UBC Faculty Research and Publications

Diets and selected lifestyle practices of self-defined adult vegetarians from a population-based sample… Bedford, Jennifer L; Barr, Susan I Apr 13, 2005

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

Item Metadata


52383-12966_2004_Article_21.pdf [ 256.24kB ]
JSON: 52383-1.0223208.json
JSON-LD: 52383-1.0223208-ld.json
RDF/XML (Pretty): 52383-1.0223208-rdf.xml
RDF/JSON: 52383-1.0223208-rdf.json
Turtle: 52383-1.0223208-turtle.txt
N-Triples: 52383-1.0223208-rdf-ntriples.txt
Original Record: 52383-1.0223208-source.json
Full Text

Full Text

ralInternational Journal of Behavioral ssBioMed CentNutrition and Physical ActivityOpen AcceResearchDiets and selected lifestyle practices of self-defined adult vegetarians from a population-based sample suggest they are more 'health conscious'Jennifer L Bedford and Susan I Barr*Address: Human Nutrition, University of British Columbia, 2205 East Mall, Vancouver, BC, CanadaEmail: Jennifer L Bedford - jbedford@interchange.ubc.ca; Susan I Barr* - sibarr@interchange.ubc.ca* Corresponding author    DietvegetarianHealth behaviorFood habitsHealth attitudes and behaviors.AbstractBackground: Few population-based studies of vegetarians have been published. Thus we compared self-reported vegetarians to non-vegetarians in a representative sample of British Columbia (BC) adults, weighted toreflect the BC population.Methods: Questionnaires, 24-hr recalls and anthropometric measures were completed during in-personinterviews with 1817 community-dwelling residents, 19–84 years, recruited using a population-based healthregistry. Vegetarian status was self-defined. ANOVA with age as a covariate was used to analyze continuousvariables, and chi-square was used for categorical variables. Supplement intakes were compared using the Mann-Whitney test.Results: Approximately 6% (n = 106) stated that they were vegetarian, and most did not adhere rigidly to a flesh-free diet. Vegetarians were more likely female (71% vs. 49%), single, of low-income status, and tended to beyounger. Female vegetarians had lower BMI than non-vegetarians (23.1 ± 0.7 (mean ± SE) vs. 25.7 ± 0.2 kg/m2),and also had lower waist circumference (75.0 ± 1.5 vs. 79.8 ± 0.5 cm). Male vegetarians and non-vegetarians hadsimilar BMI (25.9 ± 0.8 vs. 26.7 ± 0.2 kg/m2) and waist circumference (92.5 ± 2.3 vs. 91.7 ± 0.4 cm). Femalevegetarians were more physically active (69% vs. 42% active ≥4/wk) while male vegetarians were more likely touse nutritive supplements (71% vs. 51%). Energy intakes were similar, but vegetarians reported higher % energyas carbohydrate (56% vs. 50%), and lower % protein (men only; 13% vs. 17%) or % fat (women only; 27% vs. 33%).Vegetarians had higher fiber, magnesium and potassium intakes. For several other nutrients, differences byvegetarian status differed by gender. The prevalence of inadequate magnesium intake (% below Estimated AverageRequirement) was lower in vegetarians than non-vegetarians (15% vs. 34%). Female vegetarians also had a lowerprevalence of inadequate thiamin, folate, vitamin B6 and C intakes. Vegetarians were more likely than non-vegetarians to consider various health conditions and food/nutrition concerns when choosing foods.Conclusion: In this population-based study, evidence was obtained to indicate that vegetarians appear more'health conscious' than non-vegetarians, although specific differences were not always consistent by gender.Additional population-based studies are required to determine if the observed gender differences exist in otherPublished: 13 April 2005International Journal of Behavioral Nutrition and Physical Activity 2005, 2:4 doi:10.1186/1479-5868-2-4Received: 15 December 2004Accepted: 13 April 2005This article is available from: http://www.ijbnpa.org/content/2/1/4© 2005 Bedford and Barr; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.Page 1 of 11(page number not for citation purposes)populations.International Journal of Behavioral Nutrition and Physical Activity 2005, 2:4 http://www.ijbnpa.org/content/2/1/4BackgroundInterest in the dietary habits of vegetarians emerges fromresearch suggesting that vegetarians have a lower preva-lence of various chronic diseases that currently plague thedeveloped world [1,2]. It has been hypothesized thatthese findings are due to vegetarians' dietary habits, whichmore closely follow recommendations for healthy eating[3], and also to their lower BMI [4]. Recently, however, ithas been found that mortality rates did not differ amongvegetarians and similar 'health conscious' omnivores [5,6]despite vegetarians' lower age-adjusted BMI [5-7]. Thissuggests that other lifestyle behaviors commonlyobserved in health conscious individuals may be respon-sible for the observed beneficial health effects. Yet, it isimportant to note that the majority of reports of vegetari-ans' dietary intakes and lifestyle behaviors are from con-venience samples.Few population-based studies have examined vegetarians'dietary intake and habits. Kennedy and coworkers usedpopulation representative data from the Continuing Sur-vey of Food Intakes by Individuals (CSFII) to gain insightsinto the dietary patterns of vegetarians [8]. Specifically,they compared the intakes of those who consumed meaton data collection days to those who did not. However,many of those who did not consume meat were likely not'vegetarian' as the proportion of those grouped as vegetar-ian by this method (15%) was much higher than thosewho self-identified as vegetarian in the same sample(2.5%) [9].CSFII data were also analyzed to compare self-identifiedvegetarians to non-vegetarians [9]. Adult vegetarians werefound to have lower BMI as well as lower intakes of totalfat, saturated fat and cholesterol and higher intakes offiber and fruits [9]. However, differences by gender or agewere not examined, and if age or gender differences in theprevalence of vegetarian status exist, as has been observedin other population representative studies of vegetarians[10], these could confound the results. For example, ifvegetarians were more likely than non-vegetarians to beyounger and female this could impact group BMIs andintakes of nutrients that differ by gender or age. In addi-tion, differences in lifestyle behaviors were not assessed.This could be important to consider as it has been sug-gested that other lifestyle behaviors may be the determi-nant of differences observed in health conditions byvegetarian status [5,6].Thus, the aim of the present analysis was to describe andcompare the demographics, lifestyle behaviors, dietaryintake, nutritive supplement use, and food and nutritionconcerns of male and female self-defined vegetarians andsample of adults from the province of British Columbia,Canada.MethodsThe data used for this analysis were collected as part of theBritish Columbia Nutrition Survey (BCNS). Details of themethodology used for the BCNS including sampling strat-egies, survey instruments, and data entry and analysis aredescribed in detail elsewhere [11], and are summarizedbelow.ParticipantsAdults aged 19 to 84 years living in BC were recruited forthe BCNS using the BC Health Registry – a central reposi-tory of individuals who receive health services in BC.Exclusion criteria included those living in care or correc-tional facilities, military bases, or Indian Reserves, as wellas pregnant and lactating women. Less than 3% of thepopulation was excluded on these grounds. The samplewas stratified by age, sex and geographical region. Thestudy protocol was approved by the University of BritishColumbia's Behavioral Research Ethics Board, and writteninformed consent was obtained from all participants.MeasuresThe BCNS included a 24-hour recall; questions on foodhabits and choices, physical activity and demographics;and anthropometric measurements.24-Hour RecallTo obtain information on dietary intake each individualcompleted a 24-hour recall conducted by personal inter-view using the open-ended, multiple-pass technique forwhich each participant recalled all foods and beveragesconsumed on the previous day (midnight to midnight).Food models and household measures were used to esti-mate portion sizes. In one third of the sample, a second24-hour recall was conducted on a different weekday atleast one week following the first recall. Supplement datawere also obtained during the 24-hour recall. Participantswere first asked whether they took any nutritional supple-ments yesterday followed by a question about whetherthey took any supplements within the past month thatdiffered from the ones taken yesterday. The frequency(daily, weekly or monthly) and the number or amount ofeach supplement were also recorded. When possible,brand names and the drug identification number (DIN)were recorded.Food and Nutrition HabitsParticipants were read a list of health-related reasons forchoosing and avoiding foods to determine nutrition andfood concerns. Vegetarian status was assessed by askingPage 2 of 11(page number not for citation purposes)non-vegetarians from a population-based representative participants if they considered themselves to be a vegetar-ian. Those who answered 'yes' were also asked if they everInternational Journal of Behavioral Nutrition and Physical Activity 2005, 2:4 http://www.ijbnpa.org/content/2/1/4ate animal products including dairy, eggs, fish/seafood,poultry and red meat.Physical ActivityQuestions from previously validated instruments wereasked to obtain information on physical activity, includ-ing the frequency of mild, moderate and strenuous activ-ity; and motivational readiness for exercise [12,13].Demographic QuestionnaireQuestions regarding age, gender, marital status, educationand income were asked to characterize the sample.AnthropometricsWeight, without shoes, hats or any heavy clothing oritems, was measured using a weekly calibrated electronicscale. Height was measured using a setsquare and measur-ing tape, and girths were assessed using a measuring tape.Weight and height were measured and recorded once;waist circumferences was measured and recorded at leasttwice. BMI was calculated from weight and height (kg/m2).ProcedureEligible residents who chose not to participate were askedto complete a non-response questionnaire to determine ifnon-responders differed from those who agreed to partic-ipate. Those who agreed to participate were interviewed inperson by trained interviewers. Interviews lasted approxi-mately 90 minutes and most took place in participants'homes.AnalysisResponse rateIn a large population survey such as the BCNS, theresponse rate depends on whether individuals that couldnot be contacted (unresolved cases) were eligible to par-ticipate in the survey. Lower and upper bounds for theresponse rate can be calculated based on the assumptionthat all unresolved cases are eligible (lower bound), orthat all unresolved cases are ineligible (upper bound).Nutrient intake dataAll data were sent to the Bureau of Nutritional Sciences atHealth Canada and were entered into the Nutrition Sur-vey System, a software program that uses the CanadianNutrient File and a recipe database adapted from theUnited States Department of Agriculture CSFII. The Cana-dian Nutrient File was updated to reflect fortification ofgrain products with folic acid that began in Canada in1998. Data on nutritional supplements were enteredusing the DIN or by name and/or nutrient content. Themethod of estimating the distribution of usual intakesremove within-person variability from population distri-butions of nutrient intakes, yielding an adjusted distribu-tion of usual intakes for age-sex groups [11,14]. Then themonthly supplement data were expressed per day andadded to the adjusted usual intake distribution. The pro-portion of this distribution that fell below the EstimatedAverage Requirement (EAR) was used to estimate theprevalence of inadequate nutrient intakes in an age-sexgroup, for nutrients with an EAR and a symmetricalrequirement distribution [14]. This analysis could not beconducted for vitamin A as the EAR is expressed in newRetinol Activity Equivalents, whereas intake data were inRetinol Equivalents. Food intake data were also expressedas number of servings from the food groups included inCanada's Food Guide to Healthy Eating (CFGHE).Statistical analysisAll data were weighted to reflect the BC population basedon gender, age, and geographical region, and were ana-lyzed using the Statistical Package for Social Sciences(SPSS; v11.0, Chicago, Ill., 2002). One-way ANOVA withage as a covariate was used to analyze parametric data(demographics, nutrient intakes) and Chi-square analysiswas used to analyze categorical data (lifestyle behaviors,supplement usage, nutrition concerns) between vegetari-ans and non-vegetarians. The Mann-Whitney test wasused to compare differences by vegetarian status in sup-plemental nutrient intake as the data were not normallydistributed. Non-parametric tests were also applied tonutrient intake data that were not normally distributed.However, findings did not differ from those of parametricanalysis and thus ANOVA was applied to all nutrients forconsistency, and because ANOVA allowed considerationof effects of covariates. The data were also examined toassess whether demographic differences between vegetar-ians and non-vegetarians (other than age and sex) mayhave influenced the results. The significance level was setat p = 0.05 for all statistical measures.ResultsResponse rate lower and upper bounds were 42% and52%, respectively [11]. Approximately 66% of those whodeclined to participate completed the non-respondentsurvey. Using this information, it was found that BCNSparticipants were less likely to smoke (17% vs. 23%) andmore likely to use vitamin/mineral supplements (66% vs.60%) and hold university degrees than non-participants(14% vs. 9%). However, in comparison to the general BCpopulation, BCNS participants had a similar prevalence ofsmoking, although study participants were more educated(21% vs. 13% completed university) and more men weremarried (64% vs. 55%). Because of the difference in edu-cational attainment between survey participants and thePage 3 of 11(page number not for citation purposes)from food sources alone and from food sources and sup-plements used the data from duplicate 24-hour recalls topopulation, the effect of this variable on nutrient intakeswas examined. Educational attainment was associatedInternational Journal of Behavioral Nutrition and Physical Activity 2005, 2:4 http://www.ijbnpa.org/content/2/1/4Table 1: Participant demographics by vegetarian statusNon-vegetarian (n = 1711)Vegetarian (n = 106) Test Statistic1 P valueAge (years, mean ± SE) 44.8 ± 0.4 41.5 ± 1.8 F = 3.63 0.057Age Group χ2 = 9.80 0.02019 – 30 years 24.8% 38.1%31 – 50 years 40.9% 31.4%50 – 70 years 24.3% 20.0%71+ years 10.1% 10.5%Sex (% female) 49.2% 70.8% χ2 = 18.58 <0.001Marital status χ2 = 17.55 0.001Single 22.5% 35.2%Married 62.7% 50.5%Widowed 5.3% 10.5%Divorced/separated 9.5% 3.8%Education level χ2 = 0.45 0.800Secondary or less 33.3% 32.1%Technical or some university 47.6% 46.2%University graduate 19.1% 21.7%Low income 22.8% 37.2% χ2 = 9.28 0.002Weight (kg, mean ± SE)Men 83.1 ± 0.6 82.2 ± 3.0 F = 0.10 0.754Women 67.9 ± 0.6 62.5 ± 1.8 F = 7.92 0.005BMI (kg/m2, mean ± SE)Men 26.7 ± 0.2 25.9 ± 0.8 F = 0.81 0.346Women 25.7 ± 0.2 23.1 ± 0.7 F = 13.66 <0.001BMI category – men χ2 = 0.92 0.821Underweight (< 18) 0.2% 0.0%Normal weight (18-<25) 48.8% 51.6%Overweight (≥ 25-<30) 31.6% 35.5%Obese (≥ 30) 19.4% 12.9%BMI category – women χ2 = 16.34 0.001Underweight (< 18) 1.1% 0.0%Normal weight (18-<25) 59.6% 83.1%Overweight (≥ 25-<30) 21.2% 12.7%Obese (≥ 30) 18.1% 4.2%Waist Circ. (cm, mean ± SE)Men 91.7 ± 0.4 92.5 ± 2.3 F = 0.11 0.740Women 79.8 ± 0.5 75.0 ± 1.5 F = 9.66 0.002Health Condition – Men (% yes)Diabetes 5.5 3.2 χ2 = 0.31 0.580Heart disease 6.1 19.4 χ2 = 8.59 0.003Stroke 0.8 3.2 χ2 = 1.99 0.158High blood pressure 11.6 6.5 χ2 = 0.79 0.375High cholesterol 14.7 9.7 χ2 = 0.61 0.434Cancer 4.5 6.5 χ2 = 0.27 0.6062 Page 4 of 11(page number not for citation purposes)Osteoporosis 0.7 0.0 χ = 0.22 0.643None of the above 70.9 63.3 χ2 = 0.81 0.370International Journal of Behavioral Nutrition and Physical Activity 2005, 2:4 http://www.ijbnpa.org/content/2/1/4only with the intakes of vitamin C and vitamin B12: itappeared that those with more education had higherintakes of vitamin C and lower intakes of vitamin B12 [11].Of the 1817 participants, 5.8% (n = 106) identified them-selves as vegetarian. Those who identified themselves asvegetarian were asked if they 'ever' ate various animalproducts. It appeared that the majority of self-identifiedvegetarians did not adhere rigidly to a vegetarian dietarypattern: In terms of tissue protein, 74.9% consumed fishand/or seafood at least occasionally, while 57.6% con-sumed poultry and 22.4% consumed red meat at leastoccasionally. Dairy products were used at least occasion-ally by 97.3% and eggs by 92.3%.DemographicsThe demographic characteristics of vegetarian and non-vegetarian participants are presented in Table 1. Vegetari-ans tended to be younger than non-vegetarians (p =0.057), and the age group distribution differed signifi-cantly, with more vegetarians falling in the 19 to 30 yearrange. Groups differed in sex distribution with womenrepresenting over 70% of vegetarians and half of non-veg-etarians. Although the majority of both groups was mar-ried, there was a significant difference in marital statusdistribution, with vegetarians more likely to be single. Inaddition, vegetarians were significantly more likely to beof low income status although there were no differencesin education level. Accordingly, data were examined toassess the effects of marital status and income status.Female vegetarians had a significantly lower mean age-adjusted body weight and mean BMI than non-vegetari-ans, as well as a lower waist circumference. Low incomestatus and marital status did not affect these variables. Inaddition, vegetarian women were significantly less likelyto be classified as overweight or obese (17% vs. 40%).Conversely, for males, weight, BMI, and BMI category dis-sified as overweight or obese. There were no significantdifferences in age-adjusted waist circumference betweenvegetarian and non-vegetarian men.The prevalence of certain health conditions differed byvegetarian status. Male vegetarians had a higher preva-lence of heart disease while female non-vegetarians weremore likely to report cancer and hypertension.Lifestyle behaviorsFemale vegetarians were more likely than non-vegetariansto report moderate to strenuous physical activity four ormore times weekly (69% vs. 42%, χ2 = 21.69, p < 0.001),and more women vegetarians than non-vegetarians werein the 'action' or 'maintenance' stages of motivationalreadiness for exercise (76% vs. 53%, χ2 = 21.67, p <0.001). Although single women in the sample as a wholewere more active than women who were married, wid-owed, divorced or separated, when age was consideredmarital status did not affect physical activity level. Lowincome status was not associated with physical activity inwomen. In contrast, male vegetarians and non-vegetari-ans did not differ in the amount of weekly exercise: themajority of both groups participated in moderate to stren-uous physical activity less than four times a week (55% vs.52% respectively, χ2 = 0.18, p = 0.913). Men were alsosimilar in terms of the distribution of the stage of motiva-tional readiness for exercise (χ2 = 1.78, p < 0.776). On theother hand, smoking status differed between vegetariansand non-vegetarians for both men (3% vs. 18%, χ2 = 4.15,p < 0.05) and women (0% vs. 18%, χ2 = 15.85, p < 0.001).Supplement use and intakesThe majority of all groups reported nutritive supplementuse. Among men, significantly more vegetarians thannon-vegetarians reported using supplements (71% vs.51%, χ2 = 4.76, p = 0.029). However, for women, thedifference was not significant with 76% of vegetarians andHealth Condition – Women (% yes)Diabetes 4.9 1.3 χ2 = 1.97 0.160Heart disease 3.7 4.0 χ2 = 0.02 0.889Stroke 1.9 0.0 χ2 = 1.45 0.228High blood pressure 15.3 6.7 χ2 = 4.15 0.042High cholesterol 11.3 6.8 χ2 = 1.44 0.230Cancer 8.3 1.3 χ2 = 4.70 0.030Osteoporosis 6.1 8.0 χ2 = 0.44 0.506None of the above 67.7 78.4 χ2 = 3.62 0.0571. Statistical analysis consisted of ANOVA (vegetarian versus non-vegetarian, with age as a covariate) for continuous variables. Chi-square analysis was used for categorical variables.Table 1: Participant demographics by vegetarian status (Continued)Page 5 of 11(page number not for citation purposes)tribution were very similar between vegetarians and non-vegetarians, with approximately 50% of both groups clas-68% of non-vegetarians reporting supplement use (χ2 =1.88, p = 0.170). On the other hand, female vegetariansInternational Journal of Behavioral Nutrition and Physical Activity 2005, 2:4 http://www.ijbnpa.org/content/2/1/4who used supplements reported using a higher number ofsupplements than non-vegetarians (3.5 ± 0.4 (mean ± SE)vs. 2.3 ± 0.1, F = 15.33, p < 0.001), while the numbersused by males were similar (1.3 ± 0.4 vs. 1.5 ± 0.1, F =0.40, p = 0.53).The proportion of individuals using supplements of manynutrients differed significantly by vegetarian status (datanot shown). More vegetarians than non-vegetarians ofboth sexes used a supplement containing the B vitamins.However, for other nutrients, the results were split alonggender lines: more female vegetarians used supplementsof all other vitamins/minerals except for vitamin E andcalcium; whereas among males, there were no additionalsignificant differences by vegetarian status.Differences in supplemental nutrient intake were also evi-dent between vegetarians and non-vegetarians who usedsupplements (data not shown). Among women, vegetari-ans had significantly higher median supplemental intakesof calcium, iron, magnesium, potassium, niacin, folic acidand vitamins A, D and B12. Among men, vegetarians hada significantly higher median supplemental intake of vita-min C.Energy and nutrient intakesAge-adjusted energy and nutrient intakes from food arepresented by vegetarian status and gender (see AdditionalFile 1). Income status did not affect nutrient intakes [11].There were no significant differences in energy intakebetween vegetarians and non-vegetarians, but energy dis-tribution differed significantly by vegetarian status forboth sexes. Compared to non-vegetarians, both male andfemale vegetarians consumed significantly more energy ascarbohydrate. Among men, vegetarians had a significantlylower proportion of energy from protein. Conversely,female vegetarians had a significantly lower percentage ofenergy from fat.Male and female vegetarians had significantly higherintakes of fiber, magnesium and potassium. Female vege-tarians had significantly higher intakes of carbohydrate(g), phosphorus, thiamin, pantothenic acid, vitamin B6,and folate and lower intakes of saturated fat and sodium.Conversely, male vegetarians had significantly higherintakes of vitamin C and calcium, and lower intakes ofprotein (g), niacin and cholesterol.The prevalence of inadequate intakes of selected nutrientsby vegetarian status and gender are shown in Figure 1.These data are based on combined intake from food plussupplements. No differences were observed in the preva-lence of inadequate intakes of vitamin B12 or zinc, but sig-also significant differences for both genders in the preva-lence of inadequacies for thiamin, although in both casesthe prevalence of inadequacy was <10%: for men, morevegetarians were below the EAR, while for women, non-vegetarians were more likely to be below the EAR. Femalenon-vegetarians were also significantly more likely to bebelow the EAR for vitamin C, vitamin B6 and folate.Intake based on Canada's Food Guide to Healthy EatingAnalyses of dietary intake based on servings of CFGHEfood groups by vegetarian status and gender, adjusted forage, are presented in Table 2. Vegetarians of both gendershad a significantly higher number of servings of fruits andvegetables. Only female vegetarians had a significantlyhigher number of servings of grain products while onlymale vegetarians had a significantly higher number ofservings of milk products and a significantly lowernumber of servings of meat and alternatives.The proportions of participants meeting the minimumnumber of CFGHE servings by vegetarian status and gen-der are also displayed in Table 2. Vegetarians of both gen-ders were less likely to meet the minimumrecommendations for meat and alternatives. Amongwomen, vegetarians were more likely to meet the mini-mum servings of grain products, while among men, vege-tarians were more likely to meet recommendations forfruits and vegetables as well as milk products.Food and nutrition concernsFor both genders, vegetarians were significantly morelikely to report 'maintaining/improving health' as a con-sideration when choosing/avoiding foods than non-vege-tarians (men: 100% vs. 65%, p < 0.001; women: 93% vs.77%, p = 0.001). Male vegetarians were more likely thannon-vegetarians to also consider heart disease (77% vs.38%, p < 0.001) and high blood pressure (45% vs. 25%,p = 0.013) when choosing/avoiding foods. Female vege-tarians were more likely than non-vegetarians to also con-sider cancer (41% vs. 30%, p = 0.05), osteoporosis (61%vs. 38%, p < 0.001) and food allergies/intolerances (43%vs. 30%, p = 0.026), and were less likely than non-vegetar-ians to consider weight gain (46% vs. 61%, p = 0.013)when choosing/avoiding foods. Finally, more non-vege-tarians than vegetarians reported that they did not con-sider any of the aforementioned factors (maintaining/improving health, heart disease, cancer, osteoporosis,high blood pressure, weight gain, food allergies/intoler-ances) when choosing/avoiding foods (men: 26% vs. 0%,p = 0.001; women: 12% vs. 1%, p = 0.004).Vegetarians of both genders were more likely than non-vegetarians to report choosing foods because of the nutri-Page 6 of 11(page number not for citation purposes)nificantly more non-vegetarians had intakes below theEAR for magnesium in both men and women. There wereents they contain (men: 73% vs. 53%, p = 0.025; women:88% vs. 68%, p < 0.001), and to report avoiding foodsInternational Journal of Behavioral Nutrition and Physical Activity 2005, 2:4 http://www.ijbnpa.org/content/2/1/4because of their fat content (men: 77% vs. 59%, p = 0.041;women: 84% vs. 72%, p = 0.023). Male vegetarians weremore likely than non-vegetarians to also consider the typeof fat (63% vs. 36%, p = 0.003) and the amount of unsatu-rated fat (63% vs. 28%, p < 0.001) when choosing foodsand to avoid foods because of their cholesterol (55% vs.36%, p = 0.032) and saturated fat (58% vs. 38%, p =0.021) content. Female vegetarians were more likely thannon-vegetarians to also report avoiding foods because oftheir salt content (57% vs. 45%, p = 0.043). On the otherhand, more non-vegetarians than vegetarians reportedthat, when choosing foods, they did not consider any ofnutrient content, type of fat, amount of unsaturated fat, orfiber content (men: 34% vs. 16%, p = 0.035; women: 19%the fat, salt, cholesterol, sugar or saturated fat content(16% vs. 5%, 0 = 0.014).DiscussionThe purpose of this study was to examine and comparethe dietary habits and lifestyle behaviors of self-definedvegetarians and non-vegetarians from a population-basedrepresentative sample of BC adults. Approximately 6% ofthe sample, weighted to reflect the BC population,reported being vegetarian. The findings of this study sug-gest that the dietary habits, lifestyle behaviors, and food-choice motivations of self-defined vegetarians differ fromthose of non-vegetarians, and that there may be variationbetween men and women which has not previously beenPrevalence of nutrient inadequacies by vegetarian status and gender for selected nutrientsFigur  1Prevalence of nutrient inadequacies by vegetarian status and gender for selected nutrients. The prevalence of nutrient inade-quacy was estimated by determining the proportion of the usual intake distribution (from food plus supplements) that was below the Estimated Average Requirement (EAR). * Prevalence of inadequacy higher in non-vegetarians (p < 0.05). NV = non-vegetarians, V = vegetarians.% Below EAR0 5 10 15 20 25 30 35 40Female NVFemale VMale NVMale VVitamin B12ZincFolateVitamin B6ThiaminVitamin CMagnesium******Page 7 of 11(page number not for citation purposes)vs. 3%, p < 0.001). When avoiding foods, more femalenon-vegetarians reported that they did not consider any ofexamined in population-based studies.International Journal of Behavioral Nutrition and Physical Activity 2005, 2:4 http://www.ijbnpa.org/content/2/1/4Several aspects of our results warrant additional consider-ation, one of which is the small proportion of self-identi-fied vegetarians who adhered rigidly to diets free fromanimal flesh. Occasional use of seafood, poultry, or meatby a majority of those who consider themselves to be veg-etarian has also been reported in other studies [9,15]. If astrict definition of vegetarianism had been used, the prev-alence in our study would be less than 1.5% rather thanclose to 6%. Despite basing our analysis on respondents'self-definition, we still observed a number of differencesin nutrient intake and lifestyle behavior. At some level,this validates respondents' self-identification asvegetarian.Evidence for a higher level of 'health consciousness'among vegetarians in our sample was provided by find-ings of increased use of nutrient supplements, higherintakes of several nutrients (fiber, magnesium, potas-sium), higher intakes of fruits and vegetables, a consider-ably lower prevalence of smoking, and among women,higher physical activity and a lower BMI. Many of thesefindings have been reported in other studies, althoughmost reports from convenience samples have not founddifferences in smoking or exercise behavior by vegetarianstatus [7,16-21]. It is likely that convenience samplingresulted in recruitment of more 'health conscious' partici-that vegetarians have healthier lifestyle practices than thegeneral population of non-vegetarians.Vegetarians were also more likely to consider 'maintain-ing/improving health' when choosing/avoiding foods, tochoose foods for the nutrients they contain and to avoidfoods for their fat content. These findings provide addi-tional evidence of health consciousness, and are consist-ent with research reporting that health concerns andbenefits are a primary reason for adopting a vegetarianlifestyle [22,23], although we did not assess motivationfor adopting a vegetarian diet. They are also consistentwith a population-based study in the Netherlands thatfound vegetarians were more likely to report health con-siderations when purchasing food [10]. That study, how-ever, did not report nutrient intakes.A novel aspect of our analysis was that, in addition toassessing differences in nutrient intakes, we also com-pared the prevalence of inadequate nutrient intakes usingthe EAR cut-point method [14]. As assessed by the propor-tions with total usual nutrient intakes below the EAR, veg-etarians were less likely to have an inadequate intake ofmagnesium, and female vegetarians were also less likelyto have inadequate intakes of folate, vitamin C, thiaminand vitamin B6. Although there were no differences byTable 2: Canada's Food Guide servings and percentage of participants meeting minimum recommendations by vegetarian status and gender.Food Group Servings (Mean ± SE) % Meeting RecommendationNon-vegetarian Vegetarian Non-vegetarian VegetarianGrain Products1 7.7 ± 0.16 7.9 ± 0.81 70.5 67.7 5.0 ± 0.09 5.9 ± 0.31** 45.5 57.3*Fruit and Vegetables1 5.3 ± 0.13 7.3 ± 0.70** 45.5 64.5* 4.5 ± 0.12 5.6 ± 0.40** 34.0 40.0Milk and Milk Products2 1.7 ± 0.06 3.0 ± 0.29*** 33.5 67.7*** 1.4 ± 0.04 1.5 ± 0.14 25.3 33.3Meat and Alternatives3 4.5 ± 0.12 1.9 ± 0.6*** 78.1 51.6*** 2.7 ± 0.07 2.6 ± 0.24 57.8 45.3** p < 0.05, ** p < 0.01, *** p < 0.0011 Minimum serving recommendation is 5 servings.2 Minimum serving recommendation is 2 servings.3 Minimum serving recommendation is 2 servings; servings calculated as 50 g equivalentPage 8 of 11(page number not for citation purposes)pants and therefore did not detect differences. Thus ourfindings provide population-level support for the conceptvegetarian status in the proportions with zinc intakesbelow the EAR, this may not be an accurate reflection ofInternational Journal of Behavioral Nutrition and Physical Activity 2005, 2:4 http://www.ijbnpa.org/content/2/1/4zinc adequacy, as the requirement for dietary zinc may beas much as 50% greater for vegetarians [24]. Similarly,iron requirements of vegetarians are estimated to be 80%greater than those of non-vegetarians [24]. However, theadequacy of iron intakes was not assessed in our studybecause the iron requirement distribution is skewed, andtherefore the EAR cut-point method cannot be used toestimate the prevalence of inadequacy [14]. Finally,although adequacy of vitamin B12 intakes is often identi-fied as a concern for vegetarians, in our sample the preva-lence of inadequate intakes was similar by vegetarianstatus. This is likely due to the fact that almost all vegetar-ians used dairy products and eggs, as well as to the highprevalence of B vitamin supplementation amongvegetarians.Although our vegetarian sample was small, our resultsprovide suggestive evidence of gender differences. Forexample, vegetarian women had a lower age-adjusted BMIand waist circumference, and a lower prevalence of over-weight/obesity, while no differences were seen betweenvegetarian and non-vegetarian men. This may have beendue to the higher frequency of physical activity reportedby vegetarian women (but not men), as energy intake didnot differ by vegetarian status for either sex. Reports fromconvenience samples often suggest that vegetarians havelower BMI and/or a lower rate of obesity [2,7,22,25-27].Conversely, other convenience samples, in which energyintakes and physical activity were similar betweenvegetarians and non-vegetarians, did not detect differ-ences in BMI between groups [16-19,28,29]. In the popu-lation-based CSFII, self-identified vegetarians had lowerenergy intakes and age-adjusted BMI [9]. However, amajor limitation of that report was that analyses were notconducted by gender. Accordingly, if vegetarians weremore likely to be female, as observed in our sample andanother population-based sample [10], vegetarians' meanenergy intake and BMI would appear to be lower becauseof women's lower mean energy intakes and BMI.The distribution of macronutrient intakes also providedsuggestive evidence of gender differences. Carbohydrateas a percentage of energy was higher among vegetarians,as was also found in the CSFII vegetarian analysis [9] andthe majority of convenience sample studies[18,22,26,27,30,31]. Other studies have also reportedlower percentage energy from fat [8,9,22,27,32] and pro-tein [8,18,22,27,28,30-32]. In our sample, only male veg-etarians had a lower proportion of energy from proteinand only female vegetarians consumed less energy fromfat.We also observed gender differences in motivations forblood pressure when choosing/avoiding foods and toreport avoiding foods because of their cholesterol or satu-rated fat content. This is consistent with the higher preva-lence of heart disease among the male vegetarians in oursample, who we speculate may have chosen to follow avegetarian diet as a result of heart disease. Because we didnot assess motivation for adopting a vegetarian diet, thiscannot be ascertained, and in any case, the study's cross-sectional design precludes causal inferences. Female vege-tarians, on the other hand, were not more concernedabout heart disease, but were more likely to consider can-cer, osteoporosis and food allergies/intolerances whenchoosing/avoiding foods and to avoid foods because oftheir salt content. They were also less likely to considerweight gain when choosing/avoiding foods. It has beensuggested that some young women may adopt a vegetar-ian lifestyle in an effort to lose weight [33,34]; however,this does not appear to be true for our population-basedsample.While our findings suggest that variation by gender mayexist in vegetarians' dietary habits and lifestyle behaviors,the study limitations should be acknowledged. First,although the sample was considered representative of theprovince of British Columbia, it was not nationally repre-sentative, which means that inferences cannot be madeabout the Canadian population. Also, the response rate,although typical of other studies of this kind, was notoptimal. Second, the absolute number of self-identifiedvegetarians was small and therefore caution must be usedwhen interpreting the apparent gender differences. Wehad limited power to detect gender-by-vegetarian statusinteractions. Finally, data on dietary intake and lifestylebehaviors were based on self-reports, and it is known thatdietary intakes are underreported [35]. This would beproblematic if differences existed in the extent of underre-porting by vegetarian status. However, based on similarreported energy intakes of the two groups, it appearsunlikely that differential underreporting occurred.We do not believe that our observations of higher 'healthconsciousness' among vegetarians were confounded byother differences between vegetarian and non-vegetariangroups. First, although the prevalence of vegetarianismwas higher among women than men, we conducted anal-yses separately by gender. Second, vegetarians tended tobe younger than non-vegetarians, so age was included asa covariate in nutrient intake and anthropometric analy-ses. Third, although vegetarians were more likely to be sin-gle and to report low-income status, consideration ofthese differences did not affect our observations.ConclusionPage 9 of 11(page number not for citation purposes)choosing/avoiding foods. Only male vegetarians weremore likely to report considering heart disease and highTaken together, these population-based findings add fur-ther support to the concept that adult vegetarians areInternational Journal of Behavioral Nutrition and Physical Activity 2005, 2:4 http://www.ijbnpa.org/content/2/1/4more health-conscious than non-vegetarians, and thatthis difference extends to food choice and nutrition con-cerns. Additional population-based studies comparingdietary habits and lifestyle behaviors by vegetarian statusand gender are needed to determine if gender differencesobserved in our representative sample exist in other pop-ulations in the developed world.Competing interestsThe author(s) declare that they have no competinginterests.Authors' contributionsJB performed the statistical analyses and drafted the man-uscript. SB conceived of the study and participated in itsdesign, and helped to draft the manuscript. Both authorsread and approved the final manuscript.Additional materialAcknowledgementsThe BC Nutrition Survey was supported by the BC Ministry of Health Serv-ices and Health Canada.References1. Appleby PN, Thorogood M, Mann JI, Key TJ: The Oxford vegetar-ian study: an overview. Am J Clin Nutr 1999, 70:525S-531S.2. Brathwaite N, Fraser HS, Modeste N, Broome H, King R: Obesity,diabetes, hypertension, and vegetarian status among Sev-enth-Day Adventists in Barbados: preliminary results. EthnDis 2003, 13:34-39.3. Segasothy M, Phillips PA: Vegetarian diet: panacea for modernlifestyle diseases? QJM 1999, 92:531-544.4. Key TJ, Davey GK, Appleby PN: Health benefits of a vegetariandiet. Proc Nutr Soc 1999, 58:271-275.5. Key TJ, Fraser GE, Thorogood M, Appleby PN, Beral V, Reeves G,Burr ML, Chang-Claude J, Frentzel-Beyme R, Kuzma JW, Mann J,McPherson K: Mortality in vegetarians and nonvegetarians:detailed findings from a collaborative analysis of 5 prospec-tive studies. Am J Clin Nutr 1999, 70:516S-524S.6. Key TJ, Appleby PN, Davey GK, Allen NE, Spencer EA, Travis RC:Mortality in British vegetarians: review and preliminaryresults from EPIC-Oxford. Am J Clin Nutr 2003, 78:533S-538S.7. Spencer EA, Appleby PN, Davey GK, Key TJ: Diet and body massindex in 38000 EPIC-Oxford meat-eaters, fish-eaters, vege-tarians and vegans. Int J Obes Relat Metab Disord 2003, 2:728-734.8. Kennedy ET, Bowman SA, Spence JT, Freedman M, King J: Populardiets: correlation to health, nutrition, and obesity. J Am DietAssoc 2001, 101:411-420.9. Haddad EH, Tanzman JS: What do vegetarians in the UnitedStates eat? Am J Clin Nutr 2003, 78:626S-632S.10. Hoek AC, Luning PA, Stafleu A, de Graaf C: Food-related lifestyle11. Forster-Coull L, Barr SI, Levy-Milne R: British Columbia NutritionSurvey. Report on Energy and Nutrient Intakes. 2004[http:www.healthservices.gov.bc.ca/prevent/nutrition/pdf/nutrientsreport.pdf]. Health Canada, British Columbia Ministry of Health Plan-ning, & University of British Columbia12. Godin G, Shephard RJ: A simple method to assess exercisebehavior in the community. Can J Appl Sport Sci 1985, 10:141-146.13. Marcus BH, Simkin LR: The stages of exercise behavior. J SportsMed Phys Fitness 1993, 33:83-88.14. Institute of Medicine, Food and Nutrition Board: Dietary ReferenceIntakes: Applications in Dietary Assessment Washington DC: NationalAcademy Press; 2000. 15. Barr SI, Chapman GE: Perceptions and practices of self-definedcurrent vegetarian, former vegetarian and nonvegetarianwomen. J Am Diet Assoc 2002, 102:354-360.16. Ball MJ, Bartlett MA: Dietary intake and iron status of Austral-ian vegetarian women. Am J Clin Nutr 1999, 70:353-358.17. Ball MJ, Ackland ML: Zinc intake and status in Australianvegetarians. Br J Nutr 2000, 83:27-33.18. Barr SI, Broughton TM: Relative weight, weight loss efforts andnutrient intakes among health-conscious vegetarian, pastvegetarian and nonvegetarian women ages 18 to 50. J Am CollNutr 2000, 19:781-788.19. Outila TA, Karkkainen MU, Seppanen RH, Lamberg-Allardt CJ: Die-tary intake of vitamin D in premenopausal, healthy veganswas insufficient to maintain concentrations of serum 25-hydroxyvitamin D and intact parathyroid hormone withinnormal ranges during the winter in Finland. J Am Diet Assoc2000, 100:434-441.20. Davey GK, Spencer EA, Appleby PN, Allen NE, Knox KH, Key TJ:EPIC-Oxford: lifestyle characteristics and nutrient intakes ina cohort of 33 883 meat-eaters and 31 546 non meat-eatersin the UK. Public Health Nutr 2003, 6:259-269.21. Locong A: Nutritional status and dietary intake of a selectedsample of young adult vegetarians. J Can Diet Assoc 1986,47:101-106.22. Huang Y, Lin W, Cheng C, Su K: Nutrient intakes and iron statusof healthy young vegetarians and nonvegetarians. Nutr Res1999, 19:663-674.23. Kim EHJ, Schroeder KM, Houser RF, Dwyer JT: Two small surveys,25 years apart, investigating motivations of dietary choice in2 groups of vegetarians in the Boston area. J Am Diet Assoc 1999,99:598-601.24. Institute of Medicine, Food and Nutrition Board: Dietary ReferenceIntakes for Vitamin A, Vitamin K, Arsenic, Boron, Chromium, Copper, Iodine,Iron, Manganese, Molybdenum, Nickel, Silicon, Vanadium, and Zinc Wash-ington DC: National Academy Press; 2000. 25. Appleby PN, Thorogood M, Mann JI, Key TJ: Low body mass indexin non-meat eaters: the possible roles of animal fat, dietaryfibre and alcohol. Int J Obes Relat Metab Disord 1998, 22:454-460.26. Janelle KC, Barr SI: Nutrient intakes and eating behavior scoresof vegetarian and nonvegetarian women. J Am Diet Assoc 1995,95:180-186.27. Millet P, Guilland JC, Fuchs F, Klepping J: Nutrient intake and vita-min status of healthy French vegetarians andnonvegetarians. Am J Clin Nutr 1989, 50:718-727.28. Harman SK, Parnell WR: The nutritional health of New Zealandvegetarian and non-vegetarian Seventh-day Adventists:selected vitamin, mineral and lipid levels. N Z Med J 1998,111:91-94.29. Hunt IF, Murphy NJ, Henderson C: Food and nutrient intake ofSeventh-day Adventist women. Am J Clin Nutr 1988, 48:850-851.30. Faber M, Gouws E, Benade AJ, Labadarios D: Anthropometricmeasurements, dietary intake and biochemical data ofSouth African lacto-ovovegetarians. S Afr Med J 1986,69:733-738.31. Lee HY, Woo J, Chen ZY, Leung SF, Peng XH: Serum fatty acid,lipid profile and dietary intake of Hong Kong Chinese omni-vores and vegetarians. Eur J Clin Nutr 2000, 54:768-773.32. Wilson AK, Ball MJ: Nutrient intake and iron status of Austral-ian male vegetarians. Eur J Clin Nutr 1999, 53:189-194.33. Gilbody SM, Kirk SF, Hill AJ: Vegetarianism in young women:another means of weight control? Int J Eat Disord 1999, 26:87-90.34. McLean JA, Barr SI: Cognitive dietary restraint is associatedAdditional File 1Dietary intake (age-adjusted, mean ± SE) by vegetarian status and gender AClick here for file[http://www.biomedcentral.com/content/supplementary/1479-5868-2-4-S1.doc]Page 10 of 11(page number not for citation purposes)and health attitudes of Dutch vegetarians, non-vegetarianconsumers of meat substitutes, and meat consumers. Appetite2004, 42:265-272.with eating behaviors, lifestyle practices, personality charac-Publish with BioMed Central   and  every scientist can read your work free of charge"BioMed Central will be the most significant development for disseminating the results of biomedical research in our lifetime."Sir Paul Nurse, Cancer Research UKYour research papers will be:available free of charge to the entire biomedical communitypeer reviewed and published immediately upon acceptancecited in PubMed and archived on PubMed Central International Journal of Behavioral Nutrition and Physical Activity 2005, 2:4 http://www.ijbnpa.org/content/2/1/4teristics and menstrual irregularity in college women. Appe-tite 2003, 40:185-92.35. Subar AF, Kipnis V, Troiano RP, Midthune D, Schoeller DA, BinghamS, Sharbaugh CO, Trabulsi J, Runswick S, Ballard-Barbash R, SunshineJ, Schatzkin A: Using intake biomarkers to evaluate the extentof dietary misreporting in a large sample of adults: theOPEN study. Am J Epidemiol 2003, 158:1-13.yours — you keep the copyrightSubmit your manuscript here:http://www.biomedcentral.com/info/publishing_adv.aspBioMedcentralPage 11 of 11(page number not for citation purposes)


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items