Efficacy and Safety of Different Combinations of Add-on Diuretic Therapy in Acute Heart Failure: A Systematic Review and Network Meta-Analysis of Randomized Controlled Trials.

Patients hospitalized with acute heart failure (HF) may experience diuretic resistance and require an add-on agent despite increasing loop diuretic dosage. While randomized controlled trials (RCTs) have compared add-on therapy to loop diuretics only, sparse literature exists on direct comparisons between various add-on therapies. We performed a systematic review and network meta-analysis of RCTs to assess the efficacy and safety of different diuretic add-on therapies in patients hospitalized with acute HF. Any RCT evaluating the effect of add-on diuretic therapy in patients hospitalized with acute HF was eligible for inclusion. A systematic search of EMBASE and PubMed was conducted until March 29, 2024. The primary outcome was the hospital length of stay. Data was pooled using a random-effects model for direct comparisons. A network meta-analysis using frequentist methods was performed under random-effects multiple treatment comparisons. We assessed ranking probability using the surface under the cumulative ranking curve (SUCRA) method. Of the 1,103 references, 29 RCTs enrolling 8,362 patients met the eligibility and were included. For the direct comparisons, there was no significant difference in hospital length of stay (MD -0.42, 95% CI= -0.87,0.02). Ranking probability based on SUCRA indicated that acetazolamide had the highest likelihood of being the best treatment for shorter hospital length of stay (SUCRA, 0.89), followed by SGLT2i (SUCRA, 0.70). The certainty of estimates for all outcomes ranged from moderate to very low. In conclusion, the efficacy of add-on therapy was associated with reduced hospital length of stay. Albeit uncertain, the results from NMA provide initial evidence suggesting there may be optimal treatment strategies to decongest patients with HF to achieve and maintain euvolemia. However, well-designed direct comparison RCTs are needed to increase the certainty of the estimates. Protocol registered in PROSPERO (CRD42023476669).