Changing global prevalence of diuretic resistance amongst patients with heart failure: a systematic review and cumulative meta-analyses

Background Diuretic resistance (DR) is a major challenge in the management of heart failure, contributing to persistent congestion, prolonged hospitalization, and adverse outcomes. However, inconsistent DR definitions and study designs have resulted in uncertain prevalence estimates and limited guideline development. Objective To determine the pooled prevalence of DR in heart failure using studies with valid heart-failure denominators and extractable binary DR events, and to evaluate the stability and heterogeneity of prevalence estimates over time. Data Sources PubMed, EMBASE, Scopus, the Cochrane Database of Systematic Reviews, Science Citation Index, WHO International Clinical Trials Registry Platform (ICTRP), and the Database of Abstracts of Reviews of Effects were searched from 2000 through the final updated search in 2026, supplemented by reference screening. Study Selection Observational studies and randomized studies relevant to DR in adults with heart failure were considered for the systematic review. For the quantitative prevalence analysis, studies were required to include an unselected or consecutively assembled heart-failure denominator and an extractable binary DR event count based on a prespecified or study-defined criterion arising after cohort entry. Studies enrolling only patients with established DR, reporting only continuous diuretic-response measures, or lacking a valid binary DR numerator were excluded from prevalence pooling but could remain relevant to the qualitative systematic review. Two reviewers independently extracted data. Prevalence was pooled on the logit scale using an inverse-variance random-effects model with restricted maximum likelihood estimation of between-study variance and Hartung-Knapp inference. Heterogeneity, prediction intervals, cumulative meta-analysis, influence and sensitivity analyses, a generalized linear mixed-model robustness analysis, and exploratory funnel-plot asymmetry testing were undertaken. Results Twelve prevalence-eligible studies involving 24,904 patients were included in the quantitative prevalence analysis, of whom 4,861 met study-defined criteria for DR. The pooled prevalence was 23.0% (95% CI, 13.6%–36.2%), with substantial heterogeneity (I²=98.5%; τ²=0.851; Q=757.17, df=11; p<0.001). The 95% prediction interval was 3.4%–71.6%. A generalized linear mixed model yielded a similar pooled prevalence of 22.5% (95% CI, 14.5%–33.2%). Cumulative analysis showed that the estimate changed as new eligible studies accrued, reaching 23.0% after inclusion of the most recent studies. Exploratory regression analysis did not demonstrate statistically significant funnel-plot asymmetry (p=0.072). Conclusion Approximately one in four patients with heart failure meets a study-defined criterion for DR in prevalence-eligible cohorts. However, the very wide prediction interval and extreme heterogeneity show that DR frequency is strongly dependent on population, setting, timing of assessment, and operational definition. Standardized diagnostic criteria and prospectively assembled prevalence cohorts are required to improve comparability and guide clinical practice.