Objective: To study the influence of different levels of serum (s)-triglycerides in relation to s-cholesterol on the risk of myocardial infarction. Design ann subjects: A 6-13 (mean 10) year follow-up of 12510 middle-aged men. Fasting s-triglycerides and s-cholesterol were measured at the screening examination. Setting: Section of Preventive Medicine at the Department of Internal Medicine, Malmo General Hospital, an urban hospital for 240000 inhabitants in southern Sweden. Intervention: In minor groups of patients there were interventions addressing high lipid levels, high alcohol consumption, hypertension and glucose intolerance. Main outcome measure: Myocardial infarction was used as an end-point. Results: 446 myocardial infarctions occurred. The cumulative incidence rates were for the lowest triglyceride quartile 1.2%, for the second 3.2%, for the third 4.1% and for the highest 5.6%. After adjustment for age, year of screening, body mass index, diabetes, smoking, hypertension and s-cholesterol there was a significant relationship between triglycerides and the relative risk for myocardial infarction (P for trend = 0.0087). For increasing levels of triglycerides, adjusted for the above factors except cholesterol, the impact of a certain cholesterol value for the occurrence of myocardial infarction was increased (P for trend = 0.0092). Conclusions: The present study emphasizes the interaction between cholesterol and triglyceride values for the risk of myocardial infarction. It is concluded that at triglyceride values above 1.0 mmol/l and cholesterol above 6.8 mmol/l there is an increasing interaction between cholesterol and triglyceride levels that might be of importance when evaluating the cardiovascular risk of middle aged men. (C) 1999 Elsevier Science Ireland Ltd. All rights reserved.