The overall results of the Multiple Risk Factor Intervention Trial (MRFIT) showed a nonsignificant 7% lower coronary artery disease (CAD) mortality rate in the special-intervention (SI) as compared to the usual-care (UC) group. The initial results also suggested that the SI program was more effective than UC in the community in reducing the CAD mortality rate in nonhypertensive persons than in hypertensive persons, and that the SI program used was more effective in reducing CAD deaths among men without than men with electrocardiographic (ECG) abnormalities at rest. Furthermore, an unfavorable mortality trend in hypertensive SI men with ECG abnormalities at rest compared with UC men was noted (adjusted relative risk of 1.67). Further analyses in baseline-defined subgroups indicated that (1) the most common ECG abnormalities at rest were high R waves and ST-T changes, (2) the CAD mortality differential (SI/UC) was similar in the subgroup with these abnormalities and in the subgroup with other abnormalities at rest, (3) the apparent excess CAD mortality among hypertensive SI men with ECG abnormalities at rest was manifested chiefly as sudden death within 1 hour, and (4) the association between ECG abnormalities at rest and the CAD mortality rate among hypertensive men was independent of the baseline level of blood pressure or of the findings on the exercise electrocardiogram. However, CAD mortality in those with abnormalities on the electrocardiogram both at rest and during exercise was lower in the SI than the UC group. A possible explanation for the difference in outcome in the baseline-defined subgroup was an unexpectedly low UC mortality rate. However, within-group analysis revealed an interaction between ECG abnormalities at rest and diuretic treatment in the SI group, with the risk of CAD death for men prescribed diuretic drugs relative to men not prescribed diuretic drugs estimated as 3.34 among men with baseline ECG abnormalities at rest and as 0.95 among men without such abnormalities. No such effect was found in the UC group, in which men generally were prescribed lower doses of hydrochlorothiazide and chlorthalidone than SI men. However, analyses do not suggest an effect of diuretic dose or of hypokalemia on the CAD mortality rate in treated SI participants. Although subgroup analyses must be interpreted with caution, particularly those that go beyond the randomized clinical trial design by the MRFIT, these findings pose hypotheses for investigation by other researchers in systemic hypertension and may have implications for therapy.