Joy LeeFollow

Date of Award

Summer 6-23-2015

Degree Type


Degree Name

Master of Science (MS)



First Advisor

Dr. Megan A. McCrory, PhD

Second Advisor

Barbara Hopkins, MMSc, RD, LD

Third Advisor

Dr. Terryl J. Hartman, PhD, MPH, RD


Background: The role of breakfast skipping in obesity and associated co-morbidities is uncertain. Experimental studies show mixed findings while observational studies show breakfast skipping is consistently associated prospectively with weight gain or cross-sectionally with higher BMI. Relatively few studies exist on breakfast skipping in relation to metabolic syndrome (MetS). One difficulty in examining these associations is that self-reported energy intake (rEI) is often under reported, particularly among overweight and obese individuals, and most previous research on breakfast consumption and obesity has not taken these implausible rEIs into account. Additionally, there is no standard definition of breakfast, leading to difficulty comparing across studies.

Objective: We investigated the associations between the timing of morning eating, rather than breakfast skipping per se, with risks for overweight/obesity, elevated waist circumference and MetS using US national survey data. We examined these associations in both the total sample and in the plausible subsample after excluding individuals with implausible rEIs.

Methods: We included non-pregnant participants from the Continuous National Health and Nutrition Examination Survey (NHANES) cycles 2005-2006, 2007-2008, and 2009-2010 aged 20-65 years who did not perform shift work and who completed 2 multiple pass 24h dietary recalls. Participants were classified according to their BMI as being either underweight (17.0-18.5 kg/m2), normal weight (>18.5 to 24.9 kg/m2), overweight (25.0 to 29.9 kg/m2) or obese (30.0 to 60.0 kg/m2). Individuals were categorized as having MetS or not based on both Adult Treatment Panel III (ATP III) and International Diabetes Federation (IDF) criteria. Waist circumference was categorized as elevated or normal based on ATPIII criteria. Morning intake on each day was categorized as early, late or none according to time of first reported intake ≥50 kcals. Morning intake was categorized as “early” if the first intake occurred between 5:00 AM and 8:59 AM, “late” if it occurred between 9:00 AM and 11:30 AM, and none if there was no intake during either time period. Across the two days of dietary intake, six morning eating patterns were possible: 1) early intake on both days; 2) early intake on one day and late intake on the other; 3) early intake one day and no intake the other; 4) late intake both days; 5) late intake on day and no intake the other; and 6) and no morning intake on either day. The two-day average rEI was compared to estimated energy requirements (EER) using the Institute of Medicine equations to determine energy intake plausibility. The rEIs were deemed implausible if rEI was not within the ±1SD calculated range of EER (±22.69%). Statistical analyseswere performed in SPSS on both the total sample and the plausibly reporting subsample. Logistic regression and multinomial logistic regression analyses were performed to determine the associations between timing of morning eating and risk for overweight/obesity, elevated waist circumference, and MetS, controlling for age, sex, race/ethnicity, smoking status, alcohol consumption, chronic disease presence, and poverty-income ratio. In this preliminary analysis, survey design was not taken into account.

Results: There were n=4590 and 2174 participants in the total sample and plausible subsample, respectively, with median BMIs (95%CI) 27.5 kg/m2 (28.4, 28.8 kg/m2) and 27.1 kg/m2 (27.8, 28.3 kg/m2). Relative to normal weight individuals, the odds of being obese was lower in those having eaten early in the morning both days (OR 0.663, 95% CI: 0.457, 0.961, p=0.030) or late in the morning both days (OR 0.622, 95% CI: 0.419, 0.923, p=0.018) in the total sample, but not in the plausible subsample. This result may be explained by the fact that across all weight status categories, these two morning eating patterns had the lowest percentage of under reporters (21-48%) compared to the other morning eating patterns (26-73%). Also, participants in the total sample who reported no morning intake had rEIs of 66% of EER, while those who reported other morning eating patterns had rEIs between 77-89% of EER. Overall, rEI as a % of EER was much greater in the plausible sample. There were no significant associations between morning eating and waist circumference or MetS as defined by either criteria in either sample.

Conclusions: These preliminary results illustrate the importance of accounting for rEI plausibility in studies of eating patterns in relation to disease risk, and that the timing of morning eating may be unrelated to BMI, waist circumference or MetS in adults aged 20-65 years.