The influence and clinical consequences of different atrioventricular delays on ventricular filling have been studied in 30 patients (mean age 60 +/- 5 years) who implanted DDD pacemaker for complete A-V block. All patients presented a normal ejection fraction: 63.9 +/- 6.5%. In 18 subjects (Group I) an echo-Doppler examination revealed ventricular hypertrophy (mean end-diastolic wall thickness of 1.4 +/- 0.16 cm; left ventricular mass index 155 +/- 17 g/m2) and an abnormal relaxation pattern (isovolumic relaxation time 124 +/- 11: early to late peak velocity 0.6 +/- 0.03; deceleration time of the early diastolic peak 296 +/- 34 ms). Group II included the remaining 12 patients without left ventricular hypertrophy and normal filling pattern. In all 30 patients the filling pattern was reassessed following modification of the A-V delay (200, 150, 100 and 75 ms). Patients at baseline (200 ms of A-V delay) underwent an exercise test with determination of respiratory gas exchange. In Group I, 13 (72.5%) patients were classified as Weber class B (VO2 max 16.8 +/- 1.7 ml/min/kg) and 5 (27.5%) as class A (VO2 max 22.5 +/- 1.4 ml/min/kg). In Group II, all 12 patients were classified as Weber class A (VO2 max 23.1 +/- 1.1 ml/ min/kg). In Group II, changes in A-V delay caused no homogeneous variation in filling pattern. A-V delay was not modified in this group. In Group I, the reduction of A-V delay to 100 ms resulted in filling normalization. In this group A-V delay was programmed definitely to 100 ms. Graded exercise test repeated at 6 months follow-up showed an improved Weber class in 13 patients (from B to A) and greater VO2 max in the remaining 5 already in class A. We conclude that, in sequential paced patients with normal ejection fraction but abnormal relaxation pattern, a modification of A-V delay can induce filling normalization and improve cardiac functional capacity.