MATERIALS AND METHODS Two hundred patients at steady-state on long-term perhexiline were identified retrospectively. The ratio of maintenance dose to steady-state plasma concentration (dose:[Px]) was correlated with the following putative determinants via simple and multiple linear regression analyses: age, weight, left ventricular ejection fraction (LVEF), and creatinine clearance (CrCl, Cockroft-Gault formula). A Mann-Whitney U test was performed to determine if severe left ventricular systolic impairment affected maintenance dose. RESULTS Advanced age, left ventricular systolic impairment, and renal impairment were frequently encountered. Using simple linear regression, age was a negative correlate of dose:[P] (R = 0.23, P = 0.001), whereas weight (R = 0.27, P = 0.0001) and CrCl (R = 0.30, P < 0.0001) were positive correlates. Mann-Whitney U analysis showed no difference between dose: [Px] among patients with LVEF of less than 30% versus 30% or greater. Advancing age was strongly associated with decreasing weight (R = -0.45, P < 0.00001) and calculated CrCl varied directly with weight, as expected (R = 0.66, P < 0.0001). Stepwise multiple linear regression using age, LVEF, CrCl, and weight as potential predictors of dose:[P] yielded only weight as a significant determinant. DISCUSSION Perhexiline has become a "last-line" agent for refractory angina as a result of complex pharmacokinetics and potential toxicity. Use has increased predictably in the aged and infirm who have exhausted standard medical and surgical therapeutic options. Beyond genotype, the effect of patient characteristics on maintenance dose has not been explored in detail. In this study, dose requirement declined with age in a frail and wasting population as a result of weight-related pharmacokinetic factors. LVEF had no apparent effect on maintenance dose and should not be considered a contraindication to use. CONCLUSION A weight-adjusted starting dose may facilitate the safe and effective prescription of perhexiline and is calculated by 50 + 2 × weight (kg) mg/d, rounded to the closest 50 mg/day.