Abstract

INTRODUCTION: Decreasing body temperature on first follow-up visit-relative to discharge-predicts early rehospitalization in heart failure (HF). We studied whether admission-to-discharge temperature reduction was associated with increased HF rehospitalization in the ESCAPE trial. METHODS: We compared patients with or without ≥1 °C decrease in temperature from admission-to-discharge. The study endpoint was rehospitalization due to HF for up to 6 months after discharge. RESULTS: Among 354 patients (average age 57 years, 73% men) with recorded admission and discharge temperature, 22 (6.2%) had an admission-to-discharge temperature reduction ≥1 ºC. Patients with admission-to-discharge temperature reduction ≥1 ºC had higher frequency of rehospitalization for HF (68.2% vs. 44.3%, estimated odds ratio [OR] 2.697, 95% confidence interval [CI] 1.072-6.787, P = 0.029) despite a significantly higher admission temperature. On multivariate analysis, admission-to-discharge temperature reduction ≥1 ºC predicted rehospitalization for HF (OR 2.02, 95% CI 1.028-3.966, P = 0.041) after adjustment for age, BMI, baseline Na, creatinine, ejection fraction and discharge NYHA class. A standard logistic model treating temperature change as a continuous variable, and a model using a restricted cubic spline, did not demonstrate a statistically significant relationship between temperature reduction and HF rehospitalization. Subsequently, an altered logistic model was fit expressing the log odds of HF rehospitalization as a piecewise linear function of temperature decrease; this model did demonstrate statistical significance (P = 0.013) with an estimated odds ratio of 1.140 per 0.1 ºC beyond 0.5 ºC. CONCLUSION: Admission-to-discharge temperature reduction ≥1 ºC is an unfavorable prognostic sign associated with future rehospitalization due to HF.