3.0  6.2±2.6  7.3±5.1  Hunger Subscale (15)  4.8±2.1  4.5±1.4  3.5±1.3  Statistically significant differences at a level of P<O.05 between vegan and lacto-vegetarians. Differences at a level of P=O.054 between lacto-ovo and lacto-vegetarians. + Comparisons were made using the student’s unpaired t-test.  Due to these differences in restraint scores between the lacto-vegetarians and the other two vegetarian subgroups, and because of its small sample size, the lacto-vegetarians were merged with the lacto-ovo group to form a single population that was subsequently compared in eating inventory data to the vegan vegetarians. These results are summarized in Table 1 3. There were no significant outcome differences in this new analysis.  Table 13. Mean ( ± SD) total eating inventory scores and scores on subscales for the Three-Factor Eating Questionnaire, according to vegetarian subgroups. Scale  Vegan Vegetarians  Lacto Vegetarians  Number of Women  8  15  Total Eating Inventory (56)  14.4±6.5  18.1 ±7.6  0.250  Restraint Subscale (21)  5.3±3.2  7.1 ±4.9  0.353  Disinhibition Subscale (20)  4.3 ±3.0  6.5 ±3.3  0.127  Hunger Subscale (15)  4.8±2.1  4.2± 1.4  0.458  + Comparisons  were made using the student’s unpaired t-test.  93  P vaIue -  RESUL TS  5. SPINAL BONE MINERAL DENSITY  The mean spinal bone mineral density for the 45 premenopausal women was 2 (range, 0.916 to 1.50 g/cm ) for the lumbar vertebra one to lumbar 2 1.18±0.12 g/cm vertebra four group ) 2 (range, 0.923 to 1.52 glcm 4 1 (L L , and 1.20±0.13 g/cm ) for the lumbar 2 vertebrae two to four class ) 4 2 (L L . As can be seen in Table 1 4, bone mineral content and bone mineral density in all vertebral segments were found to be lower in the vegetarian group, although the only difference that was significant was bone mineral content (61.15 ±8.17 2 vs. 67.21 ±9.80 g/cm g/cm ; P<O.05). However, because weight has been shown to have 2 the greatest association with bone density and because the vegetarians tended to weigh less than the nonvegetarians, weight was entered into the analysis as a covariate. In doing so the difference in bone mineral content between vegetarians and nonvegetarians was no longer significant  ,  although it was still large. This is demonstrated in Figure 4. Furthermore, the  difference between the two groups was found to be smaller and thus the two groups were quite similar in spinal bone mineral density values.  94  RESUL TS  Table 14. Mean (± SD) spinal bone mineral density characteristics as measured by dual energy x-ray absorptiometry, in 45 premenopausal women, according to dietary group. Characteristics  All  Nonvegetarians  Vegetarians  Number of Women  45  22  23  Bone Density:  P  +  value  -H-  Bone Mineral Content  64.11±9.41  67.21 ±9.80  61.15±8.17  0.029”  0.088  Lumbar One (L ) 1  1.10±0.13  1.14±0.13  1.07±0.12  0.064  0.137  Lumbar Two (L ) 2  1.20±0.14  1.24±0.15  1.16±0.12  0.053  0.120  Lumbar Three (L ) 3  1.23 ± 0.13  1.26 ± 0.14  1.19 ± 0.12  0.071  0.151  Lumbar Four (L ) 4  1.19±0.12  1.21 ±0.13  1.16±0.11  0.150  0.294  Lumbar One to Four 4 -L 1 (L )  1.18±0.12  1.22±0.13  1.15±0.11  0.064  0.142  Lumbar Two to Four 4 -L 2 (L )  1.20±0.13  1.24±0.13  1.17±0.11  0.074  0.161  . 2 Units of measurement for bone density were in glcm “Significant difference at a level of P<O.05. +Comparisons were made using student’s unpaired t-test. -f—f—Analysis of variance with weight as a covariate.  However, when group means were Compared within the vegetarian group, it was found that the lacto-vegetarian women had significantly lower bone density measurement values (L , 1 , and 4 3 L -L than vegans, and significantly lower bone density values (mean bone mineral 1 L ) content, L , L 1 , L 2 , and 4 3 -L than the lacto-ovo vegetarians. These results are displayed in 2 L ) Table 1 5.  Furthermore, when weight was entered into the analysis, these differences  remained significant. However, due to the small sample size of the lacto-vegetarian subgroup (n=4) participating in this study, these results cannot be generalized and it is difficult to assess the true meaning of these findings.  95  __  ________  RESUL TS  BONE MINERAL CONTENT (between vegetarian and nonvegetarian women)  6721 60  6L15  50 •  -.•  30 20 10 0  I  I  ) 2 BMC (g/cm  flOfl’ye P=O.088 (when weight was entered into the  analysis as a covariate)  Figure 4. Mean (±SD) bone mineral contents of the vegetarian and nonvegetarian women measured using dual energy x-ray absorptiometry in gIcm , and controlling for weight. 2  96  RESUL TS  Table 1 5. Mean (± SD) spinal bone mineral density characteristics as measured by dual energy x-ray absorptiometry in 23 premenopausal vegetarian women, according to dietary group. Characteristics  Vegan Vegetarians  Lacto-ovo Vegetarians  Lacto-vegetarians  Number of Women  8  11  4  Bone Mineral Content  62.46±6.76’”  63.89±7.07””  50.97±6.79  Lumbar One (L ) 1  1.10±0.13’”  1.10±0.09b’”  0.92±0.06c”’”  Lumbar Two (L ) 2  1.17±0.12  1.20±0.09””’  l.O2±O.1O””  Lumbar Three (L ) 3  1.20±0.10’”  1.24±0.10””  ’d’” 09 ±O. 04 l.  Lumbar Four (L ) 4  1.15±0.09  1.20±0.12  1.07±0.12  Lumbar One to Four 4 -L 1 (L )  1.15±0.10°”  l.19±0.09”  Lumbar Two to Four 4 -L 2 (L )  1.17±0.10  1.22±0.10”  Bone Density:  1.O5±O.1Oc”c”  LJnits of measurement for bone density were in g/cm’. “Significant difference at a level of P<O.05. °°°Significant difference at a level of P<O.O1. aSignificant difference between vegan and lacto-vegetarians using student’s unpaired t-test. bSignificant difference between lacto-ovo and lacto-vegetarian women using student’s unpaired t-test. cSignificant difference between vegan and lacto-vegetarians using analysis of variance with weight as a covariate. dSignificant difference between lacto-ovo and lacto-vegetarian women using analysis of variance with weight as a covariate.  To ensure that the differences in bone values seen between the lacto-vegetarians and the other  two  subgroups were due to samp’e size, the lacto-ovo vegetarians and the lacto  vegetarians were collapsed into one group and the bone measurements were reevaluated. There were no significant differences indicated for any of the bone evaluations when using the unpaired t-test or when the analysis controlled for weight. These results are presented in Table 1 6.  97  RESUL TS  Table 16. Mean (±SD) spinal bone mineral density characteristics as measured by dual energy x-ray absorptiometry in 23 premenopausal vegetarian women, according to dietary subgroup. Characteristics  Vegan Vegetarians  Lacto Vegetarians  Number of women  8  15  Bone Density:  P  value -  +  -H-  Bone Mineral Content  62.46 ±6.76  60.44 ±8.97  0.585  0.549  Lumbar One (L ) 1  1.10±0.13  1.05±0.11  0.406  0.409  Lumbar Two (L ) 2  1.17±0.12  1.15±0.12  0.779  0.780  Lumbar Three (L ) 3  1.20±0.10  1.19±0.13  0.897  0.883  Lumbar Four (L ) 4  1.15±0.09  1.17 ±0.12  0.755  0.755  Lumbar One to Four 4 -L 1 (L )  1.15±0.10  1.15±0.12  0.874  0.864  Lumbar Two to four 4 -L 2 (L )  1.17±0.10  1.17±0.12  0.988  0.979  ‘Units of measurement for bone density were in g/cm . 2 +Comparisons were made using student’s unpaired t-test. Analysis of variance with weight as a covariate.  6. URINE ANAL YSIS  The ratios of calcium to creatinine and cortisol to creatinine could not be compared between groups because the original urine volumes provided by each subject were not available and thus true calcium, cortisol and creatinine contents could not be measured. Furthermore, there is the potential of significant differences in creatinine values between vegetarians and nonvegetarians due to dietary intake differences 95 which could not be measured here. Hence, comparisons were made only within diet groups and not between the two groups.  98  RESUL TS  Table 1 7. Mean (± SD) urinary excretion of calcium, cortisol and creatinine levels from a two hour fasting urine sample, according to dietary group. Characteristics  Nonvegetarians  Vegetarians  Number of Women  22  23  Creatinine (mmol/L)  4.20  Calcium {mmol/L)  3.50  -  2.03  0.294  0.74±0.73  0.56±0.64  0.397  Cortisol (nmol/L)  116 ±98  73±46  0.068  Calcium/Creatinine Ratio  0.17±0.10  0.19±0.17  Cortisol/Creatinine Ratio  30.47  24.60  ±  ±  16.94  3.29  P value  ±  ±  10.03  -  -  Comparisons were made using students unpaired t-test.  None of the values differed significantly between the two groups, however the cortisol concentrations tended to be higher in the nonvegetarians than vegetarians. The standard deviations however, were quite large in all values which indicates considerable inter-individual variability, much of which likely reflects differences in the volumes of the original urine samples. It could also be due to some kind of physiological difference. In the nonvegetarian group, the excretion ranges for the substances analysed were: creatinine, from 1 .32 mmol/L to 1 6.96 mmol/L; calcium, from 0.20 mmol/L to 2.80 mmol/L; and cortisol, from 34 nmol/L to 489 nmol/L.  In the vegetarian group, the ranges were as follows: creatinine, from 1 .00  mmol/L to 8.34 mmol/L; calcium, from 0.10 mmol/L to 3.1 mmol/L; and cortisol, from 22 nmol/L to 175 nmol/L. These values were compared within the vegetarian subgroup and no significant differences were found.  These results are illustrated in Table 18.  The ranges for the  excretion levels in the vegan vegetarians were from 1.17 mmol/L to 5.60 mmol/L for creatinine, from 0.10 mmol/L to 0.80 mmol/L for calcium, and from 28 nmol/L to 1 64 nmol/L  99  RESUL TS  for cortisol. The lacto-ovo subgroup’s ranges were: creatinine, from 1 .31 mmol/L to 8.34 mmol/L; calcium, from 0.1 mmol/L to 3.1 mmol/L; and cortisol, from 22 nmol/L to 175 nmol/L. The lacto-vegetarians’ ranges were similar with creatinine ranging from 1 .00 mmol/L to 6.57 mmol/L, calcium ranging from 0.1 mmol/L to 1.1 mmol/L, and cortisol’s range included 25 nmol/L and 107 nmol/L.  Table 18. Mean (± SD) urinary excretion of calcium, cortisol and creatinine levels from a twohour fasting urine sample, according to vegetarian dietary subgroup. Characteristics  Vegan Vegetarians  Lacto-ovo Vegetarians  Lacto-vegetarians  Number of Women  8  11  4  Creatinine (mmol/Ll  2.98±1.60  3.58±2.24  3.15±2.61  Calcium (mmol/L)  0.30±0.26  0.75±0.82  0.58±0.55  Cortisol (nmol/L)  79±40  77±54  52±37  Calcium/Creatinine Ratio  0.16  0.22  Cortisol/Creatinine Ratio  28.60  ±  0.22 ±  9.73  ±  22.39  0.16 ±  0.15±0.08  8.92  22.65  ±  13.64  ‘Comparisons were made using student’s unpaired t-test.  7. FACTORS RELA TING CHANGES IN THE MENSTRUAL CYCLE A) THE VEGETARIAN/NON VEGETARIAN POPULA TION Relationships between demographic, anthropometric, bone, and restraint data on the one hand and menstrual cycle characteristics on the other were assessed using correlation analyses.  These results, presented in Table 19, indicated that age (r=0.430; P=O.002),  gynecologic age (r=0.411; P<O.0O5), and dietary restraint (r=-0.251; P<O.05) were significantly correlated to the number of ovulatory cycles.  Age (r=-0.305; P<O.05) and  gynecologic age (r =-0.31 9; P< 0.02) were also significantly, though negatively correlated to  100  RESUL TS  the number of short luteal phase cycles. As with the number of ovulatory cycles, the average luteal phase length was significantly associated with age (r=O.51 5; P<O.000), gynecologic age (r=O.505; P<O.000) and dietary restraint (r=-O.251; P<O.05). Finally, age (r=O.468; P=O.OO1), gynecologic age (r=O.473; P=O.OO1), sum of skinfolds(r=-O.264; P<O.05) and total body fat (r=-O.252; P<O.05) were significantly correlated to the mean luteal phase index. None of the variables were significantly associated with the mean cycle length. Further correlation analysis of calcium and cortisol values, adjusted to lean body mass, were compared to the menstrual function data.  It was found that none of the menstrual  measurements were significantly correlated to the urine figures in either the vegetarian or the nonvegetarian groups. These data are available in Table 20.  101  RESUL TS  Table 19. Correlation coefficients for demographic, anthropometric, bone and restraint data, as related to menstrual function characteristics. Characteristics  # Ovulatory Cycles  It Short LP  Cycle Length  LP Length  LP Index  Age (years)  0.430”  -0.305’  0.147  0.515”  0.468”  Age at Menarche (yrs.)  0.132  -0.006  0.054  0.116  0.061  Gynecologic Age (yrs.)  0.411”  -0.319’  0.138  0.505”  0.473”  Exercise (hrs/wk)  -0.161  -0.049  -0.161  -0.018  0.204  Exercise Regimen (yrs.)  0.113  0.152  0.075  0.027  -0.095  Weight (kg)  -0.109  0.135  -0.070  -0.184  -0.054  Height (cm)  0.024  0.220  0.118  -0.038  -0.062  BMI (kg/rn ) 2  -0.164  -0.043  -0.194  -0.194  0.001  Sum of Skinfolds (mm)  -0.073  0.044  -0.005  -0.215  -0.264’  Body Fat (%)  -0.074  0.036  -0.011  -0.210  -0.252’  Total Body Fat (kg)  -0.092  0.088  -0.030  -0.225  -0.209  Total Muscle Mass (kg)  -0.038  0.196  -0.024  -0.075  0.061  Eating Inventory (56)  -0.164  -0.026  -0.112  -0.166  -0.038  Dietary Restraint (21)  -0.251’  -0.019  -0.181  -0.249’  -0.113  Disinhibition (20)  -0.075  -0.004  -0.118  -0.115  -0.038  Hunger (15)  0.013  -0.110  0.125  0.076  0.155  Bone Mineral Content (g/cm ) 2  0.114  0.034  0.118  0.102  0.075  1 (g/cm L ) 2  0.074  -0019  0.025  0.082  0.024  2 (g/cm L ) 2  0.039  0.041  -0.006  0.016  -0.023  3 (g/cm L ) 2  -0.006  -0.002  -0.044  -0.001  0.014  4 (g/cm L ) 2  -0.024  0.015  -0.104  -0.040  0.006  -L (g/cm 1 L 4 ) 2  0.017  0.012  -0.040  0.007  0.000  -L (g/cm 2 L 4 ) 2  0.001  0.019  -0.055  -0.010  -0.001  Signifioant correlation at a level of P<O.05. “Significant correlation at a level of PO.O1.  102  RESUL TS  Table 20. Correlation coefficients for urine values as related to menstrual function data, according to dietary group. Nonvegetarians  (n=22)  Vegetarians  (n=23)  Menstrual Cycle Data  Ca:Cr  Co:Cr”  Ca:Cr  Co:Cr  No. Ovulatory Cycles  -0.148; P=O.255  -0.148; P=O.255  0.068; P=O.379  0.162; P=O.230  No. Short Luteal Ph.  0.156; P=O.244  -0.098; P=O.332  -0.090; P=O.341  -0.109; P=O.311  Avg. Cycle Length  -0.293; P=O.093  -0.209; P=O.176  0.148; P=O.250  0.288; P=O.092  Avg. Luteal Length  -0.177; P=O.215  -0.036; P=O.437  0.048; P=O.414  0.042; PO.425  Avg. Luteal Ph. Index  -0.048; P=O.417  0.058; P=O.399  -0.075; P=O.367  -0.189; P=O.194  ca:cr  =  Co:Cr  Calcium to Creatinine ratio. Cortisol to Creatinine ratio.  =  When nutrients were evaluated in comparison to menstrual cycle data through correlation analysis, the following were found: total fibre (r=0.261; P<O.05) and vitamin A (r=0.296; P<0.03) were significantly correlated to mean cycle length; total fibre (r=0.266; P<0.05), vitamin A (r=0.338; P<0.02), folacin (r=0.317; P<0.02) and iron (r=0.248; P= 0.05) were significantly correlated to mean luteal phase length; folacin (r = 0.317; P< 0.02) and iron (r=O.292; P<0.03) were correlated to luteal phase index; and, only vitamin A (r=0.329; P<0.02) was correlated to the mean number of ovulatory cycles.  No dietary  nutrients were significantly correlated to the number of short luteal-phase cycles. These results are summarized in Table 21.  103  RESUL TS  Table 21. Correlation coefficients of total nutrient intakes to menstrual cycle data 0 in 45 premenopausal women. Daily Dietary Intakes  No. Ovulatory  No. Short Lut.  Avg. Cycle Lgth  Avg. Lut. Lgth  Avg. Lut. Index  Calories (kcal)  0.069  -0.020  -0.018  0.126  0.189  Protein (g)  -0.025  -0.105  -0.034  0.072  0.176  Carbohydrates (9)  0.105  0.019  0.068  0.136  0.154  Fibre (g)  0.214  0.093  O.261  0.266’  0.206  Total Fat (g)  0.033  -0.016  -0.069  0.092  0.166  Saturated Fat (g)  0.045  0.014  -0.075  0.048  0.096  Monounsaturated Fat (9)  0.060  -0.013  -0.066  0.104  0.177  Polyunsaturated Fat (gI  -0.029  -0.036  -0.036  0.086  0.143  Cholesterol (mg)  0.054  -0.038  -0.106  0.100  0.203  Vitamin A (RE)  0.329  0.067  0.296  0.338’  0.127  Thiamin (mgI  -0.063  -0.064  0.195  0.082  0.112  Riboflavin (mg)  0.044  -0.107  0.017  0.141  0.241  Niacin (mgI  0.107  -0.044  0.091  0.192  0.195  Vitamin B 6 (mg)  -0.015  0.041  0.124  0.068  0.121  Vitamin  -0.1 11  -0.072  -0.062  -0.07 1  0.066  Folacin (jig)  0.196  -0.079  0.177  0.342’  0.317’  Pantothenic Acid (mgI  0.072  -0.039  0.199  0.185  0.191  Vitamin C (mg)  0.188  -0.039  0.161  0.266  0.210  Calcium (mg)  0.052  -0.039  0.067  0.086  0.139  Copper (mg)  0.047  -0.067  0.131  0.178  0.205  Iron (mgI  0.053  -0.171  0.147  0.248’  0.292’  Magnesium (mgI  -0.042  -0.098  0.051  0.103  0.168  Phosphorus (mg)  0.068  -0.069  0.091  0.149  0.218  Potassium (mgI  0.080  -0.004  0.129  0.174  0.221  Sodium (mg)  0.214  0.069  0.083  0.212  0.186  Zinc (mg)  -0.012  -0.106  0.098  0.109  0.189  Protein (%)  -0.166  -0.143  -0.055  -0.108  0.005  Carbohydrate (%)  0.120  0.107  0.211  0.049  -0.075  Fat (%)  -0.068  -0.053  -0.182  -0.014  0.059  812  (jig)  Significant correlation at a level of PO.O5.  104  RESUL TS  Stepwise multiple regression analysis was performed on the menstrual function data incorporating those variables, whether demographic, anthropometric, bone, restraint or dietary data, that had been shown to be significant through correlation analysis, to identify which of the latter were most important. Overall, age (r 2  =  0.1 84; P< 0.005), levels of dietary restraint  =0.539; P,0.0005), iron (r 2 =0.280; P=0.001), vitamin A (r 2 (r = 0.423; P=O.0002), and 2 dietary fibre (r 2 =0.480; P=0.0001) were found to be significantly correlated with the mean number of ovulatory cycles.  The interaction of these variables is presented in Table 22.  =0.102; P<0.05 2 Gynecologic age (r  -  B=-0.082; R=-0.319) was the only variable  associated with the number of short luteal phase cycles.  Age (r =0.265; P=0.0003), 2  =0.470; P<0.0000) and 2 =0.373; P=0.0001), vitamin A (r 2 percent body fat (r  level of  dietary restraint (r 2 =0.524; P< 0.0000) were significantly related to the average luteal phase length, as can be seen in Table 23. Gynecologic age (r 2 =0.224; P=0.001) and percent body fat (r 2  =  0.360; P= 0.0001) were correlated to the mean luteal phase index. Their combination  is shown in Table 24. And finally, the only variable significantly correlated to the mean cycle length was vitamin A (r =0.088; P<0.05 2  -  B= 0.001; B=0.296).  Table 22. Variables correlated to the number of ovulatory cycles in 45 premenopausal women. Variable  R Square  Significant F  B  Beta (la)  Significant T  Age (years)  0.184  0.0032  0.168  0.573  0.0000  Dietary Restraint (21)  0.280  0.0010  -0.088  -0.255  0.0364  Vitamin A (RE)  0.359  0.0004  431°  0.334  0.0269  Iron (mg)  0.423  0.0002  -0.189  -0.605  0.0045  Dietary Fibre (g)  0.480  0.0001  0.07 1  0.429  0.0442  VariabIes in the stepwise multiple regression analysis equation.  105  RESUL TS  Table 23. Variables correlated to the average luteal phase length in 45 premenopausal women. Variabl&  R Square  Significant F  B  Beta (la)  Significant T  Age (years)  0.265  0.0003  0.395  0.586  0.0000  Body Fat (%)  0.373  0.0001  -0.182  -0.292  0.0146  Vitamin A (RE)  0.470  0.0000  8.77°  0.295  0.0103  Dietary Restraint (21)  0.524  0.0000  -0.196  -0.245  0.0359  Variables in the stepwise multiple regression analysis equation.  Table 24. Variables correlated to the average luteal phase index in 45 premenopausal women. Variable  R Square  Significant F  B  Beta (1)  Significant T  Gynecologic Age (yrs)”  0.224  0.0010  0.008  0.559  0.0001  Body Fat (%)  0.360  0.0001  -0.005  -0.379  0.0047  Variables in the stepwise multiple regression analysis equation.  Gynecologic age  =  age  -  age at menarche.  B) THE NORMAL/ABNORMAL MENSTRUAL FUNCTION POPULA TION  To further evaluate the influences on the menstrual cycle, subjects were divided according to their menstrual function and the two new groups were analysed. The women were classified as “normal” (n=20) and “abnormal” (n=25). The women who fell into the “normal” category consisted of those who had experienced regular cycles throughout the six months, which is to say they had ovulatory cycles with a luteal phase length no less than ten days, and those women who had only one short luteal phase cycle throughout the study period. The women who made up the “abnormal” group included those who experienced more than one short luteal phase cycle and the women who had one or more anovulatory cycles over the six month study period. When the two groups were compared using student’s unpaired t-test the only significant differences found were in their mean age (29.4± 5.4 years  106  RESUL TS  vs. 25.4 ±4.2; P<O.O1) (Table 25) and in the amount of vitamin and mineral supplements used over the six months (3.2±2.8 vs. 1.2±2.1; P=O.O1).  The “normal” group was  therefore found both to be older and to use more supplements than the “abnormal” group. No differences were detected in levels of eating restraint (Table 26), in spinal bone density (Table 27), or in nutrient intakes (Table 28).  Table 25. Mean ( ± SD) characteristics of 45 premenopausal women, according to menstrual cycle group. P value  Variable  Normal Menstrual Group  Abnormal Menstrual Group  Number of Women  20  25  Age (years)  29.4±5.4  25.4±4.2  0.008”  Age at Menarche (yrs.)  1 3.3 ± 1.5  1 3.3 ± 1.5  0.965  Exercise (hrs/wk)  3.5 ±2.1  3.6±2.1  0.826  Exercise Regimen (yrs.)  3.8 ±4.1  5.0±5.8  0.415  Weight (kg)  58.3±8.3  61.9 ±7.3  0.121  Height (cml  165.2 ±8.5  166.7 ±7.9  0.541  ) 2 Body Mass Index (kg/rn  21.3±2.5  22.3±2.0  0.160  Sum of Skinfolds (mm)  85.6 ±26.4  94.8±23.4  0.218  Body Fat (%)  24.5 ±5.9  26.6 ±5.2  0.224  Total Body Fat (kg)  14.5 ±4.8  16.7 ±4.7  0.133  Total Muscle Mass (kg)  27.9 ±4.9  29.6 ±3.6  0.203  -  Normal = women with regular cycles or with only one short luteal phase (i.e. luteal phase <10 days); abnormal more than one short luteal phase and/or one or more anovulatory cycles. “Significant difference at P<O.O1. +Comparisons were made using student’s unpaired t-test.  107  =  women with  RESUL TS  Table 26. Mean ( ± SD) total eating inventory scores and scores on subscales for the ThreeFactor Eating Questionnaire, according to menstrual cycle group. Scale  Normal Menstrual Group  Abnormal Menstrual Group  Number of Women  20  25  Total Eating Inventory (56)  18.6±7.5  19.9±8.4  0.584  Restraint Subscale (21)  7.2 ±3.6  8.5±4.8  0.313  P value -  +Comparisons were made using the student’s unpaired t-test.  Table 27. Mean (± SD) spinal bone mineral density characteristics as measured by dual energy x-ray absorptiometry, in 45 premenopausal women, according to menstrual cycle group. Characteristics  Normal Menstrual Group  Abnormal Menstrual Group  Number of Women  20  25  Bone Mineral Content  64.19±9.17  64.05 ± 9.80  0.961  Lumbar One (L ) 1  1.11 ±0.14  1 .10 ± 0.12  0.956  ) 2 Lumbar Two (L  1.19±0.14  1.21 ±0.14  0.584  Lumbar Three (L ) 3  1.22±0.15  1.23±0.12  0.744  Lumbar Four IL ) 4  1 .17 ± 0.14  1.20±0.11  0.550  Lumbar One to Four 4 -L 1 (L )  1.17 ± 0.14  1 .19 ± 0.11  0.675  Lumbar Two to Four 4 -L 2 (L )  1.19 ± 0.14  1.21 ±0.12  0.609  P value  Bone Density:  Units of measurement for bone density were in gfcm . 2 +Comparisons were made using student’s unpaired t-test.  108  RESUL TS  Table 28. Mean (±SD) total nutrient intakes for nine days, according to menstrual group. Daily Dietary Intakes  Normal Menstrual Group  Abnormal Menstrual Group  Number of Women  20  25  Calories (kcal)  2036±464  2037 ± 529  0.999  Protein (g)  66.0 ± 20.4  65.9 ± 18.7  0.990  Carbohydrates (g)  280.7 ± 64.0  294.1 ±78.3  0.536  Fibre (g)  24.4 ± 8.1  26.1 ±9.1  0.533  Total Fat (g>  75.2 ± 23.8  71.7 ± 23.5  0.618  Saturated Fat (g)  23.8 ±9.6  22.3±11.1  0.626  Monounsaturated Fat (g)  28.7 ±9.8  27.3 ±9.1  0.615  Polyunsaturated Fat (g>  15.7 ± 6.0  15.5 ± 5.4  0.916  Cholesterol (mg)  196±94  167 ± 100  0.331  Vitamin A (RE>  1759±1207  1655 ± 1142  0.768  Thiamin (mg)  1.46 ± 0.66  1.57±0.42  0.480  Riboflavin (mg)  1.57±0.38  1.56 ± 0.45  0.967  Niacin (mg)  16.5 ± 6.5  15.6±3.4  0.551  Vitamin B 6 (mg>  1.50±0.54  1.66±0.44  0.273  Vitamin B 12 (jig)  2.35 ± 1.82  2.52 ± 1.86  0.756  Folacin (jig)  315±122  304 ± 92  0.726  Pantothenic Acid (mg)  4.63 ± 1.64  4.70 ± 1.16  0.866  Vitamin C (mgI  132± 50  140±60  0.600  Calcium (mg)  833 ± 326  880±406  0.673  Copper (mgI  1.61 ±0.53  1.64±0.61  0.833  Iron (mgI  15.4±5.3  15.0±4.4  0.786  Magnesium (mgI  322±95  338±108  0.600  Phosphorus (mgI  1272 ± 385  1312±423  0.745  Potassium (mgI  2962±863  3186±817  0.378  Sodium (mgI  2485 ± 661  2500 ± 934  0.953  Zinc (mg)  9.60±4.12  9.74 ± 2.78  0.892  Protein (%I  12.9 ± 2.9  12.9 ± 2.8  0.962  Carbohydrate (%)  54.6 ± 5.4  56.8 ± 5.0  0.175  32.0 ± 5.5  29.9±4.7  0.189  Fat  (%I  Comparisons were made using student’s unpaired t-test.  109  P value  RESUL TS  8. FACTORS RELA TING TO BONE  Demographic, anthropometric, dietary restraint and menstrual characteristics were correlated to the bone evaluations. The correlation coefficients are displayed in Table 29. Weight, BMI, sum of skinfolds, percent body fat, total body fat and muscle mass were all , L 1 significantly and positively correlated to the following bone measures: L , L 2 ,L 3 , 4 4 -L and 1 L , -L Furthermore, bone mineral content was significantly correlated to weight, height, total 2 L . 4 1 and age at menarche was body fat, and skeletal muscle -mass. The association between L quite important (P=O.052). When the means for nutrient intake available through diet analyses were correlated to 12 and zinc were significantly bone data, the following was found: cholesterol, vitamin B associated with all bone data, as can be seen in Table 30.  Bone mineral content was  significantly correlated with caloric intake (r=0.266; P<O.05), protein (r=0.311; P<O.02), total fat (r=0.272; P<O.05), saturated fat (r=0.01 1; P<O.02), monounsaturated fat (r=0.271; P<O.05), cholesterol (r=0.421; P=O.002), niacin (r=0.026; P<O.05), vitamin 12 (r=0.419; P=O.002), calcium (r=0.334; P<O.02), phosphorus (r=0.327; P<O.02), B sodium (r=0.339; P<O.02), and zinc (r=0.451; P=O.OO1).  Lumbar vertebra one was  significantly correlated to cholesterol (r=0.338; P<O.02), vitamin B 12 (r=O.369; P<O.O1), zinc (r=0.334; P<O.02), and to the percentage of the diet as fat (r=O.257; P<O.05). Lumbar vertebra one was also negatively correlated to the percentage of the diet as carbohydrate (r=-0.348; P=O.1O). Cholesterol (r=0.332; P<O.02), vitamin B 12 (r=O.371;  P<O.O1), and zinc (r=0.296; P<O.05) were all positively associated with lumbar vertebra 2 (r=-O.358; two, and the percentage of the diet as carbohydrate was negatively related to L  110  RESUL TS  P<0.01). Lumbar spine three was significantly correlated to protein (r=0.254; P<0.05), cholesterol (r =0.387; P<0.005), vitamin 812 (r =0.455; P= 0.001), zinc (r =0.338; P<0.02), and to the percentages of the diet from protein (r=0.294; P<0.05) and fat (r=0.256; P<0.05). Again, the percentage of dietary intake from carbohydrates was negatively related  to the bone measure (r=-0.418; P=0.002). Only cholesterol (r=0.362; P<O.O1), vitamin 12 (r=0.387; P=0.002), zinc (r=0.267; P<0.05) and the percentage of the diet from fat B (r=0.365; P<0.01) were positively correlated to the lumbar vertebra four. The percentage of carbohydrate in the diet was negatively correlated to the latter bone evaluation (r =-0.479; -L was significantly correlated to cholesterol (r=0.368; 1 L P.<0.000). The bone range 4 12 (r =0.418; P=0.002), zinc (r=0.324; P<0.02), and to the percentages P<0.01), vitamin B of total intake for protein (r=0.248; P=0.05) and fat (r=0.284; P<0.05). Additionally, the percentage of the diet as carbohydrate was negatively correlated to that bone range (r= 0.416; P=O.002). Finally, cholesterol (r=0.375; P<0.01), vitamin (r 12 =0.420; P=0.002), B zinc (r=0.312; P<O.02) and the percentages of total dietary intake from protein (r=O.260; P<0.05) and fat (r=0.298; P<0.05) were positively correlated to the bone evaluation 4 -L 2 L ,  with percentage of the diet from carbohydrate (r=-0.434; P<0.05) negatively associated with the latter bone range. Because of the implied role of protein in increasing urinary calcium excretion, the association between the calcium to protein ratio and bone values was assessed. None of the correlation coefficients was found to be significant.  111  RESUL TS  Table 29. Correlation coefficients for bone values to demographic, anthropometric, restraint and menstrual cycle data, for the entire study population (n=45). Characteristics  BMC’  1 L  2 L  L  4 L  -L 1 L 4  -L 2 L 4  Age (years)  0.224  0.146  0.108  0.105  0.058  0.105  0.091  Age at Menarche (yrs.)  0.026  -0.246  -0.189  -0.133  -0.166  -0.186  -0.167  Gynecologic Age (yrs.)  0.228  0.230  0.175  0.151  0.112  0.168  0.147  Exercise (hrs/wk)  -0.034  0.012  -0.014  -0.116  -0.101  -0.063  -0.081  Exercise Regimen (yrs.)  0.173  0.112  0.025  -0.036  0.064  0.038  0.020  Weight (kg)  0.553’”  0.337”  0.351  0.331”  0.336”  0.353”’  0.351  Height (cm)  0.527’”  0.184  0.159  0.136  0.117  0.145  0.141  ) 2 BMI (kg/rn  0.214  0.254’  0.293’  0.291”  0.329’  0.312”  0.315”  Sum of Skinfolds (mm)  0.225  0.269”  0.327’  0.349  0.359”  0.347”  0.356’”  Body Fat (%)  0.219  0.260”  0.315’  0.341’  0.348”  0.337”  0.345”  Total Body Fat (kg)  0.386’”  0.330”  0.383”  0.388”  0.393”  0.394”’  0.400”’  Total Muscle Mass (kg)  0.553”  0.333”  0.340’  0.291”  0.270”  0.317”  0.310”  Eating Inventory (561  0.055  -0.040  -0.039  -0.072  -0.148  -0.077  -0.091  Restraint (21)  0.013  0.020  -0.011  -0.038  -0.041  -0.019  -0.033  Disinhibition (20)  -0.020  -0.123  -0.127  -0.137  -0.228  -0.158  -0.170  Hunger (15)  0.213  0.027  0.093  0.037  -0.060  0.026  0.021  No. Ovulatory Cyc.  0.114  0.074  0.039  -0.006  -0.024  0.017  0.001  No. Short LP  0.034  -0.019  0.041  -0.002  0.015  0.012  0.019  Cycle Length (days)  0.118  0.025  -0.006  -0.044  -0.104  -0.040  -0.055  LPLength (days)  0.102  0.082  0.016  -0.001  -0.040  0.007  -0.010  LP Index  0.075  0.024  -0.023  0.014  0.006  0.000  -0.001  BMC = Bone mineral content, Significance at a level of P<O.05. “Significance at a level of P<O.O1.  112  RESUL TS  Table 30. Correlation coefficients for bone values to nutrients, for 45 premenopausal women. Nutrients  BMC’  1 L  2 L  2 L  4 L  -L 1 L 4  -L 2 L 4  Calories (kcal)  0.266”  0.099  0.062  0.074  0.071  0.079  0.074  Protein (g)  0.311”  0.211  0.211  0.254”  0.190  0.223  0.228  Carbohydrate (g)  0.161  -0.030  -0.061  -0.077  -0.096  -0.069  -0.079  Fibre (g)  0.008  -0.089  -0.120  -0.194  -0.195  -0.160  -0.176  Total Fat (g)  0.272”  0.173  0.134  0.149  0.191  0.164  0.166  Saturated Fat (g)  0.342”  0.210  0.182  0.225  0.198  0.211  0.211  Monounsat. Fat (g)  0.271”  0.197  0.131  0.145  0.215  0171  0.172  Polyunsat. Fat (g)  0.030  0.022  0.026  -0.046  0.046  0.008  0.009  Cholesterol (mg)  0.421  0.338”  0.332”  0.387’”  0.362”  0.368”  0.375”  Vitamin A (RE)  -0.004  0.026  0.027  -0.059  0.018  0.005  -0.005  Thiamin (mg)  0.144  0.050  0.073  -0.028  -0.014  0.021  0.010  Riboflavin (mg)  0.367”  0.208  0.221  0.245  0.197  0.231  0.231  Niacin (mg)  0.292”  0.202  0.174  0.116  0.075  0.146  0.125  Vitamin B 6 {mg)  0.048  0.024  0.005  -0.076  -0.083  -0.035  -0.053  Vitamin B 12 (jig)  0.419”  0.369”  0.371  0.465’  0.387  0.418”  0.420”’  Folacin (jig)  0.072  0.002  -0.038  -0.131  -0.109  -0.081  -0.096  Pantothenic Acid (mg)  0.226  0.153  0.143  0.141  0.097  0.138  0.131  Vitamin C (mg)  -0.002  -0.032  -0.039  -0.109  -0.099  -0.076  -0.085  Calcium (mg)  0.334”  0.179  0.195  0.221  0.164  0.202  0.202  Copper (mg)  -0.064  -0.082  -0.152  -0.163  -0.148  -0.146  -0.160  Iron (mgI  0.059  -0.028  -0.045  -0.111  -0.135  -0.086  -0.101  Magnesium (mgI  -0.039  -0.081  -0.121  -0.129  -0.126  -0.120  -0.129  Phosphorus (mg)  0.327”  0.202  0.206  0.215  0.191  0.215  0.212  Potassium(mg)  0.116  0.077  0.050  -0.011  -0.015  0.024  0.008  Sodium (mg)  0.339”  0.246  0.231  0.207  0.166  0.219  0.209  Zinc (mg)  0.451  0.334”  0.296”  0.338”  0.267”  0.324”  0.312”  Protein (%)  0.164  0.170  0.234  0.294”  0.226  0.248  0.260’  Carbohydrate (%)  -0.234  -0.348”  -0.358’’  -0.418”  -0.479”  -0.416”  -0.434”  Fat (%)  0.160  0.257”  0.239  0.256”  0.365”  0.284  0.298”  “  “  BMC = Bone mineral content. “Significance at a level of P<O.05.  “‘Significance at a level of P<O.O1.  113  RESUL TS  Urine values were correlated to the bone evaluations in the vegetarian and nonvegetarian women. These correlations were completed separately as the original urine volumes were missing. None of the associations were found to be significant in either group, but further details are available in the following table (Tables 31).  Table 31. Correlation coefficients for bone values to urine data for the two subgroups of women. Characteristics  BMC  1 L  2 L  3 L  4 L  -L 1 L 4  -L 2 L 4  Calcium:Creatinine  -0.056  0.101  0.175  0.281  0.298  0.226  0.261  Cortisol:Creatinine  0.107  0.191  0.308  0.222  0.164  0.228  0.234  Calcium:Creatinine  0.218  0.278  0.287  0.065  -0.016  0.151  0.107  Cortisol:Creatinine  0.134  0.082  0.133  0.172  0.093  0.129  0.130  Nonvegetarians (n  =  22):  Vegetarians (n=23):  BMC  =  Bone mineral content.  Stepwise multiple regression analysis was performed on the bone density data (bone mineral content, 4 -L and 4 1 L , -L incorporating those variables that had been previously shown 2 L ) to be significant through correlation analysis, to identify which of the latter were most important.  Bone mineral content was significantly predicted by weight (r =O.306; 2  12 (r =O.435; P<O.0000) and BMI 2 2 P=O.0001), vitamin B (r = O.539; P<O.0000), as can be seen in Table 32.  12 (r Vitamin B =O.174; P<O.005) and total body fat (r 2 =O.277; 2  P=O.OO1 1) were significantly associated with 4 -L Further details are presented in Table 1 L . 33. And finally, the percentage of calories as carbohydrate (r =O.189; P<O.005) and total 2 body fat (r =O.291; P<O.OO1) were significantly correlated with the bone range 4 2 -L This 2 L . is summarized in Table 34. 114  RESUL TS  Table 32. Characteristics correlated to the bone mineral content, in 45 premenopausal women. Variable  R Square  Significant F  B  Beta (Q)  Significant T  Weight (kg)  0.306  0.0001  0.994  0.830  0.0000  Vitamin B 12 (pg)  0.435  0.0000  2.203  0.426  0.0003  BMI (kg/rn ) 2  0.539  0.0000  -1.951  -0.462  0.0042  Variables in the stepwise multiple regression analysis equation.  Table 33. Characteristics correlated to the bone density evaluation 4 -L in 45 premenopausal 1 L , women. Variabl&  R Square  Significant F  B  Beta (L)  Significant T  Vitamin B 12 (pg)  0.175  0.0043  0.024  0.354  0.0113  Total Body Fat (kg)  0.277  0.0011  0.008  0.326  0.01 93  Variables in the stepwise multiple regression analysis equation.  Table 34. Characteristics correlated to the bone density evaluation 4 -L in 45 premenopausal women. 2 L , Variable  R Square  Significant F  B  Beta (Ia)  Significant T  Carbohydrate (%)  0.189  0.0029  -0.009  -0.369  0.0080  Weight (kg)  0.291  0.0007  0.009  0.327  0.0179  Variables in the stepwise multiple regression analysis equation.  9. FACTORS RELA TING TO DIETARY RES TRA INT A) THE VEGETARIAN/NON VEGETARIAN POPULA TION  Correlation analysis of the anthropometric, demographic, bone and menstrual data to the eating inventory scores and the dietary restraint subscale revealed that weight (r=0.477; P<Q.000), BMI (r=0.544; P<O.000), sum of skinfolds (r=0.286; P<O.05), percent body  115  RESUL TS  fat (r=O.293; P<O.05), total body fat (r=O.411; P<O.005) and total skeletal muscle mass (r=O.395; P<O.005) were all significantly correlated to total eating inventory scores. Age at menarche (r=O.259; P<O.05), weight (r=O.341; P<O.02), BMI (r=O.456; P=O.OO1), total body fat (r=O.314; P<O.02), and total skeletal muscle mass (r=O.283; P=O.03) were significantly correlated to levels of dietary restraint. The relationship between the sum of skinfolds (r=O.227; P=O.067) and percent body fat (r=O.232; P=O.063) with the restraint subscale were also important. Correlation analysis of the calcium ratios, adjusted to lean body mass, to dietary restraint indicated that both total eating inventory scores (r=-O.393; P<O.05) and levels of dietary restraint (r=-O.426; P<O.05) were significantly negatively correlated to it in the nonvegetarian group only.  Total eating inventory scores (r=O.552; P<O.005) were also  significantly correlated to the cortisol ratio in the nonvegetarian group. Again, there were no significant findings for the cortisol values in the vegetarian group. Total eating inventory and levels of dietary restraint were correlated to nutrients analyzed and the results indicated that vitamin C (r=-O.276; P<O.05) and the percentage of the diet as protein (r=O.278; P<O.05) were the only nutrients significantly associated to dietary restraint. The correlation of polyunsaturated fat to restraint (r=-O.241; P=O.055) was also important, though not significant. There were no nutrients significantly associated with total eating inventory scores. When the variables that had been significantly correlated to restraint data were entered into a stepwise multiple regression analysis the following was established: (i) BMI (r 2 =0.260; P=O.0001  -  B= 1.948; g=Q.544) was significantly associated with total eating inventory  116  RES(IL TS  scores; and, (ii) BMI (r =O.208; P<O.005) and age at menarche 2 2 (r = O.282; P=O.OO1) were significantly correlated with the dietary restraint subscale. Further information regarding the latter association is available in Table 35.  Table 35. Characteristics correlated to dietary restraint, in 45 premenopausal women. Variable  R Square  Significant F  B  Beta (B)  Significant T  BMI (kg/rn ) 2  0.208  0.0017  0.896  0.464  0.0010  Age at Menarche (years)  0.282  0.0010  0.780  0.273  0.0434  Variables in the stepwise multiple regression analysis equation.  B) THE HIGH/LOW DIETARY RESTRAINT POPULA TION DIWDED BY PERCENTILES  To further analyze the factors relating to dietary restraint, the study population was divided into a highly restrained group (n=8) and a low restraint group (n=22). The highly restrained group consisted of those women who had scored above the eightieth percentile (score,  13/21) and the low restraint category contained those women who fell below the  fiftieth percentile (score,  7121) of the restraint subscale on the Three-Factor Eating  Questionnaire. When these two groups were compared using student’s unpaired t-test it was found that the low restraint group had a significantly lower mean BMI (21 .23±1.84 kg/rn 2 vs. 23.58 ± 2.43 kg/rn ; P< 0.01), more ovulatory cycles (5.14 ± 1 .25 cycles vs. 3.63 ± 2.33 2 cycles; P<0.05), a longer mean luteal phase length (9.88±2.87 days vs. 6.38±4.46 days; P<O.02), and a larger luteal phase index (0.393±0.066 vs. 0.330±0.077; P<O.05) than  their highly restrained colleagues. These results are displayed in Table 36 and in Figure 5. There were no significant differences between the two restraint groups in any of the nutrients analyzed, as is shown in Table 37. 117  RESUL TS  Table 36. Mean (± SD) characteristics of 45 premenopausal women, according to dietary restraint group (divided by percentiles). Characteristics  Low-restraint Group  Highly Restrained Group  Number of Women  22  8  Age (years)  26.8 ± 4.7  27.0 ± 6.3  0.932  Age at Menarche (yrs.)  13.1 ±1.6  14.1 ±1.5  0.116  Exercise (hrs/wk)  3.2±2.4  4.0±2.1  0.431  Exercise Regimen (yrs.)  4.9±5.6  3.8±3.3  0.604  Weight (kg)  58.7 ± 7.7  64.1 ±5.7  0.081  Height (cm)  166.1 ±9.7  165.1 ±6.2  0.786  Body Mass Index (kg/rn ) 2  21.2 ± 1.8  23.6 ± 2.4  0.00W  Sum of Skinfolds (mm)  89.2 ± 24.0  105.5 ± 28.5  0.1 30  Body Fat (%)  25.3 ± 5.5  28.9 ± 6.0  0.142  Total Body Fat (kg)  15.0±4.5  18.8±5.4  0.067  Total Muscle Mass (kg)  27.7 ±4.1  29.8 ±2.1  0.190  No. Ovulatory Cycles  5.1 ±1.3  3.6±2.3  0.029”  No. Short Luteal Phases  1.2 ± 1.2  1.3 ± 1.4  0.895  Avg. Cycle Length (days)  27.2 ±4.8  24.5 ± 4.9  0.184  Avg. Luteal Phase Lgth. (d.)  9.9±2.9  6.4±4.6  0.017”  Avg. Luteal Phase Index  0.39 ± 0.07  0.33 ± 0.08  0.034”  Bone Mineral Content (g/cm ) 2  63.78±8.17  64.01±11.68  0.952  1 (g/cm L ) 2  1.09±0.10  1.12±0.14  0.575  2 (g/cm L ) 2  1.19±0.11  1.22 ± 0.20  0.541  3 (glcm L ) 2  1.23±0.09  1.23 ±0.14  0.965  4 (g/cm L ) 2  1.19 ± 0.10  1.20±0.12  0.795  4 1 L L (g/cm ) 2  1.18±0.09  1.20 ± 0.14  0.676  4 2 L L (g/cm ) 2  1.20±0.10  1.22±0.15  0.742  Creatinine (mmol/L)  4.27 ± 3.42  3.83 ± 2.92  0.752  Calcium (mmol/L)  0.84 ± 0.86  0.51 ±0.46  0.324  Cortisol (nmol/L)  153±213  216±319  0.533  P value  “Low-restraint group are below the 50th percentile and highly restrained women scored above the 80th percentile. ““P<O.O1 +Comparisons made using student’s unpaired t-test.  ““P<O.05  118  (0  -I  —.  oc  “  I!  C, DCD  CD  Cii 0.•  CC  Co  p 01  A  a-  ICD  0  C.) CD  I C CD  a  ICD  CD  C)  C-)  cQ  D  CD —4 CD  -4, -I,  a  C-) Q) D  -4,  D  Ca  U,  *  *  w  .1_  H H  -D  C  0  -4  CD  +  0)  -4  CD C,)  -4  a.  :3-  Ca  2:  a-  D  z  5_  CD CD D  -  m  F—  2  2  —  Cl)  >  (I-)  -<  (DQ  RESUL TS  Table 37. Mean (± SD) total nutrient intakes for nine days, according to dietary restraint group (divided by percentiles). Daily Dietary Intakes  Lowrestraint Group (n=22)  Highly Restraint Group (n=8)  P value+  Calories (kcal)  2125±558  2019±484  0.637  Protein (g)  65.4±17.3  73.6 ± 24.6  0.311  Carbohydrates (gi  302.6 ± 80.4  286.6 ± 70.8  0.625  Fibre (g)  26.6±11.0  24.6±9.0  0.649  Total Fat (gi  77.4 ± 27.5  69.7 ± 16.7  0.468  Saturated Fat (g)  24.1 ±12.9  21.0±6.6  0.524  Monounsaturated Fat (g)  29.2 ± 10.7  26.9 ±8.3  0.584  Polyunsaturated Fat (g)  16.9±5.9  15.3±3.0  0.474  Cholesterol (mgi  188±105  176±98  0.781  Vitamin A (RE)  1753 ± 1262  1773±1435  0.970  Thiamin (mg)  1.55 ± 0.54  1.70±0.47  0.491  Riboflavin (mgi  1.57±0.38  1.67±0.52  0.590  Niacin (mgi  15.1 ±3.7  18.7±8.0  0.103  Vitamin B6 (mgi  1.60±0.53  1.87±0.61  0.245  Vitamin Bi 2 (jig)  2.11 ±1.60  2.84 ± 1.70  0.285  Folacin (jig)  335±132  302 ± 66  0.50 2  Pantothenic Acid (mgi  4.76 ± 1.22  5.21 ± 1.96  0.458  Vitamin C (mgi  153±59  119±42  0.147  Calcium (mgi  866±391  908±382  0.796  Copper (mgi  1.71 ±0.59  1.73±0.83  0.938  Iron (mgi  15.4±4.5  17.2±6.5  0.389  Magnesium (mg)  347±100  359±147  0.813  Phosphorus (mgi  1290±373  1462±490  0.314  Potassium (mgi  3145±883  3264 ± 1123  0.764  Sodium (mgi  2600±843  2227 ±824  0.291  Zinc (mgi  9.44±2.83  10.26±3.01  0.498  Protein (%)  12.5±3.2  14.3 ± 1.7  0.136  Carbohydrate (%)  56.3 ± 5.3  55.6 ± 3.9  0.748  Fat (%)  30.9 ± 5.7  30.2 ±3.9  0.736  +Comparisons were completed using student’s unpaired t-test.  1 20  RESUL TS  C) THE HIGH/LOW DIETARY RESTRAINT POPULA TION DIVIDED BY TERTILES When the study group was divided into tertiles: those who scored in the top tertile (n  =  1 3) and those who scored in the lowest tertile (n  =  1 2) on the restraint subscale of the  Three-Factor Eating Questionnaire, the highly restrained group had a higher age at menarche (13.92 ± 1 .19 years vs. 1 2.50 ± 1 .62 years; P< 0.02), a heavier mean weight (63.25 ± 7.26 kg vs. 55.39 ± 6.92 kg; P< 0.02), a larger BMI (23.05 ± 2.24 kg/rn 2 vs. 20.73 ± 1 .79 kg/rn ; 2 P<O.O1), more total body fat (17.31 ± 5.04 kg vs. 13.14±3.59 kg; P<0.05), and more total muscle mass (30.06 ±3.46 kg vs. 26.55 ±4.20 kg; P<0.05). These results are available in Table 38. In comparing the two groups’ dietary intakes, it was established that the less restrained group consumed more total fat (82.56±21.80 g vs. 65.92 ± 18.15 g; P<O.05) and more polyunsaturated fat (18.76±5.98 g vs. 14.33±4.04 g; P<0.05) than the highly restrained group. It is interesting to note that there was also an important difference in the percentage of the diet as protein betweenthe two groups (11.31 ± 2.77% vs. 13.30±2.13%; P=0.055), with the restrained group consuming less than their less restrained counterparts. The mean nutrients intakes are summarized in Table 39.  121  RESUL TS  Table 38. Mean (±SD) characteristics of 45 premenopausal women, according to dietary restraint group (divided by tertiles). P value+  Characteristics  Low-restraint Group  Highly Restrained Group  Number of Women  12  13  Age (years)  25.3 ± 3.8  27.9 ± 6.2  0.221  Age at Menarche (yrs.)  12.5±1.6  13.9 ± 1.2  0.019”  Exercise (hrs/wk)  3.1 ±2.6  4.2±2.0  0.228  Exercise Regimen (yrs.)  3.1 ±6.0  2.9±2.9  0.916  Weight (kg)  55.4 ± 6.9  63.3 ± 7.3  0.011’  Height (cm)  163.4±10.6  165.6 ± 6.6  0.548  Body Mass Index (kg/rn ) 2  20.7 ± 1.8  23.1 ±2.2  0.009’”  Sum of Skinfolds (mm)  81.7±20.3  96.6 ± 25.6  0.123  Body Fat (%)  23.6 ± 4.7  27.0±5.4  0.112  Total Body Fat (kg)  13.1 ±3.6  17.3±5.0  0.027W’  Total Muscle Mass (kg)  26.5 ± 4.2  30.1 ±3.5  0.03 1•  No. Ovulatory Cycles  5.1 ±1.2  4.5 ±2.1  0.383  No. Short Luteal Phases  1.1 ±1.1  1.2 ± 1.4  0.887  Avg. Cycle Length (days)  26.0 ± 5.7  25.3±4.1  0.730  Avg. Luteal Phase Lgth. Id.)  9.6±2.8  8.2±4.3  0.337  Avg. Luteal Phase Index  0.39 ± 0.06  0.37 ±0.09  0.540  Bone Mineral Content (g/cm ) 2  61.75 ± 6.89  63.84 ± 9.91  0.550  1 (g/cm L ) 2  1.11 ±0.09  1 .12 ± 0.14  0.808  ) 2 2 (g/cm L  1.21 ±0.11  1.21 ±0.18  0.950  3 (g/cm L ) 2  1.23 ± 0.09  1.22 ± 0.16  0.806  4 (gIcm L ) 2  1.21 ±0.11  1 .17 ± 0.14  0.497  ) 2 4 1 L L (g/cm  1.19±0.09  1.18 ± 0.15  0.819  ) 2 4 2 L L (g/cm  1.22±0.10  1.20 ± 0.16  0.722  Creatinine (rnmol/L)  3.70±2.10  3.36 ± 2.34  0.706  Calcium (mmol/L)  9.16±28.62  0.42 ± 0.39  0.282  Cortisol (nmol/L)  97±55  167±255  0.358  Low-restraint group fit into the lowest tertile and highly restrained women are in the highest tertile. “P<O.O1 “P<O.05 +Comparisons made using student’s unpaired t-test.  1 22  RESUL TS  Table 39. Mean (±SD) total nutrient intakes for nine days, according to dietary restraint group (divided by tertiles). Daily Dietary Intakes  Low-restraint Group (n  Calories (kcal)  2191 ±482  1923±461  0.168  Protein (g)  61.8± 14.1  65.3 ±22.7  0.645  Carbohydrates (9)  310.5±75.0  277.7 ±68.2  0.263  Fibre (9)  25.5± 11.6  24.7±7.6  0.854  Total Fat (g)  82.6±21.8  65.9± 18.2  0.049  Saturated Fat(g)  25.5±9.9  20.1 ±6.3  0.118  Monounsaturated Fat (g)  30.5 ±8.1  25.3 ±8.0  0.122  Polyunsaturated Fat (g)  18.8 ±6.0  14.3 ±4.0  O.039  Cholesterol (mgI  185 ±96  165 ±95  0.596  Vitamin A (RE)  1798 ± 1 227  1597 ± 1 175  0.679  Thiamin (mgI  1.48 ±0.46  1.50 ±0.47  0.897  Riboflavin (mgI  1.55±0.38  1.56±0.44  0.906  Niacin (mg)  14.4±3.1  17.0±7.0  0.254  Vitamin 86 (mgI  1.55 ±0.54  1.71 ±0.52  0.451  Vitamin Bi 2 (p9)  2.02 ± 1.48  2.36 ± 1.52  0.576  Folacin (jig)  320 ± 1 17  293 ± 55  0.459  Pantothenic Acid (mg)  4.66 ± 1.04  4.80 ± 1.68  0.66 1  Vitamin C (mg)  152±72  122±41  0.221  Calcium 1mg)  852±339  841 ±324  0.938  Copper (mg)  1.69 ±0.58  1.63 ±0.68  0.803  Iron (mgI  15.2 ±4.6  15.5±5.7  0.865  Magnesium (mgI  337± 103  332± 121  0.902  Phosphorus (mg)  1252 ±334  1317 ±434  0.681  Potassium (mgI  3104±1014  3088±917  0.967  Sodium (mgI  2629 ±766  2167 ±761  0.144  Zinc (mg)  9.15±2.42  9.33±2.77  0.866  Protein (%)  11.3±2.8  13.3±2.1  0.055  Carbohydrate (%)  55.7 ±5.7  66.4±4.5  0.727  Fat (%)  32.6 ±4.7  29.7 ±3.6  0.093  =  1 2)  Highly Restraint Group (n  =  13)  Significant difference at a level of P<O.05. +Comparisons were completed using student’s unpaired t-test.  1 23  P vaIue  RESUL TS  10. SUMMARY OF RESUL TS WITH REFERENCE TO THE STUDY H YPO THESES  Hypothesis 1: There would be no difference in the prevalence of subclinical menstrual dysfunction between vegetarian and nonvegetarian women.  The null hypothesis was generally confirmed as there were no significant differences in menstrual function between the vegetarian and the nonvegetarian premenopausal women. However, it is interesting to note that the vegetarian women consistently had mean values denoting more sound cycles than their nonvegetarian counterparts. That is to say that the vegetarian women experienced longer cycles lengths, longer luteal phase lengths, more ovulatory cycles and fewer short luteal phases than the nonvegetarian women. Furthermore, the P values for the difference in mean luteal phase lengths (P=O.058) and number of ovulatory cycles (P=O.072) approached significance.  Hypothesis 2: There would be no significant difference in the prevalence of subclinical menstrual disorders between restrained and unrestrained eaters.  This null hypotheses must be dismissed as there were significant differences in the number of ovulatory cycles, the mean luteal phase lengths, and the mean luteal phase indices when the women were divided into highly restrained (>80th percentile) and less restrained  124  RESUL TS  (<50th percentile).  Menstrual dysfunction increased with increasing restraint. However,  there were no significant differences in any of the menstrual cycle evaluations when the women were arranged into tertile groups.  Hypothesis 3: There would be no difference in spinal bone mineral densities between women with normal ovulatory and normal luteal phase length menstrual cycles and those with irregular cycles.  The null hypothesis was confirmed as there were no differences in any of the bone density evaluations or in the bone mineral content measures between normally ovulating women and women who had experienced one or more anovulatory cycle or more than one short luteal phase.  It is important to note however, that large inter-individual variability in  bone density among premenopausal women renders the detection of small differences unlikely.  Hypothesis 4: There would be no differences in spinal bone mineral densities between vegetarian and nonvegetarian premenopausal women.  This null hypothesis was substantiated as there were no differences in mean spinal bone mineral density measurements between the vegetarian women and the nonvegetarian  1 25  RESUL TS  women, when weight was entered into the analysis as a covariate. Initially, there was only a significant difference in bone mineral content, with the vegetarian women showing the lower mean, although when the equation controlled for weight this difference was removed.  Hypothesis 5: There would be no association between nutrient intakes and spinal bone mineral density results.  This null hypothesis must be repudiated as many nutrients were significantly correlated to the spinal bone mineral density measurements. Analysis showed that cholesterol, vitamin 12 and zinc were significantly associated with all bone measures, and additional nutrients B were individually related to discrete bone assessments.  Furthermore, when the bone  evaluations’ relationship to anthropometric, restraint, urine and menstrual data was assessed through stepwise multiple regression analysis, it was established that dietary factors always participated in the associations determined.  Moreover, vitamin B 12 was included in every  -L evaluation. 2 L instance with the exception of the 4  1 26  DISCUSSION  Chapter V  DISCUSSION  1. INTRODUCTION A brief review of the subjects involved in the study will introduce the discussion, focusing on differences in anthropometric, demographic and dietary data observed between the two subgroups. This will be followed by the major findings pertaining to each of the hypotheses: (i) the prevalence of subclinical menstrual dysfunction in vegetarian and nonvegetarian women; (ii) the prevalence of subclinical menstrual disorders in restrained and unrestrained eaters; (iii) the differences in bone density evaluations between women with normal ovulatory cycles and those with irregular cycles; (iv) the bone mineral densities of vegetarian and nonvegetarian women; and finally, (v) the association between nutrient intakes and spinal bone mineral density. Subsequently, the possible overall relationships established through this study will be reviewed and conclusions made.  2.SUBJECTS A) DEMOGRAPHIC AND ANTHROPOMETRIC DA TA Comparisons of the demographic data for the vegetarian and the nonvegetarian women demonstrated that the two groups were well matched with regard to age, age at menarche and exercise habits; however, the vegetarian women were found to be both leaner and to  127  DISCUSSION  , when 8 have a lower mean BMI than their nonvegetarian peers. Conversely, Pedersen et al. studying menstrual differences due to vegetarian and nonvegetarian diets, found their two groups to be indistinguishable with respect to BMI.  In the present study, although the  vegetarian women did report exercising more on average and consuming fewer calories than the nonvegetarian women, the differences were negligible and thus these anthropometric differences must be related to some other factor. It is possible that dietary restraint may have played a role. This will be examined in detail later. Regarding the vegetarian subgroups, the discussion section will only highlight those results that combined the lacto-ovo vegetarian women and the lacto-vegetarian women into one group. The two groups of vegetarian women, the vegan and the lacto vegetarians, were well matched for all the demographic and anthropometric variables measured.  The only  difference that approached significance was in the mean number of years each group had been practising a vegetarian diet. The vegans tended to have been vegetarians for a shorter period of time than their lacto vegetarian peers (3.1 ±2.2 years vs. 5.4±3.1 years; P=O.082). During their initial interview, when the vegetarian women were asked to classify  themselves into one of three vegetarian categories (vegan, lacto and lacto-ovo), many of them were hesitant as they had previously been one type of vegetarian and had converted or had the desire to convert to either a more strict or a more lax vegetarian diet. Although it was ensured, through screening, that all women had followed a similar diet for two years prior to the study, this classification merely pertained to the time they had been a vegetarian and not how long they had been adhering to their present vegetarian practices. Perhaps, with the vegan vegetarian diet being more restrictive, fewer women followed it or found it too difficult  128  DISCUSSION  to commit to over an extended period of time. With respect to the reasons for becoming a vegetarian, the findings were similar to . 34 those established by Sims  As in this study, she reported that reasons for practising  vegetarianism included ethical considerations, religious-philosophical convictions, health reasons, ecological concerns, and economic limitations. Similarly, in both Sims’s and the present study, health reasons were cited as the major explanatory factor in the decision to 34. adopt vegetarianism B) DIETARY INTAKES (i) Phase of the Menstrual Cycle , when examining the correlation of food intake to various stages of the 80 Lyons et al. menstrual cycle, had 18 healthy normally menstruating women weigh and record their food intake daily during one complete menstrual cycle.  Corresponding to these findings, the  comparison of dietary intakes over the three stages of the menstrual cycle in the present study did show that the women consumed extra energy in the luteal phase of the menstrual cycle. However, mean intakes of the macronutrients at various stages of the menstrual cycle ° Protein intake was greatest 8 varied greatly here in comparison to those found by Lyons et al. during ovulation here whereas that was the time that nutrient was at its lowest point in the ° 8 former study; fat intake was lowest during the follicular phase in this study, yet Lyons et al. found this took place during ovulation in their subjects; greater intakes of carbohydrate were found for both the luteal and the follicular phase when compared to the ovulatory stage in ° study, whereas the follicular phase demonstrated the lowest intake of that 8 Lyons et al.’s macronutrient in the present study. Furthermore, energy intake was at its lowest over the  129  DISCUSSION  follicular phase in this study rather than at the period coinciding with the midcycle surge in ° had found. 8 blood estrogen that occurs with ovulation as Lyons et al. 79 found energy intake to be significantly increased in the Similarly, however, Dalvit postovulatory phase of the menstrual cycle. She only considered caloric intake in the pre- and postovulatory stages though.  79 also suggested that estrogen may be an appetiteDalvit  suppressing hormone and that the change in this sex hormone’s concentration may account 79 The pattern of food for the observed variations in food intake over the menstrual cycle. intake appears to be inversely related to estrogen concentrations in the present study, and this hypothesis could thus explain these changes. Overall, these results suggest that phase of the menstrual cycle may be more important than days and should also be considered when food intake in women of reproductive age is assessed. (ii) Vegetarian and Nonvegetarian Intakes The overall differences in dietary intakes between the vegetarian and nonvegetarian subjects were in protein, dietary fibre, cholesterol and the percentages of calories from protein and carbohydrate. These differences are not surprising as the vegetarian diet tends to consist more of carbohydrates, particularly complex carbohydrates, and uses more complimentary protein sources than animal sources. Furthermore the decreased animal based food intakes in the vegetarian diet leads to the lower cholesterol and protein intakes seen in this study. , who looked at menstrual differences due to vegetarian and 8 Unlike Pedersen et al. nonvegetarian diets, no significant differences in total carbohydrate and polyunsaturated fat intakes were noted here, although the vegetarian women’s means did exceed those of their 8 findings however, the vegetarians’ nonvegetarian counterparts. Similar to Pedersen et al.’s  130  DISCUSSION  intakes of dietary fibre did surpass those of their nonvegetarian peers and their level of cholesterol intake was significantly lower.  29 also reported a difference in Pedersen et al.  magnesium intake. In the present study, the vegetarians did consume more magnesium than the nonvegetarians, and the difference approached significance (P=O.067). Interestingly, no differences in fat intakes were noted in this study, and moreover, the percent of calories in the diet from fat were identical in the two groups.  This was  unexpected, as it has previously been hypothesized that the vegetarian diet is both higher in 833 Furthermore, dietary fibre and lower in fat intake than the conventional nonvegetarian diet. 43 examined estrogen excretion patterns and plasma levels in ten vegetarian when Goldin et al. and ten omnivorous premenopausal women, dietary records indicated that the vegetarian women consumed less total fat than their omnivorous counterparts, although the difference was not significant (30% of total energy vs. 40% of total energy). It is possible however, that due to the current trends toward adopting a healthier lifestyle, all these women have chosen to limit their fat intake. Also, as the study recruitment was based solely on participant interest, it is possible that a recruitment bias arose, resulting in a nonvegetarian study population more concerned with health and diet than the norm. 33 investigated the changes in the diet from a mixed to a vegetarian Johansson et al. diet and examined the effects on nutrient intake, food choices, meal patterns and cooking methods.  Similar to what they noted, there was an increased relative consumption of  carbohydrates, particularly dietary fibre, and decreased consumption of protein and cholesterol in the vegetarian women in this study.  Furthermore, the vegetarian participants also  2 as did those in Johansson et B consumed greater amounts of vitamin C and less vitamin 1  131  DISCUSSION  33 investigation. al.’s  Because the vegetarian diet is based mainly on the consumption of  complex carbohydrates, fruit and vegetables, and less animal protein than the nonvegetarian conventional diet, these differences are not surprising. Other differences in nutrient intakes were found here that had not been noted previously. These included increased intakes of folacin and copper in the vegetarian group, and significantly lower intakes of riboflavin, niacin, sodium and zinc. These differences can be easily assigned to the different food choices and dietary components of the two groups. When the vegan vegetarian diet composition was compared to that of the lacto vegetarians, it was found that the latter group exceeded the vegan’s mean intakes for 12 and calcium, which would be anticipated considering the B saturated fat, riboflavin, vitamin , lacto vegetarians consume both animal source foods and dairy products which are rich in these nutrients. The vegans ingested significantly more dietary fibre, vitamin A, thiamin, niacin, vitamin B , folacin, pantothenic acid and carbohydrate as a percentage of total calories. 5 This is likely due to their increased consumption of fruits and vegetables, lentils, and grains compared to that of the lacto vegetarians who supply a portion of their diet from eggs and dairy products.  3. SUBCLINICAL MENS TRUAL DISTURBANCES Only 1 5 of the 45 participants (33%) maintained normal ovulatory menstrual cycles throughout the six months of study. A normal ovulatory menstrual cycle was defined as one in which ovulation occurred and that had a Iuteal phase of normal length (at least ten days) , when investigating spinal bone 10 over the six menstrual cycles studied. Similarly, Prior et al.  132  DISCUSSION  loss and ovulatory disturbances in premenopausaf women with differing levels of activity, found only 1 3 out of their 66 subjects (20%) maintained normal ovulatory menstrual cycles 8 examined menstrual over their one year study period, Interestingly, when Pedersen et al. differences due to vegetarian and nonvegetarian diets, they concluded that menstrual irregularity at a level of 4.9% was normal. However, their study relied solely on self-reported menstrual histories which were classified as regular, irregular, and amenorrheic.  No  restrictions were in place that examined the possibility of subclinical menstrual disturbances, and thus their definition of “normal” clearly differs from the criteria used to define a normal ovulatory menstrual cycle in the present study. In comparing the results for the prevalence of menstrual irregularities in this study ° study (80%) to the frequently quoted 5% of women with 1 (67%) and those from Prior et al.’s disturbed cycles (i.e., amenorrhea and oligomenorrhea), it seems feasible to conclude that the clinical definition of a “normal” menstrual cycle requires further investigation. It appears that the occurrence of occasional subclinical menstrual disturbances is normal in all women, considering more than half of this study’s sample population and more than three-quarters of ° sample population exhibited these tendencies. 1 Prior et al.’s  Further research on the  prevalence of subclinical menstrual disturbances and physiological consequences thereof is required before a clear clinical definition of a “normal” menstrual cycle pattern can be ° findings of no change in Quantitative 1 determined. From the bone perspective, Prior et al.’s Computed Tomography (QCT) bone density with one short luteal phase cycle, but a loss with more than one short luteal phase or one or more anovulatory cycles may be physiologically important in defining what is clinically “normal”.  133  DISCUSSION  There were no significant differences in the prevalence of subclinical menstrual disturbances in the vegetarian and nonvegetarian groups; however, the vegetarian women tended to have more sound cycles overall. The vegetarian women’s mean cycle lengths and luteal phase lengths were longer, their luteal phase index was greater, and they experienced more ovulatory cycles and less short luteal phase cycles than their nonvegetarian counterparts.  The differences in luteal phase length (P=O.058) and in the number of  ovulatory cycles (P=O.072) between the two groups approached significance. When considering the number of women with disturbed cycles, it was found that more of the nonvegetarian women had abnormal cycles (73% of the nonvegetarian women) than the vegetarian women (61 % of the vegetarian women). Furthermore, a larger number of the nonvegetarian women’s cycles were disturbed as well (39% of the nonvegetarian women’s cycles vs. 24% of the vegetarian women’s cycles). 8 The present study’s results contrast with previous reports. When Pedersen et al. examined the effects of different nutritional patterns on menstrual regularity (i.e., cycle length) in premenopausal vegetarian and nonvegetarian women, they found the vegetarian women to display significantly more menstrual cycle disturbances (26.5%) than their nonvegetarian counterparts  (4•9%)•8  7 reported on the effects of vegetarian Furthermore, when Pirke et al.  and nonvegetarian weight-reducing diets on the menstrual cycle of healthy, normal weight omnivorous women, they found similar results. Seven of the nine women they had placed on a vegetarian weight-reducing diet experienced anovulatory cycles, which they evaluated using plasma hormone measurement, whereas only two of the nine women placed on a s. 7 nonvegetarian weight-reducing diet displayed altered cycle characteristic  134  DISCUSSION  The limitations of the data from these studies are numerous, but most importantly: 8 did not examine subclinical menstrual disturbances and their study may have Pedersen et al. 7 not only looked at the diet in a weight reducing suffered from recruitment bias; Pirke et al. capacity but also used omnivorous women for their vegetarian subjects. The diet period in the latter study was only six weeks which does not allow time for physiological and nutritional adjustment to the new food consumption patterns seen with a change to a vegetarian diet. In the present study however, the vegetarian women had been practising a similar diet for at least two years prior to the start of the study, dietary intakes were examined over a total of nine days, the menstrual cycle data were collected over six consecutive menstrual cycles, and subclinical menstrual disturbances were monitored. Thus, it is likely that the menstrual cycle data from this study are more indicative of true menstrual regularity in vegetarian and nonvegetarian women. Further analysis performed on the menstrual cycles characteristics within the vegetarian subgroup demonstrated that the lacto vegetarian group (n  =  1 5) had a significantly  shorter mean cycle length than their vegan counterparts (n=8). Furthermore, the difference in luteal phase length also approached significance (P=O.098), being lower in the lacto vegetarians again, although the mean luteal phase index results were identical in the two groups. The differences observed in menstrual cycle characteristics within the vegetarian sample population seemed to conform to those seen between the vegetarian and nonvegetarian women.  The association of the degree of vegetarianism appeared to be  inversely related to menstrual cycle characteristics: as the diet more closely approximated a  135  DISCUSSiON  nonvegetarian conventional type, menstrual cycle characteristics tended toward disruption. In spite of these facts however, the luteal phase index results only followed this pattern when comparisons were made between the vegetarian and nonvegetarian women, and not when comparing the vegan and lacto vegetarian women. Furthermore, when the nutrients analysed were correlated to the menstrual cycle characteristics, the only significant associations found included the following positive correlations: (i) dietary fibre with cycle length and luteal length; (ii) vitamin A with the number of ovulatory cycles, cycle length and luteal length; (iii) folacin with luteal length and luteal phase index; and, (iv) iron with luteal phase length and luteal phase index. The mean intakes of these nutrients significantly differed when comparing the vegan and lacto vegetarian intakes, with the vegans’ exceeding those of their lacto vegetarian peers. However, only dietary fibre and folacin levels significantly differed when the vegetarian and nonvegetarian group intakes were compared. 6 studied the effect of doubling dietary fibre (from approximately 1 5 g to 30 Rose et al. g/day) in 62 premenopausal women with regular ovulatory menstrual cycles. The women were randomly assigned to supplement their diets with either corn, wheat or oat bran for two months. Compared to baseline, the women who received the wheat bran supplement had significantly lower levels of serum estradiol and estrone. They concluded that dietary fibre 6 can affect circulating estrogen concentrations and bioavailability. , when examining estrogen excretion patterns and plasma 43 Furthermore Goldin et al. levels in ten vegetarian and ten omnivorous premenopausal women, found that a positive correlation existed between fecal weight and fecal excretion of estrogens in both groups. The vegetarian women consumed significantly more dietary fibre per day than their omnivorous  136  DISCUSSION  counterparts, and were shown to have higher fecal weights and increased fecal excretions of estrogens. Again, these authors concluded that the greater fecal bulk and non-absorbed fibre 43 somehow shielded the estrogens excreted in the bile from deconjugation and reabsorption. The overall consensus appears to indicate that dietary fibre is inversely related to sex steroid levels. Although these cited studies do not examine the effects of dietary fibre on menstrual cycle characteristics per se, it is not unrealistic to speculate that a similar relationship would be manifested in menstrual cycle characteristics with increased intake of dietary fibre. The vegetarian women in the present study did consume significantly greater amounts of dietary fibre than their nonvegetarian peers. However, dietary fibre was found to be significantly and positively correlated with menstrual cycle characteristics, although it should be noted that this nutrient only entered a multiple regression equation for the number of ovulatory cycles and not for luteal phase length or luteal phase index. It is possible that the biological impact of dietary fibre on circulating sex steroid levels , as 41 previously demonstrated in these studies was mediated by a specific fibre subspecies 6 changes were observed with wheat bran but not with corn or oat bran in Rose et al.’s study. Unfortunately, as fibre embraces a variety of complex chemical structures, the dietary intake analysis program used offered limited information regarding the fibre content of foods 4282 providing only total dietary fibre measurements and no information on fibre subspecies. Thus, it is feasible that the differences in menstrual cycle characteristics observed between the vegetarian and the nonvegetarian women may be due in part to the effect of a specific fibre subspecies that could not be analyzed here. 16 indicated that large amounts of vitamin A ingested by amenorrheic Ulirich et al.  137  DISCUSSION  females may results in inhibited ovulation, an occurrence previously seen in rats.  This  relationship was ascribed to l’-carotene being deposited in fat tissue and in the corpus 16 Once again, this relationship was reversed here as those with the increased intakes Iuteum. of vitamin A were also those with the more sound menstrual cycles, and vitamin A entered the regression equations for both ovulatory cycles and luteal phase length. Furthermore, the vegetarians, who consumed greater amounts of vitamin A, were also found to have significantly lower body fat amounts (percent body fat and total body fat) than the nonvegetarian women. There has been no previous literature that has related either folacin or iron with menstrual function and thus the relationship remains unclear.  It is interesting to note  however, that the vegetarian women did consume significantly greater amounts of folacin than the nonvegetarian women, and this very nutrient was significantly and positively correlated with luteal phase length and luteal phase index, although it did not enter the regression equation. 1 age, as well as gynecologic age, were , 2 ’ 19 Similar to previously stated findings established as important determinants of menstrual function in all categories with the exception of cycle length. However, no differences in mean age or gynecologic age were seen either between the vegetarian and the nonvegetarian groups or within the vegetarian group itself. Thus, the difference in menstrual function appears to be related to another variable, or may act synergistically with a variety of other factors in establishing menstrual function. Interestingly, Schweiger et  j2l  stated that by a gynecologic age of six years a large  138  DISCUSSION  majority of women have ovulatory cycles with adequate luteal phases.  Yet, the present  study’s results showed that although the mean gynecologic age was approximately 14 years, a large majority of the participants did not consistently have ovulatory cycles with adequate luteal phases. Again, this emphasizes the need to reexamine the definition of a “normal” menstrual cycle. , when reviewing menstrual dysfunction in female athletes, 20 Noakes and Van Gend found that a lower percent body fat was correlated with a higher incidence of menstrual disturbances.  Unlike their findings, percent body fat was significantly and negatively  correlated with the mean luteal phase index here. Thus, as body fat decreased, luteal phase index would increase, indicating a longer luteal phase in contrast to the total cycle length and hence, a healthier cycle.  Although the vegetarian women had significantly lower mean  percent body fat values than their nonvegetarian peers, there was no significant difference in luteal phase indices between the two groups (P=O. 188). It is also important to note that the subjects participating in the present study displayed typical percent body fat evaluations ° 2 and none had very low values as might be seen in the athletes in Noakes and Van Gend’s review. From all the information provided through previous research and from the present study, the relationship between vegetarianism and the menstrual cycle remains to be elucidated. The results from this study appear to indicate the opposite of formerly reported findings  -  the vegetarian women in this study showed a tendency toward more normal  ovulatory cycles than their nonvegetarian counterparts. Furthermore, the relationship between dietary fibre and menstrual cycle irregularity was not detected here. If anything, the reverse  139  DISCUSSiON  relationship seemed more prominent. Overall, it is likely that the definition of a “normal” cycle requires reexamination, and that some additional factor, other than vegetarianism or a component of the diet, is contributing to the subclinical menstrual disturbances observed in these women. The only other factor related to any of the menstrual cycle characteristics was the level of dietary restraint, and this will be discussed in the following section.  4. DIETARY RESTRAINT A) BETWEEN VEGETARIANS AND NONVEGETARIANS  The results of the comparison between the vegetarian and the nonvegetarian women revealed that the vegetarian group had significantly lower total eating inventory scores and levels of dietary restraint than the nonvegetarian study group. No difference was noted in the scores for those individuals within the two vegetarian subgroups. Age, weight, BMI, total body fat and total skeletal muscle mass were all significantly and positively related to levels of dietary restraint.  This is logical as the nonvegetarian  women, who were found to have a significantly higher BMI and more total body fat than their vegetarian peers, also displayed signs of increased restraint. , when examining the food-related value-orientations, attitudes, and beliefs of 34 Sims vegetarians and nonvegetarians, noted that more vegetarians were satisfied with their current weight status than nonvegetarians. Furthermore, she found that the variables most positively related to vegetarian behaviour were the value-orientations of ethics, health and the belief in health foods whereas the value-orientations of social-psychological uses of food and 34 education, and weight control misconceptions were negatively related to vegetarianism.  140  DISCUSSION  Women with high scores for restraint are characterized by their psychological need to 4 Thus, these limit the amount of food they eat, and consciously eat less than they desire. restrained eaters may be experiencing increased stress in association with food intake. This is potentially relevant to the observed menstrual cycle differences between vegetarians and nonvegetarians, as central stress responses have been shown to be associated with increased 27 Shangold corticotropin-releasing hormone levels, which in turn can affect the steroid milieu. 19 postulated that chronic dietary restriction, as displayed in restrained eaters, is the et al. additional essential component necessary to convert normal menstrual cycles into more abnormal ones.  45 found that increasing levels of dietary Furthermore, Schweiger et al.  restraint resulted in shorter total cycle lengths as well as decreased luteal phase lengths. Similar findings were demonstrated in the highly restrained nonvegetarian women in this study. It appears that the nonvegetarian women in the present study were more concerned with weight control than their vegetarian peers, and thus may have experienced greater stress in association with food intake. It can be speculated that this increased stress disturbed their sex steroid milieu, thus affecting their menstrual cycle characteristics.  This is further  supported by the fact that the nonvegetarian women’s urinary cortisol values, a marker of stress, tended to be higher than that of their vegetarian peers (P=O.068). The cortisol to creatinine ratios were more similar between the two groups, but this ratio may tend to minimize any difference that existed, as urinary creatinine concentrations were inclined to be 95 Total skeletal muscle mass lower in the vegetarians, as has been reported elsewhere. evaluations of the two groups were very similar, suggesting that the tendency for higher  141  DISCUSSION  cortisol concentrations may actually reflect increased stress. Additional research, however, is needed to confirm or refute this suggestion. It is plausible that vegetarians tend to be less restrained in their dietary practices as they have already incorporated a form of restraint mechanism into their lives by adopting vegetarian practices. The moral and health restrictions they establish for themselves limit their dietary intake and food choices, and may thereby alleviate the stress experienced by the nonvegetarian women in association with eating. Overall therefore, it is conceivable that it is the food-related stress acting centrally to affect the menstrual cycle, rather than the vegetarian diet or any component thereof. Furthermore, it is likely this food-related stress that is contributing to the subclinical menstrual disturbances observed here. B) BETWEEN HIGH AND LOW DIETARY RESTRAINT GROUPS  To further examine these relationships, the study population was divided by levels of dietary restraint. When the groups were arranged by lower and upper tertiles, significant differences were found in age at menarche, weight, BMI, total body fat and total skeletal muscle mass means, but no differences were prominent in menstrual cycle characteristics. This may be due to the fact that the tertile groups were small, making detections in menstrual cycle characteristics difficult.  47 findings however, the more Similar to Tuschl et al.’s  restrained group had significantly greater weight, BMI, body fat and lean mass values despite tending to report lower caloric intakes. This may be due in part to decreased levels of energy 47 Furthermore, the restrained group consumed significantly less dietary fat than expenditure. the less restrained group in the present study, although the percent of energy as fat did not  142  DISCUSSION  differ between groups. 45 prospective cohort study on everyday eating Corresponding to Schweiger et al.’s behaviour and menstrual function in women, the present study’s population was also divided into a highly restrained group (>80th percentile) and a less restrained group (<50th percentile). The more restrained population was found to have a significantly higher BMI again. The relationship consistently appears and therefore must be relevant the body weight -  of restrained eaters is found to be elevated regardless of similar or slightly lower caloric intakes. This was established when comparing the vegetarian and nonvegetarian women, 47 when when comparing high and low restraint groups and was also noted by Tuschl et al. examining behavioural and biological correlates of restrained eating. The most meaningful results in this comparison however, were with regard to menstrual cycle characteristics: the more restrained group had significantly fewer ovulatory cycles, significantly shorter luteal phases and a significantly lower tuteal phase index. These 45 who found that increasing levels of results correspond to those noted by Schweiger et al. restraint resulted in shorter total cycle lengths as well as decreased luteal phase lengths. No differences in dietary intake were noted here though. A number of hypothesis regarding the relationship of menstrual function and dietary 32 restraint have been put forward and several are pertinent to this study. Schweiger et al. found that luteal phase function was more easily disrupted than estradiol production by inadequate nutrition, and that caloric intake and stress interact with exercise to bring about disturbances to luteal phase hormone production.  Although there were no significant  differences in levels of exercise or caloric intakes here, levels of cortisol to creatinine ratios,  143  DISCUSSION  an indication of stress, were significantly correlated to total eating inventory scores in the nonvegetarian more restrained group. Furthermore, according to the theory that individuals have a biologically determined set-point for weight, the restrained women consumed less total calories than required to maintain an “ideal” weight, although the difference in actual energy 4 intakes between the two restraint groups was negligible. 47 noted, restrained eaters permanently practice an inconspicuous form As Tuschl et al. of dieting. This was apt to be true of the highly restrained women in the present study as 19 speculated that chronic dietary restriction is the well. Furthermore, because Shangold et al. additional essential component necessary to convert a normal cycle to an abnormal one, as previously stated, it seems conceivable to conclude that dietary restraint contributes to the menstrual cycle differences seen in these women.  7 hypothesized that Also, Pirke et al.  intermittent dieting with alternating losses and gains of approximately 5% of body weight might be able to induce menstrual disturbances without leading to a weight deficit. Because highly restrained women are consciously controlling their food intake, it is plausible that they experience these body weight changes which in turn affected their menstrual status. Although only weight stable women were recruited for the present study, it is plausible that these small fluctuations could have gone unnoticed in the restrained participants here, and yet still play a role in menstrual status. 45 suggested that a possible mechanism contributing to the reduced Schweiger et al. caloric intake in restrained eaters may be the association of luteal phase length with increased energy expenditure.  The authors hypothesized that women with shortened luteal phase  lengths would therefore have decreased energy expenditure and develop restrained eating as  144  DISCUSSION  an adaptation to their predicament. 47 indicated, the energy expenditure in these restrained Furthermore, as Tuschl et al. women may have adapted to intermittent dieting or a constantly reduced energy supply, thereby resulting in lower caloric intakes without changes in body weight. It thus appears that these restrained women are trapped in a vicious circle: they feel compelled to constantly limit their food intake which results in small weights fluctuations and in turn, an adapted and reduced energy expenditure requiring even greater restraint to maintain or achieve a desired weight. It is likely, however, when examining the results of the two separate analyses of dietary restraint groups, that there are a number of factors, such as weight, stress and diet, that act synergistically in affecting menstrual status. As discussed, dietary restraint can result in altered menstrual cycle status, but can also act directly on bone through the increased cortisol levels associated with the stress response to food intake. These increased cortisol levels can act on bone by decreasing levels of vitamin . Although restraint has been 49 D, or by interfering with the osteoblast role in bone formation associated with increased levels of cortisol secretion, no differences in bone evaluations were apparent here. The highly restrained and less restrained women, whether divided by tertiles or percentiles, were found to have similar bone values in both instances in the present study. Perhaps, due to the large inter-individual variability seen in bone assessments, a larger sample population is required to be able to establish a distinction in this data, or perhaps bone evaluations conducted over time in a longitudinal study would provide more insight into this relationship. Previous research has only indicated bone values to be significantly different in  145  DISCUSSION  extreme eating disorder patients, such as in women with anorexia nervosa or bulimia , but none have examined the relationship of bone and levels of dietary restraint. 50 nervosa  5. BONE DENSITY AND THE MENS TRUAL CYCLE To further evaluate the influence of the menstrual cycle on bone mineral density, the study’s participants were divided into menstrual cycle groups.  The women who had  maintained normal cycles were grouped with those who had experienced only one short luteal phase cycle throughout the study and were defined as “normal”, and those women who had experienced more than one short luteal phase cycle or one or more anovulatory cycles were arranged into another category defined as “abnormal”. Comparisons of these two groups showed that they only differed in relation to their mean age and in their use of vitamin and mineral supplements. Those classified as “normal” were older and used vitamin and mineral supplements more frequently than did their “abnormal” counterparts.  Interestingly, it has  frequently been noted that women experiencing more irregular menses tend to be younger counterparts. 1.16.21.29 than their “normal” 1 The two groups were very similar with respect to bone values. When Prior et at)° studied spinal bone loss and ovulatory disturbances in runners and sedentary controls, they found the luteal phase index to be the strongest explanatory variable for the change in spinal bone density. Bone density evaluations were not found to be correlated to any menstrual cycle data here, including luteal phase index; however, perhaps if these women were followed over time and their bone densities reevaluated, a difference in the bone changes between the two groups would be detected.  Overall, though likely due to the large inter-individual  146  DISCUSS/ON  variability in bone measures, the cross-sectional nature of this aspect of the study, and the small sample size, the two groups were found to be similar in bone density evaluations. Although previous research has demonstrated that menstrual function during a specified time period appears to be related to changes in bone over that same time period°, the relationship of menstrual status with bone density was not apparent here. The present study however, obtained only a single measure for bone; thus data are not available with regard to bone change in these women. There does appear to be a synergistic effect of other factors such as diet and anthropometrics that play an important role in this association though. This postulate will be discussed further in the following section.  6. BONE DENSITIES OF VEGETARIAN AND NON VEGETARIA N WOMEN  No significant differences were found when comparing the bone mineral densities of both the vegetarian and the nonvegetarian women, as well as when comparing those of the vegan and the lacto vegetarians. Interestingly however, the vegetarian group consistently displayed lower mean bone values for all measured sites. These differences, however, were largely eliminated when body weight was entered into the analysis as a covariate. 66 compared urinary hormonal concentrations and spinal bone When Lloyd et al. densities of premenopausal vegetarian and nonvegetarian women, they found that the vegetarian women had higher intakes of dietary fibre and carbohydrates, lower protein intakes, and similar fat contents to those of the nonvegetarians’ three-day intakes. This generally agrees with the findings in the present study. Additionally, as established here, 66 found that although the difference was not significant, bone density values Lloyd et al.  147  DISCUSSION  tended to be lower in the vegetarian women. These findings were reproduced in the present 66 reported menstrual cycle disturbances occurred study, with the exception that Lloyd et al. more frequently in the vegetarian group, while the prevalence was more common in the nonvegetarian group in this study. The present study also found that the percentages of total energy as carbohydrate and fat were significantly correlated to the bone data, in a negative and a positive direction respectively, however dietary fibre was not significantly correlated to any of the bone measures and protein was only significantly and positively associated with bone mineral content. 61 findings, body composition factors were significantly Similar to Heinrich et al.’s correlated to bone density  -  total skeletal muscle mass.  weight, BMI, sum of skinfolds, percent and total body fat and It is interesting to note that many of the correlated body  composition variables are the same as those that originally differed between the vegetarian and nonvegetarian subjects: BMI, sum of skinfolds, percent body fat and total body fat. The fact that these factors significantly differed between the two subgroups and were positively correlated to the bone evaluations, along with the lower bone values seen in the vegetarian women suggests that these variables act synergistically with dietary factors in affecting bone density. This will be discussed further in the following section.  7. NUTRIENT INTAKES AND BONE DENSITY  2 and zinc were significantly and positively correlated with all B Cholesterol, vitamin 1 the bone data. Furthermore, the percent of the diet as carbohydrate was significantly but negatively correlated to the bone values with the exception of bone mineral content.  148  DISCUSSION  Interestingly, intakes of these same nutrients differed significantly between the vegetarian and the nonvegetarian women, with the latter group exceeding vegetarians’ intakes for 2 and zinc, but consuming less calories as carbohydrates. Thus, the B cholesterol, vitamin 1 2 and zinc , 1 relationship of these values suggests that as consumption of cholesterol, vitamin B increase so will bone mineral density, yet as the proportion of the diet from carbohydrate increases bone mineral density decreases.  Correlation results, however, do not establish  causation, and these differences may not have a direct influence on bone density. 2 and zinc in bone formation have yet to be , 1 Possible roles of cholesterol, vitamin B 39 found that dietary carbohydrate was negatively elucidated. However, Leuenberger et al. associated with bone density, independently of sex steroid hormones. They concluded that low calcium intakes or decreased bioavailability, resulting from the presence of certain types of carbohydrates, including dietary fibre and organic acids, could account for a deleterious effect on bone density. In the present study, the vegetarian women had significantly greater fibre intakes than their nonvegetarian peers, but fibre was not correlated with bone density. Calcium intakes, while somewhat lower in the vegetarian women, did not differ significantly between groups, and were also not consistently related to bone. Furthermore, bone density did not differ significantly between these two dietary groups, although the values tended to be lower in the vegetarian women. Although the vegan vegetarian subjects consumed significantly more dietary fibre and significantly less calcium than their lacto vegetarian counterparts, their bone mineral density evaluations were similar. This is perhaps due to the small sample size of these subgroups in addition to the difficulty in detecting small differences in bone density due to the large inter  149  DISCUSSION  individual variabilities.  Moreover, these women had been following the vegan diet for a  relatively short time in comparison to their age. Further investigation through longitudinal studies is needed to clarify this relationship. In the present study, there does not appear to be a relationship among dietary fibre, calcium and bone content, and though the vegetarian women displayed lower overall mean bone mineral densities, more research examining changes in bone content is required to further define this relationship between vegetarianism and bone density.  8. POSSIBLE RELA TIONSHIPS ELUCIDA TED FROM THIS RESEARCH  Two main associations appear to have been established through this study. One is related to the relationship of dietary restraint and the menstrual cycle and the other is related to vegetarianism and bone. These will be summarized in the following paragraphs. A) DIETARY RESTRAINT AND THE MENS TRUAL CYCLE  Although none of the dietary variables was significantly correlated to menstrual cycle characteristics, and no significant differences were detected in the menstrual cycle characteristics of vegetarian and nonvegetarian women, there was a tendency for the vegetarian women to have more stable menstrual cycles.  This finding was contrary to  previous findings. 78 A brief review of the significant differences observed between the different groups in the present study will provide groundwork for the following theory.  The nonvegetarian  women had higher BMI and percent body fat evaluations than the vegetarian women. Furthermore, the nonvegetarians were significantly more restrained. Moreover, these same  150  DISCUSSION  women tended to have less sound cycles than their vegetarian peers. Percent body fat was inversely correlated to menstrual cycle function, thus as percent body fat increased, as with the nonvegetarian women, the menstrual cycles became more disturbed. Furthermore, total body fat and BMI were positively correlated to dietary restraint. Accordingly, as BMI increased, such as was established in the nonvegetarian group, levels of dietary restraint also increased. Dietary restraint is accompanied by heightened weight awareness which may in turn increase stress in association with eating. Although urinary cortisol excretion results were inconclusive, cortisol excretion was certainly not lower among the nonvegetarian women and may have been somewhat higher. Higher cortisol excretion, reflective of increased stress, may have disturbed the steroid milieu and affected ovulatory function in these women. Furthermore, when the women were divided into restraint groups, the more restrained women once again exhibited higher BMI as well as significantly altered menstrual cycle characteristics. Dietary restraint results in constant control of food intake, the perception of chronic energy restrictions, and may be associated with small weight fluctuations.  This  inadequate nutrition has been shown to disrupt luteal function without affecting sex steroids 19 Luteal phase analyses were significantly different between restraint to as great an extent. groups and the differences between dietary groups approached significance. There were no differences noted in bone values when the women were divided by dietary restraint levels or by menstrual function data.  It is possible that the inadequate  nutrition prominent in women with high levels of dietary restraint is such that it disrupts the luteal function without affecting the sex steroids to a large enough extent to observe a  151  DISCUSSION  difference in bone mineral density values. Furthermore, this could account for the fact that although the nonvegetarian women were significantly more restrained than the vegetarian women, their menstrual functions did not differ significantly. Alternately, there may be real effects on bone density when studies are conducted over time. Information on the stability of the degree of dietary restraint is also needed, as differences in bone density would be anticipated only if restraint remained at a given level. Overall, it appears that the menstrual cycle disturbances seen in this study resulted from a synergistic effect of dietary restraint and anthropometric variables on menstrual function. It remains unclear which of these factors could have initially influenced the others: did the women have altered menstrual cycle characteristics, including shorter luteal phase lengths, to begin with, which resulted in decreased energy expenditure, altering body composition and thus increasing dietary restraint to maintain or achieve their ideal body weight? Or, were the women restrained to begin with due to social pressures concentrated on body image, which resulted in chronic dieting and small fluctuations in weight, thus decreasing energy expenditure, altering body composition, increasing the stress associated with food intake, increasing levels of cortisol and altering menstrual cycle characteristics? This relationship requires further investigation to establish the exact links occurring here. B) VEGETARIAN/SM AND BONE When regression analysis was performed on the bone density measures, one of the body composition values and one of the nutrients were always identified as the most important variables in the relationship. Although the bone density value difference between the vegetarian and the nonvegetarian women were not significant they were consistent and  152  DISCUSSION  potentially important, with the vegetarians displaying lower means on all bone density considerations. Additionally, when the fact that large inter-individual variabilities exist making small differences difficult to detect is taken into account, it seems plausible that the difference in bone density noted is worthy of additional investigation. BMI, sum of skinfolds, percent and total body fat were all found to be positively correlated with bone density evaluations. These same anthropometric variables significantly differed between the vegetarian and nonvegetarian women, with the former group displaying the lower means in the present study. 2 and zinc were positively correlated to bone B Furthermore, cholesterol, vitamin 1 density. The vegetarian women were shown to consume significantly lower amounts of these nutrients than their nonvegetarian peers. Additionally, the proportion of total calories as carbohydrate was negatively correlated to bone density, and consumption of this nutrient was significantly greater in the vegetarian group. The correlation of cholesterol, vitamin  812,  zinc and percentage of calories as  carbohydrate to bone density appears to highlight the relationship between certain nutrients 39 found that carbohydrate was inversely associated with and bone density. Leuenberger et al. lower bone density values independent of sex steroids, insinuating that there may be a direct link between this nutrient and bone. Furthermore, they concluded that low calcium intakes, or decreased bioavailability of calcium could account for the deleterious effects on bone. Calcium bioavailability was said to be affected by the presence of certain types of carbohydrates, including dietary fibre and organic acids. The calcium intakes of the vegetarian women tended to be lower than those of the nonvegetarian women, and the vegetarian  153  DISCUSSION  women also ingested significantly greater amounts of dietary fibre. It is therefore possible that the tendency for lower bone values demonstrated in the vegetarian women could be a result of their increased carbohydrate and fibre intakes which interfered with the bioavailability of calcium, thereby affecting bone density without affecting the sex steroid milieu or menstrual function.  Furthermore, their decreased intakes of calcium likely intensified the  effect of components of the diet, perhaps such as carbohydrate and dietary fibre, on bone density. Overall, the relationship of vegetarianism and bone data remains to be elucidated. Further research is needed on changes in bone density and the association of the changes to diet.  9. IMPLICA TIONS FOR FUTURE RESEARCH  The implication for future research stemming from the present study’s findings are the following: 1. Further research in the area of subclinical menstrual disorder prevalence in the female population is required to properly define and establish criteria determining truly normal menstrual cycle characteristics. 2. More prospective research is needed in the area of dietary restraint and bone, to establish whether the menstrual cycle disturbances exhibited by highly restrained women are predisposing them to increased risk of bone loss and osteoporosis in later life. 3. Again, more research in the area of the effect of particular components of the  154  DISCUSSION  vegetarian diet on bone is essential to clarify the relationship between diet and bone. 4. As research has shown insulin to be positively correlated with ovulation in , and as vegetarians have been shown to consume greater amounts of 96 animals carbohydrate as well as displaying healthier menstrual cycles, the role of insulin in vegetarians and its subsequent effect on menstrual function should be investigated.  10. STUDY DRAWBACKS  Although this study was well controlled and executed, there are a few changes that would be highly recommended in future endeavours in this area. These include: 1. The database used for analysing dietary intakes did not provide any information regarding fibre subspecies, which, had it been available, might have provided some insightful information on the role of fibre on the menstrual cycle, the sex steroid milieu, and on bone density. Furthermore, the database did not provide information on vitamin D intake either, which is a nutrient of particular interest when comparing vegetarian and nonvegetarian diets, and is also directly involved in calcium absorption in bones. 2. Had the original urine volumes been available, insightful information regarding the creatinine excretion levels in vegetarian and nonvegetarian might have been established, and cortisol excretion data may have been interpreted more clearly.  155  CONCLUSION  CONCLUSION  This study initially set out to determine if there was a relationship among vegetarianism, the menstrual cycle, dietary restraint and bone mineral density. Overall, one relationship was established through this investigation: that of dietary restraint and subclinical menstrual disturbances. The results from this research support the concept that dietary restraint results in heightened subclinical menstrual disturbances. Furthermore, anthropometric measurements were shown to be positively correlated with dietary restraint and subclinical menstrual disturbances. It is therefore feasible that dietary restraint, stress and anthropometric variables act synergistically in affecting menstrual function.  The exact relationships remain to be  determined however. Additionally, it was noted that vegetarian women tended to have slightly lower bone densities than their nonvegetarian counterparts. This difference in bone densities was largely eliminated however, when weight was entered into the analysis as a covariate.  The  relationship of certain components of the vegetarian diet, such as dietary fibre and carbohydrate, affecting bone still requires investigation however. Both of these associations, dietary restraint on subclinical menstrual cycle disturbances and vegetarianism and bone density, have meaningful nutritional implications. On one hand, controlling dietary restraint would require a complete modification in society’s manner in approaching body image and food.  However, nutritional intervention for the affiliation of  components of the diet and bone density is a more practical and workable objective. If, in 156  CONCL US/ON  fact, vegetarians are at increased risk of bone loss and osteoporosis, it is imperative that intervention and education be arranged to arrest any further bone loss. It would be beneficial to at least to have useful information regarding this relationship in place and available to the vegetarian population. It is important to note however, that with the current trends toward adopting a healthier lifestyle, many people are increasing their carbohydrate intake and decreasing fat intakes so as to achieve a more sound lifestyle. Thus, it may well be best to target the whole population, providing specific information to dietary restraint, menstrual cycle characteristics and bone health.  1 57  REFERENCES  REFERENCES 1. Goldin, B.R., Adlercreutz, H., Gorbach, S.L., Woods, M.N., Dwyer, J.T., Conlon, T., Bohn, E., Gershoff, S.N. The relationship between estrogen levels and diets of Caucasian American and Oriental immigrant women. American Journal for Clinical Nutrition 44:945-953, 1 986. 2. Barr, S.l. Nutrition knowledge of female varsity athletes and university students. Journal of the American Dietetic Association 87:1660-1664, 1987. 3. Stunkard, A.J., Messick, S. The Three-Factor Eating Questionnaire to measure dietary restraint, disinhibition and hunger. Journal of Psychosomatic Research 29(1):71-83, 1985. 4. Herman, C.P., Mack, D. 43:647-660, 1975.  Restrained and unrestrained eating.  Journal of Personality  5. Rose, D.P., Boyar, A.P., Cohen, C., Strong, L.E. Effect of a low-fat diet on hormone levels in women with cystic breast disease. I. Serum steroids and gonadotropins. Journal of the National Cancer Institute 78(4):623-626, 1 987. 6. Rose, D.P., Goldman, M., Connoly, J.M., Strong, L.E. High-fibre diet reduces serum estrogen concentrations in premenopausal women. American Journal of Clinical Nutrition 54:520-525, 1991. 7. Pirke K.M., Schweiger, U., Laessle, R., Dickhaut, B., Schweiger, M., Waechtler, M. Dieting influences the menstrual cycle: vegetarian versus nonvegetarian diet. Fertility and Sterility 46(6):1083-1088, 1986. 8. Pedersen, A.B., Bartholomew, M.J., Dolence, L.A., Aljadir, L.P., Netteburg, K.L., Lloyd, T. Menstrual differences due to vegetarian and nonvegetarian diets. American Journal of Clinical Nutrition 53:879-885, 1991. 9. Clarvitt, S.R. Stress and menstrual dysfunction in medical students. 29(4):404-409, 1988.  Psychosomatics  10. Prior, J.C., Vigna, Y.M., Schechter, M.T., Burgess, A.E. Spinal bone loss and ovulatory disturbances. New England Journal of Medicine 323:1221-1 227, 1990. 11. Drinkwater, B.L., Bruemner, B., Chesnut, C.H. Menstrual history as a determinant of current bone density in young athletes. Journal of the American Medical Association 263(4):545-548, 1990. 1 2. Treasure, J.L., Russell, G.F.M. Reversible bone loss in anorexia nervosa. British Medical Journal 295:474-475, 1987. 1 58  REFERENCES  The possible involvement of endogenous opioid peptides and 1 3. De Crée, C. catecholestrogens in provoking menstrual irregularities in women athletes. International Journal of Sports Medicine 11 (5):329-348, 1 990. 14. Hill, P., Garbaczewski, L., Helman, P., Huskisson, J., Sporangisa, E., Wynder, E.L. Diet, lifestyle, and menstrual activity. American Journal of Clinical Nutrition 33:1192-1198, 1980. 1 5. Vollman, R.F. The menstrual cycle. Major problems in obstetrics and gynecology. W.B. Saunders Co.: Philadelphia, Vol. 7, 1977. 1 6. Ullrich, l.H. Athletic menstrual disorders. West Virginia Medical Journal 83:604-609, 1987. 17. Lein, A. The cycling female Francisco, 1979.  -  her menstrual rhythm.  W.H. Freeman and Co.: San  1 8. Lloyd, T., Myers, C., Buchanan, J.R., Demers, L.M. Collegiate women athletes with irregular menses during adolescence have decreased bone density. Obstetrics and Gynecology 72(4):639-642, 1988. 19. Shangold, M., Rebar, R.W., Wentz, A.C., Schiff, I. Evaluation and management of Journal of the American Medical Association menstrual dysfunction in athletes. 263(12):1665-1 669, 1990. 20. Noakes, T.D., Van Gend, M. Menstrual dysfunction in female athletes. South African Medical Journal 73:350-355, 1988. 21. Schweiger, U., Schwingenschloegel, M., Laessle, R., Schweiger, M., Pfister, H., Pirke, K., HoehL C. Diet-induced menstrual irregularities: effects of age and weight loss. Fertility and Sterility 48(5):746-751, 1987. 22. De Crée, C. Endogenous opioid peptides in the control of the normal menstrual cycle and their possible role in athletic menstrual irregularities. Obstetrical and Gynecological Survey 44(1O):720-732, 1989. 23. Baker, E., Demers, L. Menstrual status in female athletes: correlation with reproductive hormones and bone density. Obstetrics and Gynecology 72(5):683-687, 1988. 24. Li, T.C., Cooke, l.D. Evaluation of the luteal phase. Human Reproduction 6(4):484-499, 1991. 25. Prior, J.C., Vigna, Y.M., Schulzer, M., Hall, J.E., Bonen, A. Determination of luteal phase length by quantitative basal temperature methods: Validation against midcycle LH peak. Clinical and Investigative Medicine 13(3):123-131, 1990. 1 59  REFERENCES  26. Baer, J.T., Taper, L.J. Amenorrheic and eumenorrheic adolescent runners: dietary intake and exercise training status. Journal of the American Dietetic Association 92(1)89-91, 1992. 27. Hotta, M., Shibasaki, T., Masuda, A., Imaki, T., Demura, H., Ling, N., Shizume, K. The responses of plasma adrenocorticotropin and cortisol to corticotropin-releasing hormone (CRH) and cerebrospinal fluid immunoreactive CRH in anorexia nervosa patients. Journal of Clinical Endocrinology and Metabolism 62(2):3 1 9-324, 1 986. 28. Frisch, R.E. (Abstract)  Amenorrhea, vegetarianism, and/or low fat. The Lancet 1:1024, 1984.  29. Fisher, E.C., Nelson, M.E., Frontera, W.R., Turksoy, R.N., Evans, W.J. Bone mineral content and levels of gonadotropins and estrogens in amenorrheic running women. Journal of Clinical Endocrinology and Metabolism 62(6):1232-1236, 1986. 30. Fardy, H.J. Women in sport. Australian Family Physician 17(3):183-186, 1988. 31. Pirke, K.M., Schweiger, U., Broocks, A., Tuschl, R.J., Laessle, R.G. Luteinizing hormone and follicle stimulating hormone secretion patterns in female athletes with and without menstrual disturbances. Clinical Endocrinology 33:345-353, 1990. 32. Schweiger, U., l—lerrman, F., Laessle, R., Riedel, W., Schweiger, M., Pirke, K-M. Caloric intake, stress, and menstrual function in athletes. Fertility and Sterility 49(3):447-450, 1988. 33. Johansson, G., Callmer, E., Gustafsson, J-A. Changing from a mixed diet to a Scandinavian vegetarian diet: effects on nutrient intake, food choice, meal pattern and cooking methods. Euronean Journal of Clinical Nutrition 46:707-716, 1992. 34. Sims, L.S. Food related value-orientations, attitudes, and beliefs of vegetarians and nonvegetarians. Ecology of Food and Nutrition 7:23-35, 1978. 35. Adlercreutz, H. Effect of diet and exercise on hormones: implications for monitoring training in women. Clinical Journal of Snorts Medicine 1(3):149-153, 1991. 36. Woods, M.N., Gorbach, S.L., Longcope, C., Goldin, B.R., Dwyer, J.T., Morrill-LaBrode, A. Low-fat, high-fiber diet and serum estrone sulfate in premenopausal women. American Journal of Clinical Nutrition 49:1179-1183, 1989. 37. Hill, P., Garbaczewski, L., Haley, N., Wynder, E.L. Diet and follicular development. American Journal of Clinical Nutrition 39:771-777, 1984. 38. Calberg, K.A., Buckman, M.T., Peake, G.T., Riedsel, M.L. A survey of menstrual function in athletes. Euronean Journal of Arplied Physiology 51:211-222, 1983. 160  REFERENCES  39. Leuenberger, P.K., Buchanan, J.R., Myers, C.A., Lloyd, T., Demers, L.M. Determination of peak trabecular bone density: interplay of dietary fiber, carbohydrate, and androgens. American Journal of Clinical Nutrition 50:955-961, 1989. 40. Watkin, V.A., Myburgh, K.H., Noakes, T.D. Low nutrient intake does not cause the menstrual cycle interval disturbances seen in some ultramarathon runners. Clinical Journal of Snorts Medicine 1(3):154-161, 1991. 41. Lloyd, T., Buchanan, J.R., Bitzer, S., Waldman, C.J., Myers, C., Ford, B.G. Interrelationships of diet, athletic activity, menstrual status, and bone density in collegiate women. American Society for Clinical Nutrition 46:681-684, 1987. 42. Barr, S.l., Kwan, S., Janelle, K.C. Nutrient analysis using computer programs: comparison of a Canadian and an American database. Journal of the Canadian Dietetic Association. (in press for 1 994) 43. Goldin, B.R., Adlercreutz, H., Gorbach, S.L., Warram, J.H., Dwyer, J.T., Swenson, L., Woods, M.N. Estrogen excretion patterns and plasma levels in vegetarian and omnivorous women. The New England Journal of Medicine 307(25):1542-1547, 1982. 44. Hagerty, M.A., Howie, B.J., Tan, S., Shultz, T.D. Effect of low- and high-fat intakes on the hormonal milieu of premenopausal women. American Journal of Clinical Nutrition 47:653659, 1988. 45. Schweiger, U., Tus’hl, R.J., Platte, P., Broocks, A., Laessle, R.G., Pirke, K.M. Everyday eating behaviour and menstrual function in young women. Fertility and Sterility 57(4):771775, 1992. 46. Nieman, D.C. Vegetarian dietary practices and endurance performance. American Journal of Clinical Nutrition 48:754-761, 1988. 47. Tuschl, R.J., Laessle, R.G. Kotthaus, B.C., Pirke, K.M. Behavioral and biological correlates of restrained eating. Annals of the New York Academy of Science 575:580-581, 1989. 48. Gindoff, P.R., Ferin, M. Endogenous opioid peptides modulate the effect of corticotropin releasing factor on gonadotropin release in the primate. Endocrinology 1 21 (3):837-842, 1987. 49. Heaney, R.P. Thinking straight about calcium. The New England Journal of Medicine 328(7):503-505, 1993.  161  REFERENCES  50. Joyce, J.M., Warren, D.L., Humphries, L.L., Smith, A.J., Coon, J.S. Osteoporosis n women with eating disorders: comparison of physical parameters, exercise, and menstrual status with SPA and DPA evaluation. The Journal of Nuclear Medicine 31 (3):325-331, 1990. 51. Cann, C.E., Martin, M.C., Genant, H.K., Jaffe, R.B. Decreased spinal mineral content in amenorrheic women. Journal of the American Medical Association 251(5):626-629, 1984. 52. Davies, M.C., Hail, M.L., Jacobs, H.S. Bone mineral loss in young women with amenorrhea. British Medical Journal 301:790-793, 1990. 53. Snead, D.B., Weftman, A., Weitman, J.Y., Evans, W.S., Veidhuis, J.D., Varma, M.M., Teates, C.D., Powling, E.A, Rogol, A.D. Reproductive hormones and bone mineral density in women runners. Journal of Applied Physiology 72(6):2149-2156, 1992. 54. Risser, W.L., Lee, E.J., Leblanc, A., Poindexter, H.B.W., Risser, J.M.H., Schneider, V. Bone density in eumenorrheic female college athletes. Medicine and Science in Sports and Exercise 22(5):570-574, 1990. 55. Recker, R.R., Davies, K.M., Hinders, S.M., Heaney, R.P., Stegman, M.R., Kimmel, D.B. Journal of the American Medical Association Bone gain in young adult women. 268(17):2403-2408, 1992. 56. Corvol, M., Blanchard, 0., Tsagris, L. Bone and cartilage responsiveness to sex steroid hormones. Journal of Steroid Biochemistry and Molecular Biology 43(5):415418, 1992. 57. Frederick, L., Hawkins, S.T. A comparison of nutrition knowledge and attitudes, dietary practices, and bone densities of postmenopausal women, female college athletes, and nonathietic college women. Journal of the American Dietetic Association 92(3):299-305, 1992. 58. Prior, J.C. Progesterone as a bone-trophic hormone. Endocrine Review 1990.  11:386-398,  59. Sowers, M.R., Shapiro, B., Gilbraith, M.A., Jannausch, M. Health and hormonal characteristics of premenopausal women with lower bone mass. Calcified Tissue International 47:130-1 35, 1990. 60. Vico, L., Prallet, B., Chappard, D., Pallot-Prades, B., Rupier, R., Alexandre, C. Contributions of chronological age, age at menarche, and menopause, and anthropometric parameters to axial and peripheral bone densities. Osteoporosis International 2(3):1 53-1 58, 1992.  1 62  REFERENCES  61. Heinrich, C.H., Going, S.B., Pamenter, R.W., Perry, C.D., Boyden, T.W., Lohman, T.G. Bone mineral content of cyclically menstruating female resistance and endurance trained athletes. Medicine and Science in Sports and Exercise 22(5):558-563, 1990. 62. Snow-Harter, C., Bouxsein, M.L., Lewis, B.T., Carter, D.R., Marcus, R. Effects of resistance and endurance exercise on bone mineral status of young women: a randomized exercise intervention trial. Journal of Bone and Mineral Research 7(7):761-769, 1992. 63. Cann, C.E., Cavanaugh, D.J., Schnurpfiel, K., Martin, M.C. Menstrual history is the primary determinant of trabecular bone density in women runners. Medicine and Science in Sports and Exercise 20(2 Supp.)4:S59, 1988. (Abstract) Does calcium supplementation prevent 64. Riis, B., Thomsen, K., Christiansen, C. postmenopausal bone loss? The New England Journal of Medicine 316(4):173-177, 1987. 65. Reid, l.R., Ames, R.W., Evans, M.C., Gamble, G.D., Sharpe, S.J. Effect of calcium supplementation on bone loss in postmenopausal women. The New England Journal of Medicine 328:460-464, 1993. 66. Lloyd, T., Schaeffer, J.M., Walker, MA., Demers, L.M. Urinary hormonal concentrations and spinal bone densities of premenopausal vegetarian and nonvegetarian women. American Journal of Clinical Nutrition 54:1005-1010, 1991. 67. Brown, B.W.J., Hollander, M. Statistics. A Biomedical Introduction. New York: John Wiley & Sons, 1977. 68. Yates, A., Shisslak, C.M., Allender, J., Crago, M., Leehey, K. Comparing obligatory to nonobligatory runners. Psychosomatics 33(2)1 80-1 89, 1 992. 69. Medda, F., Nieddu, R., Fiore, R., Marcello, C., Lecca, U. Hormonal profile and androgen status during the menstrual cycle in women with acne. Clinical and Experimental Obstetrics and Gynecology 15(4):137-142, 1988. 70. Dikovics, A. Nutritional Assessment Philadelphia, 1987.  -  Case study methods.  George F. Stickley Co.:  71. Keys, A. (Chairman) Recommendations concerning body measurements for the characterization of nutritional status. Human Biology 28(2):111-123, 1956. 72. Health Promotion Directorate. Canadian guidelines for healthy weights. Health and Welfare: Ottawa, 1988.  Minister of  73. Gibson, R.S. Principles of Nutritional Assessment. Oxford University Press: New York, 1990. 1 63  REFERENCES  74. Roche, A.F., Siervogel, R.M., Chumlea, W.C., Reed, R.B., Valadian, I., Eichorn, D., Monographs in Pediatrics: Serial Chancies in Subcutaneous Fat McCammon, R.W. Thicknesses of Children and Adults. S. Karger: Basel, Vol. 1 7, 1 982. 75. Lohman, T.G., Roche, A.F., Martorell, R. (eds.) Anthropometric Standardization Reference Manual Human Kinetics Books: Champaign, IL., 1988, pp.39-71. 76. Jackson, A.S., Pollock, M.L., Ward, A. Generalized equations for predicting body density of women. Medicine and Science in Sports and Exercise 12(3):175-182, 1980. 77. Martin, A.D., Spenst, L.F., Drinkwater, D.T., Clarys, J.P. Anthropometric estimation of muscle mass in men. Medicine and Science in Sports and Exercise 22(5):729-733, 1990. 78. Vollman, R.F. The Menstrual Cycle: Major Problems in Obstetrics and Gynecology. W.B. Saunders Co.: Philadelphia, Vol. 7, 1977. 79. Dalvit, S.P. The effect of the menstrual cycle on patterns of food intake. American Journal of Clinical Nutrition 34:1811-1815, 1981. 80. Lyons, P.M., Truswell, A.S., Mira, M., Vizzard, J., Abraham, S.F. Reduction of food intake in the ovulatory phase of the menstrual cycle. American Journal of Clinical Nutrition 49:1164-1168, 1989. 81. Mertz, W., Kelsay, J.L. Rationale and design of the Beltsville one-year dietary intake study. The American Journal of Clinical Nutrition 40:1323-1326, 1984. 82. ESHA Research. The Food Processor II. Nutrition and Diet Analysis System. Salem, OR: ESHA Research, 1990. 83. Cioffi, L.A., James, W.P.T., Van Itallie, T.B. ed. The body weight regulatory system: Normal and disturbed mechanisms. Raven Press : New York, 1 981. 84. Sievänen, H., Oja, P., Vuori, I. Precision of dual-energy x-ray absorptiometry in determining bone mineral density and content of various skeletal sites. Journal of Nuclear Medicine 33:1137-1142, 1992. Dual-energy radiographic absorptiometry for bone 85. Sartoris, D.J., Resnick, D. densitometry: current status and perspective. American Journal of Roentgenologv 1 52:241 246, 1989.  -  Fasting urinary calcium/creatinine and 86. Goulding, A., Rae, B.G., Bland, R. hydroxyproline/creatinine values in young women with amenorrhea and in matched eumenorrheic controls. The New Zealand Medical Journal 102(881):629-630, 1989.  164  REFERENCES  87. Fong, A.K.H., Kretsch, M.J. Changes in dietary intake, urinary nitrogen, and urinary volume across the menstrual cycle. American Journal of Clinical Nutrition 57:43-46, 1 993. 88. Kurtzman, N.A., Rogers, P.W. A Handbook of Urinalysis and Urinary Sediment. Springfield, IL: Charles C. Thomas, 1974. 89. Kathol, R. Environmental effects on creatinine, and on urinary-free cortisol. Biological Psychiatry 30:1173-1174, 1991. (Letter) 90. Lutz, J., Tesar, R. Mother-daughter pairs: spinal and femoral bone densities and dietary intakes. American Journal of Clinical Nutrition 52:872-877, 1990. 91. SPSS Inc. SPSS Data Entry II for the IBM PC/XT/AT and PS/2. Chicago, IL: SPSS Inc., 1987. 92. Norusis, M.J. SPSS/PC+ Statistics 4.0 for the IBM PC/XT/AT and PS/2. Chicago, IL: SPSS Inc., 1990. 93. Fleiss, J.L. Statistical methods for rates and proportions. 2nd ed. John Wiley & Sons: New York, 1981. 94. Lozy, M.E. Dietary variability and its impact on nutritional epidemiology. Chronic Diseases 36(3):237-249, 1983.  Journal of  95. Delanghe, J., De Slypere, J-P., De Buyzere, M., Robbrecht, J., Wieme, R., Vermeulen, A. Normal reference values for creatine, creatinine, and carnitine are lower in vegetarians. Clinical Chemistry 35(8):1 802-1 803, 1989. 96. Flowers, B., Martin, M.J., Cantley, T.C., Day, B.N. Endocrine changes associated with a dietary-induced increased in ovulation rate (flushing) in gilts. Journal of Animal Science 67(3):771-778, 1989. 97. Lohman, T.G. Body composition methodology in sports medicine. Physician and Sports Medicine 10(12):46-58, 1982. 98. Basiotis, P.P., Welsh, S.0., Cronin, F.J., Kelsay, J.L., Mertz, W. Number of days of food intake records required to estimate individual and group nutrient intakes with defined confidence. Journal of Nutrition 117:1638-1641, 1987. 99. Chalmers, F.W., Clayton, M.M., Gates, L.0., Tucker, R.E., Wertz, A.W., Young, C.M., Foster, W.D. The dietary record how many and which days? Journal of the American Dietetic Association 28:7 1 1-717, 1952. -  165  REFERENCES  100. Guthrie, H.A., Crocetti, A.F. Variability of nutrient intake over a 3-day period. Journal of the American Dietetic Association 85(3):325-327, 1985. 101. Overton, T.R., Wheeler, G.D. Bone mass measurements in the distal forearm using dualenergy x-ray absorptiometry and gamma-ray computed tomography: a longitudinal, in vivo comparative study. Journal of Bone and Mineral Research 7(4):375-381, 1992. 102. Rikimaru, T., Oozeki, T., Ichikawa, M., Ebisawa, H., Fujita, Y. Comparisons of urinary creatinine, skeletal muscle mass, and indices of muscle protein catabolism in rats fed ad fib/turn, with restricted food intake, and deprived of food. Journal of Nutritional Sciences and Vitaminology 35:199-209, 1989. 103. Nordin, B.E.C. Diagnostic procedures in disorders of calcium metabolism. Endocrinology 8:55-67, 1978.  Clinical  104. Jones, C.A., Refsal, K.R., Lippert, A.C., Nachreiner, R.F., Schwacha, M.M. Changes in adrenal cortisol secretion as reflected in the urinary cortisol/creatinine ratio in dogs. Domestic Animal Endocrinology 7(4):559-572, 1990.  1 66  APPENDICES  APPENDIX A RECRUITMENT ADVERTISEMENT  FEMALE VOLUNTEERS NEEDED!! FOR A STUDY ON DIET, MENSTRUAL REGULARITY AND BONE... *  If you are between 20 and 40 years of age, you are invited to join a study at the University of British Columbia on dietary habits, the menstrual cycle and bone. I’m interested in women who have regular menstrual cycles (occur at intervals of 21-35 days) and who eat meat (beef or pork) 3 or more times a week (for example, in sandwiches, burgers, casseroles, and so on). *  In exchange for completing questionnaires, food records and having your bone density measured, you would be given an analysis of your diet, an assessment of your menstrual status, a copy of your bone density results and pertinent recommendations. *  If you’re interested, and a non-smoker who hasn’t used oral contraceptives for at least six months, please leave a message for Christina at 228-1606, and I’ll get back to you.  1 67  ________________________________________ ________________________________________ __________________________________ __________________________  __________  APPENDICES  APPENDIX B SUBJECT CONSENT FORM School of Facndy acid Nucricion,al Sciences Diicc.cc Progracii 2205 Eaci Mall Vancouver BC Caciada ‘6T iWS  DIET AND THE MENSTRUAL CYCLE IN PREMENOPAUSAL WOMEN  The investigators in this study are Dr. Susan Barr. Associate Professor of Nutrition, Dr. Jerilynn Prior, Associate Professor of Medicine, and Ms. Christina Janelle, a Master of Science student majoring in Human Nutrition. Messages for Christina can be left at 228-1606. arid Or Barr can be reached at 822-6766. The purpose of the study is to assess whether the type of diet a woman follows Ivegetarian or non. vegetarian) has an influence on characteristics of the menstrual cycle. Your role in the Study includes recording your basal body temperature for six consecutive menstrual cycles, keeping three, 3-day records of your food intake, filling Out S short questionnaire on your eating behavior, providing a urine collection and coming in for measurements that include height, weight and skirifold thicknesses. The bone mineral density of your spine would also be measured This procedure involves a very low level of radiation exposure, comparable to what you would receive if you spent several hours out-of-doors. There are no side effects from this measurement. You will be provided with a copy of your personal results, as well as a summary of the overall results of the study The total amount of time required of you will not exceed the duration of six menstrual cycles, and only a few minutes a day will be required for record-keeping. Measurement of height, weight and skinfolds will take about fifteen minutes, and measurement of bone density, to be done at the Vancouver General Hospital, takes less than half an hour If at any time you have questions or inquiries about these procedures, or any Other aspect of the study, please do not hesitate to contact Christina Janelle at 228-1 606. or Dr. Barr at 822-6766. Your identity as well as any personal information resulting from this research will be strictly confidential. The information that you provide to the investigators will only be available to them, and will be stored under assigned numbers, rather than names. Of course, you are under no obligation to participate in this study, and you may, at any time. withdraw from the study or refuse further participation This will have no consequences, and your identity will remain confidential.  I have read the consent form and fully understand the procedures involved. I consent to participate in this study, and acknowledge receipt of a copy of this form Date:  Name (print):  (signature): Witness  (print  name):  (signature):  1 68  APPENDICES  APPENDIX C COPY OF SKINFOLD TECHNIQUES USED The literature is vague, but the general impres sion is that most investigators measured the thick ness of an oblique fold along the line of the anterior axillary fold. Hertzberg et al. (1963) measured the thickness of a vertical fold. The same site for both males and females is desirable. Although the midpoint between the anterior axillary fold and the nipple is used most com.monly for males, it is not appropriate for females. Because of the vanabi.lity in the size of the mammary gland, it is difficult to use the nipple as a reference point to locate the site. Also, in most cases it would be difficult to exclude mammary tissue from measurements at the site described under (a). The recommended site allows the measurement to be made while a woman wears a two-piece bathing suit or bra.  dominal wall with normal respiration. The subject stands erect with body weight evenly distributed on both feet. Children stand on a platform to al low the measurer appropriate access to the skinfold site. Select a site 3 cm lateral to the midpoint of the umbilicus and 1 cm inferior to it (see Figure ii). The decision whether to measure to the left or right of the umbilicus should be consistent within a study. Raise a horizontal skinfold with the left hand and measure its thickness to the nearest 0.1 cm (see Figure 12).  Reliability Intrameasurer reliability coefficients are very high. ranging from .91 to .97 (Pollock et al., 1975. 1976). The standard error of measurement (SEM) gener ally averages 1 to 2 mm. Data from 68 adults showed a correlation of .96 between trials mea sured on separate days with a SEM of 1.45 mm (Pollock, unpublished data, 1985). lntermeasurer correlations are generally above .9. but the SEM may vary as much as 3 to 5 mm with inexperienced measurers, or when the site is not standardized (Lohman et al., 1984). Jackson et a!. (1978) reported a correlation among measurers of .98 with a SEM of 2.1 mm. An intermeasurer correlation of .93 with a SEM of 1.7 mm has been recorded (Pollock, unpublished data, 1985).  Figure 11  Location of abdominal skinfold site  ?ii1 11 i.  Sources of Reference Data  Children none reported  Figure 12 Measurement of abdominal skinfold  Adults none reported  Purpose  Abdominal Skinfold Recommended Technique For the measurement of abdominal skinfold thick. ness, the subject relaxes the abdominal wall mus culature as much as possible during the procedure and breathes normally. The subject may be asked to hold his or her breath near the end of expira tion if there is bothersome movement of the ab  1 69  The abdominal skinfold is measured commonly and has been included in many studies of body fatness and in many regression equations (Loh man, 1981). Abdominal skinfold thickness changes markedly with weight reduction (Després et al.. 1985). It is relatively easy to access, is relatively large. differs considerably among subjects, and is reasonably reproducible with the recommended technique.  APPENDICES  APPENDIX C (copy of skinfold techniques used continued.) 62  Harrison, Buskirk. Carter. Johnston, Lohman. l’ollock. Roche, and Wilmore  Literature Several locations have been used for measurement of the abdominal skinfold. These include adjacent to the umbilicus; level of the umbilicus but 5 cm to the left of it; slightly inferior to the umbilicus and 1 cm to the right of it, and a quarter of the dis tance between the umbilicus and the anterior su perior iliac spine (Edwards, 1950; Lohman, 1981; Paii±ková & Zdenek, 1972; Sk&lj et at., 1953; Weiner & Lourie, 1981). Most have measured horizontal fold (Behnke & Wilmore, 1974), but others have measured a vertical fold (Sinning et at., 1985; Steinkamp et at., 1965). Some subjects have a “crease” in the region of the umbilicus that precludes selection of a single site for all, and in the obese it is difficult to raise a discrete skinfold.  Figure 13 Diagram to illustrate the location of the suprailiac skinfold in the midaxilar’ line superior to the iliac crest  Reliability WU more and Behnke (1969) reported a test-retest correlation of .979 for measurements made 1 day apart in young men. An intrameasurer technical error of 0.89 mm was reported by Zavaleta and Malina (1982). Sources of Reference Data Children  None reported Adults None reported  Suprailiac Skinfold  Figure 14 Subject position for measurement of the supradiac skirifold  Recommended Technique The supraiiac skinfold is measured in the midax illary line immediately superior to the iliac crest (see Figure 13). The subject stands with feet together and in an erect position. The arms hang by the sides or, if necessary, they can be abducted slightly to improve access tc the Site (see Figure 14). In those unable tostand, the measurement can be made with the subject supine. An oblique skinfold is grasped Just posterior to the midaxillary line following the natural cleavage lines of the skin. It  is aligned inferomedially at 45° to the horizontal (see Figure 14). The caliper jaws are applied about  1 cm from the fingers holding the skinfold, and the thickness is recorded to the nearest 0.1cm (see Figure 15).  1 70  Purpose Suprailiac skmfold thicknesses are commonly used as indices of body fatness together with other skinfold thicknesses (Durnin & Womersley. 1974) Suprailiac skinfold thicknesses are useful in the study of subcutaneous adipose tissue distribution, which is important in regard to risk of disease (Lapidus et al.. 3984; Larsson et al.. 1984) Literature In most studies, the subjects stood for the measurement of suprailiac skinfold thicknesses. Considerable variation regarding the location and direction of the suprailiac skinfold occurs in the  APPENDICES  APPENDIX C (copy of skinfold techniques used continued) Skinfold Thtcknesses  63  proach to skinfold measurement of this manual. The use of a vertical fold (Behnke & Wilmore, 1974>, horizontal fold (Johnston et a]., 1974), or ob lique folds at more anterior locations (Pollock et al., 1975; Ross and Marfell-Jones, 1983) is common (see Figure 16). The recommended site of measurement is very similar to the site sometimes described for the waist skinfold (Behnke & Wilmore, 1974; Brown & Jones, 1977; Skrlj et a1., 1953). Because of this similarity, the waist skinfold procedure is not described separately. Reliability  Figure 15  Measurement of the suprailiac skinfold.  literature. Thicknesses at the various locations ap pear highly correlated with each other and with body density (Sinning & Wilson. 1984), so that no one position appears to offer unique information. Relatively large systematic differences in thick nesses among locations emphasize the need to standardize the technique for the measurement of the supraiiac skinfold. The selection of a site on the midaxilary line superior to the iliac creast has the advantage of being easily located in reference to anatomical landmarks. The direction of the fold parallel to the cleavage lines of the skin matches the general ap  Wilmore and Behnke (1969) reported a test-retest correlation of 970 for values recorded I day apart in young men. Technical errors of 1 53 mm in children and youth (Johnston et a). 1974) and of 1.7 mm in adults (l-laas & Regal. 1981) have been reported. In each study. the errors for supraiiac skinfold thicknesses were larger than those for other skinfold sites. Intrameasurer tech nical errors of 0.3 to 1.0 rnrrl have been reported by others (Buschang, 1980; Meleski. 1980; Zavaleta & Malina, 1982). Sources of Reference Data Children Baker et a)., 1958 Ferris et al., 1979 Johnston et al., 1974 (horizontal fold) Montoye, 1978 Schutte, 1979 Zavaleta, 1976 Adults Katch & Michael, 1968  Thigh Skinfold Recommended Technique  I SIondord,ze, SuroIoc Sde 2 Alter Pollocts et CI 3 After Ross ond MorfelI-Jone  Location of recommended suprailiac site in reference to other frequently measured suprailiac Figure 16 Sites  The thigh skinfold site is located in the midline of the anterior aspect of the thigh, midway between the inguinal crease and the proximal border of the patella (see Figure 17). The subject flexes the hip to assist location of the inguinat crease. The prox imal reference point is on the inguinal crease at the midpoint of the long axis of the thigh. The distal reference point (proximal border of the patella) is located while the knee of the subject is extended.  171  APPENDICES  APPENDIX C (copy of skin fold techniques used continued...) 64  Harrison, Buskirk, Carter. ohnston, Lohman, Pollock. Roche, and Wilmor  hydrostatic weighing (Wilmore & Behnke, 1969, 1970). Thigh skinfold thickness has been selected by regression analysis as one of the skinfold mea sures included in equations to predict body den sity from arithropometric values.  Inguinot  .\çCeose  Mid-Thigh Skintotd ProKimot Border of Potetto  Figure 17 Location of the midthigh skinfold site  The thickness of a vertical fold is measured while the subject stands. The body weight is shifted to the other foot while the leg on the side of the measurement is relaxed with the knee slightly flexed and the foot flat on the floor (see Figure 18). If the maintenance of balance is a problem, the sub ject holds the top of the measurer’s shoulder, a counter top, or high-backed chair. For patients confined to a bed or wheelchair, the thigh skinfold is measured while the patient is supine. The caliper jaws are applied about 1 cm distal to the fingers holding the fold; the thickness of the fold is recorded to the nearest 0.1 cm. Purpose  Literature A few early studies refer to both anterior and posterior thigh skinfold sites, but most refer only to the anterior site: thus, further discussion will relate to this site only. Although description of the thigh skinfold site appears to be standardized among many studies, considerable variation can be found (Lohman et al., 1984). The most common description of the thigh skinfold site is on the anterior aspect of the thigh, midway between the hip and knee (Wilmore & Behnke, 1969; Zuti & Golding, 1973). Sloan et al. (1962) used the midpoint from the inguinal  crease to the proximal margin of the pate[la. Others give a more general description, such as halfway down the rectus femoris muscle (Young et at. 1962). The investigators mentioned previously described their measurements as being made with the subject in the standing position, leg relaxed. Some mea sure with the leg flexed 90° at the knee by placing the foot on a box. This technique is recommended by Ross and Marfell-Jones (1984). All investigators measure thigh skinfold thicknesses with a vertical fold aligned in the long axis of the thigh.  Thigh skinfoid thicknesses have moderate to high correlations with body density determined by  Reliability Irttrameasurer reliability coefficients are very high, ranging from .91 to .98 (Pollock et al., 1976; Wil more & Behnke, 1969; Zuti & Golding, 1973), although the standard error of measurement (SEM) generally averages between 1 to 2 mm. Recent data on 68 adults showed a correlation of .985 between trials taken on separate days with a SEM of 1.4 mm (Pollock et al., unpublished data, 1985). Others have reported intrameasurer technical errors of 0.5 to 0.7 mm (Meleski, 1980; Zavaleta, 1976). Intermeasurer correlations are generally above .9, but the SEM may be as much as 3 to 4 mm with inexperienced measurers or when the sites are not standardized (Lohman et at, 1984). Jackson et al. (1978) reported a correlation among measurers of .97 and a SEM of 2.4 mm for measurers of varying experience who had trained together. In an unpub lished study, Pollock (1986) showed an intermeasurer con-elation of .975, with a SEM of 2.1 mm.  Figure 18 Measurement of midthigh skinfold.  172  APPENDICES  APPENDIX C (copy of skin fold techniques used continued...) ‘4.n,.i.’  fold thicknesses The absolute median error for all these skinfold thicknesses was 1.0 to 1.5mm using a Lange caliper read to the nearest 0.5 mm. A testretest correlation of .98 has been reported (Perez. 1981). In subjects with a wide range of ages, the intrameasurer correlations ranged from .94 to .99 (Carter, 1986). Sources of Reference Data Children Johnston et al 1974 Malina & Roche, 1983 Ross & Ward. 1984 Zavaleta. 1976 Adults Cl,iusr et al 1972 Rocs & \‘\ard, lqh 4  Figure 25  \1,irtcd rind in 1  ,.r  ‘c’n’  LiilJ  ‘i’  .  is marked ,n the ,iterii 251.  Triceps Skinfold Recommended Technique The triceps skintold is measured in the midline 01 the posterior aspect of the arm, over the triceps muscle, at a point midi’av hetis’een the lateral prolection of the acromion process of the scapula and the interior margin ot the olecranon pruci.’ss ot the ulna The level ut measurement is deter. mined b’ measuring the distance between the lateral prolection of the acromial process and the inferior border of thç olecranon process of the ulna. O using a tape measure, with the elbox’ flexed to 9 (see Figure 24). The tape is placed with its zero mark on the acromion and stretched along the up. per arm, extending below the elbow The midpoint  Figure 24 ck int’iild  Lucation  cit  mid,iriii  tcvet  ‘..  liv  ,irn’  i..’e  The suhect is ntv,i’urvr ‘oanding. c’\cc’Ln ,‘r infants and the hidic.iope.i Flic’ skint,ld 5 me,i cured is oh the trni li,inin c’-eR and cctninsr able at th suhlvcl’s iJ,’ l’ta:c 2oi. The cancer is held in th right ni rid !‘hx’tnc ,iscrer st,irixl”. behind the subleCt and pioc e lie palm sit his or her left hand on the uh c’:t .irnr prc\imai to 11w 1 marked level, nit h t h,’hci nib .i rid index 1 dtrected inreriurli Fl:’.ric’rs ‘ciinrcld is ptcked up with the len thumb and ndC\ I rigor. appro\( matelv 1 cm proximal to tIn.’ marked es el. and thy tips of the calipers are spiied to the skinfold at the marked level (sex’ Fnur 210 (lie sac iii measurement must he in lie rttd(ins’ posneriiri v .  1cr triceps Figure 26  1 73  ‘ts’,i,irs’ns’i.I  ,‘  :r,_’n” —kui.ld  APPENDICES  APPENDIX C (copy of skinfold techniques used continued..,) 68  I Iarris.ii, tjuLtrk  C.rrter, lilhgistt,n, (,ihnr,i  III’ L, k’h’  when the palm ma directed anteriorly. Particular problems will be laced when measuring the obese and muscular subjects with little fat at this site. It necessam-v in the case of obese subjects, an assis tant may pick up the (old with two hands, but this gives larger readings than if one hand is used (Da mon. 1965)  mci I’ mImmim’  Biceps SkinIoLd Recommended Technique Biceps skinfold thickness is measured as the thick ness of a vertical fold raised on the anterior aspect of the arm, over the belly of the biceps muscle (see Figure 27). The skinfold is raised 1 cm superior to the line marked for the measurement of triceps  Purpose The triceps skin(old is measured more commonly than an’ other. partly because it is so accessible. It is closely correlated with percentage of body tat and m.vith total bc’mdv fat but is less well ccmrrelated with blood pressure than are trunk skinfolds. It is olten included ri studies of tat patterning Literature The level ot th site is marked with the arm flexed at a right angle at the elbow, and the skmnfold is measured with the arm hanging loosely at the side Positioning is not crucial. except that the subject should he re’aed and the palm directed anterior ly so that the posterior midline can be determined Most measure subjects ri a standing positmnn. though nonarnhulatorv patients may be measured “hen upmne lntant may be measured king down, or being held on someone’s lap. When Supine or sitting positions are used, the recm’im. mended technique can still be applied with ltttle modification Reliability In general. measurement en-or increases m”ith the age of the subject and with u’icreastng levels cit fatness. lntermeasuj-er technical errors vary from 0.8 to 1 89 mm (Johnston et at, 1974; )ohnston icc Mack. 19851 Intrameasurer technical errors vat-v from 0.4 to 0.8 mm (Johnston et al., 1974. 1q75; Malina & Buschang, 1984. Martorell et al.. 1975)  Figure 27  Locaticmn of biceps skmntc’IJ  5it  skinfold thickness and arm c;rcurnterencc’, on a vertical line loining the anterior border cit the acrommon and the center cml the antecuhital Ioss,m (see Ftgure 28). The subject stands, facing the mea surer. with the upper extremity relaxed at the side, and the palm directed anteriorly The caliper aws are applied at the marked level see Figure 28). The thickness of the skinfcmld is recorded to the nearest 0 1 cm  Sources of Reference Data Cli i ldremm Frisancho, 1981 Iohnston et al 19R1 A mimi Its Frisanc i, 1981 Icmhnstc,n et al 1981 .  Figure 28  174  Measurement  imt  hiccp ,kmriktd  APPENDICES  APPENDIX C (copy cf skin fold techniques used continued...) Circumtt’rcnces  ate position. The subject lies supine while one measurer elevates the leg and dorsally flexes the loot to approximately a right angle. The minimum ankle circumference is then measured as described above.  51  Sources of Reference Data Children Huenemann et al., 1974 McCammon, 1970 Snyder et al., 1975  Ankle circumference is a measure of frame size and is useful in the design of clothing, especially footwear.  Adults National Aeronautics and Space Administration, 1978 Clauser et al., 1972 White & Churchill, 1971  Literature  Arm Circumference  Purpose  Ankle circumference is measured using a tape that is sufficiently flexible to conform to the irregular shape at the level of measurement. It has been measured with the subject sitting on a table, with the feet placed in a chair high enough to form a right angle at the knee (O’Brien & Shelton, 1941). Others have measured children sitting on a table with the leg extended and relaxed (Snyder et al., 1975). The degree of dorsiflexiori of the foot and support of weight by the loot are important con siderations. Dorsiflexion to more than 90° is ac companied by marked contraction of the tibialis anterior and the anterior extensur muscles of the lower leg. The associated elevation of the tendons of these muscles from the surface of the ankle dis torts the cross-sectional shape of the ankle and in creases its circumference at the level of measurement. Ankle circumference has been measured so that the superior border of the tape passes over the tip of the medial malleolus (O’Brien & Shelton, 1941; Randall & Baer, 1951). Ankle circumference mea sured in this way is not highly correlated (r = .69) with that measured by the recommended method (O’Brien & Shelton, 1941). The minimum circum ference is recommended because reliability is known to be high, and it is the technique that has been used in studies of body composition (Wil more & Behnke, 1969).  Recommended Technique For this measurement the subject stands erect. with the arms hanging freely at the sides of the trunk and with the palms lacing the thighs. The subject wears loose clothing s’ithout sleeves to allow total exposure of the shoulder area, If the midpoint of the upper arm has been marked for the measurement of triceps or biceps skinfolds. this should be used as the level for the measurement of arm circumference. To locate the midpoint, the subject’s elbow is flexed to 90’ with the palm facing superiorly. The measurer stands behind the subject and locates the lateral tip of the acromion by palpatinglaterally along the superior surface of the spinous process of the scapula. A small mark is made at the identified point. The most distal point on the acromial process is located and marked. A tape is placed so that is passes over these two marks, and the midpoint between them is marked (see Figure 19).  Reliability Huenemann and co-workers (1974) measured ankle circumference in 2 subjects 20 times during a 4-week period. The standard deviations of repli. cate measurements for the 2 subjects were 0.11 and 0.12cm for the right side and 0.12 and 0.13cm for the left side. Wilmore and Behnke (1969) reported a test-retest correlation of 99 irs young men mea sured on successive days  175  figure 19 arm.  Location of the midpoint of the upper  APPENDICES  APPENDIX C (copy of skin fold techniques used continued...) 52  Callaway. Chumlea, Bouchard. Himes. Lohman. Martin. Mitchell. Mueller. Roche, and Seefeldi  With the arm relaxed and the elbow extended and hanging just away from the side of the trunk and the palm facing the thigh, place the tape around the arm so that it is touching the skin, but not compressing the soft tissues. The tape is posi tioned perpendicular to the long axis of the arm at the marked midpoint, and the circumference is recorded to the nearest 0.1 cm (see Figure 20).  Reliability Bray et al (1978) reported intermeasurer errors of selected circumferences and skinfold thicknesses in lean and obese patients. They found less varia bility with the circumference measurements than with skinfold thicknesses. The intermeasurer varia tion in obese patients after a 2-week interval was 2.1% (± 0.10 SEMi for arm circumference. Hall et a!. (1980) calculated the measurer error for arm cir cumference as 1.54 cm’. Martorell et al. (1975) reported that in preschool children, arm circum ference had a total measurement standard devia tion of 0.24 cm; 56% of the total variance was due to intrameasurer variance. lntrameasurer techni. cal errors of 0.1 to 0.4 mm and an intermeasurer  technical error of 0 3 mm have been reported (Brown. 1984; Buschang. 1980; Malina. 1968; Malina & Buschang, 1984. Meleski. 1980; Zavaleta  & Malina, 1982). Sources of Reference Data Children Frisancho, 1974, 1981  McCammon. 1970 Adults Figure 20  Measurement of arm circumference  Purpose Arm circumference provides an index of body energy stores and protein mass. Although it can be used as an independent measure, it is often combined with skinfold thicknesses to calculate arm-muscle circumference and the areas of arm muscle and adipose tissue (Gurney & JeUiffe, 1973; Heymsfield et al., 1984). Low values are inter preted as evidence of protein-energy malnutrition (Blackburn et al., 1977). The recommended measurement is made with muscles relaxed. Arm circumference can be mea sured with the elbow flexed and the biceps con tracted when there is particular interest in muscle development. This measurement is called arm  National Aeronautics and Space Administration. 1978 Frisancho, 1974, 1981  Bishop et al., 1981  Forearm Circumference Recommended Technique  circumference-flexed,  For the measurement of forearm circumference. the subject stands with the arms hanging down ward but slightly away from the trunk, with the palms facing anteriorly (see Figure 21). The tape is placed loosely around the proximal part of the forearm, perpendicular to its long axis, and moved up and down until the level of the maximum cir cumference is located (see Figure 2.2). At this level the measurement is recorded to the nearest 0.1 cm, with the tape in Contact ‘.‘ith the skin but not com pressing the soft tissues  Literature  Purpose  If possible, the subject should stand, but the arm circumference can be measured with the subject sitting erect with the back straight and the head in the Fran.kfort Plane.  Forearm circumference is used with other body  measurements in some equations to predict body density from anthropometric data (Boileau et al., 1981; Jackson & Pollock, 1978; Katch & McArdle.  176  APPENDICES  APPENDIX D BASAL BODY TEMPERATURE RECORD FORM  CODE LAST NA  liii CALIRCAR DAY  FIRST D4TT1AL  II_...j  DATE 1I4  CAY  [_iiiiiiiil  lililil  CYCLE  rERRATvRE  R(STRUAL CYCLE DAY  COOT  (Ith.  £  ‘  BASAL BODY TEMPERATURE RECORD FORM  177  APPENDICES  APPENDIX E BASAL BODY TEMPERATURE MEASUREMENTS  Directions: I require your basal body temperature for six consecutive menstrual cycles. This will help me identify any possible subclinical menstrual disturbances you may have. Please start your temperature measurements on the first day of bleeding of your menses. This is considered DAY 1 of your menstrual cycle. So, if your menstrual cycle lasts 28 days, then DAY 28 will be the last day of your cycle, and the next day you will start your second menstrual cycle record, again on DAY 1. Please use a new basal body temperature record form each cycle. Please take your basal body temperature orally, for at least five minutes, every morning between 6:00 a.m. and 8:00 a.m., immediately after awakening and before rising. You will be provided with a low-reading digital Celsius thermometer to do this. The reading will be to two decimal places. Please record the result as it is shown. Also, it is crucial that you record the results on the attached form IMMEDIATELY! Please fill out the attached form by entering the calendar day (e.g., 01/09/92), the day of your menstrual cycle (e.g., DAY 1), your temperature reading for that day (e.g., 37.31 °C), and any pertinent comments (e.g., insomnia, late rising, fever, illness, etc...) I would greatly appreciate it if you would mail me the form after the completion of each cycle, so that I may record that data. You have been provided with six self-addressed stamped envelopes for this. If you are unsure about any of these instructions, or have hesitate to contact me or leave me a message at 822-6766.  Good luckH!  Christina Janelle  1 78  jjy  questions, please. do not  Co  -  C  0  I  I  ‘:2  1 V  —  2  —  —  I  0  ‘1  C  0  m  m  -l  m  C  -n  z x  m  0  APPENDICES  APPENDIX G DIETARY INTAKE RECORDS  Directions: I will need three 3-day records of your daily food intake. To do this, please record, at the time of consumption, all foods and beverages consumed, on the attached dietary intake forms. Please use a new dietary intake record form for each day of recording. These three 3-day records do not have to be done consecutively. It will likely be easiest for you if you do one record every second cycle. However, I do need the records from different stages of your menstrual cycle. Please do one 3-day record two to five days after menstruation has stopped, do the second record around the time of expected ovulation (identified by changes in cervical mucous amount and consistency), and the third about five to ten days before your next menstruation. When you record the food or beverage eaten, please include the amounts, either by measurements (e.g., cups, weight, teaspoon, tablespoon, etc.), size (e.g., with a ruler, the size of a piece of cake), or by count (e.g., one slice of whole wheat bread, 2 large eggs, etc.). Also, include all relevant information on that food, like type of cheese, bread, name brands of store bought foods, etc. For example, if for lunch you have a cheese sandwich, a tossed salad, a Rice Krispies square, and an apple juice, you should record it like this: 2 slices white bread 1 Tablespoon mayonnaise 2 ounces white cheddar cheese 1 leaf iceberg lettuce 2 thin slices of tomato Y2 cup iceberg lettuce 1 medium carrot 5 ¾” slices of cucumber 1 Tablespoon oil & vinegar dressing 1 2”x2” Rice Krispies square 8 ounces unsweetened apple juice Therefore, you should include EVERYTHING salad dressing, butter/margarine, mayonnaise, water, alcohol, pop (even diet!), sauces, gravies, etc... Also, there is a place to record any vitamin or mineral supplements taken, if any. If you do use any vitamin or mineral supplements, could you please record the type, amount and name brand of each. -  180  APPENDICES  APPENDIX G (dietary intake record instructions continued...) It is very important that you record each item you consume IMMEDIATELY! Please, do not wait until the end of the day, because chances are you won’t remember everything. Also, please do not alter your normal eating patterns. I am not judging you by what you eat. Even if you eat something that you would not on a regular basis, please still include it on the record form. Your honesty will benefit both of us, and will be crucial when I analyze your diet, assess your menstrual status, and provide you with recommendations. Please send me your dietary intake record forms (the set of three days) after you have completed it. You have been provided with a self-addressed stamped envelope in which you should include your completed basal body temperature record form and, every other cycle, your completed three-day intake record forms. If you are unsure about any of these instructions, or have jy questions, please do not hesitate to contact me or leave me a message at 822-6766.  Good luck!  Christina Janelle  181  APPENDICES  APPENDIX H THREE-FACTOR EATING QUESTIONNAIRE  One point is given for each item in Part I and for each numbered question in Part II. The correct answer for the true/false items is underlined and beside it is the number of the factor that it measures, i.e., factor #1, #2, or #3. The direction of the question in Part II is determined by splitting the responses at the middle. If the item is labelled +“, those responses above the middle are given a zero. Vice versa for those with a For example, if the question is marked - + -, anyone scoring 3 or 4 would receive one point, and anyone scoring 1 or 2 would receive a zero. “.  Part I 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 1 2. 1 3. 1 4. 1 5. 1 6. 1 7. 1 8. 1 9. 20. 21. 22. 23. 24. 25. 26. 27.  When I smell fresh baked bread, I find it very difficult to keep from eating, even if I have just finished a meal. (I 3) I usually eat too much at social occasions, like parties and picnics. (I 2) I am usually so hungry that I eat more than three times a day. (j 3) When I have eaten my quota of calories, I am usually good about not eating any more. (I 1) Dieting is so hard for me because I just get too hungry. (I 3) I deliberately take small helpings as a means of controlling my weight. (I 1) Sometimes things taste so good that I keep on eating when I am no longer hungry. (I 2) Since I am often hungry, I sometimes wish that while I am eating, an expert would tell me that I have had enough or that I can have something more to eat. (I 3) When I feel anxious, I find myself eating. (I 2) Life is too short to worry about dieting. (f 1) Since my weight goes up and down, I have gone on reducing diets more than once. (I 2) I often feel so hungry that I just have to eat something. (I 3) When I am with someone who is overeating, I usually overeat too. (I 2) I have a pretty good idea of the number of calories in common food. (I 1) Sometimes when I start eating, I just can’t seem to stop. (I 2) Since I know how guilty I will feel afterwards, I rarely go on eating binges. (I 1) It is not difficult for me to leave something on my plate. (E 2) At certain times of the day, I get hungry because I have gotten used to eating then. (I 3) Sometimes I get so nervous that I just have to eat something. (I 2) While on a diet, if I eat food that is not allowed, I consciously eat less for a period of time to make up for it. (I 1) Being with someone who is eating often makes me hungry enough to eat also. (I 3) When I feel blue, I often overeat. (I 2) I enjoy eating too much to spoil it by counting calories or watching my weight. (E 1) When I see a real delicacy, I often get so hungry that I have to eat right away. (I 3) I often stop eating when I am not really full as a conscious means of limiting the amount that I eat. (Ii) I get so hungry that my stomach often seems like a bottomless pit. (I 3) My weight has hardly changed at all in the last ten years. (f 2) 182  APPENDICES  APPENDIX H (Three-Factor Eating Questionnaire continued..) 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38.  I am always hungry so it is hard for me to stop eating before I finish the food on my plate. (j3) When I feel lonely, I console myself by eating. (I 2) I consciously hold back at meals in order not to gain weight. (I 1) I sometimes get very hungry late in the evening or at night. (3 3) I eat anything I want, anytime I want. (f 1) Without even thinking about it, I take a long time to eat. (E 2) I count calories as a conscious means of controlling my weight. (I 1) I do not eat some foods because they make me fat. (I 1) I am always hungry enough to eat at any time. (I 3) I pay a great deal of attention to changes in my figure. (I 1) While on a diet, if I eat a food that is not allowed, I often then splurge and eat other high calorie foods. (I 2)  Part II 39.  How often are you dieting in a conscious effort to control your weight? (+ 1) 1 rarely  40.  41.  2 sometimes  1 zero lbs.  2 <5 lbs.  3 <10 lbs.  4 <15 lbs.  5 <20 lbs.  6 <35 lbs.  7 <50 lbs.  8 <75 lbs.  9 <150 lbs.  10 >150 lbs.  Would a weight fluctuation of 5 lbs. affect the way you live your life? (+ 1) 2 slightly  3 moderately  4 very much  How often do you feel hungry? (+ 3) 1 only at mealtimes  43.  always  How many pounds over your desired weight were you at your maximum weight? (+ 2)  1 not at all 42.  4  3 usually  2 sometimes btwn meals  3 often btwn meals  What is your maximum weight loss within one month? (+ 2) 1 <2 lbs.  2 <4 lbs.  3 <6 lbs.  4 <8 lbs.  5 <10 lbs.  6 <12 lbs.  7 <15 lbs.  8 <20 lbs.  9 <30 lbs.  10 >30 lbs.  183  4 almost always  APPENDICES  APPENDIX H (Three-Factor Eating Questionnaire continued...) 44.  Do feelings of guilt about overeating help you to control your food intake? (+ 1) 1 never  45.  2 slightly difficult  49.  4 extremely  3 moderately  How frequently do you skip a meal as a conscious means of limiting your eating? (+ 1) 2 seldom  3 at least once a week  4 almost everyday  What is your maximum weight gain within a week? (+ 2) 1 zero lbs.  2 <1 lb.  3 <2 lbs.  4 <3 lbs.  6 <5 lbs  7 <6 lbs.  8 <9 lbs.  9 >9 lbs.  5 <4Ibs  How frequently do you avoid “stocking up” on tempting foods? (+ 1) 1 almost never  50.  4 very difficult  3 moderately difficult  2 slightly  1 almost never 48.  4 always  How conscious are you of what you are eating? (+ 1) 1 not at all  47.  3 often  How difficult would it be for you to stop eating halfway through dinner and not to eat for the next four hours? (+3) 1 easy  46.  2 rarely  2 seldom  3 usually  4 almost always  In a typical week, how much does your weight fluctuate (maximum-minimum)? (+ 2) 0 zero lbs.  1 <1 lb.  2 <2 lbs.  4 <4 lbs.  5 <5 lbs.  6 >5 lbs.  184  3 <3 lbs.  APPENDICES  APPENDIX H (Three-Factor Eating Questionnaire continued...) 51.  How likely are you to shop for low calorie foods? (+ 1) 1 unlikely  52.  3 often  4 always  2 slightly likely  3 moderately likely  4 very likely  2 seldom  3 at least once a week  4 almost everyday  How likely are you to consciously eat less than you want? (+ 1) 1 unlikely  56.  2 rarely  How frequently do you skip dessert because you are no longer hungry? (- 3) 1 almost never  55.  4 very likely  How likely are you to consciously eat slowly in order to cut down on how much you eat? (+ 1) 1 unlikely  54.  3 moderately likely  Do you eat sensibly in front of others and splurge alone? (+ 2) 1 never  53.  2 slightly likely  2 slightly likely  3 moderately likely  4 very likely  Do you go on eating binges even though you are not hungry? (+ 2) 1 never  2 rarely  3 sometimes  185  4 at least once a week  APPENDICES  APPENDIX H (Three-Factor Eating Questionnaire continUed...) 57.  On a scale of 0 to 10, where 0 means no restraint in eating (eat whatever you want, whenever you want it) and 10 means total restraint (constantly limiting food intake and never “giving in”), what number would you give yourself? (+ 1) eat whatever you want, whenever you want it 1 2 usually eat whatever you want, whenever you want it 3 4 often eat whatever you want, whenever you want it 5 6 often limit food intake, but often “give in” 7 8 usually limit food intake, rarely “give in” 9 10 constantly limiting food intake, never “giving in”  To what extent does this statement describe your eating behaviour? “I start dieting in the 58. morning, but because of any number of things that happen during the day, by evening I have given up and eat what I want, promising myself to start dieting again tomorrow.” (+ 2) 1 not like me  2 little like me  3  pretty good description of me  186  4 describes me perfectly  APPENDICES  APPENDIX I  TWO-HOUR URINE COLLECTION  Directions: I require a two-hour urine collection from you. This will help me understand your calcium and hormone balance. The evening before you collect the urine sample, please do not eat anything after dinner. You may drink water but should not eat any food or drink anything other than water until after the urine collection is completed the next morning. In the morning of the collection, when you get up, empty your bladder and flush away the urine. It is important to take note of the time now because this will be the start time of your two hour test. The test starts with your bladder empty. For the next two hours, drink several glasses of water and collect ALL of your urine in the container provided to you. End the test by emptying your bladder exactly two hours after you started collecting the urine. Please take your urine sample to the Vancouver General Hospital lab as soon as possible. If you cannot go to the lab immediately, keep the urine sample in the refrigerator or in a cool place. If you unsure about any of these instructions, or have y questions, please do not hesitate to contact me or leave me a message at 822-6766.  Good Luck!  Christina Janelle  187  APPENDICES  APPENDIX J  INTERVIEW FORM CODE  ANTHROPOMETR.Jcs NAME: (please print)  TEL:  AGE:  HOME #:  ETHNIC GROUP: OCCUPATION: AGE AT MENkRCHE:  (How old were you when...)  EXERCISE TYPE:  HOW MUCH  (hr/wk)  HOW LONG HAVE YOUR MAINTAINED THESE EXERCISE HABITS?_____________ ARE YOU ON ANY MEDICATION?  (PLEASE LIST NAME)  TRIALS WEIGHT:  (0.1 kg)  HEIGHT:  (0.1 cm)  BMI  AVF  ) 2 (kg/n  TRICEPS:  (0.5  ABDOMEN:  (0.5 mm)  miii)  SUPP..AILIAC: (0.5 nun) ANTERIOR THIGH: (0.5 mm) SUM OF SKINFOLDS:  MID ARM CIRCTJM.: (0.1 cm) FOREARM GIRTH:  188  ___________________  APPENDICES  APPENDIX K  VEGETARIAN ADDITIONAL INTERVIEW FORM VEGETARIANS NAME: (please print) CODE:  What type of vegetarian are you?  lacto—ovo vegetarian (include dairy products and eggs in diet)  How  (Please circle one)  lacto vegetarian  vegan vegetarian  (include dairy but no eggs)  (no dairy or eggs in diet)  long have you been a vegetarian? 2  years  (Please circle one)  5—10 years  2-5 years  >  10 years  Please complete the following sentence: “I chose to become a vegetarian because  any, if family, your of members what indicate Please by circling those of the following vegetarians as well, they have long indicate how please Also, correspond. vegetarians. spouse/partner grandfather (maternal/paternal) grandmother (maternal/paternal) father mother brother(s) (number: ) sister(s) (number: ) COMMENTS  189  years: years: years: years: years: years: years:  are that been  APPENDICES  APPENDIX L  BASIS FOR METHODOLOGIES CHOSEN  1. A NTHROPOMETRICS  The use of anthropometry, including gross body weight, height, skinfold thicknesses, circumferences and skeletal widths offer promise as a field method to approximate and evaluate fat stores, when used correctly. 97 These measures are relatively fast, non’ 7071 invasive, and require a minimum of equipment compared to laboratory techniques. 73 They are particularly important values since the variable showing the largest individual differences is body fat. 71 For some populations, an estimate of each subject’s fat content can be derived using equations that have been successfully cross-validated on other populations of the same age, sex, activity level and racial background. 97 Body fat measurements were important in this study to assess their role in menstrual irregularity. 73 A) WEIGHT  Gross body weight gives a crude indication of overall fat and muscle stores, though it is not reliable when used alone. 7097 It should be noted that body weight may be affected by variations in tissue hydration with the menstrual cycle, though anthropometric measurements are generally not corrected to account for this effect. 73 B) HEIGHT  Height is the most common factor used to determine ideal weight, yet, as with weight, it is not a reliable predictor of body fatness. 7097 190  APPENDICES  The shortcomings of using weight and height as a basis for estimation for percent body fat are well documented, however, these two variables can still give a useful indication of the fat content of the body for some purposes, and they are needed for other measuremen 97 ts. C) BODY MASS INDEX  Body Mass Index (BMI =Wt/Ht ) has been found to be the most satisfactory measure 2 based on weight and height that is available.  It has been found to have greater content  validity than skinfold thickness measurements, but similar concurrent validity.  BMI also  presents a high degree of precision, reliability, accuracy, and client acceptability, because it is easy to measure and is not time consuming. 72 The BMI continuum relates morbidity and mortality to weight. separated into zones  -  The continuum is  zone B, a BMI between 20 and 25, is associated with the lowest  mortality rate and represents a good weight for most people.” DI SKINFOLD MEASUREMENTS  In order to interpret intelligently the biological significance of a departure of the actual weight from a standard body weight, it is necessary to estimate the relative contribution of the adipose tissue. 71 Approximately one-third to 50% of the total body fat of a well nourished person is located in the subcutaneous tissue 70 layer. 7 ’ 3 Distribution of this subcutaneous fat is not uniform and has led to the evaluation of fat pads in several locations, including the area over the triceps, biceps, abdomen, thigh, subscapular, and subcostal areas. ° Ratios or sums 7 using the different skinfold measurements are then used in the assessment of body composition for individuals or populations.” 73’ No consensus exists on the best single, or combination of, skinf old sites with which  191  APPENDICES  APPENDIX L (basis for methodologies continued...) to assess body fat. The most appropriate sites depend on age, sex, and race of subject 73 studied. Triceps skinfold is most commonly evaluated, particularly because it is an easily accessible and readily identifiable area.  The abdominal skinfold thickness can change  markedly with weight reduction and is a relatively large measure. Also, it has been included in many regression equations. The subscapular skinfold has, however, been identified as a better index of adiposity by some researchers, because it has the advantage of being a uniform layer of subcutaneous fat and does not require precise location. 70 The thigh skinfold thickness has been shown to have moderate to high correlations with body density as determined by hydrostatic weighing and is a particularly important location for body fat measurement in women due to that sex’s typical gynoid fat deposition. 7175 Stepwise or step-up multiple regression analysis has often been used in developing regression equations for the anthropometric estimation of body composition.  Successive  anthropometric variables were added to the analysis based on how much they improved the 97 prediction.  In general, three to five variables contribute significantly to the prediction  equations, but additional variables make little difference. Thus, most prediction equations recommend  not more than five sites for measurement in self-assessment and/or  epidemiological studies on a prospective or retrospective basis. 707297  The skinfold  measurements used in this study include the triceps skinfold, the abdominal skinfold, the suprailiac skinf old, and the anterior thigh skinfold. These were often identified as appropriate measures for total body fat assessment in women, and are less invasive than others identified  192  APPENDICES  APPENDIX L (basis for methodologies continued...) as appropriate for female-like (gynoid) fat deposition. Also, they include two a limb skinfold (triceps and thigh) and two body skinf old (abdominal and suprailiac), which helps account for differing distribution of subcutaneous fat. 73 However, the literature consistently predicts a 2% to 3% error in percent body fat from anthropometry for a given sample of subjects. This error is low enough to be of practical usefulness if these results can be applied to other samples of the same population. 97 Furthermore, the highest standard errors were found for the age category of 40 years and above, and the percent category of below 20% body fat. 68  Errors can be even larger if  equations derived from one population are applied to another population 97 altogether. The errors can be attributed to three major effects: measurement errors, alterations in the composition and physical properties of certain tissues and the use of invalid assumptions in the derivation of body composition from anthropometric measuremen 73 ts. Therefore, by adding other anthropometric variables to the skinfold measurements, such as height-weight indices, the accuracy and reliability of body density predictions can be greatly improved. 707297 BMI and skeletal widths are of intermediate value, whereas skinfolds and circumferences or some combination of skinfolds, circumferences and skeletal widths are the most valuable anthropometric approaches. 97 El MID ARM FATAREA  The mid arm fat area represents a quick and easily obtainable two-dimensional model of fat stores, and is considered to be more indicative than the triceps skin fold. 73 ’ 70 Furthermore, the muscular development of an individual is best characterized in terms of the  193  APPENDICES  APPENDIX L (basis for methodologies continued...) limb circumferences, where direct anthropometric measurements must be used. 71  The  quantity of subcutaneous fat is dependent not only on the depth of the fat pad but also on the circumference of the muscle it covers. Therefore, for a given triceps skin fold, the larger the mid arm muscle circumference the greater the quantity of total fat stores. F) MUSCLE MASS  Skeletal muscle mass accounts for about 50% of the adipose tissue free mass, however there is considerable inter-individual variability. Commonly, one muscle group is selected with the assumption that (i) local anthropometry reflects the mass of that muscle group, and (ii) the mass of the muscle group is directly related to total muscle mass. Forearm circumference has been shown to be the best predictor of muscle mass and the correlation coefficient with muscle mass remained high without being corrected for skinfold measurements. Arm circumference alone, however, in estimating muscle mass, has been shown to consistently underestimate values, therefore by using it in conjunction with skinfold thickness corrected thigh circumference (correlation coefficient =0.990), a highly correlated muscle mass can be evaluated. Unfortunately, no equation specific for women has yet been generated  68  2. BA SAL BODY TEMPERA TURE  Basal body temperature provides a quantitative, non-invasive and valid technique to document menstrual irregularities, particularly shortened luteal phases and anovulatory 25 The basal body temperature of the adult healthy women describes a biphasic, cycles.  194  APPENDICES  APPENDIX L (basis for methodologies continued...) sinuous curve between two menstruations, and is an integrated function of endocrine ovarian 78 activity. The temperature runs at a low level during the postmenstrual phase, which is the interval in days between the date of the onset of the preceding menstruation and the date of the following ovulation (follicular phase).  During the postmenstrual phase the daily  temperature may oscillate over ±0.3°C or more, but from the early to the late postmenstrual phase the basal body temperature basically runs at a near horizontal pattern with a low grade descending slope. 78 Temperature reaches elevated levels during the premenstrual phase, which is defined as the interval from the date of ovulation to the date of the onset of the succeeding menstruation (luteal phase).  The shift to the elevated premenstrual temperature may be  preceded by a dip in the curve and the two phases are joined together by a phase of ascending inflection caused by an acute rise of 0.3 to 0.5°C of the temperature, within one to three days. During the elevated phase the temperature usually follows a nearly horizontal pattern as well, seldom with more than ±0.2°C variability. At the end of the premenstrual phase, three to zero days before the onset of the following menstruation, the curve deflects acutely to again attain the postmenstrual level. Frequently, however, the postmenstrual level is first reached on days two to five of the following cycle. 78 These daily temperatures, however, do not follow exactly a smooth geometric sine curve. From day to day, there is a degree of positive and negative scatter of the temperatures with a pattern that is highly specific for the individual women. This scatter does not eliminate  195  APPENDICES  APPENDIX L (basis for methodologies continued...) the S-shape course of the basal body temperature curve, however it does make the comparison between curves of different women difficult. 78 In menstrual cycles of 25 days and longer the length of the postmenstrual phase by basal body temperature varies in linear regression with the length of the cycles. In shorter menstrual cycles the regression curve becomes curvilinear.  However, the length of the  premenstrual phase varies independently of the length of the menstrual cycles, and, as mentioned previously, averages around 14 days. 78 Basal body temperature curves without a premenstrual phase of elevated temperature are called monophasic basal body temperature curves. These daily temperatures follow a horizontal course with irregular, accidental fluctuations of ±0.1 to ±0.2°C. By the absence of an elevated premenstrual temperature, monophasic temperature curves can be distinguished definitely from biphasic curves, and they are equated with anovulatory cycles. 78  3. DIETARY INTAKE  There is no ideal method for assessing food or nutrient intakes since none of the current methods are devoid of systematic errors, or prevent alterations in the food habits of the subjects. 73 Precise estimates (within 10% of the “true average”) of food and nutrient intakes are important in assessing the dietary status of a group of individuals, and food records have been identified as the best 94 alternative. 9 ’ 8 They consist of detailed, quantitative listings of all foods consumed by an individual over a given period. 99  Unfortunately, no  consensus has been reached regarding the number, spacing, and selection of record days  196  APPENDICES  APPENDIX L (basis for methodologies continued...) required to properly characterize the actual or usual food and/or nutrient intakes of subjects by food record. 739899 The number of days included in an estimated record varies between three, five or seven days. 73  A seven-day weighed record is usually considered appropriate for the  estimation of average intakes of individuals, however, respondent burden is high and problems with compliance may arise. 739499 Regression analysis has shown that records from the first two days of record keeping are more valid for assessing group comparisons than those from the last three days, because of deterioration in accuracy of recording.  Consequently,  compromises often have to be made between the collection of precise data on usual nutrient intakes of individuals and a high response rate. 73 Guthrie et  al.100  indicated that the mean  intake for three days can be considered a sensitive indicator of the usual intake for individuals. This investigation therefore used three sets of three-day dietary records from all subjects. In doing so, valid record keeping was maintained, and there was a total of nine days of food records, sufficient to accurately estimate true average intake of dietary fibre for groups of individuals, according to Basiotis et al. 98 79 demonstrated that food intakes in young women varies during the menstrual Dalvit cycle, being higher in the ten days following ovulation, the luteal phase, than in the ten days preceding it, the follicular phase. Therefore, the three 3-day records were also staggered to capture different phases of the menstrual cycle the follicular, the luteal and during ovulation. -  It has been demonstrated that group mean intakes, as well as individual usual intakes may vary with the day of the week, such as decreased energy intakes by college females, or  197  APPENDICES  APPENDIX L (basis for methodologies continued..) decreased dietary fibre consumption by both sexes, on weekends. 739499 Ideally, weekend days should be proportionately included in the dietary survey period for each subject, to account for potential day-of-the-week effects on food and nutrient intakes, however, it was too difficult to stagger both weekend/weekday and menstrual phases 73 simultaneously. Thus, since the main focus here is on the menstrual cycle, the weekend/weekday variability was eliminated by having all records collected on weekdays whenever possible. Random and/or systematic errors may have occurred during the measurement of food and/or nutrient intakes, however, random errors can be minimized by increasing the number of observations, though systematic errors are not easily eliminated. Through written and verbal instruction, respondent bias should have been removed, and by properly following guidelines respondent memory lapse and incorrect estimations of portion size should not have 7394 interfered. For a group of individuals true variability arises because dietary intakes differ among individuals (inter-subject) as well as within one individual (intra-subject) over time.  If  measurement errors are reduced, the precision of food records is a function of the overall true variability in nutrient intakes, and this is largely determined by the inter- and intra-subject 73100 variation.  4. EA TING RES TRA INT A SSESSMENT  The concept of “restrained eating” refers to the tendency of some people to restrict their food intake in order to control their body weight. Restrained eating has been assessed  198  APPENDICES  APPENDIX L (basis for methodologies continued...) by a 1 0-item Restraint Scale that predicted food 3 intake. However, there were many problem identified with that scale, particularly when applied to normal weight individuals. A second questionnaire was developed, looking at normal weight individuals who consciously restricted their food intake. Inconsistencies were also identified in this instance as well. 83 Therefore ’ 3 a third 58-item questionnaire, called the Three-Factor Eating Questionnaire was developed from these two, with improvements, and was used in this study to identify restrained eaters from unrestrained eaters. In developing this questionnaire, three stable factors related to restrained eating emerged: (i) “cognitive restraint of eating”; (ii) tendency toward disinhibition” or “emotional lability”; and, (iii) “hunger’. These are titled Factors #1, #2, and #3, respectively. Factor #2 predicts weight changes during depression, with the higher the disinhibition score, the greater the weight gain, and Factor #3 correlates with binge severity. 383  5. DUAL ENERGY X-RA Y ABSORPT!OMETRY Bone densiometry is a sensitive non-invasive method in determining bone mineral density that plays an important role in both the detection and the management of abnormal bone loss in patients at risk for 8485 osteoporosis. Dual energy x-ray absorptiometry appears to be an improved densiometric method that has significant advantages over existing techniques, and should allow for more accurate and precise measuremen 85 ts. The main advantages associated with the use of an x-ray source, over other options’ radioisotope, is its intensity, higher resolution, greater speed, and improved 85 precision. 1 ’ 01 It provides the  199  APPENDICES  APPENDIX L (basis for methodologies continued...) ability to detect small changes in bone mineral with a precision of better than 1  %85  Also,  since modern commercial dual energy x-ray absorptiometry scanners are practically stable throughout their operating life, and since their precision and accuracy rely on high technology and advanced digital signal processing techniques, a scanners’ effect on overall precision remains relatively small and constant. 84  6. URINE ANAL YSIS  Creatinine excretion has been found to be a reliable approximate parameter to estimate total skeletal muscle mass.  Rikimaru et al.’° 2 observed that total weight of muscle mass  closely correlated with urinary creatinine excretion, even in energy-restricted groups, and would therefore be a suitable measure even for restrained eaters. In humans, creatine is in part synthesized endogenously, but it is also present in the diet, mostly originating from meat, thus dietary intake of creatine in vegetarians is extremely limited. Because creatinine forms in vivo, by hydrolysis of creatine and phosphocreatine, Delanghe et at. 95 found that daily creatinine losses were significantly lower in their female vegetarians than those obtained for their age-matched reference population, and thus concluded that vegetarians have a considerable decrease in creatinine production rate. They suggested urinary analyte quantities be expressed per time unit rather than per mass unit. 95 Fasting urinary calcium does not come from the diet but from the tissues presumably -  bone or bone cells. An elevated value generally means an increase in net bone resorption (the difference between formation and resorption) which is seen mainly in post-menopausal  200  APPENDICES  APPENDIX L (basis for methodologies continued...) osteoporosis and primary hyperparathy 103 roidism. Hence the importance in measuring urinary calcium levels. Cortisol to creatinine ratio measurements reflect changes in adrenal cortisol secretion, to differentiate basal from elevated cortisol secretions, and were also calculated for all 4 ° 1 subjects. Goulding et al. 86 used a fasting urinary calcium to creatinine ratio to assess bone loss in premenopausal amenorrheic women, and found this value to be significantly higher in these women.  Because measurements of calcium to creatinine provide useful information  concerning bone metabolism they were calculated. The normal range of calcium to creatinine on a molar basis is 0.100.45.b03 A low fasting calcium to creatinine ratio generally means a low net bone resorption, as in osteomalacia and renal failure. 103 Two-hour fasting urine collections were used to measure urinary cortisol, calcium and creatinine. Elevated levels of cortisol are associated with increased stress with food intake, such as with restrained eaters. 27 Cortisol to creatinirie ratios can also be calculated as an indication of changes in adrenal cortisol secretions, which may be seen due to stress 27104 Urinary creatinine is measured as an index for estimating the approximate total response. amount of skeletal muscle mass , and calcium to creatinine ratios are used to assess bone 102 loss, as well as subsequent skeletal responses to treatment. 86  201  _  APPENDICES  APPENDIX M ETHICS APPROVAL FORM The University of British Coiumbia Office of Research Services  C92-311  CLINICAL SCREENING COMMI9PEE FOR RESEARCH AND OTHER STUDIES INVOLVING HUMAN SUBJECTS CERTIFICATE  of  APPROVAL  INVESTIGATOR:  Barr,  UBC DEPT:  Family & NuCr Sci  INSTITUTION:  Univ-IJBC;VGH  TITLE:  Subclinical menstrual disturbances: are they more common in premenopausal vegetarian women than in nonvegetarian women?  NUMBER:  C92-311  CO-INVEST:  Janelle,  APPROVED:  OCT 20 1992  S.  K.C.  Prior.  J.  The protocol describing the above-named project has been reviewed by the Committee and the experimental proced ures were found to be acceptable on ethical grounds for researc h involving human subjects.  Dr. B. cGilliv ay, Chairman Clinical Screening Committee  (  4  Dr. R.D. pratley / Directo Research Services  THIS CERTIFICATE OF APPROVAL IS VALID FOR THREE YEARS FROM THE ABOVE APPROVAL DATE PROVIDED THERE IS NO CHANGE IN THE EXPERIMENTAL PROCEDURES  202  APPENDICES  APPENDIX N Information regarding vegetarian subjects, re: ethnic group, employment, type of exercise, etc.. D It  Ethnic Group  Occupation  Type of Exercise  (hrs/wk)  (years)  10  Caucasian  None  Walk, Ski  2  20  11  Caucasian  Student  Weights, Dance  4  10  13  Caucasian  Student  Run, Swim, Bike  7  2  14  Caucasian  Waitress  Bike  6  2  15  Caucasian  Student  None  0  0  16  Caucasian/Asian  Student  Bike  6  0.5  17  Caucasian  Clerk  Walk, Swim  5  20  18  Caucasian  Student  Run, Swim, Bike  7  0.3  19  Asian  Student  Aerobics  3  10  21  Caucasian  Student  Weights, Swim  4  2  22  Caucasian  Secretary  Aerobics, Bike  3  8  23  Caucasian  Student  Aerobics, Run  3  0.3  24  Caucasian  Self-employed  Aerobics, Wghts  5  6  25  Caucasian  Accountant  Run, Weights  3  3  26  Caucasian  Fitness Program  Bike  7  10  27  Caucasian  Student  Swim, Bike  4  0.3  28  Black/Cauc/Asian  Student  Run, Swim, Wts  6  1  32  Caucasian  Student  Run, Swim, Bike  5  1 .5  33  Caucasian/Jewish  Musician  Swim, Weights  7  6  34  Caucasian  Researcher  None  0  0  35  Caucasian  C.A.  Walk  2  10  38  Caucasian  None  Walk  4  2  39  Caucasian/Jewish  Student  None  0  0  203  APPENDICES  APPENDIX N (continued...)  Information regarding nonvegetarian subjects, re: ethnic group, employment, type of exercise, etc.. ID #  Ethnic Group  Occupation  Type of Exercise  (hrs/wk)  (years)  40  Caucasian  Physiotherapist  Swim, Bike  4  10  41  Caucasian  Student  Run, Swim, Bike  4  4  42  Caucasian  Student  Aerobics, Wghts  4  10  46  Caucasian  Student  Run, Aerobics  4  10  47  Caucasian  Dietitian  Run, Ski  5  2  48  Caucasian  Student  Aerob, Run, Wts  5  3  49  Caucasian  Health Educator  Walk, Weights  2  0.3  50  Caucasian  Student  Run, Bike, Wts  7  4  51  Caucasian  Student  None  0  0  53  Caucasian  Student  Bike, Weights  2  0.5  55  Caucasian  Student  Walk  3  6  56  Caucasian  Research Tech.  None  0  0  57  Caucasian  Student  Bike  2  12  58  Asian  Student  Swim, Aerb, Wt  3  5  59  Asian  Data Analyst  None  0  0  62  Asian  Student  Aerobics  2  0.5  63  Caucasian  Student  Swim, Weights  4  0.5  64  Caucasian  Student  Swim, Aerobics  5  5  65  Caucasian  Student  Bike  2  1 .5  66  Caucasian  Lab. Tech.  Aerb, Swm, Bike  5  10  67  Cauc/Jewish  Lab. Tech.  Walk, Run, Wts  4  2  68  Caucasian  Program Work  Walk  3  0.5  204  APPENDICES  APPENDIX 0 Mean (± SD) characteristics of 62 premenopausal women, categorized by remaining volunteers and “drop-out’ subjects. Characteristics  Volunteers  Drop-outs  Number of Women  45  17  Age (years)  27.2 ±5.1  24.3 ±4.1  0.040  Age at Menarche (yrs)  13.3±1.5  13.7±2.1  0.485  Exercise (hrs/wk)  3.5±2.1  4.8±2.1  0.043  Exercise Regimen (yrs)  4.5±5.1  4.3±4.5  0.881  Weight (kg)  60.3 ±7.9  58.0± 6.7  0.300  Height (cm)  166.0±8.1  165.3±6.5  0.725  Body Mass Index (kg/rn ) 2  21.8±2.2  21.2± 1.9  0.322  Triceps Skinfold (mm)  20.7 ±6.8  19.6 ±6.4  0.566  Abdomen Skinfold (mm)  23.5±7.6  22.1 ±8.1  0.514  Suprailiac Skinfold (mm)  13.0±6.3  13.0±7.0  0.997  Thigh Skinfold (mm)  33.6 ±8.0  30.1 ±8.9  0.146  Sum of Skinfolds (mm)  90.7 ±24.9  84.7 ±27.2  0.413  Body Density (kg/rn ) 3  1.04 ±0.01  1.04 ±0.01  0.349  Body Fat (%)  25.7 ±5.5  24.1 ±6.0  0.347  Total Body Fat (kg)  15.7 ±4.8  14.2 ±4.9  0.291  Mid Arm Circumference (cm)  28.1 ±2.5  27.3±2.3  0.281  Mid Arm Fat Area (mm ) 2  2594 ±967  2391 ±946  0.461  Forearm Girth (cm)  24.7± 1.3  24.2± 1.2  0.186  Thigh Circumference (cm)  51.3 ±3.6  50.7 ±3.0  0.526  Corrected Thigh Circ. (cm)  40.7 ±2.9  41.2 ±2.5  0.569  Total Muscle Mass (kg)  28.8 ±4.3  28.7 ±3.7  0.876  Eating Inventory Score (56)  19.3±8.0  17.0±8.6  0.329  Restraint Score (21)  8.4 ±4.6  7.1 ±4.8  0.302  Disinhibition (20)  5.9 ±3.7  5.4±4.1  0.619  Hunger (15)  5.0±2.4  4.6±2.6  0.602  Significant difference at P<O.05.  P value -  -f-Comparisons were made using student’s unpaired t-test.  205  APPENDICES  APPENDIX P Vegetarian women’s summarized responses to the question “I became a vegetarian because...”, as well as additional information pertaining to the vegetarian subgroup. ID #  Type of Vegetarian  Years  Reasons  10  Lacto-ovo  5-10  M,H  11  Vegan  2  R,H  13  Lacto-ovo  2  M,H,V,E  14  Vegan  2  M,V  15  Lacto-ovo  2  M  16  Lacto-ovo  2  M  17  Lacto  2  M,H  18  Lacto-ovo  2-5  V  19  Vegan  2-5  H  21  Vegan  2  H  22  Vegan  2  M,H,V  23  Lacto  2-5  H  24  Lacto  2-5  H  25  Lacto-ovo  2-5  H  Father, Moth, Sis (2)  26  Vegan  5-10  M,H,V  Partner, Sisters (2)  27  Lacto-ovo  5-10  M,H  28  Vegan  2-5  H  29  Lacto-ovo  2-5  M  Brother  0.5  32  Lacto-ovo  2  M,H  Partner, Sister  2, 1  33  Lacto-ovo  >10  H,V  34  Vegan  2  H  35  Lacto-ovo  >10  M  38  Lacto  2-5  H  Father, Moth, Sis (2)  1, 1, 1, 1  39  Lacto-ovo  >  H  Mother, Sister  20, 15  M  =  moral; H  =  health; E  =  10  economic; V = environmental; H  Family Mem. Veg.  Yrs. for Fam. -  Grandmother, Sister  38, 2  -  -  -  -  -  Partner  4 -  Partner, Mother (Law)  2, 3 -  Sister  3 -  -  -  -  7, 3 -  -  -  -  -  =  religious.  206  APPENDICES  APPENDIX Q Mean (±SD) nutrient intakes for the three stages of the menstrual cycle for 45 women. Daily Dietary Intakes  Follicular  Ovulatory  Luteal  Calories (kcal)  1917±508’  2076±589  2099±628”  Protein (g)  60.1 ±21.8’  69.6±24.1  67.9±23.9”  Carbohydrates (g)  282.8 ± 77.4  287.8 ± 81.2  297.6 ± 91.7  Fibre (g)  24.9 ±9.5  25.4± 10.7  26.3± 10.7  Total Fat (g)  65.1 ±27.4’  75.8 ±29.2  76.1 ±29.3”  Saturated Fat (g)  19.7±10.3’  24.0± 11.2  24.9±15.2”  Monounsaturated Fat (g)  25.3 ± 12.2  28.6 ± 11.9  28.9 ± 11.4”  Polyunsaturated Fat (g)  14.4±6.8  16.3 ±8.0  15.4±5.1  Cholesterol (mg)  151 ±100  186± 109  204±148”  Vitamin A(RE)  1447±1518  1710±1319  2076±2127  Thiamin (mg)  1.42 ±0.57  1.67 ±0.88  1.47 ±0.50  Riboflavin (mg)  1.44 ± O.44  1.64 ± 0.53  1.60 ± 0.52”  Niacin (mg)  15.0±6.7  17.0±6.1  16.1 ±6.4  Vitamin B (mg)  1.47±0.59’  1.66±0.59  1.61 ±0.60  12 (jig) Vitamin B  2.22±2.21  2.26± 1.50  2.90 ±3.14  Folacin (jig)  284±91  344±206  302± 126  Pantothenic Acid (mg)  4.19±1.35’  5.04±1.86  4.90±1.79”  Vitamin C (mg)  130±68  139 ±76  146 ±84  Calcium (mg)  783 ±374’  891 ±428  935±462”  Copper (mg)  1.45±0.51’  1.64±0.66  1.71 ±0.68”  Iron (mg)  14.2±5.5  15.2±5.0  16.0±5.9”  Magnesium(mg)  301 ±110  339± 124  334±113  Phosphorus(mg)  1180±431’  1353±430  1365±511”  Potassium (mg)  2871 ±990  3085 ±933  3253±965”  Sodium(mg)  2409±1030  2556±912  2692±1134  Zinc (mg)  8.59±3.12’  10.00±3.37  10.56±6.83”  Significant difference between the follicular and the ovulatory phases at a level of P<O.05. “Significant difference between the follicular and the luteal phases at a level of P<O.05. +Comparisons made using paired t-test.  207  APPENDICES  APPENDIX R  Mean (±SD) nutrient intakes for weekdays vs. weekend days for 45 premenopausal women. Daily Dietary Intakes  Weekdays  Weekend Days  P value  Calories (kcal)  2021 ±673  2080±773  0.486  Protein (g)  65.3 ±24.8  67.9 ±39.0  0.464  Carbohydrate (g)  286.8 ± 100.6  284.4 ± 96.4  0.658  Fibre (g)  25.6 ± 12.2  24.7 ± 12.9  0.556  Total Fat (g)  71.5 ±35.2  77.9 ±41.5  0.150  Saturated Fat (g)  22.4± 14.5  24.8±15.6  0.171  Monounsaturated Fat (g)  27.2 ± 14.8  29.8 ± 17.7  0.164  Polyunsaturated Fat (g)  15.2±9.5  16.4± 10.1  0.312  Cholesterol (mg)  176±155  195±180  0.313  Vitamin A (RE)  1668±2042  1886±2331  0.394  Thiamin (mg)  1.52 ±0.85  1.51 ±0.73  0.859  Riboflavin (mg)  1.55 ±0.58  1.60 ±0.74  0.539  Niacin (mg)  15.8±7.4  16.6±9.2  0.383  Vitamin 86 (mg)  1.55 ±0.71  1.70 ±0.86  0.109  Vitamin  2.39±2.75  2.66±3.37  0.430  Folacin (jig)  314± 191  293± 147  0.337  Pantothenic Acid 1mg)  4.70 ±2.11  4.61 ±2.35  0.735  VitaminC(mg)  135±95  144±102  0.418  Calcium (mg)  858 ±467  876±552  0.761  Copper (mg)  1.57±0.77  1.78±1.06  0.048  Iron (mgI  15.0 ±6.8  15.8 ±7.5  0.351  Magnesium (mg)  325± 146  343±180  0.343  Phosphorus (mg)  1292±514  1307±662  0.823  Potassium (mg)  3051 ±1094  3183± 1410  0.354  Sodium (mgI  2530± 1332  2447± 1295  0.602  Zinc (mgI  9.36 ±3.86  10.80± 12.65  0.084  812  (jig)  Significant difference at a level of P<O.05. +Comparisons were completed using student’s unpaired t-test.  208  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0087266/manifest

Comment

Related Items