UBC Faculty Research and Publications

Association of breakfast consumption with body mass index and prevalence of overweight/obesity in a nationally-representative… Barr, Susan I; DiFrancesco, Loretta; Fulgoni, Victor L Mar 31, 2016

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

Item Metadata


52383-12937_2016_Article_151.pdf [ 724.97kB ]
JSON: 52383-1.0308588.json
JSON-LD: 52383-1.0308588-ld.json
RDF/XML (Pretty): 52383-1.0308588-rdf.xml
RDF/JSON: 52383-1.0308588-rdf.json
Turtle: 52383-1.0308588-turtle.txt
N-Triples: 52383-1.0308588-rdf-ntriples.txt
Original Record: 52383-1.0308588-source.json
Full Text

Full Text

RESEARCH Open AccessAssociation of breakfast consumption withbody mass index and prevalence ofoverweight/obesity in a nationally-representative survey of Canadian adultsSusan I. Barr1*, Loretta DiFrancesco2 and Victor L. Fulgoni III3AbstractBackground: This study examined the association of breakfast consumption, and the type of breakfast consumed,with body mass index (BMI; kg/m2) and prevalence rates and odds ratios (OR) of overweight/obesity amongCanadian adults. These associations were examined by age group and sex.Methods: We used data from non-pregnant, non-lactating participants aged ≥ 18 years (n = 12,377) in theCanadian Community Health Survey Cycle 2.2, a population-based, nationally-representative, cross-sectionalstudy. Height and weight were measured, and BMI was calculated. Breakfast consumption was self-reportedduring a standardized 24-h recall; individuals were classified as breakfast non-consumers, consumers of breakfaststhat included ready-to-eat cereal (RTEC) or as other breakfast consumers. Mean BMI and prevalence and OR ofoverweight/obesity (BMI ≥ 25) were compared among breakfast groups, with adjustment for sociodemographicvariables (including age, sex, race, marital status, food security, language spoken at home, physical activitycategory, smoking, education level and supplement use).Results: For the entire sample, mean BMI was significantly lower among RTEC-breakfast consumers than otherbreakfast consumers (mean ± SE 26.5 ± 0.2 vs. 27.1 ± 0.1 kg/m2), but neither group differed significantly frombreakfast non-consumers (27.1 ± 0.3 kg/m2). Similar results were seen in women only, but BMI of men did notdiffer by breakfast category. Overweight/obesity prevalence and OR did not differ among breakfast groups forthe entire sample or for all men and women separately. When examined by sex and age group, differenceswere inconsistent, but tended to be more apparent in women than men.Conclusion: Among Canadian adults, breakfast consumption was not consistently associated with differencesin BMI or overweight/obesity prevalence.Keywords: Breakfast, Overweight and obesity, National survey, Body mass indexIntroductionIt is widely believed that breakfast consumption, versusnon-consumption, protects against overweight and obesity[1–3]. Empirical support is provided by a large number ofobservational studies, summarized in several systematicreviews and meta-analyses [2, 4, 5]. The most compre-hensive of these included an analysis of 88 study groupsand yielded a pooled odds ratio (OR) of 1.55 (95 % CI:1.46, 1.65) for the likelihood of being overweight orobese among breakfast non-consumers versus breakfastconsumers [2]. There was a tendency for funnel-plotasymmetry (p = 0.086), suggesting a possibility of publi-cation bias. Among these studies, many were conductedwith children and adolescents, rather than adults.Moreover, relatively few examined variability within apopulation associated with age, sex or the consumptionof different types of breakfasts, although several didassess associations with breakfasts containing or notcontaining ready-to-eat cereal (RTEC).* Correspondence: susan.barr@ubc.ca1University of British Columbia, 2205 East Mall, Vancouver V6T 1Z4, CanadaFull list of author information is available at the end of the article© 2016 Barr et al. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Barr et al. Nutrition Journal  (2016) 15:33 DOI 10.1186/s12937-016-0151-3There is also an extensive literature indicating thatbreakfast consumption is associated with improved nu-trient intakes and adequacy [6–11]. In a population-representative sample of Canadian adults, we previouslyobserved that breakfast consumers (versus non-consumers)had higher nutrient intakes and a lower prevalence of nutri-ent inadequacy [9]. We also noted that intakes of severalkey nutrients were higher (and prevalence of inadequacylower) in those who consumed breakfasts containing RTECcompared to those who consumed other breakfasts [9].Thus, using data from adults in the same population-representative sample of Canadians (in which the over-all prevalence of overweight/obesity among adults was59 % [12]), we sought to assess whether breakfast con-sumption and the type of breakfast consumed (with orwithout RTEC) were associated with body mass index(BMI; kg/m2) and the prevalence rates and OR of over-weight/obesity. Further, we examined whether associa-tions varied by age group and sex.MethodsData sourceThe present study is a secondary analysis of data collectedin the Canadian Community Health Survey, Cycle 2.2(CCHS 2.2), a cross-sectional, nationally-representativesurvey conducted by Statistics Canada in 2004 [13, 14]. Thetarget population for CCHS 2.2 represented approximately98 % of the Canadian population, and included individualsliving in private dwellings in the 10 Canadian provinces.The multistage stratified cluster sampling plan wasdesigned to be representative in terms of age, sex, geog-raphy and socioeconomic status [13, 14]. Data collec-tion was completed in person by trained interviewers,who received extensive standardized training in all proce-dures [14]. Survey components included a 24-h dietaryrecall, a general health questionnaire to assess socio-demographic and lifestyle variables, and measured heightand weight (which were used to calculate BMI) [14]. Theresponse rate for the survey was 76.5 %, and the surveyweights included a non-response adjustment. Ethicalapproval for population surveys conducted by StatisticsCanada, such as CCHS 2.2, is based on the authority ofthe Statistics Act of Canada [15].Analytical sampleFor this analysis, we included data from CCHS 2.2respondents aged 18 years and above who were notpregnant or lactating, had measured values for height andweight and completed a valid 24-h recall (n = 12,337). The24-h recall was conducted using a modification of theAutomated Multiple Pass Method [13, 14]. In the first“pass”, respondents were asked to list all foods and bever-ages consumed on the day before the survey. Foods andbeverages could be listed in any order; there was norequirement to recall foods in a time sequence. Subse-quent “passes” obtained additional details about each fooditem listed, including the amount consumed and what therespondent called the eating occasion (e.g., breakfast,lunch, dinner, a snack). Thus, for our analysis, “breakfast”included any foods or beverages consumed during the 24-h recall at an eating occasion that the respondent identi-fied as breakfast. Those who did not identify any items asbeing consumed at breakfast were classified as breakfastnon-consumers (i.e., "breakfast skippers"). Those who con-sumed RTEC as a component of breakfast were classifiedas RTEC breakfast consumers, and those whose breakfastsdid not include RTEC were classified as other breakfastconsumers. Approval to conduct the analyses reported inthis paper was obtained from the Statistics Canada Re-search Data Centre program [16], project number 11-SSH-UTO-2848.Statistical analysisStatistical Analysis Software (SAS), version 9.2 (Cary, NC)and SUDAAN, version 10.0 (RTI International, ResearchTriangle Park, NC) were used to analyze the data.SUDAAN was used to create variance estimates andstandard errors (SE) of proportions. All analyses were ad-justed for the complex CCHS 2.2 sampling design usingappropriate sample weights and, when necessary, theMISSUNIT option in SUDAAN was used due to a num-ber of cases with only one stratum within a primary sam-pling unit. This option then calculates the variancecontribution using the difference from the overall mean ofthe population. Means, percentages and standard errorswere obtained using PROC DESCRIPT. Covariate-adjusted mean BMI values were compared among thethree breakfast groups using analysis of variance (i.e.,using PROC REGRESS). Overweight/obesity prevalencewas defined as the proportion with BMI ≥ 25.0, and wascompared among the breakfast groups using a t-test.Covariates included age, sex, race, household food se-curity (reflecting minimal or no limitations to house-hold food access in the context of financial resourceconstraint) [17], marital status, language spoken athome, physical activity category, smoking, level of edu-cational attainment and supplement use. A p value of <0.05 (Bonferroni-adjusted p < 0.0167) was used to assesssignificance of differences by breakfast group. Finally,adjusted OR and 95th percentile confidence limits foroverweight/obesity were calculated to compare the twogroups of breakfast consumers to breakfast non-consumers.ResultsDemographic characteristicsAs reported previously, the weighted proportions whodid not consume breakfast, consumed RTEC breakfastsBarr et al. Nutrition Journal  (2016) 15:33 Page 2 of 9and consumed other breakfasts were 11, 20 and 69 %,respectively [9]. Significant differences in demographiccharacteristics were observed among groups (Table 1).Specifically, breakfast non-consumers were younger andless likely to be married or living common-law than thetwo groups of breakfast consumers. They were the leastlikely to use dietary supplements and to be food secure,and the most likely to smoke. Breakfast non-consumerswere also more likely to be male than other breakfastconsumers, but did not differ significantly from RTECbreakfast consumers in that regard. RTEC breakfast con-sumers were the most likely to use dietary supplementsand be food secure, and were more likely than the othertwo groups to be white. Other breakfast consumers wereintermediate in the proportions that smoked, used diet-ary supplements and were food secure. They were lesslikely to speak English at home than RTEC breakfastconsumers.Body mass indexMean values for BMI by age group, sex and breakfast statusare displayed in Table 2. For the entire sample (both sexescombined), mean BMI was significantly lower amongRTEC breakfast consumers than among other breakfastconsumers, but neither of these groups differed significantlyfrom breakfast non-consumers. The same pattern of differ-ences was observed among those aged 51–70 years and ≥71 years, whereas among adults ≤ 50 years, BMI did not dif-fer by breakfast group. When the sexes were examined sep-arately, no differences by breakfast group were detected forthe entire sample of men or for men up to 50 years of age.Among men aged 51–70 years, BMI was significantly lowerin RTEC breakfast consumers than breakfast non-consumers, with other breakfast consumers having anintermediate value that did not differ from either of theother two groups. Among men aged ≥ 71 years, BMI wassignificantly lower in RTEC consumers than in otherbreakfast consumers. In this age group, mean BMI ap-peared lowest in breakfast non-consumers, but high vari-ability meant that differences with the other breakfastgroups were not significant. Among the entire group ofwomen, BMI was lower in women who consumed RTECbreakfasts versus other breakfasts, with breakfast non-consumers having a value that did not differ from the othergroups. This same pattern was observed among womenaged ≥ 71 years. In contrast, among women aged 51–70years, BMI was significantly lower among breakfast non-consumers than among other breakfast consumers, withRTEC breakfast consumers having an intermediate valuethat did not differ from the other two groups.Prevalence of overweight and obesityThe prevalence of overweight and obesity by age groupand for adults of all ages is shown in Fig. 1 for bothsexes combined and for men and women separately.Overall, no significant differences in overweight/obesityprevalence by breakfast group were seen among adultsas a whole, nor were differences observed in all men orall women. When the combined sexes were examined byage group (Panel a in Fig. 1), the only significant differ-ence was among those aged 18–30 years, where over-weight/obesity prevalence was significantly higher inbreakfast non-consumers than in those who consumedRTEC breakfasts (50 % vs 37 %, respectively). Those whoconsumed other breakfasts had an intermediate preva-lence (42 %) that did not differ from either of the othertwo groups. No differences were observed in men in anyof the age groups (Panel b in Fig. 1). In women (Panel cin Fig. 1), prevalence of overweight/obesity was signifi-cantly higher in breakfast non-consumers aged 18–30years (49 %) than in both groups of breakfast consumers(31 % in RTEC breakfast consumers and 36 % in otherbreakfast consumers). Conversely, among women aged51–70 years, the prevalence of overweight/obesity wasTable 1 Demographic data for Canadian adults aged ≥ 18 years by breakfast statusdMeasures All (n = 12,337) No breakfast (n = 1445) RTEC breakfast (n = 2799) Other breakfast (n = 8093)Age (y) 46.1 ± 0.3 37.6 ± 0.7a 48.4 ± 0.6b 46.7 ± 0.3bMale (%) 47.4 ± 0.9 53.6 ± 2.5a 47.4 ± 1.9a,b 46.4 ± 1.1bMarried/common law (%) 63.7 ± 0.8 52.7 ± 2.4a 61.2 ± 2.0b 66.1 ± 0.9bPost-secondary graduate (%) 52.1 ± 0.9 47.5 ± 2.6 51.9 ± 2.0 52.9 ± 1.1Physically inactive (%) 56.5 ± 0.9 59.2 ± 2.5 52.2 ± 1.9 57.4 ± 1.1Dietary supplement use (%) 42.5 ± 0.9 33.4 ± 2.3a 48.4 ± 2.0b 42.1 ± 1.1cFood secure (%) 93.1 ± 0.4 89.2 ± 1.4a 95.5 ± 0.7b 92.9 ± 0.6cSmoker (%) 21.2 ± 0.7 35.1 ± 2.4a 11.2 ± 1.3b 22.2 ± 0.9cWhite (%) 84.2 ± 0.8 78.4 ± 2.4a 91.2 ± 1.4b 83.0 ± 1.0aEnglish spoken at home (%) 62.2 ± 1.0 70.4 ± 2.6a,b 70.0 ± 2.1a 58.5 ± 1.2ba,b,cMeans with different superscripts are significantly different, p < 0.05 (Bonferroni-adjusted p < 0.0167)dData are from the Canadian Community Health Survey Cycle 2.2 (2004) and are shown as weighted mean ± SE. No Breakfast = no food or beverages reported asbreakfast; RTEC Breakfast = breakfast that included ready-to-eat cereal (RTEC); Other Breakfast = any other type of breakfastBarr et al. Nutrition Journal  (2016) 15:33 Page 3 of 9Table 2 BMI (kg/m2) in Canadians by age group, sex and breakfast statuscAll (n = 12,241) Male (n = 5204) Female (n = 7037)Breakfast group Breakfast group Breakfast groupAge n None RTEC Other None RTEC Other None RTEC OtherWeighted mean ± standard error18–30 y 2947 26.0 ± 0.5 24.7 ± 0.4 25.1 ± 0.2 25.7 ± 0.5 25.1 ± 0.4 25.5 ± 0.3 26.4 ± 0.8 24.2 ± 0.6 24.8 ± 0.331–50 y 3129 27.0 ± 0.4 27.2 ± 0.4 27.2 ± 0.2 27.6 ± 0.7 27.7 ± 0.5 27.3 ± 0.2 26.4 ± 0.5 26.7 ± 0.6 27.1 ± 0.451–70 y 3618 28.1 ± 0.6a,b 27.4 ± 0.3a 28.4 ± 0.2b 29.9 ± 0.7a 27.6 ± 0.4b 28.4 ± 0.3a,b 26.0 ± 0.7a 27.3 ± 0.5a,b 28.2 ± 0.3b≥71 y 2547 27.4 ± 0.7a,b 26.5 ± 0.2a 27.5 ± 0.2b 26.0 ± 0.8a,b 26.6 ± 0.3a 27.5 ± 0.3b 27.8 ± 0.8a,b 26.5 ± 0.3a 27.5 ± 0.3bAll ≥ 18 y 12,241 27.1 ± 0.3a,b 26.5 ± 0.2a 27.1 ± 0.1b 27.5 ± 0.4 26.8 ± 0.3 27.2 ± 0.2 26.7 ± 0.4a,b 26.1 ± 0.3a 27.0 ± 0.3ba,bMeans within an age group and sex category (all, male, female) that do not share a common superscript letter differ significantly, p < 0.05 (Bonferroni-adjusted p < 0.0167)cValues are weighted mean ± SE (data from CCHS 2.2). Adjusted for age, sex, race, supplement use, food security, language spoken at home, physical activity category, smoking, education level and marital status.None = no food or beverages reported as breakfast; RTEC Breakfast = breakfast that included ready-to-eat cereal (RTEC); Other Breakfast = any other type of breakfastBarretal.NutritionJournal (2016) 15:33 Page4of9significantly higher among those who consumed otherbreakfasts than among breakfast non-consumers (67 %versus 48 %, respectively), while RTEC breakfast con-sumers had an intermediate prevalence (56 %) that didnot differ from either of the other two groups.Table 3 presents adjusted OR for overweight/obesityfor each of the two breakfast groups, compared to break-fast non-consumers (reference group). Data are shown forall adults and for men and women separately, and also forall ages combined and by separate age groups. There wereno significant differences in OR for the entire group ofadults aged ≥ 18 years, for all men aged ≥18 years or forall women aged ≥ 18 years. When examined by age group,the odds of overweight/obesity were lower among alladults aged 18–30 years who consumed RTEC breakfasts.There were no differences in odds among men of any agegroup. Among women aged 18–30 years, both groups ofbreakfast consumers had lower odds of overweight/obesitythan breakfast non-consumers, whereas among those aged51–70 years, consumers of other breakfasts had higherodds of overweight/obesity than non-consumers.DiscussionIn this population-based study of Canadian adults, neitherbreakfast consumption (versus non-consumption) nor thetype of breakfast consumed (whether or not RTEC wasincluded) was consistently associated with BMI or theprevalence of overweight/obesity. For the overall adultpopulation, mean BMI of breakfast non-consumers andthose who consumed other breakfasts was almost identical(27.1 ± 0.3 and 27.1 ± 0.1 kg/m2, respectively). While meanBMI of RTEC breakfast consumers (26.5 kg/m2) was sig-nificantly lower than that of other breakfast consumers,the difference of 0.6 kg/m2 reflects a difference of only1.7 kg at the mean population height of 1.68 m. Further-more, the prevalence of overweight/obesity and the ORfor being overweight/obese did not differ among the threebreakfast groups for the adult population as a whole:Overweight/obesity prevalence was close to 60 % in allgroups, and adjusted OR (and 95 % CI) for consumers ofRTEC breakfasts and other breakfasts were 0.95 (0.72,1.26) and 1.04 (0.81, 1.34), respectively, relative to break-fast non-consumers. When these associations were exam-ined by sex and age (which has not been done in themajority of previous studies), they were not consistent.Our results can be compared to those of otherpopulation-based, cross-sectional studies of adults[11, 18–25], the majority of which report data onoverweight/obesity prevalence. Significantly higher adjustedOR for overweight and/or obesity were observed among alladults who skipped breakfast in studies conducted inTaiwan [18] and Sweden [19], and among both maleand female breakfast skippers in a study conducted inSpain [20]. In contrast, OR were not significantly differ-ent between breakfast skippers and consumers in stud-ies conducted in Serbia [21] or the United States [22].Three studies, all conducted in the United States, ex-amined associations with the type of breakfast con-sumed [11, 22, 23]. One study identified 12 breakfastpatterns, including no breakfast [11]. Compared to break-fast non-consumers, the OR for overweight/obesity werelower among consumers of five types of breakfasts, butwere similar among consumers of the other six breakfasttypes (data comparing all breakfast consumers to non-consumers were not provided). In the second study [22],OR for overweight/obesity did not differ between break-fast consumers and non-consumers; however, female butnot male consumers of RTEC breakfasts had lower ORfor overweight/obesity compared to consumers of otherbreakfasts. The third study [23] examined young adultsABCFig. 1 Prevalence (SE) of overweight/obesity (BMI ≥ 25) amongCanadian adults, by sex and age group. Values within an agegroup and sex category (all, male, female) without a commonsuperscript letter differ significantly, p < 0.05 (Bonferroni-adjustedp < 0.0167). RTEC Breakfast = breakfast that includes ready-to-eatcereal (RTEC). Other Breakfast = any other type of breakfastBarr et al. Nutrition Journal  (2016) 15:33 Page 5 of 9aged 20–39 years and found the OR for overweight/obesitywas lower in RTEC breakfast consumers, as compared toboth breakfast skippers and other breakfast consumers. Re-sults by sex were not reported. Our study is most compar-able to the two latter studies [22, 23], in that differenceswere assessed among those consuming no breakfast, RTECbreakfast and other breakfasts. However, in contrast toSong et al. [22], we found no difference in OR for over-weight/obesity between consumers of RTEC and otherbreakfasts among the entire group of women. And in con-trast to Deshmukh-Taskar et al. [23] who studied youngadults, in our study the OR was lower among young femaleconsumers of both RTEC and other breakfasts when com-pared to breakfast skippers.A smaller number of studies, all of which used data fromdifferent waves of the United States National Health andNutrition Examination Survey (NHANES), report on meanBMI by breakfast intake [11, 23–25]. Cho et al. [24], usingdata from NHANES III (1998–2004), found that meanBMI was lower among those who consumed RTEC, cookedcereal or quick breads for breakfast than among those whoskipped breakfast or consumed breakfast based on meatand eggs. The analysis of Kant et al. [25], with data fromNHANES 1999–2004, reported that BMI was lower inwomen who consumed breakfast, but not in men.Deshmukh-Taskar et al. [23], with data from NHANES1999–2006, reported that BMI was lower among youngadults who consumed RTEC, as compared to breakfastskippers or other breakfast consumers. Finally, the study byO’Neil et al. [11], using data from NHANES 2001–2008,found lower BMI among consumers of four of 11 breakfasttypes compared to those who did not eat breakfast, butsimilar BMI among consumers of the other seven breakfasttypes. We observed a lower BMI among those who con-sumed RTEC breakfasts compared to those consumingother breakfasts, but neither group differed from breakfastnon-consumers.Variability in the associations between breakfast andweight status also exists within studies. For example,breakfast intake was associated with lower BMI or ORfor overweight/obesity in women but not men [22, 25],whereas in another large study [20], the OR for over-weight/obesity among breakfast skippers versus con-sumers were very similar in men and women (1.58 and1.53, respectively). Our study appears to be the first toexamine weight status in association with breakfast byboth sex and age group. Although we did detect somedifferences, for the most part these were observed betweenconsumers of RTEC breakfasts and other breakfasts, ratherthan between breakfast consumers and non-consumers.The one exception was in women aged 18–30 years, whereprevalence and OR for overweight/obesity were lowerTable 3 Odds ratios for BMI ≥ 25 kg/m2 by breakfast consumption group among Canadian adultsbNo breakfast RTEC breakfast Other breakfastAge n Odds ratio (Reference) Odds ratio 95 % CI Odds ratio 95 % CIAll18–30 y 2947 1.0 0.57 0.36, 0.88a 0.72 0.49, 1.0531–50 y 3129 1.0 1.51 0.91, 2.50 1.22 0.82, 1.8451–70 y 3618 1.0 1.07 0.62, 1.83 1.28 0.78, 2.08≥ 71 y 2547 1.0 0.67 0.26, 1.71 0.80 0.31, 2.02All ≥ 18 y 12,241 1.0 0.95 0.72, 1.26 1.04 0.81, 1.34Men18–30 y 1345 1.0 0.70 0.38, 1.29 0.96 0.56, 1.6431–50 y 1463 1.0 1.51 0.72, 3.15 1.14 0.60, 2.1551–70 y 1510 1.0 0.75 0.32, 1.74 0.59 0.30, 1.18≥ 71 y 886 1.0 0.84 0.24, 1.91 1.01 0.29, 3.53All ≥ 18 y 5204 1.0 1.03 0.68, 1.57 1.02 0.70, 1.49Women18–30 y 1602 1.0 0.43 0.23, 0.80a 0.54 0.34, 0.85a31–50 y 1666 1.0 1.67 0.84, 3.34 1.39 0.83, 2.3051–70 y 2108 1.0 1.46 0.69, 3.11 2.34 1.17, 4.68a≥ 71 y 1661 1.0 0.58 0.18, 1.81 0.70 0.23, 2.16All ≥ 18 y 7037 1.0 0.90 0.62, 1.29 1.09 0.80, 1.48a95 % CI for Odds Ratio excludes 1.0bData from CCHS 2.2. Adjusted for age, sex, physical activity, race, smoking, marital status, supplement use, food security and language spoken at home. Breakfastnon-consumers (No Breakfast) were the reference group; RTEC Breakfast = breakfast that included ready-to-eat cereal (RTEC); Other Breakfast = any other typeof breakfastBarr et al. Nutrition Journal  (2016) 15:33 Page 6 of 9in both groups of breakfast consumers compared tonon-consumers. However, in other age groups therewas no evidence for this trend; indeed, among womenaged 51–70 years, those who consumed other break-fasts had a significantly higher OR when compared tobreakfast non-consumers.The reasons for different findings across and withinstudies are difficult to ascertain. Most of thepopulation-based studies described above did not adjustfor energy intake [11, 18–21, 24]; in some cases thismay have been because data on energy intake were notavailable [18, 20, 21]. We have previously reported theenergy intakes of adult participants in CCHS [9], andalthough intakes were lower among breakfast non-consumers compared to the two breakfast groups, inthe present analysis we chose not to adjust for energyintake as it is on the causal pathway to overweight/obesity. Authors of several other studies have made thesame choice [11, 19, 24], while at least one study pre-sented results with energy intake included or excludedas a covariate [22] and others included it [23, 25]. Over-all, this adjustment did not appear to differentiate be-tween studies that did or did not detect differences inBMI or obesity prevalence among breakfast groups.Previous studies also differed to some extent in termsof adjusting for other sociodemographic variables, butall controlled for a substantial number of these vari-ables, as did our study. Furthermore, differences in var-iables that were controlled (e.g., marital status, alcoholconsumption) were not consistently associated withstudy results, suggesting that the extent of statisticaladjustment is unlikely to explain the different results. Itis possible that cultural differences related to breakfastmay play a role, yet studies conducted in countries withdifferent cultures (e.g., Taiwan, Sweden, Spain) [18–20]reported similar findings, and studies conducted in thesame country (e.g., the United States) [22, 25] were notalways consistent.Taken together, the lack of consistent patterns of dif-ferences in weight status between breakfast consumersand non-consumers, or between consumers of RTECbreakfasts and other breakfasts, appears to argue againsta physiologically-based causal relationship. It has beensuggested that breakfast consumption may serve as amarker for a healthier lifestyle [3, 26, 27] and that break-fast consumers believe that eating breakfast helps withweight control [3], which may contribute to the associa-tions observed in some studies. To date, the few ran-domized trials that have been conducted have notprovided convincing evidence that breakfast consump-tion has beneficial effects on weight status [28–30].It is possible that future research may establish thatspecific types of breakfast are beneficial for long-termweight management or have other health benefits. Forexample, among adolescents who habitually skipbreakfast, high-protein breakfasts resulted in improvedshort-term appetite control and satiety [31–33]. Thosefindings, however, appear to contrast with population-based cross-sectional studies reporting that breakfastscharacterized as high in grains and fruit juice, RTECor cooked cereal were associated with reduced OR ofoverweight/obesity, whereas breakfasts characterizedas high in eggs or meat (and thus higher in protein)were not [11, 24]. Nevertheless, irrespective of whetherbreakfast itself (or a certain type of breakfast) affectsweight status, the overall benefits of breakfast consump-tion in terms of nutrient intake and diet quality shouldnot be overlooked [6–11].Strengths of this study include the large population-representative sample, use of measured values forheight and weight, examination of associations by sexand age group, and consideration of potentially con-founding variables. Limitations are that the data wereself-reported and that a single 24-h recall may not re-flect habitual patterns of breakfast intake. However,the differences in sociodemographic variables that weobserved among breakfast groups suggest that manyof those classified in a given breakfast group mayhave consistently skipped breakfast or consumed agiven type of breakfast. We used the conventionalBMI cut-point of ≥25 kg/m2 to define overweight/obesity, and some research indicates that for olderadults, BMI in the overweight range is associated withincreased health and longevity [34–37]. We alsoassessed only two types of breakfasts, and a recentstudy examined weight status of consumers of 11 dif-ferent types of breakfasts, as compared to breakfastnon-consumers [11]. Nevertheless, like our study, thatstudy also observed variability in the associations be-tween breakfast and weight status, supporting the con-cept that breakfast per se may not have a consistentimpact on weight. Finally, the cross-sectional natureof our data means that causation cannot be inferred.This, however, would be more of a concern if we werereporting strong associations, rather than theirabsence.ConclusionsAmong this large population-representative sample ofCanadian adults, breakfast consumption was not consist-ently associated with BMI or overweight/obesity pre-valence. Our findings, in conjunction with otherobservational and experimental evidence, suggest that itmay be inappropriate to promote weight-managementbenefits of breakfast consumption per se. Nevertheless,it is still possible that particular types of breakfast con-sumption may be helpful for weight management;Barr et al. Nutrition Journal  (2016) 15:33 Page 7 of 9future long-term randomized trials appear necessaryin this regard. In the meantime, the consistently-reported nutritional contributions of breakfast shouldnot be neglected.AbbreviationsBMI: body mass index; CCHS: Canadian Community Health Survey;CI: confidence interval; NHANES: National Health and Nutrition ExaminationSurvey; OR: odds ratio; RTEC: ready-to-eat cereal; SE: standard error; y: years.Competing interestsAt the time of the study, Susan Barr, Loretta DiFrancesco and Victor FulgoniIII had consulting agreements with Kellogg Canada Inc. Loretta DiFrancescocarries on business as Source! Nutrition® and provides scientific andregulatory affairs consulting to various food and beverage companies,commodity groups and nutrition organizations. Victor Fulgoni III, as Senior VicePresident of Nutrition Impact LLC, performs consulting and database analysesfor various food and beverage companies and related entities.Authors’ contributionsSB, LD and VF designed the study. VF was primarily responsible for the dataanalysis; SB wrote the first draft of the manuscript; and LD and VF providedcritical input. All authors read and approved the final manuscript.AcknowledgmentsThe authors thank Carmina Ng and Dave Haans at the Statistics CanadaToronto Region Research Data Centre for their support in accessing theCCHS 2.2 data. Although the research and analysis are based on data fromStatistics Canada, the opinions expressed do not represent the views ofStatistics Canada.Financial supportThe research was supported by Kellogg Canada Inc., but the sponsor had norole in the implementation of the study, or the analysis or interpretation ofthe data.Author details1University of British Columbia, 2205 East Mall, Vancouver V6T 1Z4, Canada.2Source! Nutrition, 303-2511 Bloor Street West, Toronto M6S 5A6, Canada.3Nutrition Impact LLC, 9725 D Drive N, Battle Creek, MI 49014-8514, USA.Received: 17 December 2015 Accepted: 22 March 2016References1. Casazza K, Fontaine FR, Astrup A, Birch LL, Brown AW, Bohan Brown MM, et al.Myths, presumptions, and facts about obesity. N Engl J Med. 2013;368:446–54.2. Brown AW, Bohan Brown MM, Allison DB. Belief beyond the evidence: usingthe proposed effect of breakfast on obesity to show 2 practices that distortscientific evidence. Am J Clin Nutr. 2013;98:1298–308.3. Reeves S, Halsey LG, McMeel Y, Huber JW. Breakfast habits, beliefs andmeasures of health and wellbeing in a nationally representative UK sample.Appetite. 2013;60:51–7.4. Mesas AE, Munoz-Pareja M, Lopez-Garcia E, Rodriguez-Artalejo F. Selectedeating behaviours and excess body weight: a systematic review. Obes Rev.2012;12:106–35.5. Horikawa C, Kodama S, Yachi Y, Heianza Y, Hirasawa R, Ibe Y, et al. Skippingbreakfast and prevalence of overweight and obesity in Asian and Pacificregions: a meta-analysis. Prev Med. 2011;53:260–7.6. Williams P. Breakfast and the diets of Australian adults: an analysis of datafrom the 1995 National Nutrition Survey. Int J Food Sci Nutr. 2005;56:65–9.7. Deshmukh-Taskar PR, Radcliffe JD, Liu Y, Nicklas TA. Do breakfast skippingand breakfast type affect energy intake, nutrient intake, nutrient adequacy,and diet quality in young adults? NHANES 1999–2001. J Am Coll Nutr. 2010;29:407–18.8. Gibson SA, Gunn P. What’s for breakfast? Nutritional implications of breakfasthabits; insights from the NDNS dietary records. Nutr Bull. 2011;36:78–86.9. Barr SI, DiFrancesco L, Fulgoni III VL. Consumption of breakfast and the typeof breakfast consumed are positively associated with nutrient intakes andadequacy of Canadian adults. J Nutr. 2013;143:86–92.10. Barr SI, DiFrancesco L, Fulgoni III VL. Breakfast consumption is positivelyassociated with nutrient adequacy in Canadian children and adolescents.Br J Nutr. 2014;112:1273–83. Corrigendum Br J Nutr 2015;113:190.11. O’Neil CE, Nicklas TA, Fulgoni III VL. Nutrient intake, diet quality, and weight/adiposity parameters in breakfast patterns compared with no breakfast inadults: National Health and Nutrition Examination Survey 2001–2008. J AcadDiet Nutr. 2014;114 Suppl 3:S27–43.12. Tjepkema M. Nutrition: Findings from the Canadian Community HealthSurvey. Issue No. 1. Measured Obesity. Adult obesity in Canada: Measured heightand weight. Component of Statistics Canada Catalogue no. 82-620-MWE2005001.http://www.statcan.gc.ca/pub/82-620-m/2005001/pdf/4224906-eng.pdf. Accessed17 Feb 2016.13. Health Canada. Canadian Community Health Survey Cycle 2.2 Nutrition,2004. A guide to accessing and interpreting the data. Health Canada. 2006.http://www.hc-sc.gc.ca/fn-an/surveill/nutrition/commun/cchs_guide_escc-eng.php. Accessed 3 Nov 2015.14. Statistics Canada. Canadian Community Health Survey – Nutrition(CCHS), Detailed information for 2004 (Cycle 2.2). Statistics Canada.2007. http://www23.statcan.gc.ca/imdb/p2SV.pl?Function=getSurvey&Id=7498. Accessed 10 Feb 2016.15. Statistics Canada. Statistics Act 1985, c. S-19 amended by 2005, c.38.Statistics Canada. 2005. http://laws-lois.justice.gc.ca/eng/acts/S-19/FullText.html. Accessed 3 Nov 2015.16. Statistics Canada. The Research Data Centres (RDC) Program. Statistics Canada.2014. http://www.statcan.gc.ca/rdc-cdr/index-eng.htm. Accessed 3 Nov 2015.17. Health Canada. Canadian Community Health Survey Cycle 2.2 Nutrition(2004). Income-related household food security in Canada. Health Canada2007. http://www.hc-sc.gc.ca/fn-an/surveill/nutrition/commun/income_food_sec-sec_alim-eng.php#metho25. Accessed 17 Feb 2016.18. Huang C-J, Hu H-T, Fan Y-C, Liao Y-M, Tsai PS. Associations of breakfastskipping with obesity and health-related quality of life: evidence from anational survey in Taiwan. Int J Obes. 2010;34:720–5.19. Berg C, Lappas G, Wolk A, Strandhagen E, Toren K, Rosengren A, et al.Eating patterns and portion size associated with obesity in a Swedishpopulation. Appetite. 2009;52:21–6.20. Marin-Guerrero AC, Gutierrez-Fisac JL, Guallar-Castillon P, Banegas JR. Eatingbehaviours and obesity in the adult population of Spain. Br J Nutr. 2008;100:1142–8.21. Grujic V, Cvejin MM, Nikolic EA, Dragnic N, Jovanovic VM, Kvrgic S, et al.Association between obesity and socioeconomic factors and lifestyle.Vojnosanit Pregl. 2009;66:705–10.22. Song WO, Chun OK, Obayashi A, Cho S, Chung CE. Is consumption ofbreakfast associated with Body Mass Index in US adults? J Am Diet Assoc.2005;105:1373–82.23. Deshmukh-Taskar P, Nicklas TA, Radcliffe JD, O’Neil CE, Liu Y. Therelationship of breakfast skipping and type of breakfast consumed withoverweight/obesity, abdominal obesity, other cardiometabolic risk factorsand the metabolic syndrome in young adults: The National Health andNutrition Examination Survey (NHANES): 1999–2006. Public Health Nutr.2013;11:2073–82.24. Cho S, Dietrich M, Brown CJP, Clark CA, Block G. The effect of breakfast typeon total daily energy intake and body mass index: Results from the ThirdNational Health and Nutrition Examination Survey (NHANES III). J Am CollNutr. 2003;22:296–302.25. Kant A, Andon MB, Angelopoulos JT, Rippe JM. Association of breakfastenergy density with diet quality and body mass index in American adults:National Health and Nutrition Examination Surveys, 1999–2004. Am J ClinNutr. 2008;88:1396–404.26. Chen J, Cheng J, Liu Y, Tang Y, Sun X, Wang T, et al. Associations betweenbreakfast eating habits and health-promoting lifestyle, suboptimal healthstatus in Southern China: a population based, cross sectional study. J TranslMed. 2014;12:348.27. Keski-Rahkonen A, Kaprio J, Rissanen A, Virkkunen M, Rose RJ. Breakfastskipping and health-compromising behaviors in adolescents and adults.Eur J Clin Nutr. 2003;57:842–53.28. Dhurandhar EJ, Dawson J, Alcorn A, Larsen LH, Thomas EA, Cardel M, et al.The effectiveness of breakfast recommendations on weight loss: a randomizedcontrolled trial. Am J Clin Nutr. 2014;100:507–13.29. Schlundt DG, Hill JO, Sbrocco T, Pope-Cordle J, Sharp T. The role ofbreakfast in the treatment of obesity: a randomized clinical trial. Am J ClinNutr. 1992;55:645–51.Barr et al. Nutrition Journal  (2016) 15:33 Page 8 of 930. Betts JA, Richardson JD, Chowdhury EA, Holman GD, Tsintzas K, ThompsonD. The causal role of breakfast in energy balance and health: a randomizedcontrolled trial in lean adults. Am J Clin Nutr. 2014;100:539–47.31. Hoertel HA, Will MJ, Leidy HJ. A randomized crossover, pilot studyexamining the effects of a normal protein vs. high protein breakfast onfood cravings and reward signals in overweight/obese “breakfast skipping”,late-adolescent girls. Nutr J. 2014;13:80.32. Leidy JJ, Racki EM. The addition of a protein-rich breakfast and its effects onacute appetite control and food intake in ‘breakfast-skipping’ adolescents.Int J Obes. 2010;34:1125–33.33. Leidy HJ, Ortinau LC, Douglas SM, Hoertel HA. Beneficial effects of a higher-protein breakfast on the appetitive, hormonal, and neural signals controllingenergy intake regulation in overweight/obese, “breakfast-skipping”,late-adolescent girls. Am J Clin Nutr. 2013;97:677–88.34. Flicker L, McCaul KA, Hankey GJ, Jamrozik K, Brown WJ, Byles JE, et al. Bodymass index and survival in men and women aged 70 to 75. J Am GeriatrSoc. 2010;58:234–41.35. Rolland Y, GallinivA CC, Schott A-M, Blain H, Beauchet O, et al. Body-compositionpredictors of mortality in women aged ≥75 y: data from a large population-based cohort study with a 17-y follow-up. Am J Clin Nutr. 2014;100:1352–60.36. Murayama H, Liang J, Bennett JM, Shaw BA, Botoseneanu A, Kobayashi E,et al. Trajectories of body mass index and their associations with mortalityamong older Japanese: do they differ from those of Western populations?Am J Epidemiol. 2015;182:597–605.37. Veronese N, Cereda E, Solmi M, Fowler SA, Manzato E, Maggi S, et al.Inverse relationship between body mass index and mortality in oldernursing home residents: a meta-analysis of 19,538 elderly subjects. ObesRev. 2015;16:1001–5.•  We accept pre-submission inquiries •  Our selector tool helps you to find the most relevant journal•  We provide round the clock customer support •  Convenient online submission•  Thorough peer review•  Inclusion in PubMed and all major indexing services •  Maximum visibility for your researchSubmit your manuscript atwww.biomedcentral.com/submitSubmit your next manuscript to BioMed Central and we will help you at every step:Barr et al. Nutrition Journal  (2016) 15:33 Page 9 of 9


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