Interventions for preventing cardiomyopathy due to anthracyclines: a Bayesian network meta-analysis

Background: The relative efficacy of interventions for primary prevention of anthracycline-associated cardiotoxicity is unknown. Methods: We conducted a systematic review of randomized controlled trials for primary prevention of anthracycline-associated cardiotoxicity in adult cancer patients. We used hierarchal outcome definitions in the following order of priority: (1) composite of heart failure or decline in left ventricular ejection fraction, (2) decline in ejection fraction, or (3) heart failure. Data were analyzed using a Bayesian network meta-analysis with random effects. Results: A total of 16 trials reported cardiotoxicity as a dichotomous outcome among 1918 patients, evaluating dexrazoxane, angiotensin antagonists, beta-blockers, combination angiotensin antagonists and beta-blockers, statins, Co-enzyme Q-10, prenylamine, and N-acetylcysteine. Compared with control, dexrazoxane reduced cardiotoxicity with a pooled odds ratio (OR) of 0.26 (95% credible interval [CrI] 0.11-0.74) and had the highest probability (33%) of being most effective. No other agent was demonstrably better than placebo. Angiotensin antagonists had an 84% probability of being most effective in a sensitivity analysis excluding one outlying study (OR 0.06 [95% CrI 0.01-0.24]). When the outcome was restricted to heart failure, dexrazoxane was associated with an OR of 0.12 (95% CrI 0.06-0.23) relative to control and had 58% probability of being most effective, while angiotensin antagonists had an OR of 0.18 (95% CrI 0.05-0.55). Available data suggested that dexrazoxane and angiotensin antagonists did not affect malignancy response rate or risk of death. Conclusion: Moderate quality data suggest that dexrazoxane, and low quality data suggest angiotensin antagonists, are likely to be effective for cardiotoxicity prevention.