Background: Chemotherapy is the standard treatment for advanced non-small-cell lung cancer, and myelosuppression is a common side-effect. We aimed to assess whether haematological toxic effects could be a biological measure of drug activity and a marker of efficacy.
Methods: We analysed data for 1265 patients who received chemotherapy (vinorelbine, gemcitabine, gemcitabine and vinorelbine, cisplatin and vinorelbine, or cisplatin and gemcitabine) within three randomised trials. Primary landmark analyses were restricted to 436 patients who received all six planned chemotherapy cycles and who were alive 180 days after randomisation. Neutropenia was categorised on the basis of worst WHO grade during chemotherapy: absent (grade 0), mild (grade 1-2), or severe (grade 3-4). All statistical analyses were stratified by treatment allocation. Analyses were repeated in the out-of-landmark group (829 patients), stratifying by treatment allocation and number of chemotherapy cycles. The primary endpoint was overall survival.
Findings: In the landmark group, hazard ratios of death were 0.65 (0.46-0.93) for patients with severe neutropenia and 0.74 (0.56-0.98) for those with mild neutropenia. Median survival after the landmark time of 180 days was 31.4 weeks (95% CI 25.7-39.6) for patients without neutropenia compared with 42.0 weeks (32.7-59.7) for patients with severe neutropenia, and with 43.7 weeks (36.6-66.0) for those with mild neutropenia (severe vs mild vs no neutropenia p=0.0118). Findings were much the same for the out-of-landmark group.
Interpretation: Neutropenia during chemotherapy is associated with increased survival of patients with advanced non-small-cell lung cancer, and its absence might be a result of underdosing. Prospective trials are needed to assess whether drug dosing guided by the occurrence of toxic effects could improve efficacy of standard regimens.