Cisplatin nephrotoxicity: a multivariate analysis of potential predisposing factors

STUDY OBJECTIVE: To evaluate the usefulness of biologic and pharmacologic parameters for early identification of cisplatin-induced renal dysfunction. DESIGN: Prospective evaluation of 62 consecutively admitted patients with cancer. SETTING: Cancer center. PATIENTS: Sixty-two consecutive patients with cancer (52 men, 10 women; mean age 61.9 yrs). INTERVENTIONS: Patients received cisplatin as a single short intravenous infusion every 3 weeks. One hundred twenty-one cycles were analyzed. The dosage in the first cycle ranged between 61 and 105 mg/m2 (mean 84 mg/m2). All patients received a standard hydration protocol. MEASUREMENTS AND MAIN RESULTS: Renal function was evaluated for each cycle before treatment (day 0) and before next cycle (day 21) based on the estimated creatinine clearance (Clcr). For each cycle, the weighted relative decrease (WD) of Clcr was calculated (WDClcr = 100 x Clcr (day 0) - Clcr (day 21)]/Clcr (day 0)](2). Total and ultrafilterable (UF) platinum were measured as a single-sample assay taken 16 hours after the end of cisplatin administration. The mean WDClcr was 0.07 min/100 ml (range -1.0 to +1.7 min/100 ml). The intensity of renal dysfunction evaluated by WDClcr was independent of cisplatin dosage, age, sex, body surface area, initial Clcr, and cycle number. Of interest, total and UF platinum concentrations were significantly correlated to WDClcr: the higher the platinum concentration, the greater the intensity of renal dysfunction. In stepwise regression analysis, UF platinum concentration was the only selected factor. The best prediction of UF platinum was obtained by stepwise regression including cisplatin dosage, initial Clcr, and cycle number (r=0.58, p<0.0001). CONCLUSION: We consider our results to be a first step toward a clinical strategy to identify patients at risk for renal dysfunction after cisplatin treatment.