Observational studies included in the meta-analysis.

Abbreviations: LC, lung cancer; RR, relative risk; C-C, case control; Co, cohort; NR, not reported.
*1, age; 2, sex; 3, comorbidity score;4, other lipid-lowering therapy; 5, duration of follow-up; 6, history of neoplasia; 7, number of physician visits; 8, body mass index; 9, smoking status; 10, diabetes; 11, prior hospitalizations, 12, use of diuretics; 13, use of angiotensin-converting enzyme inhibitor; 14, use of calcium channel blockers; 15, use of nonsteroidal anti-inflammatory drugs; 16, hormone replacement therapy; 17, use of cardiovascular drugs; 18, alcohol use; 19, race; 20, education; 21, study center; 22, interview year; 23, inflammatory bowel disease; 24, benign mammary dysplasia; 25, arthritis; 26, use of gastroprotective drugs; 27, estrogen use; 28, obesity; 29; tobacco abuse; 30, mammography; 31, gynecologic examination; 32, Papanicolaou smear; 33, colonoscopy; 34, stool occult blood; 35, distinct generic medicines taken; 36, prior nursing home stay; 37, weight; 38, thyroid disease; 39, hypertension; 40 cardiovascular disease; 41, renal failure;42, chest pain; 43, mental illness; 44, alcoholism; 45, lung disease; 46, gastrointestinal disease; 47, prostate disease; 48, total cholesterol; 49, aspirin use ; 50,calendar year; 51,Townsend score, 52, any other cancer; 53,corticosteroids; 54, asthma; 55, Cox2-inhibitors; 56, physical activity; 57, history of elevated cholesterol; 58, tuberculosis.
†The risk estimate was calculated by post hoc analysis.