Predictors of in-hospital mortality in acute heart failure: Results from a national registry in Mauritania

Acute heart failure (AHF) carries substantial in-hospital mortality, yet prognostic data from sub-Saharan Africa, and particularly from Mauritania, are virtually absent. We aimed to identify independent predictors of in-hospital mortality at the only dedicated cardiology centre in Mauritania.

The MATURE registry prospectively enrolled consecutive adults hospitalised for AHF at the Centre National de Cardiologie, Nouakchott, between 1 January and 31 May 2024. Because deaths were few, a pre-specified parsimonious multivariable logistic-regression strategy avoided overfitting; discrimination was assessed by the area under the receiver-operating-characteristic curve (AUC) and calibration by the Hosmer-Lemeshow test.

Of 307 patients (median age 61 years [IQR 51-70]; 64.8% male), in-hospital mortality was 5.5% (n = 17). Non-survivors had lower systolic blood pressure (100 vs 120 mmHg, p = 0.001), higher heart rate (114 vs 95 bpm, p = 0.001) and higher creatinine (14 vs 10 mg/L, p = 0.002); cardiogenic shock (23.5% vs 3.4%, p = 0.005) and intensive-care admission (70.6% vs 35.9%, p = 0.008) were more frequent. In the final model, systolic blood pressure (adjusted OR 0.96 per mmHg, 95% CI 0.94-0.99, p = 0.002) and serum creatinine (adjusted OR 1.02 per mg/L, 95% CI 1.00-1.03, p = 0.030) were independent predictors. Discrimination was moderate (AUC 0.75) and calibration acceptable (p = 0.31).

In this first Mauritanian AHF cohort, in-hospital mortality was comparable to international registries. Low systolic blood pressure and renal dysfunction were the strongest independent predictors of death, offering simple bedside risk stratification where advanced diagnostics are unavailable.