Validez pronóstica de la escala MEESSI-AHF para la predicción de mortalidad a 30 días en pacientes con insuficiencia cardíaca aguda del Hospital General Guasmo Sur, 2022–2026.

Abstract

Background: Acute heart failure requires early risk stratification. MEESSI-AHF estimates 30-day all-cause mortality using 13 variables available at emergency department presentation; nevertheless, transportability should be tested before applying the model to a different population. Objective: To externally validate the prognostic performance of MEESSI-AHF among adults hospitalized at Hospital General Guasmo Sur from 2022 to 2026. Methods: Retrospective cohort of 467 unique patients represented by their first recorded episode. The original complete-model equation was reproduced, and discrimination, calibration, Brier score, risk categories, and clinically relevant cutoffs were assessed. Results: Fifty deaths at 30 days were verified (10.7%; 95% CI: 8.2–13.8%). The area under the receiver operating characteristic curve was 0.665 (95% CI: 0.581–0.748). Mortality increased from 7.8% in the low-risk group to 33.3% in the very-high-risk group (trend p<0.001), but mean predicted risk was lower than observed risk in the low- and intermediate-risk groups. Calibration-in-the-large was 1.095 and the calibration slope was 0.469; the Hosmer–Lemeshow test indicated lack of fit (p<0.001). Excluding 2022 yielded an AUC of 0.667. Conclusions: MEESSI-AHF preserved a prognostic gradient but showed modest discrimination and systematic underprediction in this cohort. It should not be used without local recalibration as the sole basis for clinical decisions.