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.
2007 / Cross-sectional study
Researchers compared 278 long-term multiple dietary supplement users with 176 single multivitamin/mineral users and 602 non-users from NHANES datasets using online questionnaires and physical examinations. Serum nutrient concentrations generally increased with greater supplement intake. After adjusting for age, gender, income, education, and BMI, greater supplement use was associated with more favorable levels of homocysteine, C-reactive protein, HDL cholesterol, and triglycerides, as well as lower odds of prevalent elevated blood pressure and diabetes. The authors note these findings should be confirmed in other groups of heavy supplement users.
Shows an association, not proof of cause and effect.
Block G, Jensen CD, Norkus EP, et al. Nutr J, 2007. PMID 17958896
2006 / Case-control study
A case-control study compared 95 long-term lacto-ovo-vegetarians with 107 non-vegetarian controls, all of normal weight (BMI 18.6–25.0 kg/m²). Glucose, insulin, and HOMA-IR were significantly lower in vegetarians (glucose 4.47 +/- 0.05 vs. 4.71 +/- 0.07 mmol/l; insulin 4.96 +/- 0.23 vs. 7.32 +/- 0.41 mU/l; HOMA-IR 0.99 +/- 0.05 vs. 1.59 +/- 0.10). HOMA-IR increased significantly with age (starting at 31–40 years) only among subjects on a traditional western diet, whereas vegetarians maintained low values across age groups.
Shows an association, not proof of cause and effect.
Valachovicová M, Krajcovicová-Kudlácková M, Blazícek P, et al. Eur J Nutr, 2006. PMID 15940383
2006 / Ecological study
Researchers conducted an ecological study across 37 world areas with a reported 3% yearly increase in type 1 diabetes (T1D) incidence, assessing per caput daily supply of milk, meat, and cereals from 1961 to 2000 using FAO Food Balance Sheets. While milk and cereal supplies remained almost unchanged, meat supply increased by over 31%, with an absolute mean T1D increase of +0.32 cases per 100,000 per year. A significant positive correlation with milk supply was observed from 1961 to 2000, whereas correlations with meat and cereals became significant in 1983 and 2000.
Shows an association, not proof of cause and effect.
Muntoni S, Muntoni S Ann Nutr Metab, 2006. PMID 16276070
2006 / Prospective cohort study
In 529 consecutive patients with type 2 diabetes followed for 5 years, researchers evaluated the predictive value of HbA1c and blood glucose measured fasting, after breakfast, after lunch, and before dinner on cardiovascular events. Cardiovascular events occurred in 77 subjects (19% of men and 9.4% of women). When all blood glucose values and HbA1c were analyzed simultaneously, only blood glucose after lunch predicted cardiovascular events; the hazard ratio comparing the third tertile to the first and second was 5.54 (95% CI, 1.45-21.20) in women and 2.12 (95% CI, 1.04-4.32) in men (P < 0.01).
Shows an association, not proof of cause and effect.
Cavalot F, Petrelli A, Traversa M, et al. J Clin Endocrinol Metab, 2006. PMID 16352690
2006 / Ecological comparative study
Researchers compared the consumption of cow milk proteins A1 and B beta-casein among 2-year-olds and 11- to 14-year-olds in Iceland and Scandinavia (Norway, Denmark, Sweden, and Finland) and evaluated it against type 1 diabetes incidence (0-14 years). Among 2-year-olds, consumption differed significantly between Iceland and Scandinavia for A1 (p = 0.034) and the sum of A1 and B beta-casein (p = 0.021), and A1 consumption correlated with type 1 diabetes incidence across the countries (r = 0.9; p = 0.037). No significant difference in consumption of A1 or the sum of A1 and B beta-casein was observed for 11- to 14-year-olds.
Shows an association, not proof of cause and effect.
Birgisdottir BE, Hill JP, Thorsson AV, et al. Ann Nutr Metab, 2006. PMID 16407643
2006 / Cross-sectional comparative study
Ninety-eight healthy Taiwanese women (49 Buddhist lactovegetarians and 49 omnivores) were evaluated for dietary intake, anthropometrics, hormone secretion, and glycaemic control. Vegetarians had lower BMI, smaller waist circumference, significantly lower fasting insulin (median: 35.3 vs. 50.6 pmol/l), and lower plasma glucose (mean: 4.7 vs. 4.9 mmol/l). Homeostasis model assessment insulin resistance was significantly lower in vegetarians (median: 1.10 vs. 1.56), with BMI and diet independently contributing 18% and 15% of the variation in insulin resistance, respectively.
Shows an association, not proof of cause and effect.
Hung CJ, Huang PC, Li YH, et al. Br J Nutr, 2006. PMID 16441925
2006 / Prospective cohort study
A prospective study followed 84,555 women without chronic disease for 20 years, documenting 4,496 new cases of type 2 diabetes. Potato and french fry intake were both positively associated with type 2 diabetes risk; multivariate relative risk (RR) comparing the highest to lowest quintile was 1.14 (95% CI: 1.02, 1.26; P for trend = 0.009) for potatoes and 1.21 (95% CI: 1.09, 1.33; P for trend < 0.0001) for french fries. The RR was 1.18 (95% CI: 1.03, 1.35) for 1 daily potato serving and 1.16 (95% CI: 1.05, 1.29) for 2 weekly servings of french fries, with higher risk observed in obese women and when substituting 1 serving/day of potatoes for whole grains (RR: 1.30; 95% CI: 1.08, 1.57).
Shows an association, not proof of cause and effect.
Halton TL, Willett WC, Liu S, et al. Am J Clin Nutr, 2006. PMID 16469985
2006 / Prospective cohort study
Researchers followed 85,031 healthy women aged 34-59 years from 1980 to 2000, documenting 4,599 incident cases of type 2 diabetes over 20 years. No association was found between total, dietary, supplemental, or nonheme iron intake and type 2 diabetes risk. However, heme iron intake was positively associated with risk, with relative risks across increasing quintiles of cumulative intake of 1.00, 1.08, 1.20, 1.27, and 1.28 (P trend < 0.0001). Comparing women consuming ≥2.25 mg/day to those with <0.75 mg/day, the multivariate relative risk was 1.52 (95% CI 1.22-1.88), with significant associations observed in both overweight and lean women.
Shows an association, not proof of cause and effect.
Rajpathak S, Ma J, Manson J, et al. Diabetes Care, 2006. PMID 16732023
2006 / Cross-sectional study
Cross-sectional associations between serum concentrations of six persistent organic pollutants (POPs) and diabetes prevalence were examined in 2,016 adult participants from NHANES 1999-2002. After adjusting for age, sex, race and ethnicity, poverty income ratio, BMI, and waist circumference, diabetes prevalence was strongly positively associated with lipid-adjusted serum concentrations of all six POPs. When classified by the sum of category numbers of the six POPs, adjusted odds ratios were 1.0, 14.0, 14.7, 38.3, and 37.7 (P for trend < 0.001). The association was consistent in stratified analyses and stronger in younger participants, Mexican Americans, and obese individuals.
Shows an association, not proof of cause and effect.
Lee DH, Lee IK, Song K, et al. Diabetes Care, 2006. PMID 16801591
2006 / Cross-sectional comparative study
Researchers compared adult Pima Indians and non-Pima populations in the Sierra Madre mountains of Mexico to U.S. Pima Indians in Arizona using oral glucose tolerance tests, obesity assessments, and physical activity evaluations. The age- and sex-adjusted prevalence of type 2 diabetes in Mexican Pima Indians (6.9%) was less than one-fifth that in U.S. Pima Indians (38%) and similar to non-Pima Mexicans (2.6%). Obesity prevalence was also substantially lower in Mexican Pimas (7% in men, 20% in women) compared to U.S. Pimas, while physical activity levels were much higher in both Mexican groups.
Shows an association, not proof of cause and effect.
Schulz LO, Bennett PH, Ravussin E, et al. Diabetes Care, 2006. PMID 16873794
2006 / Prospective cohort study
In the Nurses' Health Study II, 13,110 women completed a food frequency questionnaire in 1991, with 758 incident cases of GDM documented between 1992 and 1998. After adjusting for age, parity, pre-pregnancy BMI, and other covariates, comparing the highest to the lowest quintile of the Western pattern yielded an RR of 1.63 (95% CI 1.20-2.21; p trend=0.001), while comparing the lowest to highest quintile of the prudent pattern had an RR of 1.39 (95% CI 1.08-1.80; p trend=0.018). Each additional serving per day was associated with an RR of 1.61 (95% CI 1.25-2.07) for red meat and 1.64 (95% CI 1.13-2.38) for processed meat.
Shows an association, not proof of cause and effect.
Zhang C, Schulze MB, Solomon CG, et al. Diabetologia, 2006. PMID 16957814
2006 / Cross-sectional autopsy study
Researchers evaluated the relation of Pathobiological Determinants of Atherosclerosis in Youth (PDAY) risk scores—based on sex, age, serum lipoproteins, smoking, hypertension, obesity, and hyperglycemia—to early and advanced atherosclerotic lesions in deceased individuals aged 15 to 34 years autopsied after external death. Risk scores derived from modifiable risk factors were associated with the prevalence of AHA grade 1 lesions in the left anterior descending coronary artery and with the extent of fatty streaks in the right coronary artery and abdominal aorta. These scores were also associated with higher-severity microscopic and gross lesions across the examined vessels.
Shows an association, not proof of cause and effect.
McMahan CA, Gidding SS, Malcom GT, et al. Pediatrics, 2006. PMID 17015535
2005 / Case-control study
A case-control study compared 21 matched vegans and 25 omnivores matched for sex, age, and BMI to examine associations with insulin resistance and intramyocellular lipid (IMCL). Groups did not differ in body fat, activity levels, energy intake, or HOMA %S. Vegans had significantly lower systolic blood pressure (-11.0 mmHg, P=0.027), fasting triacylglycerol (-0.7 mmol/l, P<0.001), fasting glucose (-0.4 mmol/l, P=0.05), and soleus IMCL levels (-9.7, P=0.01). Vegans also had higher HOMA %B (32.1%, P=0.005) and higher intakes of carbohydrate, nonstarch polysaccharides, and polyunsaturated fat, alongside a lower dietary glycaemic index.
Shows an association, not proof of cause and effect.
Goff LM, Bell JD, So PW, et al. Eur J Clin Nutr, 2005. PMID 15523486
2005 / Case report
An obese male with severe type 2 diabetes, mixed hyperlipidemia, chloracne, headaches, and numbness from organochlorine toxicity was treated with dietary olestra. After 2 years of including olestra in his diet, he achieved successful weight loss without symptom worsening and reverted to normoglycemia and normolipidemia. The authors suggest olestra may have assisted weight loss and diabetes improvement by increasing fecal elimination of organochlorines.
Shows an association, not proof of cause and effect.
Redgrave TG, Wallace P, Jandacek RJ, et al. J Nutr Biochem, 2005. PMID 15936651
2004 / Case-cohort study
A case-cohort study evaluated 580 incident cases of diabetes and 566 noncases from the Atherosclerosis Risk in Communities study over 9 years of follow-up. Comparing the fourth versus first quartile of plasma free fatty acids (FFAs), relative risks for incident diabetes were increased in a basic adjusted model (hazard ratio 1.68, 95% CI 1.20-2.34) and after further adjustment for baseline fasting glucose, insulin, BMI, waist circumference, triglycerides, and an inflammation score (1.63, 95% CI 1.04-2.57). Relative risks were lowest among normal-weight participants and highest among the obese, though a formal test for interaction between FFAs and BMI was not statistically significant.
Shows an association, not proof of cause and effect.
Pankow JS, Duncan BB, Schmidt MI, et al. Diabetes Care, 2004. PMID 14693970
2004 / Comparative cross-sectional study
Researchers compared insulin sensitivity indices between 19 Chinese vegetarians and 17 omnivores using an insulin suppression test. Omnivores showed higher steady-state plasma glucose (105.4+/-10.2 vs 80.3+/-11.3 mg/dl, P<0.001), higher fasting insulin (4.06+/-0.77 vs 3.02+/-1.19 microU/ml, P=0.004), and higher HOMA-IR (6.75+/-1.31 vs 4.78+/-2.07, P=0.002) compared to vegetarians. HOMA %S was higher in vegetarians (264.3+/-171.7% vs 159.2+/-31.7%, P=0.018), while HOMA %beta did not differ significantly between groups. A linear relationship was observed between years on a vegetarian diet and SSPG (r=-0.541, P=0.017).
Shows an association, not proof of cause and effect.
Kuo CS, Lai NS, Ho LT, et al. Eur J Clin Nutr, 2004. PMID 14749752
2004 / Cross-sectional cohort study
Researchers analyzed the relation of plasma BNP and N-ANP to body mass index in 3389 Framingham Study participants without heart failure. Multivariable-adjusted mean BNP levels decreased across lean, overweight, and obese men (21.4, 15.5, and 12.7 pg/mL; trend P<0.0001) and women (21.1, 16.3, and 13.1 pg/mL; trend P<0.001). Obese individuals had higher odds of low BNP (OR 2.51 in men; OR 1.84 in women) and low N-ANP (OR 4.81 in men; OR 2.85 in women) compared with lean individuals. Diabetes was also associated with low natriuretic peptide levels, with additive negative effects alongside obesity.
Shows an association, not proof of cause and effect.
Wang TJ, Larson MG, Levy D, et al. Circulation, 2004. PMID 14769680
2004 / Population-based case-control study
A population-based case-control study assessed 295 hepatocellular carcinoma (HCC) cases and 435 age-, gender-, and race-matched controls between 1984 and 2002. Heavy alcohol consumption and cigarette smoking were independently associated with a statistically significant 2- to 3-fold increased risk of HCC after adjusting for hepatitis B and C serology, while diabetes was associated with an OR of 2.7 (95% CI, 1.6-4.3). Synergistic interactions on HCC risk were found between heavy alcohol and diabetes (OR = 4.2; 95% CI, 2.6-5.8), heavy alcohol and viral hepatitis (OR = 5.5; 95% CI, 3.9-7.0), and diabetes and viral hepatitis (OR = 4.8; 95% CI, 2.7-6.9).
Shows an association, not proof of cause and effect.
Yuan JM, Govindarajan S, Arakawa K, et al. Cancer, 2004. PMID 15329910
2004 / Standardised case-control study
A standardized case-control study (INTERHEART) in 52 countries evaluated modifiable risk factors in 15,152 myocardial infarction cases and 14,820 controls. Significant associations with acute myocardial infarction were observed for smoking (OR 2.87), raised ApoB/ApoA1 ratio (OR 3.25), hypertension (OR 1.91), diabetes (OR 2.37), abdominal obesity (OR 1.62 for middle vs lowest tertile), psychosocial factors (OR 2.67), daily fruit and vegetable intake (OR 0.70), regular physical activity (OR 0.86), and regular alcohol intake (OR 0.91). Together, these nine factors accounted for 90% of the population attributable risk in men and 94% in women globally.
Shows an association, not proof of cause and effect.
Yusuf S, Hawken S, Ounpuu S, et al. Lancet, 2004. PMID 15364185
2003 / Case-control study
Dietary intake, insulin sensitivity, and postprandial triglyceride metabolism were compared between 25 NASH patients and 25 matched controls. NASH patients consumed significantly more saturated fat (13.7% +/- 3.1% vs. 10.0% +/- 2.1% total kcal, P = 0.0001) and cholesterol (506 +/- 108 vs. 405 +/- 111 mg/d, P = 0.002), and less polyunsaturated fat (10.0% +/- 3.5% vs. 14.5% +/- 4.0% total fat, P = 0.0001), fiber (12.9 +/- 4.1 vs. 23.2 +/- 7.8 g/d, P = 0.000), vitamin C (84.3 +/- 43.1 vs. 144.2 +/- 63.1 mg/d, P = 0.0001), and vitamin E (5.4 +/- 1.9 vs. 8.7 +/- 2.9 mg/d, P = 0.0001). The insulin sensitivity index (ISI) was significantly lower in NASH patients, and postprandial triglycerides were higher. Saturated fat intake correlated with ISI, metabolic syndrome features, and postprandial triglyceride increases.
Shows an association, not proof of cause and effect.
Musso G, Gambino R, De Michieli F, et al. Hepatology, 2003. PMID 12668986
2003 / Prospective cohort study
Researchers prospectively investigated the link between plasma cholesterol ester (CE) and phospholipid (PL) fatty acid composition and incident diabetes in 2,909 adults aged 45-64 years followed for 9 years. After adjusting for multiple confounders, diabetes incidence was positively associated with total saturated fatty acids in plasma CE and PL, with rate ratios across quintiles of 1.00, 1.36, 1.16, 1.60, and 2.08 (P = 0.0013) in CE and 1.00, 1.75, 1.87, 2.40, and 3.37 (P < 0.0001) in PL. In CE, diabetes was also positively associated with palmitic, palmitoleic, and dihomo-gamma-linolenic acids, and inversely associated with linoleic acid. In PL, diabetes incidence was positively linked to palmitic and stearic acids.
Shows an association, not proof of cause and effect.
Wang L, Folsom AR, Zheng ZJ, et al. Am J Clin Nutr, 2003. PMID 12816776
2003 / Ecological correlation study
Researchers analyzed dietary intake data from 4,701 adolescents aged 10 to 16 years across 11 European countries alongside national type 1 diabetes incidence rates. The incidence of type 1 diabetes correlated positively with intake of total fat (r = 0.674; p = 0.023), saturated fatty acids (r = 0.714; p = 0.047), and combined fruits and vegetables (r = 0.786; p = 0.036), though fruit or vegetable intake alone did not correlate. Cow's milk (r = 0.829; p = 0.042) and animal product consumption (r = 0.999; p = 0.001) correlated with incidence when Icelandic data were excluded, while sugar intake showed a borderline negative correlation (r = -0.721; p = 0.068).
Shows an association, not proof of cause and effect.
Thorsdottir I, Ramel A Ann Nutr Metab, 2003. PMID 14520022
2002 / Pretest/posttest study
Employees (n = 442) from six worksites self-selected into an 8-week facilitator-based video version of the Coronary Health Improvement Project (CHIP) evaluated via a pretest/posttest design. Demographic and biometric data were assessed at baseline and 8 weeks. All sites demonstrated significant and meaningful reductions in body weight, body mass index, total cholesterol, low-density lipoprotein cholesterol, triglycerides, and fasting blood glucose. Men demonstrated greater improvement than women, and those with higher baseline risks experienced the greatest reductions; the abstract notes future research is needed to determine persistence.
Shows an association, not proof of cause and effect.
Aldana SG, Greenlaw R, Diehl HA, et al. J Occup Environ Med, 2002. PMID 12227675
2001 / Prospective cohort study
Researchers followed 84,941 female nurses from 1980 to 1996 to examine the combined effects of diet and lifestyle factors on type 2 diabetes risk. Over 16 years, 3,300 new cases of type 2 diabetes were documented. Overweight or obesity was the single most important predictor, but lack of exercise, poor diet, smoking, and alcohol abstinence were also associated with significantly increased risk. Women adhering to all five low-risk factors (3.4% of the cohort) had a relative risk of diabetes of 0.09 (95% CI, 0.05 to 0.17) compared with the rest of the cohort.
Shows an association, not proof of cause and effect.
Hu FB, Manson JE, Stampfer MJ, et al. N Engl J Med, 2001. PMID 11556298