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.
2006 / Cross-sectional survey
Using the 2004 Behavioral Risk Factor Surveillance System telephone survey, 299,658 adults reported their past-month alcohol use and past-year sunburn history. Approximately 33.5% of respondents reported a sunburn in the past year. Heavier average alcohol use and binge drinking were positively associated with sunburn prevalence and frequency, with adjusted odds ratios among binge drinkers of 1.39 (95% CI 1.31-1.48) for prevalence and 1.29 (95% CI 1.20-1.38) for number of sunburns. Limitations included the cross-sectional design and reliance on participant self-report.
Shows an association, not proof of cause and effect.
Mukamal KJ J Am Acad Dermatol, 2006. PMID 17010736
2006 / Population-based case-control study
A population-based case-control study assessed marijuana use and the risk of lung and upper aerodigestive tract cancers among 1,212 cases and 1,040 matched controls. Although >=30 joint-years was positively linked to each cancer type except pharyngeal cancer in crude analyses, no positive associations were observed after adjusting for confounders including cigarette smoking. Adjusted odds ratios (and 95% confidence limits) for >=60 versus 0 joint-years were 1.1 (0.56, 2.1) for oral, 0.84 (0.28, 2.5) for laryngeal, and 0.62 (0.32, 1.2) for lung cancer; for >=30 versus 0 joint-years, adjusted odds ratios were 0.57 (0.20, 1.6) for pharyngeal and 0.53 (0.22, 1.3) for esophageal cancer. The authors noted potential selection bias or exposure measurement errors, concluding results suggest the association is not strong and may be below practically detectable limits.
Shows an association, not proof of cause and effect.
Hashibe M, Morgenstern H, Cui Y, et al. Cancer Epidemiol Biomarkers Prev, 2006. PMID 17035389
2006 / Retrospective study combined with in vitro experiments
Data from 60 patients with head, neck, or pelvic cancer topically treated with green or black tea extracts were analyzed retrospectively alongside cellular assays. Tea extracts supported the restitution of skin integrity in patients. In vitro, the extracts inhibited 26S proteasome function, suppressed release of IL-1beta, IL-6, IL-8, TNFalpha, and PGE2, altered NF-kappaB activity in a caspase-dependent manner, and slightly protected macrophages from ionizing radiation.
Shows an association, not proof of cause and effect.
Pajonk F, Riedisser A, Henke M, et al. BMC Med, 2006. PMID 17140430
2006 / Cross-sectional comparative study
Researchers compared plasma growth factors and hormones among 21 sedentary adults consuming a low-protein, low-calorie diet for 4.4 +/- 2.8 y, 21 BMI-matched endurance runners, and 21 sedentary Western diet controls. Both the low-protein, low-calorie group and runners showed lower insulin, free sex hormones, leptin, and C-reactive protein, and higher sex hormone-binding globulin compared to the Western diet group (all P < 0.05). Plasma IGF-I and the IGF-I to IGFBP-3 ratio were lower in the low-protein, low-calorie group (139 +/- 37 ng/mL and 0.033 +/- 0.01) than in runners (177 +/- 37 ng/mL and 0.044 +/- 0.01) and Western diet controls (201 +/- 42 ng/mL and 0.046 +/- 0.01; P < 0.005).
Shows an association, not proof of cause and effect.
Fontana L, Klein S, Holloszy JO Am J Clin Nutr, 2006. PMID 17158430
2006 / Cross-sectional dietary assessment and database development
Researchers constructed a database of 6 flavonoid classes and lignans for 50 food group items using literature and USDA data to enhance intake estimates. They applied this database to responses from a modified Block FFQ completed in 1996–1997 by 1,500 women without breast cancer on Long Island, New York. Total flavonoid and lignan content ranged from 0 to 129 mg/100 g of food, with tea, cherries, and grapefruit identified as the richest sources. Estimated intakes from highest to lowest were flavan-3-ols, flavanones, flavonols, lignans, isoflavones, anthocyanidins, and flavones, with each class exhibiting a wide range of intake levels.
Shows an association, not proof of cause and effect.
Fink BN, Steck SE, Wolff MS, et al. Nutr Cancer, 2006. PMID 17176218
2005 / Prospective cohort study
Researchers evaluated associations between glyphosate exposure and cancer incidence among 57,311 licensed pesticide applicators in the Agricultural Health Study. Glyphosate exposure was assessed as ever used, cumulative lifetime exposure days, and intensity-weighted cumulative exposure days. Glyphosate exposure was not associated with overall cancer incidence or with most of the cancer subtypes evaluated. However, there was a suggested association with multiple myeloma incidence that warrants follow-up as more cases accrue.
Shows an association, not proof of cause and effect.
De Roos AJ, Blair A, Rusiecki JA, et al. Environ Health Perspect, 2005. PMID 15626647
2005 / Prospective cohort study (50-year follow-up)
A total of 34,439 male British doctors reporting smoking habits in November 1951 were followed for 50 years until November 2001, death, emigration, or censoring. Eleven of the 13 cancer types classed as liable to be caused by smoking were significantly related to smoking, with findings for the other two suggesting they might be. Among 13 types with previously sparse or inconsistent data, the 638 combined deaths were clearly unrelated to smoking. Findings tended to support no relationship for prostate cancer, while colorectal cancer showed a weak relationship that could not be attributed to alcohol confounding in a small subset.
Shows an association, not proof of cause and effect.
Doll R, Peto R, Boreham J, et al. Br J Cancer, 2005. PMID 15668706
2005 / Observational cohort study
In Norway, late-summer serum calcidiol concentrations are roughly 50% higher than in late winter due to solar radiation. Researchers evaluated whether season of diagnosis influenced the 18-month survival rate of colon cancer in men and women. The 18-month survival rate was dependent on the season of diagnosis, with higher calcidiol concentrations at diagnosis appearing to correlate with an increased survival rate. The authors propose calcidiol or calcitriol may act positively alongside conventional colon cancer therapies.
Shows an association, not proof of cause and effect.
Moan J, Porojnicu AC, Robsahm TE, et al. J Photochem Photobiol B, 2005. PMID 15708515
2005 / Cross-sectional national telephone survey
A US national telephone survey of 6,369 individuals, including 1,245 current smokers, evaluated risk perceptions and beliefs about smoking harms. Smokers underestimated their relative risk compared to non-smokers and believed they possessed a lower risk of lung cancer than the average smoker. Perceived risk of cancer barely increased with daily cigarettes smoked, and over half agreed that exercise undoes most smoking effects.
Shows an association, not proof of cause and effect.
Weinstein ND, Marcus SE, Moser RP Tob Control, 2005. PMID 15735301
2005 / Case report and hypothesis
A 3-year-old girl with acute lymphocytic leukemia and low blood zinc received zinc adjuvant (3.18 mg/kg body weight/day; 18 mg at breakfast and 18 mg at supper) from the start of chemotherapy through 3 years of maintenance therapy. Bone marrow blast cells decreased from 95+% to an observed zero count in both hips within 14 days, without relapse, accompanied by normalized hemopoietic functions and no observed side effects. The authors noted identical results occurred in 13 other children whose parents chose zinc adjuvant, and they hypothesize zinc may strengthen chemotherapy and immune function.
Shows an association, not proof of cause and effect.
Eby GA Med Hypotheses, 2005. PMID 15823699
2005 / Hospital-based case-control study
Researchers conducted a hospital-based case-control study of 128 cases and 238 non-cancer controls to evaluate the relationship between colorectal cancer, the CD36 gene A52C polymorphism, and meat consumption. Relative to the A/A genotype with low meat consumption, the odds ratio (OR) for the C/C genotype with moderate meat consumption was 8.30 (95% CI=2.15-32.00); no individuals with the C/C genotype were in the high meat consumption group. In the low vegetables/fruit consumption group, the OR for the C/C genotype versus the A/A genotype was 3.03 (95% CI=1.12-7.90).
Shows an association, not proof of cause and effect.
Kuriki K, Hamajima N, Chiba H, et al. Nutr Cancer, 2005. PMID 15860439
2005 / Cross-sectional study
Fasting serum concentrations of angiogenic factors and endostatin were evaluated in 58 lean and 42 overweight and obese participants. VEGF, VEGF-C, VEGF-D, soluble VEGF receptor-2, angiopoietin-2, angiogenin, and endostatin (but not HGF) were significantly elevated in overweight and obese subjects. Positive correlations with BMI remained significant for VEGF-C, VEGF-D, sVEGF-R2, angiopoietin-2, angiogenin, and endostatin after adjusting for gender and age.
Shows an association, not proof of cause and effect.
Silha JV, Krsek M, Sucharda P, et al. Int J Obes (Lond), 2005. PMID 15953938
2005 / Case report
A 68-year-old cancer patient presented with a reduced Glasgow Coma Score, seizures, and severe lactic acidosis requiring intubation shortly after taking her first 3 g dose of amygdalin alongside 4800 mg of vitamin C per day. She responded rapidly to hydroxocobalamin treatment, and the adverse reaction was rated probable on the Naranjo scale. An interaction where vitamin C increases the conversion of amygdalin to cyanide and depletes cysteine stores was presented as a plausible explanation.
Shows an association, not proof of cause and effect.
Bromley J, Hughes BG, Leong DC, et al. Ann Pharmacother, 2005. PMID 16014371
2005 / Ecological correlation study
Researchers correlated cancer incidence rates (1993-1997) from 40 countries with food intake data (1961-1997). Meat intake correlated most strongly with breast cancer incidence (r=0.827; SMRA [R]=0.862), followed by milk (0.817) and cheese (0.751). Milk was most closely correlated with ovarian cancer (r=0.779) and corpus uteri cancer (r=0.814). In stepwise multiple-regression analysis, milk plus cheese made the greatest contribution to ovarian cancer ([R]=0.767) and corpus uteri cancer ([R]=0.861).
Shows an association, not proof of cause and effect.
Ganmaa D, Sato A Med Hypotheses, 2005. PMID 16125328
2005 / Ecologic analyses and economic modeling
Ecologic analyses and published literature estimates were used to model the health and economic burdens of insufficient UVB/vitamin D versus excess UV irradiance in the United States. An estimated 50,000-63,000 people in the US and 19,000-25,000 in the UK die prematurely from cancer each year due to insufficient vitamin D. In 2004, the estimated US economic burden of vitamin D insufficiency was $40-56 billion, while the burden for excess UV irradiance was $6-7 billion. The authors state that further research is required to confirm these estimates.
Shows an association, not proof of cause and effect.
Grant WB, Garland CF, Holick MF Photochem Photobiol, 2005. PMID 16159309
2005 / Case report
A 32-year-old woman with metastatic breast cancer presented unresponsive, in shock, hypothermic, tachycardic, and with fixed dilated pupils after taking Internet-purchased 'vitamin supplements'. Over the next 6 hours, she required mechanical ventilation and inotropes, and developed diabetes insipidus. Her medications included 'vitamin B 17' (amygdalin), and her serum thiocyanate level was markedly elevated. She recovered fully over 8 hours with supportive therapy alone.
Shows an association, not proof of cause and effect.
O'Brien B, Quigg C, Leong T Eur J Emerg Med, 2005. PMID 16175068
2005 / Prospective cohort study
Researchers analyzed data from 40,559 postmenopausal women who completed a 61-item food-frequency questionnaire between 1987 and 1998, with 1,868 breast cancer cases identified. Principal components analysis identified three dietary patterns: vegetable-fish/poultry-fruit, beef/pork-starch, and traditional southern. After confounder adjustment, the traditional southern pattern was significantly associated with a reduced risk of invasive breast cancer (relative hazard = 0.78; 95% CI = 0.65, 0.95; P for trend = 0.003), whereas the other two patterns showed no significant association. Specific foods within the traditional southern pattern linked to lower risk were legumes, low mayonnaise-salad dressing intake, and possibly cabbage.
Shows an association, not proof of cause and effect.
Velie EM, Schairer C, Flood A, et al. Am J Clin Nutr, 2005. PMID 16332665
2004 / Population-based prospective cohort study
A population-based prospective study followed 18,684 men and 20,381 women aged 40-59 years from 1990 to 2001, identifying 486 histologically confirmed gastric cancer cases (358 men, 128 women). Salt intake quintiles showed a dose-dependent association with gastric cancer risk in men after adjusting for confounders (P for trend <0.001), but not in women (P for trend = 0.48). Highly salted foods like salted fish roe and salted fish preserves were strongly associated with risk in both sexes, though stratifying by study area attenuated associations with salt and salted foods.
Shows an association, not proof of cause and effect.
Tsugane S, Sasazuki S, Kobayashi M, et al. Br J Cancer, 2004. PMID 14710219
2004 / Retrospective cohort study
A retrospective cohort study evaluated 7,303 Swedish women hospitalized for anorexia nervosa prior to age 40 years (excluding 31 diagnosed with cancer prior to first discharge) using national registry data from 1965 to 1998. Compared with the Swedish general population, women hospitalized for anorexia nervosa had a 53% (95% CI, 3%-81%) lower incidence of breast cancer. Nulliparous women with anorexia nervosa had a 23% lower incidence (95% CI, 79% higher to 75% lower), while parous women with anorexia nervosa had a 76% (95% CI, 13%-97%) lower incidence.
Shows an association, not proof of cause and effect.
Michels KB, Ekbom A JAMA, 2004. PMID 15010444
2004 / Prospective cohort study
Researchers prospectively examined serum iron, transferrin saturation (TS), copper, and zinc levels and cancer mortality among 6,244 US adults from NHANES II. Over 83,664.4 person-years of follow-up, 307 cancer deaths occurred. In multivariable-adjusted models comparing the highest (>90th percentile) to the reference level (10th-50th percentile), adjusted relative risks (RRs) were 1.86 (95% CI: 1.07-3.22) for iron, 1.82 (95% CI: 1.10-3.02) for TS, 1.89 (95% CI: 1.07-3.32) for copper, and 0.84 (95% CI: 0.53-1.33) for zinc. Analyzed continuously, each 100 microg/dl increase in iron (RR 1.66; 95% CI: 1.03-2.68), 10% increase in TS (RR 1.17; 95% CI: 1.01-1.36), and 100 microg/dl increase in copper (RR 1.98; 95% CI: 1.12-3.50) were associated with increased cancer mortality risk, while zinc was not significantly linked (RR 0.57; 95% CI: 0.16-1.96).
Shows an association, not proof of cause and effect.
Wu T, Sempos CT, Freudenheim JL, et al. Ann Epidemiol, 2004. PMID 15036223
2004 / Multicountry ecologic study
Researchers conducted an ecologic study using linear and multiple linear regression analyses across 32 predominantly Caucasian countries to examine associations between prostate cancer mortality, dietary supply data, and solar UV-B radiation. The strongest risk factor for prostate cancer mortality was animal products, with the nonfat portion of milk and alcohol being somewhat weaker. The strongest risk reduction factors were onions, other protective vegetable products, and solar UV-B radiation, with dietary supply data from 1979-81 yielding the highest correlations.
Shows an association, not proof of cause and effect.
Grant WB Eur Urol, 2004. PMID 15036670
2004 / Case-control study
Researchers compared the percentage of equol producers between prostate cancer patients and controls in Japan, Korea, and the United States. The percentage of equol producers among patients versus controls was 29% and 46% in Japan (P = 0.004) and 30% and 59% in Korea (P = 0.001). Among Americans, active isoflavone levels were markedly lower and the percentage of equol producers was lower (17% for patients and 14% for controls) compared to Japanese and Korean participants.
Shows an association, not proof of cause and effect.
Akaza H, Miyanaga N, Takashima N, et al. Jpn J Clin Oncol, 2004. PMID 15067102
2004 / Prospective cohort study
Researchers prospectively assessed ejaculation frequency across life periods in 29,342 men from 1992 to 2000. During 222,426 person-years of follow-up, 1,449 new cases of total prostate cancer were documented. Compared to 4-7 ejaculations per month, reporting 21 or more ejaculations monthly was associated with multivariate relative risks for total prostate cancer of 0.89 (95% CI, 0.73-1.10) for ages 20-29, 0.68 (95% CI, 0.53-0.86) for ages 40-49, 0.49 (95% CI, 0.27-0.88) for the previous year, and 0.67 (95% CI, 0.51-0.89) averaged across a lifetime. Similar associations were found for organ-confined prostate cancer, but ejaculation frequency was not statistically significantly associated with advanced prostate cancer.
Shows an association, not proof of cause and effect.
Leitzmann MF, Platz EA, Stampfer MJ, et al. JAMA, 2004. PMID 15069045
2004 / Prospective cohort study
Researchers analyzed 7,674 participants aged 30 to 74 from NHANES II (1976-1980) followed for mortality through 1992. After adjusting for multiple factors including smoking, physical activity, BMI, alcohol, and diabetes, higher WBC was significantly associated with total cancer mortality (highest vs. lowest quartile RR 1.66; 95% CI, 1.08-2.56; P trend = 0.03). Stratified analyses showed an association with non-lung cancer (P trend = 0.04) but not lung cancer (P trend = 0.18). Among never smokers, a 1 SD increase in WBC (2.2 x 10^9 cells/L) was associated with higher total (RR 1.32; 95% CI, 1.05-1.67) and non-lung cancer mortality (RR 1.30; 95% CI, 1.03-1.63).
Shows an association, not proof of cause and effect.
Erlinger TP, Muntner P, Helzlsouer KJ Cancer Epidemiol Biomarkers Prev, 2004. PMID 15184263