Gemcitabine-cisplatin: a schedule finding study.

PURPOSE To evaluate the tolerability of four alternating cisplatin-gemcitabine schedules. A secondary aim was to evaluate the clinical efficacy of this combination. PATIENTS AND METHODS Forty-one patients with advanced solid tumors received alternating sequences with a 4- and 24-hour interval of cisplatin and gemcitabine. Gemcitabine 800 mg/m2 was administered as a 30-min infusion on day 1, 8 and 15, and cisplatin 50 mg/m2 over 1 hour on day 1 and 8; in case of the 24-hour time interval the second drug was administered one day later. Four cisplatin-gemcitabine schedules were studied: gemcitabine four hour before cisplatin (10 patients), or vice versa (14 patients) and gemcitabine twenty-four hours before cisplatin (9 patients) or vice versa (8 patients). The sequence of drug administration was reversed in the second cycle of therapy in each individual patient, enabling the evaluation of sequence-dependent side effects. Twenty-six patients had received prior chemotherapy, of which twenty-one platinum-based. RESULTS The main toxicity was myelosuppression. Overall, grade 3 and 4 thrombocytopenia was observed in 27 out of 41 patients (66%) and was not schedule dependent. No serious bleeding occurred. Leukopenia was significantly different between the 4 alternating schedules (P = 0.01); gemcitabine 24 hours before cisplatin was significantly less toxic compared to both cisplatin 4 hours and 24 hours before gemcitabine (P = 0.01 and P = 0.003, respectively). Furthermore, paired analysis of the 4-hour and 24-hour data sets showed that leukopenia was significantly more serious when cisplatin preceded gemcitabine (P = 0.005). Although most patients received prior treatment, both prior chemotherapy and radiotherapy were not related to toxicity. Overall, grade 3 and 4 leukopenia occurred in 19 out of 41 patients (46%). Anemia (Hb < or = 6.0 mmol/l) was not sequence dependent and was observed in 63% of patients. Myelotoxicity was cumulative between cycles and caused frequent omission of gemcitabine on day 15. Overall, in 51% of administered cycles there was no omission of gemcitabine. A mean of 3.5 therapy cycles was administered. Non-hematological toxicity was moderate, consisting mainly of grade 1 and 2 nausea/vomiting and fatigue, and was not schedule dependent. Recently, we described that the schedule in which cisplatin was administered 24 hours before gemcitabine produced the best pharmacological profile. Based on this and because toxicity was manageable, the schedule cisplatin 24 hours prior to gemcitabine was chosen for phase II evaluation. Nine out of thirty-six evaluable patients had an objective response. These responses were observed in head and neck squamous-cell carcinoma (HNSCC), non-small-cell lung cancer (NSCLC), melanoma, adenocarcinoma of unknown origin, ovarian and esophageal carcinoma. CONCLUSIONS Myelosuppression was the most important toxicity. Leukopenia was schedule dependent: gemcitabine before cisplatin was less toxic than the reversed sequence, in this respect. Some encouraging responses were seen in patients with esophageal cancer. Currently, a phase II study with cisplatin 24 hours before gemcitabine is ongoing in patients with advanced upper gastro-intestinal tumors.