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.
2019 / Prospective cohort study
Researchers tracked 106,966 postmenopausal women from the Women's Health Initiative to examine links between fried food consumption and mortality, observing 31,558 deaths over follow-up. Consuming at least one serving per day of total fried food compared with none had an adjusted hazard ratio of 1.08 (95% CI 1.01 to 1.16) for all-cause mortality and 1.08 (0.96 to 1.22) for cardiovascular mortality. At least one serving per week of fried chicken was associated with all-cause (HR 1.13, 95% CI 1.07 to 1.19) and cardiovascular mortality (HR 1.12, 95% CI 1.02 to 1.23); fried fish/shellfish had HRs of 1.07 (1.03 to 1.12) and 1.13 (1.04 to 1.22), respectively. Total or individual fried food consumption was not generally associated with cancer mortality.
Shows an association, not proof of cause and effect.
Sun Y, Liu B, Snetselaar LG, et al. BMJ, 2019. PMID 30674467
2019 / Cohort / biomonitoring study
Internal doses of the foodborne genotoxic carcinogen glycidol were determined in 50 children and 12 adults by measuring the hemoglobin adduct N-(2,3-dihydroxypropyl)valine. Hemoglobin adduct levels demonstrated a fivefold variation among the children, with an estimated mean intake of 1.4 μg/kg/day, which was about two times higher than the European Food Safety Authority's estimate for children. Non-smoking adults showed exposure levels similar to children, while smoking adults had higher levels. Using a multiplicative risk model, the estimated lifetime cancer risk calculated from the children's in vivo doses was 200/10⁵, exceeding what is considered acceptable.
Shows an association, not proof of cause and effect.
Aasa J, Vryonidis E, Abramsson-Zetterberg L, et al. Toxics, 2019. PMID 30717263
2019 / Cross-sectional study
Researchers analyzed blood specimens from 95 self-selected female subjects using liquid PCR, Sanger sequencing, and ELISA to detect bovine leukemia virus (BLV) DNA and antibodies. BLV DNA was identified in buffy coat cells of 33/95 (38%) subjects. Plasma antibodies were detected in 30/95 (32%) for IgG, 55/95 (58%) for IgM, and 30/95 (32%) for IgA. No significant correlation was observed between the presence of antibodies and the presence of BLV DNA.
Shows an association, not proof of cause and effect.
Buehring GC, DeLaney A, Shen H, et al. BMC Infect Dis, 2019. PMID 30940091
2019 / Prospective cohort study
In an analysis of 410,701 NIH-AARP participants followed for 15.6 years, 76,921 deaths occurred. In fully adjusted models, consuming potatoes seven or more times per week was not associated with overall mortality (HR C4 vs C1 = 1.02, 95% CI: 0.97, 1.06), nor with mortality from cancer, heart disease, respiratory disease, or diabetes. The only statistically significant association was between French fry consumption and cancer-related mortality (HR C4 vs C1 = 1.27, 95% CI: 1.02, 1.59), though uncontrolled confounding could not be ruled out.
Shows an association, not proof of cause and effect.
Hashemian M, Murphy G, Etemadi A, et al. PLoS One, 2019. PMID 31063480
2019 / Prospective cohort study
This prospective study followed 68,327 women and 44,664 men for up to 26 years, documenting 22,469 incident cancer cases. In pooled multivariable analysis, participants consuming five or more servings of mushrooms per week had no significantly different risk of total cancer compared to those who almost never consumed mushrooms (HR, 1.06; 95% CI, 0.98-1.14). Mushroom consumption was not associated with 16 site-specific cancers, although a marginal positive association was observed for lung cancer (P trend = 0.05).
Shows an association, not proof of cause and effect.
Lee DH, Yang M, Keum N, et al. Cancer Prev Res (Phila), 2019. PMID 31164344
2019 / Prospective cohort study
The NIH-AARP Diet and Health Study evaluated the intake of nuts and peanut butter using a food-frequency questionnaire among individuals aged 50-71 years. Over a median follow-up of 15.5 years, 64,464 deaths occurred. Nut consumption showed a significant inverse association with overall mortality (HR C4 vs C1 = 0.78, 95% CI = 0.76, 0.81, p ≤ 0.001) and was associated with reduced mortality from cancer, cardiovascular, respiratory, infectious, renal, and liver diseases, but not diabetes or Alzheimer's disease. Peanut butter intake was not significantly associated with all-cause mortality (HR C4 vs C1 = 1.00, 95% CI = 0.98, 1.04, p = 0.001) or cause-specific mortality.
Shows an association, not proof of cause and effect.
Amba V, Murphy G, Etemadi A, et al. Nutrients, 2019. PMID 31269682
2019 / Prospective cohort study
Researchers evaluated 12-year changes (1986–1998) in 3 plant-based diet indices (overall PDI, healthful PDI, and unhealthful PDI) and subsequent mortality (1998–2014) in NHS and HPFS participants. Compared with stable scores, the greatest increases in scores had pooled multivariable-adjusted HRs for total mortality of 0.95 (95% CI, 0.90-1.00) for PDI, 0.90 (95% CI, 0.85-0.95) for healthful PDI, and 1.12 (95% CI, 1.07-1.18) for unhealthful PDI. A 10-point increase in score was associated with CVD mortality reductions of 7% (95% CI, 1-12%) for PDI and 9% (95% CI, 4-14%) for healthful PDI, but an 8% higher risk (95% CI, 2-14%) for unhealthful PDI. No consistent associations were found between diet changes and cancer mortality.
Shows an association, not proof of cause and effect.
Baden MY, Liu G, Satija A, et al. Circulation, 2019. PMID 31401846
2019 / In silico and laboratory genomic analysis
Researchers analyzed 109 replication-competent genomic DNA sequences isolated from cow milk and dairy products (97 in the BMMF2 group and 4 in the BMMF1 group). In silico analyses placed these agents between bacterial plasmids and circular ssDNA viruses. A member of the BMMF1 group, H1MBS.1, was recovered in a distinctly modified form from human colon tissue via laser microdissection. The authors suggest this highlights a previously unrecognized, broad spectrum of potential pathogens.
Shows an association, not proof of cause and effect.
de Villiers EM, Gunst K, Chakraborty D, et al. Emerg Microbes Infect, 2019. PMID 31409221
2019 / Prospective cohort study
Researchers assessed lifestyle factors in 1072 stage I-III colorectal cancer survivors at diagnosis, 6 months, and 2 years post-diagnosis using a 2018 WCRF/AICR lifestyle score (range 0-7). Over 2 years, the largest dietary changes were reductions in sugary drinks (-45 g/day) and red and processed meat (-62 g/week), while BMI (+0.4 kg/m2) and waist circumference (+2 cm) increased slightly, and fiber intake decreased (-1 g/day). No statistically significant changes occurred for fruit and vegetable, alcohol, or ultra-processed food intake, nor for physical activity or smoking. Half of participants made simultaneous changes that improved one component while worsening another, with overall lifestyle scores shifting from 3.4 ± 0.9 at diagnosis to 3.5 ± 0.9 at 2 years.
Shows an association, not proof of cause and effect.
van Zutphen M, Boshuizen HC, Kok DE, et al. J Cancer Surviv, 2019. PMID 31646463
2019 / Prospective cohort study
In the Japan Public Health Center-based Prospective Cohort of 70,696 participants followed through 2016 (12,381 deaths), researchers examined associations between protein intake and mortality. Animal protein intake showed no clear link with total or cause-specific mortality. Higher plant protein intake was associated with lower total mortality (HR 0.87 [95% CI, 0.78-0.96] for quintile 5 vs 1; P for trend = .01) and lower CVD-related mortality (HRs ranging from 0.84 to 0.70; P for trend = .002). Isocaloric substitution of 3% energy from plant protein for red meat protein was linked to lower total, cancer-related, and CVD-related mortality.
Shows an association, not proof of cause and effect.
Budhathoki S, Sawada N, Iwasaki M, et al. JAMA Intern Med, 2019. PMID 31682257
2019 / Three prospective cohort studies
Researchers analyzed 168,153 women and 49,602 men followed for up to 32 years, documenting 51,438 total deaths (12,143 cardiovascular and 15,120 cancer deaths). Compared with the lowest intake (average 0.8 servings/day), the highest total dairy consumption (average 4.2 servings/day) was associated with a higher risk of total mortality (HR 1.07, 95% CI 1.04 to 1.10). Whole milk was significantly associated with higher risks of total (HR per 0.5 additional serving/day 1.11), cardiovascular (HR 1.09), and cancer mortality (HR 1.11). In substitution analyses, substituting nuts, legumes, or whole grains for dairy was associated with lower mortality, whereas replacing dairy with red and processed meat was associated with higher mortality.
Shows an association, not proof of cause and effect.
Ding M, Li J, Qi L, et al. BMJ, 2019. PMID 31776125
2018 / Pilot observational study
This pilot study investigated optimistic bias in risk perception and health-related behaviors for cancer, respiratory disorders, and cardiovascular diseases in a sample of young adults. The participants showed an optimistic bias related to cancer and cardiovascular diseases. The findings suggest that optimistic bias is linked to specific behavioral patterns widespread in young adults, such as cigarette smoking and alcohol consumption.
Shows an association, not proof of cause and effect.
Masiero M, Riva S, Oliveri S, et al. J Health Psychol, 2018. PMID 27624614
2018 / Retrospective cohort study
A retrospective chart review of 523 metastatic cancer patients examined serum vitamin B12 as a prognostic factor. Median survival time was 1.8 months in the high B12 group (>911 pg/mL) compared to 5.1 months in the normal B12 group (211-911 pg/mL) (p < 0.001). Among patients without liver lesions, median survival was 2.1 months in the high group versus 6.1 months in the normal group (p < 0.001). Multivariate analysis identified serum vitamin B12 as an independent prognostic factor for overall survival (HR: 1.62; 95% CI: 1.34-1.96, p < 0.001).
Shows an association, not proof of cause and effect.
Oh HK, Lee JY, Eo WK, et al. Nutr Cancer, 2018. PMID 29220583
2018 / Case-control study
Researchers examined archival breast tissue sections from 216 women to evaluate the presence of bovine leukemia virus (BLV) and human papillomavirus (HPV) DNA using in situ PCR and/or DNA hybridization. Women diagnosed with breast cancer, as well as women with premalignant breast pathology and no history of breast cancer, were significantly more likely to have BLV DNA in their breast tissue compared with women with benign diagnoses. HPV status was not associated with malignancy, premalignant breast disease, or the presence of BLV DNA.
Shows an association, not proof of cause and effect.
Baltzell KA, Shen HM, Krishnamurthy S, et al. Cancer, 2018. PMID 29266207
2018 / Systematic analysis of population data (GBD study)
Using Global Burden of Disease Study data from 1990 to 2016, researchers estimated disease burden, mortality, and risk factors across US states. Between 1990 and 2016, overall US death rates dropped from 745.2 to 578.0 per 100,000 persons, with persistent state-level disparities in life expectancy. In 2016, six risk factors individually accounted for more than 5% of risk-attributable DALYs: tobacco consumption, high BMI, poor diet, alcohol and drug use, high fasting plasma glucose, and high blood pressure. Top risk factors for attributable DALYs across states were tobacco (32 states), high BMI (10 states), or alcohol and drug use (8 states).
Shows an association, not proof of cause and effect.
US Burden of Disease Collaborators, Mokdad AH, Ballestros K, et al. JAMA, 2018. PMID 29634829
2018 / Prospective cohort study with life table modeling
Researchers evaluated 5 low-risk lifestyle factors (never smoking, BMI 18.5-24.9, ≥30 min/d moderate-to-vigorous physical activity, moderate alcohol, and high diet quality) using NHS and HPFS cohorts with up to 34 years of follow-up (42,167 deaths). Comparing adults with 5 versus zero low-risk factors, multivariable-adjusted hazard ratios were 0.26 (95% CI, 0.22-0.31) for all-cause, 0.35 (95% CI, 0.27-0.45) for cancer, and 0.18 (95% CI, 0.12-0.26) for cardiovascular mortality. Projected life expectancy at age 50 was 14.0 years longer for women (43.1 vs 29.0 years) and 12.2 years longer for men (37.6 vs 25.5 years) with all 5 low-risk factors compared to zero.
Shows an association, not proof of cause and effect.
Li Y, Pan A, Wang DD, et al. Circulation, 2018. PMID 29712712
2018 / Case report
A 67-year-old male with lung cancer receiving paclitaxel/carboplatin presented with asthenia, anorexia, jaundice, and choluria after self-medicating with dietary products including Chlorella (520 mg/day) and Curcuma longa (15 g/day). Analysis revealed the Chlorella supplement was contaminated with cyanobacteria containing 1.08 μg of Microcystin-LR (MC-LR) per gram, leading to consumption of approximately 0.01 μg MC-LR/kg/day. Toxic events began only after starting the dietary supplements, representing a case of acute toxic hepatitis likely related to pharmacokinetic interaction with turmeric and contaminated Chlorella.
Shows an association, not proof of cause and effect.
Costa ML, Rodrigues JA, Azevedo J, et al. Toxicon, 2018. PMID 29857089
2018 / Pooled cohort analysis
Researchers analyzed pooled data from two randomized clinical trials (N = 1,129 and N = 2,196) and a prospective cohort (N = 1,713) to examine the relationship between serum 25(OH)D levels and breast cancer risk over a median of 4.0 years. Across the pooled cohort of 5,038 women, 77 breast cancer cases were diagnosed. Women with 25(OH)D concentrations ≥60 ng/ml had an 82% lower incidence rate compared to those with <20 ng/ml (RR = 0.18, P = 0.006). In multivariate Cox regression adjusting for age, BMI, smoking, calcium supplement intake, and study of origin, concentrations ≥60 ng/ml were associated with an 80% lower risk of breast cancer versus <20 ng/ml (HR = 0.20, P = 0.03).
Shows an association, not proof of cause and effect.
McDonnell SL, Baggerly CA, French CB, et al. PLoS One, 2018. PMID 29906273
2018 / Nested case-control and prospective biomarker evaluation study
Researchers evaluated antibody levels against red meat-derived Neu5Gc and Neu5Gc-glycans in samples from the Nurses' Health Studies, HPFS, and EPIC. ELISA assays showed reproducible antibody responses, but levels did not correlate with coronary artery disease risk, red meat intake, or colorectal cancer (CRC) risk for individual epitopes. However, when total antibody responses against all Neu5Gc-glycans were combined via a sialoglycan microarray, individuals in the top quartile of total anti-Neu5Gc IgG concentrations had an increased risk of CRC compared to the bottom quartile (multivariate OR: 2.98, 95% CI: 0.80, 11.1; P for trend = 0.02).
Shows an association, not proof of cause and effect.
Samraj AN, Bertrand KA, Luben R, et al. PLoS One, 2018. PMID 29912879
2018 / Prospective cohort study
Using baseline and genetic data from 498,134 UK Biobank participants followed for 10 years (14,225 deaths), researchers evaluated associations between coffee intake and mortality, including modification by genetic caffeine metabolism scores. Compared to non-drinkers, HRs for all-cause mortality were 0.94 for <1 cup/day, 0.92 for 1 cup/day, 0.88 for 2-3 cups/day, 0.88 for 4-5 cups/day, 0.84 for 6-7 cups/day, and 0.86 for 8+ cups/day. Similar inverse associations were observed for instant, ground, and decaffeinated coffee, as well as common causes of death. These relationships persisted across caffeine metabolism genetic scores without evidence of effect modification (P = .17 for heterogeneity).
Shows an association, not proof of cause and effect.
Loftfield E, Cornelis MC, Caporaso N, et al. JAMA Intern Med, 2018. PMID 29971434
2018 / Multi-centre case-control study
A multi-centre case-control study in Punjab and Haryana evaluated dietary patterns among 400 women with newly diagnosed breast cancer and 354 matched controls. Participants were categorized as non-vegetarians, lacto-vegetarians, or lacto-ovo-vegetarians, with 62% identified as lacto-ovo-vegetarians. Breast cancer risk was lower in lacto-ovo-vegetarians compared to non-vegetarians (OR 0.6, 95% CI 0.3-0.9) and lacto-vegetarians (OR 0.4, 95% CI 0.3-0.7). The authors note that the unexpected difference between lacto-ovo-vegetarian and lacto-vegetarian patterns could be due to egg-consumption patterns and requires further confirmation.
Shows an association, not proof of cause and effect.
Shridhar K, Singh G, Dey S, et al. Int J Environ Res Public Health, 2018. PMID 30200632
2018 / Population-based prospective cohort study
In the NutriNet-Santé prospective cohort study followed from May 10, 2009, to November 30, 2016, researchers evaluated the link between an organic food score (0-32 points across 16 products) and cancer risk among 68,946 participants. During follow-up, 1,340 first incident cancer cases occurred. High organic food scores were inversely associated with overall cancer risk (HR for quartile 4 vs quartile 1, 0.75; 95% CI, 0.63-0.88; P for trend = .001; absolute risk reduction, 0.6%; HR for a 5-point increase, 0.92; 95% CI, 0.88-0.96).
Shows an association, not proof of cause and effect.
Baudry J, Assmann KE, Touvier M, et al. JAMA Intern Med, 2018. PMID 30422212
2017 / Prospective cohort study
Using data from the EPIC study, researchers evaluated whether dietary fiber intake modified the association between alcohol consumption and breast cancer risk among 334,850 women followed for an average of 11.0 years. Over 3,670,439 person-years, 11,576 incident breast cancer cases were diagnosed. For women with low fiber intake (<18.5 g/day), the hazard ratio of breast cancer per 10 g/day of alcohol intake was 1.06 (1.03-1.08), compared to 1.02 (0.99-1.05) among those with high fiber intake (>24.2 g/day; test for interaction p = 0.011). The modulating effect was stronger for fiber from vegetables, and women with high alcohol and low fiber intake had the highest breast cancer risk.
Shows an association, not proof of cause and effect.
Romieu I, Ferrari P, Chajès V, et al. Int J Cancer, 2017. PMID 27599758
2017 / Mendelian randomization study
Researchers conducted a Mendelian randomization study using two caffeine-associated genetic variants (rs4410790 and rs2472297) as proxies for coffee intake in 46,687 men. A combined genetic risk score showed no clear evidence of association with prostate cancer risk (OR: 1.01, 95% CI: 0.98, 1.03) or high-grade versus low-grade disease (OR: 1.01, 95% CI: 0.97, 1.04). However, there was some evidence linking the score to higher odds of nonlocalized versus localized disease (OR: 1.03, 95% CI: 1.01, 1.06). Among men with prostate cancer, the genetic score showed no clear association with all-cause mortality (HR: 1.00, 95% CI: 0.97, 1.04) or prostate cancer-specific mortality (HR: 1.03, 95% CI: 0.98, 1.08).
Shows an association, not proof of cause and effect.
Taylor AE, Martin RM, Geybels MS, et al. Int J Cancer, 2017. PMID 27741566