8,456 peer-reviewed studies, each with a short note written from its published abstract: what kind of study it was, who took part, what was measured and what the authors reported. Observational studies show associations, not proof.
Patterns in how people eat and what happens to them. Shows an association, not proof.
2013 / Cross-sectional and longitudinal observational study
Researchers analyzed the gut microbial composition in 123 non-obese and 169 obese Danish individuals. Individuals were categorized into two groups differing by gut bacterial richness; those with low bacterial richness (23% of the population) exhibited more marked overall adiposity, insulin resistance, dyslipidaemia, and a more pronounced inflammatory phenotype compared to those with high bacterial richness. Furthermore, obese individuals in the lower bacterial richness group gained more weight over time.
Shows an association, not proof of cause and effect.
Le Chatelier E, Nielsen T, Qin J, et al. Nature, 2013. PMID 23985870
2013 / Cross-sectional study
A cross-sectional study of 71,751 subjects from the Adventist Health Study 2 examined nutrient intakes across nonvegetarian, semi-vegetarian, pesco vegetarian, lacto-ovo vegetarian, and strict vegetarian diets using a 204-item food frequency questionnaire. Nonvegetarians had the lowest intakes of plant proteins, fiber, beta carotene, and magnesium, and the highest intakes of saturated, trans, arachidonic, and docosahexaenoic fatty acids. Strict vegetarians had lower tails of some nutrient distributions suggesting inadequate intakes by a portion of subjects. Mean BMI was highest in nonvegetarians (28.7, SD=6.4) and lowest in strict vegetarians (24.0, SD=4.8), while energy intake was close to 2,000 kcal/day across patterns except for semi-vegetarians (1,707 kcal/day).
Shows an association, not proof of cause and effect.
Rizzo NS, Jaceldo-Siegl K, Sabate J, et al. J Acad Nutr Diet, 2013. PMID 23988511
2013 / Prospective cohort study with matched controls
Researchers tracked the dynamics of organohalogenated contaminants (OHCs) and their metabolites in obese individuals before weight loss and after 3, 6, and 12 months, comparing them with matched lean controls. At baseline, obese individuals had significantly lower concentrations of major PCBs, their metabolites, and PCP compared to controls. During weight loss, increasing serum levels were observed for all OHCs, with release from adipose tissue appearing dependent on the octanol-water partition coefficient. The authors noted that weight loss influences the release and metabolism of OHCs, raising potential concern over increased levels of endocrine-disrupting compounds.
Shows an association, not proof of cause and effect.
Dirtu AC, Dirinck E, Malarvannan G, et al. Environ Sci Technol, 2013. PMID 24074050
2013 / Cross-sectional validation analysis
Researchers assessed the validity of caloric intake data from NHANES across 1971 to 2010 using 28,993 men and 34,369 women. Historical rEI/BMR values were 1.31 (95% CI: 1.30-1.32) for men and 1.19 (95% CI: 1.18-1.20) for women, while disparity values (rEI-TEE) were -281 kcal/day (95% CI: -299 to -264) and -365 kcal/day (95% CI: -378 to -351), respectively. Across the 39-year history, energy intake data on 67.3% of women and 58.7% of men were not physiologically plausible, with the greatest mean disparity values seen in obese men (-716 kcal/day) and obese women (-856 kcal/day).
Shows an association, not proof of cause and effect.
Archer E, Hand GA, Blair SN PLoS One, 2013. PMID 24130784
2013 / Cross-sectional analysis
Data from 23,168 US adults in NHANES 1999-2010 were analyzed to assess dietary fiber intake and cardiometabolic risks. Mean dietary fiber intake averaged 15.7-17.0 g, consistently below recommended adequate intake levels. Comparing the highest to lowest quintiles of fiber intake, adjusted predicted marginal risk ratios were 0.78 (95% CI: 0.69-0.88) for metabolic syndrome, 0.66 (95% CI: 0.61-0.72) for inflammation, and 0.77 (95% CI: 0.71-0.84) for obesity. Fiber intake was associated with lower inflammation across all racial/ethnic groups, but significant associations with obesity or metabolic syndrome were observed only among whites.
Shows an association, not proof of cause and effect.
Grooms KN, Ommerborn MJ, Pham DQ, et al. Am J Med, 2013. PMID 24135514
2013 / Content analysis and advertising exposure assessment
Researchers combined a content analysis of the nutritional quality and messaging in ready-to-eat cereal television advertisements with syndicated advertising exposure data for children and adults. Children viewed 1.7 ready-to-eat cereal ads per day, and 87% of these ads promoted high-sugar products, whereas adults viewed half as many ads with an equal distribution of high- and low-sugar cereals. Messages in high-sugar ads viewed by children were significantly more likely to communicate unrealistic and contradictory ideas: 91% ascribed extraordinary powers to products, and 67% depicted both healthy and unhealthy eating behaviors.
Shows an association, not proof of cause and effect.
LoDolce ME, Harris JL, Schwartz MB J Health Commun, 2013. PMID 24175878
2013 / Cross-sectional cohort analysis
In 54,311 adults from the NutriNet-Santé cohort, cluster analysis identified 5 consumer profiles, including regular (RCOP, 14%) and occasional (OCOP, 51%) organic product consumers. RCOP were more educated, more physically active, consumed more plant foods, fewer sweet and alcoholic drinks, less processed meat and milk, and had healthier nutrient profiles with closer adherence to dietary guidelines. After adjusting for confounders including guideline adherence, RCOP participants had a significantly lower probability of overweight (-36% in men, -42% in women) and obesity (-62% in men, -48% in women; P<0.0001) compared to non-interested consumers.
Shows an association, not proof of cause and effect.
Kesse-Guyot E, Péneau S, Méjean C, et al. PLoS One, 2013. PMID 24204721
2013 / Cross-sectional survey
A self-administered questionnaire was used to collect data from 8,159 female high school students aged 11-19 years to assess associations between happiness, leisure time, and health-related behaviors. Multivariate linear regression analysis showed significant associations between happiness and weight, regular exercise, exposure to second-hand tobacco smoke, daily fruit or vegetable consumption, and leisure time usage. Happiness was significantly associated with lower BMI, regular physical activity, absence of exposure to second-hand smoke, higher consumption of fruits and vegetables, and spending leisure time with family (all P < 0.005).
Shows an association, not proof of cause and effect.
Fararouei M, Brown IJ, Akbartabar Toori M, et al. J Adolesc, 2013. PMID 24215965
2013 / Cross-sectional and prospective cohort analysis
Using data from the Adventist Health Study-2, researchers examined vegetarian dietary patterns in relation to prevalent (4,237 of 65,981) and incident (1,184 of 41,212) hypothyroidism using multivariate logistic regression. In the prevalence analysis, a vegan versus omnivorous diet was non-significantly linked to reduced risk (OR 0.89, 95% CI: 0.78-1.01), whereas a lacto-ovo diet was linked to increased risk (OR 1.09, 95% CI: 1.01-1.18). Overweight and obesity also increased prevalence odds (OR 1.32 and 1.78, respectively). In the incidence study, a vegan diet tended to be protective but was not statistically significant (OR 0.78, 95% CI: 0.59-1.03).
Shows an association, not proof of cause and effect.
Tonstad S, Nathan E, Oda K, et al. Nutrients, 2013. PMID 24264226
2013 / Cross-sectional analysis of survey data
Using NHANES 2003-2008 data, researchers assessed the relationship between urinary triclosan levels and BMI in adults using presence versus absence and quartiles of exposure. Compared with undetectable levels, detectable triclosan was associated with a 0.9-point increase in BMI (p<0.001). Relative to the lowest quartile, the 2nd, 3rd, and 4th quartiles of urinary triclosan were associated with BMI increases of 1.5 (p<0.001), 1.0 (p=0.002), and 0.3 (p=0.33), respectively. Its metabolite, 2,4-dichlorophenol, was strongly correlated (ρ=0.4) but showed a weaker association with BMI, and no other likely confounder was identified.
Shows an association, not proof of cause and effect.
Lankester J, Patel C, Cullen MR, et al. PLoS One, 2013. PMID 24278238
2012 / Longitudinal observational study
This 6-year longitudinal observational study evaluated whether changes in sleep duration were associated with objective adiposity indicators in adults aged 18-64 years. Baseline short sleepers (≤6 h/day) who maintained short sleep (n=20) were compared to those who increased sleep to 7-8 h/day (n=23) and controls who slept 7-8 h/day throughout (n=173). Short sleepers who maintained short sleep had a significantly greater increase in BMI (difference: 1.1±0.36 kg m^-2, P<0.05) and fat mass (difference: 2.4±0.64 kg, P<0.05) over 6 years than those who increased their sleep duration. No significant difference in adiposity changes was observed between controls and short sleepers who increased sleep duration.
Shows an association, not proof of cause and effect.
Chaput JP, Després JP, Bouchard C, et al. Int J Obes (Lond), 2012. PMID 21654631
2012 / Cross-sectional survey analysis
Using weighted clinic visit data from the National Ambulatory Medical Care Survey (2005-2007), researchers analyzed preventive care patterns across weight groups. Obese patients were significantly less likely than normal-weight patients to receive breast exams (OR 0.8), mammograms (OR 0.7), pap smears (OR 0.7), pelvic exams (OR 0.8), and rectal exams (OR 0.7). They were also less likely to receive tobacco education (OR 0.7), injury prevention education (OR 0.7), or psychotherapy referrals (OR 0.6), but received significantly more diet, exercise, and weight reduction education.
Shows an association, not proof of cause and effect.
Hernandez-Boussard T, Ahmed SM, Morton JM Obesity (Silver Spring), 2012. PMID 21818146
2012 / Prospective cohort study
Researchers tracked 2,820 men and 2,950 Taiwanese women for an average of 6.5 years (up to 6.8 years, 38,077 person-years) following the 2002 TwSHHH survey, identifying 225 all-cause deaths. Bean consumption was favourable for metabolic syndrome components in men (except fasting glucose) and for waist circumference and HbA1c in women. A bean-free diet was linked to higher metabolic syndrome scores in men (adjusted OR 1.83) and women (adjusted OR 1.45). In Cox regression models, a bean-free diet was associated with increased all-cause mortality in women (HR 1.98, 95% CI 1.03, 3.81) but not in men (HR 1.28, 95% CI 0.76, 2.16).
Shows an association, not proof of cause and effect.
Chang WC, Wahlqvist ML, Chang HY, et al. Public Health Nutr, 2012. PMID 21899787
2012 / Pre-post program evaluation
This study evaluated biometric changes in 5,070 participants of a 30-day volunteer-delivered Coronary Health Improvement Project across North America between January 2006 and October 2009. Significant overall reductions (p < 0.001) were recorded in body mass (-3.2%), systolic and diastolic blood pressure (-4.9% and -5.3%), total cholesterol (-11.0%), LDL cholesterol (-13.0%), triglycerides (-7.7%), and fasting plasma glucose (-6.1%). Larger reductions occurred in higher-risk subgroups, including a 19.8% average reduction in participants presenting with cholesterol >280 mg/dl, a 16.1% decrease in LDL for baseline LDL >190 mg/dl, and a 44.1% reduction in triglycerides for baseline levels >500 mg/dl.
Shows an association, not proof of cause and effect.
Rankin P, Morton DP, Diehl H, et al. Am J Cardiol, 2012. PMID 21944675
2012 / Comparative pilot study
This pilot study assessed response to non-surgical periodontal therapy in 30 obese patients with chronic periodontitis, comparing 15 patients with previous bariatric surgery and ≥40% excess weight loss for ≥6 months to 15 obese controls without surgery or weight loss. All participants received scaling, root planing, and oral hygiene instructions. Improvements after periodontal therapy were statistically significantly greater in the bariatric surgery group compared with the obese control group (P <0.05). Mean reductions for the surgery versus control groups were 0.45 mm versus 0.28 mm for probing depth, 0.44 mm versus 0.30 mm for clinical attachment level, 16% versus 15% for percentage of bleeding on probing sites, and 1.03 versus 0.52 for gingival index.
Shows an association, not proof of cause and effect.
Lakkis D, Bissada NF, Saber A, et al. J Periodontol, 2012. PMID 22014172
2012 / Repeated cross-sectional study (BRFSS)
Using 1994-2007 Behavioral Risk Factor Surveillance System data, researchers assessed trends in 4 healthy lifestyle characteristics: healthy weight, not smoking, consuming fruits and vegetables, and engaging in physical activity. Between 1994 and 2007, nonsmoking increased by 4% while having a healthy weight decreased by 10%, with little change in fruit and vegetable intake or physical activity. Overall healthy lifestyle prevalence increased minimally over time; in 2007, prevalence was 6% in the Northeast and West, compared to 4% in the South and Midwest.
Shows an association, not proof of cause and effect.
Troost JP, Rafferty AP, Luo Z, et al. Am J Public Health, 2012. PMID 22095344
2012 / Cross-sectional survey analysis
Data from 14,515 US adults in NHANES 2003-2008 were analyzed to estimate the prevalence of cardiovascular (CV) health behaviors and factors across age, sex, and race/ethnic groups. Fewer than 1% of adults achieved ideal CV health across all seven metrics. Among health behaviors, nonsmoking was most prevalent (range, 60.2%-90.4%), whereas an ideal Healthy Diet Score was least prevalent (range, 0.2%-2.6%). Prevalences of ideal body mass index and ideal physical activity were higher in younger adults, while ideal cholesterol, blood pressure, and fasting blood glucose were lower in older adults.
Shows an association, not proof of cause and effect.
Shay CM, Ning H, Allen NB, et al. Circulation, 2012. PMID 22095826
2012 / Repeated cross-sectional study (NHANES 1999-2008 secondary data analysis)
Using NHANES data from 1999 through 2008, researchers assessed parental recall of doctor notification of child overweight among 4,985 children aged 2 to 15 years with a BMI in the 85th percentile or higher. Overall, 22% of parents recalled being told by a health professional that their child was overweight, with recall being more likely among nonwhite and poor children. The percentage increased from 19.4% in 1999-2004 to 23.2%, reaching 29.1% in 2007-2008, a trend that remained significant in multivariate analysis.
Shows an association, not proof of cause and effect.
Perrin EM, Skinner AC, Steiner MJ Arch Pediatr Adolesc Med, 2012. PMID 22147758
2012 / Prospective cohort study
Researchers examined the association between baseline meat intake and BMI change over 14 years among 3,902 older Dutch adults. Total fresh meat intake was not linked to prospective BMI change in men (-0.06 kg/m² for highest vs. lowest quintile, P = 0.75) or women (0.26 kg/m², P = 0.20). However, men with the highest beef intake had a significantly lower increase in BMI after 14 years (0.60 kg/m²). Conversely, higher BMI increases after 14 years were linked to higher pork intake in women (0.47 kg/m²) and higher chicken intake in both sexes (0.36 kg/m²).
Shows an association, not proof of cause and effect.
Gilsing AM, Weijenberg MP, Hughes LA, et al. J Nutr, 2012. PMID 22223576
2012 / Cross-sectional calibration sub-study
In a calibration sub-study of 500 white participants from the Adventist Health Study-2, dietary patterns were categorized into vegan, lacto-ovo vegetarian, partial vegetarian, and omnivorous. Vegan vegetarians had significantly lower systolic (beta = -6.8 mmHg, P < 0.05) and diastolic BP (beta = -6.9 mmHg, P < 0.001) than omnivores, with similar findings in lacto-ovo vegetarians (beta = -9.1 mmHg, P < 0.001 and beta = -5.8 mmHg, P < 0.001). The odds ratio of hypertension compared with omnivores was 0.37 (95% CI 0.19, 0.74) for vegans, 0.57 (95% CI 0.36, 0.92) for lacto-ovo vegetarians, and 0.92 (95% CI 0.50, 1.70) for partial vegetarians. Effects were reduced after adjustment for BMI.
Shows an association, not proof of cause and effect.
Pettersen BJ, Anousheh R, Fan J, et al. Public Health Nutr, 2012. PMID 22230619
2012 / Cross-sectional analysis
A cross-sectional study of 39 primary care physicians and 199 patients evaluated the association between patient BMI and estimation of physician respect using questionnaires and audio-recorded visits. For each 5 kg/m2 increase in BMI, the odds of overestimating physician respect significantly increased (OR 1.32, 95% CI 1.04-1.68, p=0.02). Few patients underestimated physician respect, and independent evaluators of audiotapes found no differences in ratings of physician respectfulness. The authors consider the results preliminary.
Shows an association, not proof of cause and effect.
Gudzune KA, Huizinga MM, Beach MC, et al. Patient Educ Couns, 2012. PMID 22240006
2012 / National cross-sectional survey
A national cross-sectional survey of 500 primary care physicians evaluated the relationship between physician BMI and obesity care practices and beliefs. Normal-BMI physicians were more likely to discuss weight loss with obese patients than overweight/obese physicians (30% vs. 18%, P = 0.010) and had higher confidence in providing diet (53% vs. 37%, P = 0.002) and exercise counseling (56% vs. 38%, P = 0.001). Physicians were more likely to record an obesity diagnosis (93% vs. 7%, P < 0.001) or initiate a weight loss conversation (89% vs. 11%, P <= 0.001) if they perceived the patient's weight met or exceeded their own body weight.
Shows an association, not proof of cause and effect.
Bleich SN, Bennett WL, Gudzune KA, et al. Obesity (Silver Spring), 2012. PMID 22262162
2012 / Comparative cross-sectional study (two studies)
Two studies investigated dietary restraint and eating disorder symptoms across different dietary patterns. In Study 1, semi-vegetarians reported the highest levels of eating-related pathology, whereas true vegetarians and vegans appeared healthiest regarding weight and eating. Study 2 found little evidence for more eating-related pathology in semi-vegetarians compared to omnivores, noting that higher scores on traditional assessments appeared artificially inflated by items assessing avoidance of specific foods considered normative in a vegetarian diet.
Shows an association, not proof of cause and effect.
Timko CA, Hormes JM, Chubski J Appetite, 2012. PMID 22343135
2012 / Cross-sectional comparative study
A cross-sectional study compared 169 healthy Chinese lacto-vegetarians with 126 healthy omnivore men aged 21-76 years. Compared with omnivores, lacto-vegetarians had remarkably lower body mass index, systolic and diastolic blood pressure, fasting blood glucose, triglycerides, total cholesterol, LDL cholesterol, apolipoprotein B, γ-glutamyl transferase, serum creatinine, uric acid, and total cholesterol/HDL cholesterol ratio. Vegetarians also showed higher HOMA beta cell function, a higher insulin secretion index, and thinner carotid intima-media thickness. Among participants aged 24-55 years, these findings corresponded to lower cardiovascular risk points and a lower probability of developing cardiovascular disease in 5-10 years.
Shows an association, not proof of cause and effect.
Yang SY, Li XJ, Zhang W, et al. Nutr Clin Pract, 2012. PMID 22412169