OBJECTIVE To investigate whether the apparent discrepancy between the efficacy of the combination of isosorbide dinitrate (ISDN) and hydralazine demonstrated in the first V-HeFT trial (V-HeFT I) and that in V-HeFT II could be explained by pharmacokinetic differences in the study drug formulations, and to compare the pharmacokinetic profile of the fixed-dose combination of ISDN/hydralazine (FDC ISDN/HYD; BiDil) formulation used in A-HeFT with that of the V-HeFT study drug formulations. METHODS A bioequivalence study was performed (n = 18-19 per group) comparing the ISDN and hydralazine formulations used in V-HeFT I, V-HeFT II and A-HeFT in healthy volunteer men and women aged 18-40 years. In phase A of the study, subjects received a reference solution of hydralazine hydrochloride/ISDN (37.5mg/10mg) orally. Slow acetylators were identified and randomised into three groups in phase B to receive a single oral dose of identical amounts of hydralazine hydrochloride/ISDN (37.5mg/10mg) from either (i) a hydralazine capsule plus an ISDN tablet (the V-HeFT I formulation); (ii) a hydralazine tablet plus an ISDN tablet (the V-HeFT II formulation); or (iii) FDC ISDN/HYD (the A-HeFT formulation). Blood/plasma concentrations of hydralazine and ISDN were determined from the blood samples taken between 0 and 36 hours. RESULTS In phase B, the maximum observed concentrations (C(max)) were 65.9 +/- 53.9, 28.2 +/- 15.8 and 51.5 +/- 54.3 ng/mL of unchanged hydralazine, and 23.1 +/- 12.3, 21.7 +/- 13.4 and 26.7 +/- 18.7 ng/mL of ISDN for the V-HeFT I, V-HeFT II and A-HeFT formulations, respectively. The area under the blood/plasma concentration-time curve (AUC) values were 32.6 +/- 13.4, 23.3 +/- 15.1 and 32.6 +/- 18.5 ng x h/mL of hydralazine, and 24.4 +/- 9.0, 24.8 +/- 8.0 and 23.5 +/- 6.3 ng x h/mL of ISDN for the V-HeFT I, V-HeFT II and A-HeFT formulations, respectively. For comparison of bioequivalence, the C(max) and AUC were normalised to 65kg bodyweight, and point estimates and 90% confidence intervals of the C(max) ratios, AUC ratios and ratios of the AUC in phase B normalised for clearance by the AUC in phase A (AUCR) were calculated. The three formulations were not bioequivalent based on the C(max) and AUC comparisons. CONCLUSIONS The blood concentrations of hydralazine obtained with the tablet formulation tested in V-HeFT II were markedly lower than those obtained with the capsule formulation tested in V-HeFT I or the FDC ISDN/HYD single tablet used in A-HeFT. The apparently modest effect on survival observed in V-HeFT II could be explained in part by the poor hydralazine bioavailability of the tablet preparation used in this trial. ISDN exposures were similar in the two trials. The ISDN-hydralazine formulation used in V-HeFT II was not bioequivalent to the formulation used in V-HeFT I or to the FDC ISDN/HYD that had demonstrated a significant survival benefit in A-HeFT.