Risk of adverse cardiovascular events (CVE) and incident diabetes mellitus (DM) in patients (pts) with prostate cancer (PC) treated with androgen deprivation therapy (ADT): A meta-analysis of adjusted observational results.

Abstract

e15192 Background: ADT is a mainstay treatment in pts with PC and is supposedly associated to an unfavorable metabolic and cardiovascular profile. Recently, a meta-analysis of randomized controlled trials (RCT) found no association between ADT and increased risk of CVE (JAMA 2011). However, no conclusive data were available about ADT association with metabolic changes and adverse CVE or DM because of pts selection in RCT. Therefore, we performed a meta-analysis of adjusted observational results in order to look for DM and CVE onset in an ADT unselected population. Methods: Medline, Cochrane Library and Biomed Central were searched for articles addressing adverse events related to ADT in patients with PC. Selection criteria were: not RCT, pts assigned to ADT or not, adjusted risk of CVE and DM according to ADT. Exclusion criteria were: duplicate publication, comparison of two different strategies of ADT (different drugs or duration). Cardiovascular death was the primary endpoint; non-fatal myocardial infarction (MI), stroke/transient ischemic attack (TIA) and new DM onset were secondary endpoints. Random effects model with generic inverse variance weighting was used to estimate adjusted risks as odds ratios (OR) with 99% confidence interval (CI). Results: We selected 12/2100 screened studies. We included 208643 pts of which 102177 received ADT. At a follow up of 5 years (4-7.5), ADT did not result as an independent risk factor for cardiovascular death (OR= 1.04; 99% CI= 0.94-1.14). No increased risk of MI (OR= 1.13; CI= 0.86-1.48) or of stroke/TIA (OR= 1.11; CI= 0.78-1.57) were detected. Incident DM was more frequent among ADT pts (OR= 1.32; CI= 1.14-1-53). Even in 7205 pts with previous CVE (2450 received ADT), ADT was not associated with an increased risk of overall death (OR= 1.23; CI= 0.87-1.75). Meta-regression analysis showed no significant interactions between duration of ADT and cardiovascular death or incident DM. Conclusions: In non-selected pts,ADT appears to increase the risk of incident DM, but not of CVE or stroke/TIA. Moreover, overall mortality is not increased in pts with a history of CVE.