Myocardial performance index (MPI), or Tei index, is measured by Doppler echocardiography in clinical practice. MPI has been shown to be useful in evaluating left ventricular (LV) performance and predicting prognosis in cardiac patients. However, the effects of LV load and contractile states on MPI remain to be thoroughly investigated. In 14 anesthetized dogs, we obtained LV pressure-volume relationship with use of sonomicrometry and catheter-tip manometry. MPI was determined from the time derivative of LV volume and pressure. LV end-systolic pressure-volume ratio (Ees'), effective arterial elastance (Ea) and LV end-diastolic volume (Ved) were used as indices of LV contractility, afterload and preload, respectively. Hemodynamic conditions were varied over wide ranges [heart rate (HR), 66-192 bpm; mean arterial pressure, 71-177 mmHg] by infusing cardiovascular agents, by inducing ischemic heart failure and by electrical atrial pacing. Multiple linear regression analysis of pooled data (66 data sets) indicated that MPI (0.6-1.8) significantly correlated with Ees' [1.5-17.5 mmHg · ml(-1), p<0.0001, standard partial regression coefficient (β) =-0.66], Ea (3.6-21.9 mmHg · ml(-1), p<0.001, β = 0.4) and Ved (11-100 ml, p<0.0001, β = -0.69). MPI directly correlated with the time constant of isovolumic relaxation (19-66 ms, p<0.05), but not with HR or LV diastolic-stiffness (all p>0.1). Theoretical analysis also indicated that MPI decreases following the increases in LV contractility and in preload, while it increases in response to an increase in LV afterload. We conclude that MPI sensitively detects changes in LV contractility. However, MPI is also affected by changes in LV afterload and preload.