To examine gender-specific associations between sleep duration and sleep complaints and incident myocardial infarction (MI). DESIGN: Cohort study. SETTING: A representative population sample of middle-aged subjects in Germany. PARTICIPANTS: The study was based on 3508 men and 3388 women (aged 45 to 74 years) who participated in one of the 3 MONICA (Monitoring trends and determinants on cardiovascular diseases) Augsburg surveys between 1984 and 1995, who were free of MI and angina pectoris at baseline and were followed up until 2002. INTERVENTIONS: N/A. MEASUREMENTS AND RESULTS: A total of 295 cases of incident MI among men and 85 among women occurred during a mean follow-up period of 10.1 years. Compared with women sleeping 8 hours, the multivariable adjusted hazard ratio (HR) of MI among women sleeping < or =5 hours was 2.98 (95% CI, 1.48-6.03), and among women sleeping > or =9 hours 1.40 (95% CI, 0.74-2.64); the corresponding HRs among men were 1.13 (95% CI, 0.66-1.92) and 1.07 (95% CI, 0.75-1.53). In multivariable analysis the relative risk of an incident MI for men and women with difficulties maintaining sleep was 1.12 (95% CI, 0.84-1.48) and 1.53 (95% CI, 0.99-2.37), respectively, and for men and women with difficulties initiating sleep the relative risk was 1.16 (95% CI, 0.82-1.63) and 1.30 (95% CI, 0.81-2.06), respectively. CONCLUSIONS: Modest associations between short sleep duration and difficulties maintaining sleep and incident MI were seen in middle-aged women but not men from the general population.