Our recent clinical studies on the negative dP/dt upstroke pattern suggested that the left ventricular pressure (LVP) deviated from the exponential curve during isovolumic relaxation period (IRP) in diseased hearts. To examine this further two types of monoexponential curve fitting were done in various heart diseases (normal (N): 8, angina pectoris (AP): 8, myocardial infarction (MI): 13, hypertrophic cardiomyopathy (HCM): 10, dilated cardiomyopathy (DCM): 8). LVP was measured by a Millar's catheter-tip transducer, and four types of time constant (T1-T4) were derived: T1 was calculated by an exponential curve fitting e-t/T1 + B (B is constant), T2 by the ratio Pm/peak (-) dP/dt (Pm is LVP at peak (-) dP/dt), T3 by the best exponential curve fit e-t/T3 + B + C (C is constant), and T4 by the ratio (Pm-C)/peak (-) dP/dt. If the exponential curve fitting was reasonable, the relation between T1 and T2, or T3 and T4, should be on the line of identity. The result was as follows: T2 = 1.4 T1 - 6.1 (r = 0.84) and T4 = 1.1 T3 - 10.7 (r = 0.94). Additionally, C (mmHg) in MI (-26 +/- 15), HCM (-36 +/- 19) and DCM (-32 +/- 20) were lower (p less than 0.05) than in N (-13 +/- 7). These findings suggest that the LVP during IRP could deviate from the monoexponential curve and that careful attention should be given to calculate the time constant in diseased hearts.