To examine the associations between a biomarker of overall passive exposure to tobacco smoke (serum cotinine concentration) and risk of coronary heart disease and stroke. Prospective population based study in general practice (the British regional heart study). 4729 men in 18 towns who provided baseline blood samples (for cotinine assay) and a detailed smoking history in 1978-80. Major coronary heart disease and stroke events (fatal and non-fatal) during 20 years of follow up. 2105 men who said they did not smoke and who had cotinine concentrations < 14.1 ng/ml were divided into four equal sized groups on the basis of cotinine concentrations. Relative hazards (95% confidence intervals) for coronary heart disease in the second (0.8-1.4 ng/ml), third (1.5-2.7 ng/ml), and fourth (2.8-14.0 ng/ml) quarters of cotinine concentration compared with the first (> or = 0.7 ng/ml) were 1.45 (1.01 to 2.08), 1.49 (1.03 to 2.14), and 1.57 (1.08 to 2.28), respectively, after adjustment for established risk factors for coronary heart disease. Hazard ratios (for cotinine 0.8-14.0 nu > or = 0.7 ng/ml) were particularly increased during the first (3.73, 1.32 to 10.58) and second five year follow up periods (1.95, 1.09 to 3.48) compared with later periods. There was no consistent association between cotinine concentration and risk of stroke. Studies based on reports of smoking in a partner alone seem to underestimate the risks of exposure to passive smoking. Further prospective studies relating biomarkers of passive smoking to risk of coronary heart disease are needed.
