A Randomized Multicenter Study of Optimal Circadian Time of Vinorelbine Combined with Chronomodulated 5‐Fluorouracil in Pretreated Metastatic Breast Cancer Patients: EORTC Trial 05971

Studies in animals synchronized with an alternation of 12 h of light and 12 h of darkness have showed that hematological and systemic toxicities could be reduced if vinorelbine were administered 19 or 23 hours after light onset (HALO), corresponding to 17:00 and 21:00 h in diurnally active humans. This trial aimed to define the least toxic time of vinorelbine administration in metastatic breast cancer patients. Initially, the study treatment consisted of three courses of vinorelbine of 30 mg/m²/d on D1 and D6 and chronomodulated 5‐fluorouracil of 850 mg/m² from D2 to D5 every 21 days. Ninety metastatic breast cancer patients were randomized to receive vinorelbine at one of the eight possible dosing times. Further to the recommendations of the Independent Data Monitoring Committee, the vinorelbine dose was reduced to 25 mg/m²/d midway through the study. The primary objective of the study was detection of the least toxic time based on the incidence of grade 3–4 (G3–4) neutropenia. To show a significant result, the 90% confidence interval width of the least toxic time had to be<6 h. The least toxic time detection based on the incidence of other toxicities was also analyzed. The time of least drug toxic was estimated using a logistic regression model assuming that the logit transformation of the toxicity rate follows a sinusoidal distribution over 24 h. The bootstrap technique was used to obtain the 90% confidence interval. The least toxic time of G3–4 neutropenia was observed at 21:00 h with a non‐significant 90% CI. Secondary endpoint analyses indicated the least toxic time could differ when based on other toxicity parameters (e.g., a significant least toxic time of 17:00 h was observed for G3–4 leucopenia), in agreement with animal data. The least toxic time of 10:30 h was estimated for any G3–4 gastrointestinal toxicity. This results of this study do not allow us to recommend an optimal time for vinorelbine administration. It has highlighted, however, the inherent methodological difficulties in the conduct of such a trial in the human setting. It indicates that future optimal time‐finding trials should have tolerability and/or activity as the primary endpoint in place of a particular toxicity. The randomized optimal time‐finding design may be used to identify the best time of chemotherapy administration. (Author correspondence: bcoudert@dijon.fnclcc.fr)