Atrioventricular delay (AVD) is critical in patients with DDD pacemakers (PM). Echo/Doppler evaluation of AVD providing the longest left ventricular filling time (FT) or the highest cardiac output (CO) is used for AVD optimization. Recently myocardial performance index (MPI) has been shown to improve by optimizing AVD. The aim was to compare the CO, FT, MPI derived optimal AVD, and to analyze systolic and diastolic performance at every optimal AVD. Twenty-five patients, 16 men 68 +/- 11 years, ejection fraction >or= 50%, with a DDD PM for third-degree AV block, without other major cardiomyopathies, underwent echo/Doppler AVD optimization. CO, FT, and MPI derived optimal AVDs were identified as the AVDs providing the highest CO, the longest FT, and the minimum MPI, respectively. Isovolumic contraction and relaxation time (ICT, IRT), ejection time (ET), ICT/ET, and IRT/ET ratios were also evaluated at every optimal AVD. CO, FT, and MPI derived optimal AVDs were significantly different (148 +/- 36 ms, 116 +/- 34 ms, and 127 +/- 33 ms, respectively). ICT/ET was similar at CO, FT, and MPI derived optimal AVD (0.22 +/- 0.10, 0.23 +/- 0.11, and 0.21 +/- 0.10, respectively). IRT/ET ratio was similar at FT and MPI derived optimal AVDs (0.34 +/- 0.15 and 0.33 +/- 0.15, respectively) and significantly shorter (P < 0.02) than at CO derived optimal AVD (0.40 +/- 0.15). Different methods indicate different optimal AVDs. However analysis of systolic and diastolic performance shows that different AVDs result in similar systolic or diastolic performance. At MPI optimized AVD, a high CO combined with the most advantageous conditions of both isovolumic contraction and relaxation phases is achieved.