Purpose: Cisplatin (CDDP) and etoposide (VP16) are considered major standard cytotoxic drugs for small cell lung cancer (SCLC). The present systematic review had as its objective the evaluation of their role, as components of chemotherapy regimens, on survival.
Methods: Published randomised clinical trials (from 1980 to 1998) were selected comparing, in SCLC patients, chemotherapy regimens, given as first-line therapy. One arm (the experimental arm) had to include CDDP and/or VP16, while another had to omit the same drug(s). Trials quality was assessed by two published scores (Chalmers and European Lung Cancer Working Party (ELCWP)). For each individual trial, the hazard ratio (HR) of the survival distributions was estimated on the basis of reported statistics or, if not available, by extracting, from the survival graphical representations, the data required to construct the difference between expected and observed numbers of events as calculated in the log-rank statistic. A combined hazard ratio was obtained by the Peto method (a value < 1 meaning a benefit for CDDP and/or VP16).
Results: Thirty-six trials eligible for our systematic review were identified, classified into four groups (I-IV): group I, 1 trial testing a CDDP-based regimen (without VP16) against another arm not including either CDDP or VP16; group II, 17 trials testing a VP16-based regimen (without CDDP) against a regimen without VP16 and CDDP; group III, nine trials comparing a regimen including CDDP and VP16 with a regimen using neither drug; and, finally, group IV, nine trials comparing a regimen based on both drugs with a regimen based on VP16 only. Overall, Chalmers and ELCWP scores correlated well (r(S) = 0.76, P < 0. 001) and had respective median scores of 50.3 and 63.7%. The number of eligible patients did not have a significant impact on the scores as well as the trials group, the trial positivity (a positive trial defined as showing itself a statistically significant survival benefit for the experimental regimen), overall or in categories, and the year of publication. Combined hazard ratios with 95% confidence intervals were: 0.70 (0.41-1.21) for group I, 0.72 (0.67-0.78) for II, 0.57 (0.51-0.64) for III, and 0.74 (0.66-0.83) for IV, showing a survival benefit in favour of regimens including etoposide alone or in combination with cisplatin, justifying with high significance levels the use of each of these drugs. Overall survival benefits could also be shown for regimens including CDDP (HR = 0.61; confidence interval (CI), 0.57-0.66), as well as for those including VP16 (HR = 0. 65; CI, 0.61-0.69). Robustness of these results has to be confirmed with appropriate randomised trials.