The Early Prediction of Patient Outcomes in Acute Heart Failure: A Retrospective Study.

BACKGROUND

Acute heart failure (AHF) is a major cause of hospitalizations, posing significant challenges to healthcare systems. Despite advancements in management, the rate of poor outcomes remains high globally, emphasizing the need for timely interventions. This study aimed to identify early admission-based factors predictive of poor outcomes in hospitalized AHF patients, in order to contribute to early risk stratification and optimize patient care.

METHODS

This retrospective single-center study analyzed routine data of adult patients hospitalized for AHF at a public university teaching hospital in Switzerland. Outcomes included in-hospital death, intensive care (ICU) treatment, and length of hospital stay (LOHS). Potential predictors were limited to routine parameters, readily available at admission. Missing predictor data was imputed and predictors were identified by means of multivariable regression analysis.

RESULTS

Data of 638 patients (median age 84 years, range 45-101 years, 50% female) were included in the study. In-hospital mortality was 7.1%, ICU admission rate 3.8%, and median LOHS was 8 days (IQR 5-12). Systolic blood pressure ≤ 100 mmHg (Odds ratio (OR) 3.8, p = 0.009), peripheral oxygen saturation ≤ 90% or oxygen supplementation (OR 5.9, p < 0.001), and peripheral edema (OR 2.7, p = 0.044) at hospital admission were identified as predictors of in-hospital death. Furthermore, a stroke or transient ischemic attack in the patient's history (OR 3.2, p = 0.023) was associated with in-hospital death. ICU admission was associated with oxygen saturation ≤ 90% or oxygen supplementation (OR 22.9, p < 0.001). Factors linked to longer LOHS included oxygen saturation ≤ 90% or oxygen supplementation (IRR 1.2, p < 0.001), recent weight gain (IRR 1.1, p = 0.028), and concomitant chronic kidney disease (IRR 1.2, p < 0.001).

CONCLUSIONS

This study validated established predictors of AHF outcomes in a Swiss cohort, highlighting the predictive value of poor perfusion status, fluid overload, and comorbidities such as chronic kidney disease. The identified predictors imply potential for developing tools to improve rapid treatment decisions. Future research should focus on the prospective external validation of the identified predictors and the design and validation of risk scores, incorporating these parameters to optimize early interventions and reduce adverse outcomes in AHF.