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.
2010 / Genome-wide association study with targeted follow-up
Researchers analyzed associations between body mass index and approximately 2.8 million SNPs in up to 123,865 individuals, followed by targeted follow-up of 42 SNPs in up to 125,931 additional individuals. The study confirmed 14 known obesity susceptibility loci and identified 18 new loci associated with body mass index (P < 5 × 10⁻⁸), including a copy number variant near GPRC5B. Several identified loci map near hypothalamic regulators of energy balance (MC4R, POMC, SH2B1, BDNF) and near the incretin receptor GIPR.
Shows an association, not proof of cause and effect.
Speliotes EK, Willer CJ, Berndt SI, et al. Nat Genet, 2010. PMID 20935630
2010 / Retrospective case series
A retrospective study evaluated 16 patients (15 female, median age 19 years) hospitalized between 2004 and 2009 with psychotic symptoms who used herbal slimming products adulterated with sibutramine or its analogs. Reported psychotic features included auditory hallucination in 10 (63%), visual hallucination in 6 (38%), persecutory ideas in 6 (38%), delusions in 4 (25%), and suicidal ideation in 2 (13%). Analysis of 20 products revealed sibutramine in 16, N-desmethyl-sibutramine in 2, N-bisdesmethyl-sibutramine in 1, and other adulterants such as phenolphthalein in 9. The authors noted an association between the use of these products and psychotic features, calling for further studies.
Shows an association, not proof of cause and effect.
Chen SP, Tang MH, Ng SW, et al. Clin Toxicol (Phila), 2010. PMID 20969504
2010 / Case series
This study reports three cases of acute liver injury induced by weight-loss herbal supplements. One patient consumed Hydroxycut, and the other two took Herbalife supplements. Liver biopsies in all three patients revealed findings consistent with drug-induced acute liver injury.
Shows an association, not proof of cause and effect.
Chen GC, Ramanathan VS, Law D, et al. World J Hepatol, 2010. PMID 21173910
2009 / Cross-sectional comparison study
Researchers compared abdominal fat distribution, dietary intake, and biochemical data between non-obese and obese patients with NAFLD. Obese patients had significantly higher waist circumference, total abdominal fat, subcutaneous fat, and leptin levels, while visceral fat, adiponectin, and TNF-alpha did not significantly differ. Non-obese patients showed significantly lower immunoreactive insulin and HOMA-IR. Notably, dietary cholesterol intake was significantly higher and polyunsaturated fatty acid (PUFA) intake was significantly lower in the non-obese group compared with the obese group.
Shows an association, not proof of cause and effect.
Yasutake K, Nakamuta M, Shima Y, et al. Scand J Gastroenterol, 2009. PMID 19058085
2009 / Prospective cohort study
A prospective cohort study evaluated 252 women over 20 months to examine how changes in fiber intake relate to weight and body fat changes. Nearly 50% of the women gained weight and fat over the follow-up period. Each 1 g increase in total fiber consumed was associated with a 0.25 kg decrease in weight (P = 0.0061) and a 0.25 percentage point decrease in body fat (P = 0.0052). Controlling for confounders did not alter these relationships, except for changes in energy intake, which weakened the associations by 24-32%.
Shows an association, not proof of cause and effect.
Tucker LA, Thomas KS J Nutr, 2009. PMID 19158230
2009 / Case report
A 49-year-old female physician presented with peripheral edema, weight gain, and relative hypertension reported as caused by consuming licorice candy cigars containing glycyrrhizic acid (GZA) from natural licorice extract. The patient's response to GZA resolved spontaneously. The authors note that natural licorice extract can be found in products such as black licorice, chewing gum, herbal teas, soft drinks, tobaccos, and certain herbal remedies.
Shows an association, not proof of cause and effect.
Johns C CJEM, 2009. PMID 19166646
2009 / Population-based study
Using the 1988–1994 Third National Health and Nutrition Examination Survey, excess deaths were calculated for standard BMI levels and comparable levels of percentage body fat, waist circumference, hip and arm circumferences, waist-hip ratio, skinfolds, and waist-stature ratio. For levels equivalent to BMI <18.5, excess deaths ranged from 0.3% (waist-hip ratio) to 2.4% (percentage body fat), with all except waist circumference, waist-hip ratio, and waist-stature ratio significantly >0 (P < 0.05). For BMI 25 to <30 equivalents, excess deaths were 0.1% for percentage body fat and negative for other variables. For BMI >=30 equivalents, estimates ranged from -1.7% to 1.5% and none differed significantly from zero. All-cause, obesity-related, and other mortality estimates showed no significant or systematic differences between BMI and other measures.
Shows an association, not proof of cause and effect.
Flegal KM, Graubard BI Am J Clin Nutr, 2009. PMID 19190072
2009 / Prospective cohort study
Participants in the Japan Collaborative Cohort (JACC) Study were followed for an average of 12.5 years to assess all-cause mortality across combinations of six lifestyle factors (not currently smoking, not heavily drinking, walking >=1 h/day, sleeping 6.5-7.4 h/day, eating green-leafy vegetables almost daily, and BMI 18.5-24.9). During follow-up, 8,497 deaths occurred. Compared with having 0-2 healthy lifestyle factors, having all 6 was linked to an age-adjusted hazard ratio of 0.42 (95% CI: 0.32-0.56) in men and 0.49 (95% CI: 0.39-0.60) in women. The authors estimated that achieving a 1-point increment in lifestyle scores could have reduced death rates by 24.7% in men and 18.5% in women.
Shows an association, not proof of cause and effect.
Tamakoshi A, Tamakoshi K, Lin Y, et al. Prev Med, 2009. PMID 19254743
2009 / Retrospective case-control study
This retrospective case-control study evaluated weight gain dynamics and HPA-axis activity in 14 women with stress-related obesity (SRO), 21 with nonstress-related obesity (NSRO), and 21 normal-weight controls. Anthropometry, diet, and physical activity were similar between the obese groups. However, the SRO group had significantly higher 24-h urinary free cortisol (41.1 +/- 14.3 microg) compared with NSRO (26.6 +/- 17.6 microg) and controls (21.1 +/- 9.8 microg; P < 0.001). Time to achieve maximum weight gain was significantly shorter (P < 0.001) and the weight gain/time ratio was significantly higher (P < 0.001) in SRO versus NSRO, with a tendency toward a significant correlation between urinary cortisol and the weight gain/time ratio in the SRO group.
Shows an association, not proof of cause and effect.
Vicennati V, Pasqui F, Cavazza C, et al. Obesity (Silver Spring), 2009. PMID 19300426
2009 / Cross-sectional analysis
Data from 2,516 participants aged 15-23 years were analyzed to compare current (4.3%), former (10.8%), and never (84.9%) vegetarians. Current vegetarians in younger and older cohorts showed healthier dietary intakes regarding fruits, vegetables, and fat, and young adult current vegetarians were less likely to be overweight or obese than never vegetarians. However, current vegetarian adolescents and young adults were more likely to report binge eating with loss of control compared to nonvegetarians. Additionally, former vegetarians were more likely than never vegetarians (among adolescents) and more likely than both current and never vegetarians (among young adults) to engage in extreme unhealthful weight-control behaviors.
Shows an association, not proof of cause and effect.
Robinson-O'Brien R, Perry CL, Wall MM, et al. J Am Diet Assoc, 2009. PMID 19328260
2009 / Secondary analysis of cross-sectional survey data
Using NHANES 2003-2004 data with two 24-hour recalls from 3,098 children and adolescents, researchers modeled the impact of beverage substitution on total energy intake (TEI). Each additional 8-oz serving of sugar-sweetened beverages (SSBs) was linked to a net increase of 106 kcal/d (95% CI, 91 to 121 kcal/d), while water (+8 kcal/d) and diet drinks (+47 kcal/d) showed no net increases. Replacing SSBs with water was linked to a significant decrease in TEI, where each 1% of beverage replacement corresponded to 6.6-kcal lower TEI. The authors estimated that replacing all SSBs with water could lead to an average reduction of 235 kcal/d.
Shows an association, not proof of cause and effect.
Wang YC, Ludwig DS, Sonneville K, et al. Arch Pediatr Adolesc Med, 2009. PMID 19349562
2009 / Cross-sectional study
Researchers analyzed 22,434 men and 38,469 women in the Adventist Health Study-2 to evaluate the prevalence of type 2 diabetes across various vegetarian diet patterns. Mean BMI was lowest in vegans (23.6 kg/m2) and increased incrementally in lacto-ovo vegetarians (25.7 kg/m2), pesco-vegetarians (26.3 kg/m2), semi-vegetarians (27.3 kg/m2), and nonvegetarians (28.8 kg/m2). Type 2 diabetes prevalence was 2.9% in vegans, 3.2% in lacto-ovo vegetarians, 4.8% in pesco-vegetarians, 6.1% in semi-vegetarians, and 7.6% in nonvegetarians. After multivariate adjustment including BMI, the odds ratios for diabetes compared with nonvegetarians were 0.51 (95% CI 0.40-0.66) for vegans, 0.54 (0.49-0.60) for lacto-ovo vegetarians, 0.70 (0.61-0.80) for pesco-vegetarians, and 0.76 (0.65-0.90) for semi-vegetarians.
Shows an association, not proof of cause and effect.
Tonstad S, Butler T, Yan R, et al. Diabetes Care, 2009. PMID 19351712
2009 / Retrospective cross-sectional study
Researchers analyzed 3640 consecutive 18F-FDG PET-CT scans from 1972 patients to identify depots of brown adipose tissue, validated by UCP1 immunostaining in surgical biopsies. Positive scans were detected in 7.5% of women (76 of 1013) and 3.1% of men (30 of 959), showing a female:male ratio greater than 2:1 (P<0.001). Detection was inversely correlated with age (P<0.001), outdoor temperature (P=0.02), beta-blocker use (P<0.001), and, among older patients, body-mass index (P=0.007).
Shows an association, not proof of cause and effect.
Cypess AM, Lehman S, Williams G, et al. N Engl J Med, 2009. PMID 19357406
2009 / Comparative risk assessment modeling study
Using national health surveys and mortality statistics alongside effect estimates from systematic reviews, researchers estimated deaths attributable to 12 modifiable risk factors in the US in 2005. Smoking (467,000 deaths; 95% CI 436,000-500,000) and high blood pressure (395,000 deaths; 95% CI 372,000-414,000) were the leading causes. Overweight-obesity (216,000 deaths) and physical inactivity (191,000 deaths) each accounted for nearly 1 in 10 deaths. Among dietary factors, high dietary salt (102,000 deaths), low dietary omega-3 fatty acids (84,000 deaths), and high dietary trans fatty acids (82,000 deaths) had the largest estimated mortality effects.
Shows an association, not proof of cause and effect.
Danaei G, Ding EL, Mozaffarian D, et al. PLoS Med, 2009. PMID 19399161
2009 / Observational study using survey data
Using data from the National Longitudinal Survey of Youth and the Medical Expenditure Panel Survey, researchers examined who pays the healthcare costs associated with obesity among workers. They found that the incremental healthcare costs associated with obesity are passed on to obese workers with employer-sponsored health insurance as lower cash wages. In contrast, obese workers without employer-sponsored insurance did not show a wage offset compared to non-obese counterparts. Furthermore, a substantial portion of the lower wages among obese women previously attributed to labor market discrimination was explained by higher health insurance premiums.
Shows an association, not proof of cause and effect.
Bhattacharya J, Bundorf MK J Health Econ, 2009. PMID 19433210
2009 / Repeated cross-sectional survey analysis
Researchers compared adherence to five healthy lifestyle habits in US adults aged 40-74 years between NHANES 1988-1994 and NHANES 2001-2006. Over the 18-year period, adults with a BMI >=30 kg/m2 increased from 28% to 36% (P < .05), physical activity >=12 times/month decreased from 53% to 43% (P < .05), eating >=5 fruits and vegetables a day decreased from 42% to 26% (P < .05), and moderate alcohol use increased from 40% to 51% (P < .05). Smoking rates were relatively unchanged (26.9% to 26.1%). Overall adherence to all five healthy habits decreased from 15% to 8% (P < .05).
Shows an association, not proof of cause and effect.
King DE, Mainous AG, Carnemolla M, et al. Am J Med, 2009. PMID 19486715
2009 / Retrospective cohort study
Electronic medical records were examined to estimate the proportion of patients with select cancers who developed cachexia based on four separate clinical definitions. Among 8,541 cancer patients, cachexia was identified in 2.4% using the ICD-9 cachexia code, 5.5% using expanded diagnoses, 6.4% using prescription medication records, and 14.7% using a criterion of >=5% weight loss. The authors conclude that the proportion of patients with cachexia varied considerably depending on the definition employed, highlighting the need for a standard operational definition.
Shows an association, not proof of cause and effect.
Fox KM, Brooks JM, Gandra SR, et al. J Oncol, 2009. PMID 19587829
2009 / Prospective cohort study
The EPIC-Potsdam study followed 23,153 German participants for a mean of 7.8 years to assess the relationship between four healthy lifestyle factors (never smoking, BMI < 30, physical activity >=3.5 h/wk, and a healthy diet) and chronic diseases. During follow-up, 2,006 participants developed incident diabetes, myocardial infarction, stroke, or cancer. After adjusting for age, sex, and educational/occupational status, participants with all four healthy factors had a 78% lower risk of chronic disease (95% CI, 72% to 83%) compared to those with zero factors (diabetes: 93% lower; myocardial infarction: 81% lower; stroke: 50% lower; cancer: 36% lower).
Shows an association, not proof of cause and effect.
Ford ES, Bergmann MM, Kröger J, et al. Arch Intern Med, 2009. PMID 19667296
2009 / Cross-sectional study
In a cross-sectional study of American Samoan (n = 723) and Samoan (n = 785) adults, researchers assessed neo-traditional and modern dietary patterns in relation to metabolic syndrome. The neo-traditional pattern was linked to significantly higher serum HDL-cholesterol in American Samoa (P-trend = 0.05) and smaller abdominal circumference in both polities (P-trend = 0.004 and 0.01), with non-significant inverse associations with metabolic syndrome. The modern pattern was significantly positively associated with metabolic syndrome in Samoa (prevalence ratio = 1.21 for quintile 5 vs. 1; 95% CI: 0.93-1.57; P-trend = 0.05) and with increased serum triglycerides in both polities (P < 0.05).
Shows an association, not proof of cause and effect.
DiBello JR, McGarvey ST, Kraft P, et al. J Nutr, 2009. PMID 19710163
2009 / Longitudinal cohort study
In a longitudinal study, 85 overweight Latino youth aged 11-17 years were assessed over approximately 2 years (mean 1.5 ± 0.5 y) for changes in diet, body composition, and glucose and insulin indexes. Increases in total dietary fiber and insoluble fiber were inversely associated with changes in visceral adipose tissue (r = -0.29, P = 0.02 and r = -0.27, P = 0.03, respectively). Youth who decreased total dietary fiber (mean decrease 3 g · 1000 kcal⁻¹ · d⁻¹) had a 21% increase in visceral adipose tissue compared with a 4% decrease in those who increased fiber intake (P = 0.02). No other changes in dietary variables were related to changes in adiposity or metabolic variables.
Shows an association, not proof of cause and effect.
Davis JN, Alexander KE, Ventura EE, et al. Am J Clin Nutr, 2009. PMID 19793854
2009 / Sibling-comparison observational cohort study
Anthropometry and fasting blood levels were evaluated in 49 mothers who lost 36 +/- 1.8% body weight sustained for 12 +/- 0.8 yr and their 111 children (54 born before maternal surgery [BMS] and 57 born after [AMS]) aged 2.5-26 yr. AMS children had lower birth weight (2.9 +/- 0.1 vs. 3.3 +/- 0.1 kg, P = 0.003) and lower macrosomia prevalence (1.8 vs. 14.8%, P = 0.03). At follow-up, AMS offspring exhibited a 3-fold lower prevalence of severe obesity (11 vs. 35%, P = 0.004), higher insulin sensitivity (HOMA-IR 3.4 +/- 0.3 vs. 4.8 +/- 0.5, P = 0.02), improved lipid profiles (cholesterol/HDL 2.96 +/- 0.11 vs. 3.40 +/- 0.18, P = 0.03; HDL 1.50 +/- 0.05 vs. 1.35 +/- 0.05 mmol/liter, P = 0.04), lower C-reactive protein (0.88 +/- 0.17 vs. 2.00 +/- 0.34 microg/ml, P = 0.004), lower leptin (11.5 +/- 1.5 vs. 19.7 +/- 2.5 ng/ml, P = 0.005), and higher ghrelin (1.28 +/- 0.06 vs. 1.03 +/- 0.06 ng/ml, P = 0.005) than BMS offspring.
Shows an association, not proof of cause and effect.
Smith J, Cianflone K, Biron S, et al. J Clin Endocrinol Metab, 2009. PMID 19820018
2009 / Mathematical modeling and comparative population survey analysis
Researchers modeled weight changes from estimated energy intake changes in US children and adults between the 1970s and 2000s using equations derived from doubly labeled water studies. For children, the measured weight gain was 4.0 kg, which matched the predicted weight gain of 4.0 kg. For adults, measured weight gain was 8.6 kg, whereas predicted weight gain was 10.8 kg. The authors concluded that increased energy intake appears to be more than sufficient to explain US population weight gain, noting that reversing to 1970s mean weights would require reducing energy intake by approximately 2000 kJ/d (500 kcal/d) for adults and 1500 kJ/d (350 kcal/d) for children, or large compensatory increases in physical activity (eg, 110-150 min of walking/d).
Shows an association, not proof of cause and effect.
Swinburn B, Sacks G, Ravussin E Am J Clin Nutr, 2009. PMID 19828708
2009 / Mathematical modeling study
Researchers calculated nationwide US food waste energy content by comparing the US food supply to food intake estimated from a validated mathematical model relating body weight to food consumption. US per capita food waste progressively rose by approximately 50% since 1974, exceeding 1400 kcal per person per day (150 trillion kcal per year). Food waste accounted for over one-fourth of total freshwater consumption and roughly 300 million barrels of oil annually.
Shows an association, not proof of cause and effect.
Hall KD, Guo J, Dore M, et al. PLoS One, 2009. PMID 19946359
2008 / Postprandial observational challenge and correlation study
Serum adiponectin was assessed in 110 male non-diabetic subjects following an oral glucose tolerance test and a standardised mixed, fat-containing meal. Adiponectin slightly decreased after both test meals, reaching a significant decrease 5 and 6 hours after the lipid load and 2 hours after the glucose load. Postprandial adiponectin showed stronger correlations with most metabolic syndrome traits than fasting adiponectin, including HDL (r 0.30 vs r 0.23), postprandial insulin AUC (r -0.20 vs r -0.16), and BMI (r -0.23 vs r -0.20). In multivariate analysis, postprandial triacylglycerols were the only independent predictor of adiponectin, with no significant association observed for the -11388 G/A promoter polymorphism.
Shows an association, not proof of cause and effect.
Rubin D, Helwig U, Nothnagel M, et al. Br J Nutr, 2008. PMID 17663805