Impact of bleomycin (BLM) administration on the development of pulmonary toxicity in advanced germ cell tumors (GCT) receiving first-line chemotherapy (CT): A meta-analysis of randomized studies.

Abstract

482 Background: Little is known about the impact of BLM administration on the development of pulmonary toxicity (PT) in patients (pts) with advanced GCT. This information is critical to further enhance the successful delivery of effective and safe CT in these pts. Methods: Literature was searched for randomized trials of 1st-line CT ( ≥ 3 cycles) for GCT. Descriptive statistics were used to summarize data, dichotomized by BLM-containing vs non containing CT. Univariable and multivariable analyses by random effect models evaluated the predictive role of BLM administration on the development of all grade and G3-4 PT, adjusted for the length of follow-up (FUP, continuous), time-frame of treatment ( ≤ 1997 versus > 1997) % of pts with lung metastases (LM), % of primary mediastinal GCT (PMGCT), IGCCCG risk group. Results: 53 arms of 26 phase II and III trials totalling 6,498 pts were selected: 40 with BLM (n = 5,093) and 13 without BLM (n = 1,405). Median % of pts with LM was 44% (range 22-84%) and 31% (19-79%) in BLM-containing and non containing arms (Wilcoxon test p = 0.505). Median % of PMGCT was 15% (range 0-48%) and 2% (0-13%), respectively (p = 0.146). Median BLM total dose and No. of administrations were 360mg (range 120-375) and 12 (2-24). Median FUP was 52 and 53 months, respectively. The pooled probability of all grade PT was 11.7% (95% CI: 8.4-16.0%) and 1.7% (95% CI: 0.7-4.2%) in BLM-containing and non containing arms. Univariably, BLM administration was associated with the incidence of both all grade (odds ratio [OR]: 7.57, 95%CI: 2.84-20.18, Wald test p < 0.001) and G3-4 PT (OR: 5.19, 95%CI: 1.57-17.16, p = 0.007). Multivariably, BLM administration was significantly associated with the incidence of all grade PT (OR: 4.14, 95%CI: 1.36-12.59, p = 0.012) and trended to significance for G3-4 toxicity (OR: 2.24, 95%CI: 0.91-5.51, p = 0.080). Conclusions: In this meta-analysis, BLM administration was significantly associated with the development of PT in pts with GCT. The quantification of the magnitude of incidence of PT provided here might be useful to optimize the delivery of non-PEB CT in the 1st-line setting.