BACKGROUND Information about prognostic utility is limited for numerous electrocardiogram (ECG) abnormalities and is particularly scarce in women with coronary heart disease (CHD) or at increased risk of CHD occurrence. METHODS This study used a prospectively planned observational post-hoc analysis of a negative randomised trial designed for other purposes. METHODS ECGs of 9789 postmenopausal women were analysed at a core laboratory. ECG abnormalities were determined and evaluated for mortality risk on top of established clinical risk factors and simultaneously with all other ECG abnormalities. RESULTS During a median follow-up of 5.6 years, 500 women sustained coronary death. Normal versus abnormal baseline ECGs were associated with an annual rate of 0.53% versus 1.28% coronary deaths. Ten clinically common ECG abnormalities, including left atrial abnormality, fragmented QRS, and Cornell voltage-only left ventricular hypertrophy, emerged as independent significant predictors of coronary death, eight of them also predicted all-cause mortality. Each ECG abnormality acted as a risk multiplier for a coexisting ECG abnormality. Two or three of any of the ECG abnormalities simultaneously present on the baseline ECG identified subsets with hazard ratios for coronary death of 3.3 or 5.5 respectively. Consideration of ECG abnormalities significantly improved risk stratification by common clinical parameters. CONCLUSIONS The presence of ECG abnormalities provides independent risk information over and above that of established risk factors, both for women with CHD or at increased risk of occurrence of CHD. Normal ECG assures a low mortality risk regardless of whether CHD exists or not. The data derived could be applied to corresponding postmenopausal women in daily clinical practice.