Background: In congestive heart failure (CHF), a low body temperature at hospital admission predicts in-hospital mortality. We hypothesized that a postdischarge reduction in body temperature predicts early CHF rehospitalization and death.
Methods: We reviewed the records of 198 patients discharged after CHF hospitalization. We categorized the patients as hypothermic or normothermic (cutoff point, 36.3 degrees C/97.4 degrees F) according to body temperature at discharge. We classified the 2 groups according to the direction of temperature change between discharge and the first follow-up visit: normothermic/non-decreasing temperature (N+), normothermic/decreasing temperature (N-), hypothermic/non-decreasing temperature (H+), and hypothermic/decreasing temperature (H-).
Results: Ninety-three patients (47%) had decreasing temperatures, and 105 patients (53%) had non-decreasing temperatures. Kaplan-Meier analysis revealed a significant intergroup difference in survival (P = .01) and rehospitalization time (P = .005). On logistic regression, a decreasing temperature was significantly associated with rehospitalization within 180 days (odds ratio, 4.01; 95% confidence interval, 1.63-10.02; P = .003). On Cox regression, the hazard ratios for death were 3.19 (P = .07), 6.49 (P = .004), and 5.17 (P = .07), for the N-, H+, and H- groups, respectively, versus the N+ group. For rehospitalization time, the hazard ratios were 7.02 (P = .01), 4.24 (P = .08), and 13.43 (P = .005) for the N-, H+, and H- groups, respectively, versus the N+ group.
Conclusion: Decreasing body temperatures can predict readmission, decreased time to rehospitalization, and (in combination with hypothermia) decreased survival.