Association of Low Body Temperature and Poor Outcomes in Patients Admitted With Worsening Heart Failure

A Substudy of the Efficacy of Vasopressin Antagonism in Heart Failure Outcome Study With Tolvaptan (EVEREST) Trial

Saeed Payvar; John A. Spertus; Alan B. Miller; S. Ward Casscells; Peter S. Pang; Faiez Zannad; Karl Swedberg; Aldo P. Maggioni; Kimberly J. Reid; Mihai Gheorghiade


Eur J Heart Fail. 2013;15(12):1382-1389. 

In This Article


This is the first pre-specified analysis of a large clinical trial to report the association of low body temperature and worse outcomes in patients hospitalized due to worsening HF. Among 4108 patients followed for an average of 9.9 months after discharge, lower body temperature at randomization was associated with increased risk of CV death or HF rehospitalization in a linear fashion. Lower body temperature was modestly associated with multiple other indicators of poor prognosis; nonetheless, it had an independent association with worse outcomes after adjustment for age, gender, race, SBP, EF, BUN, and serum sodium.

In the setting of HF, the finding of cold extremities has been used, in combination with signs of congestion, to identify haemodynamic profiles that predict outcomes in patients admitted with HF.[17] Additionally, these haemodynamic profiles have been used to tailor therapy in the care of HF patients.[18] This study extends these previous observations by focusing upon body temperature as a continuous estimator of risk and showing its independent prognostic significance. Although body temperature is readily available in most settings, lower than normal values are often ignored; and this study encourages further attention to this simple and readily available finding.

The findings of this study complement the findings of previous studies on the association of low body temperature and outcome in HF patients. Reporting a retrospective analysis on 291 HF patients hospitalized for HF exacerbation, Casscells and colleagues found that low body temperature at admission (< 35.3°C) was associated with significantly higher in-hospital mortality, with an adjusted HR of 4.46, which was far larger than the effect we observed.[8] In a retrospective analysis of the Acute and Chronic Therapeutic Impact of a Vasopressin Antagonist in Congestive Heart Failure (ACTIV in CHF) trial, low body temperature at randomization (< 35.8°C) appeared to be a strong predictor of 60-day mortality after adjustment for other covariates, with an adjusted HR of 3.07.[12] Similarly, in an analysis of the data of 61 522 HF patients from the National Heart Care Project, patients with body temperatures below 36°C at admission had increased in-hospital (adjusted RR 1.28, P < 0.001) and 1-year mortality (adjusted RR 1.14, P < 0.001), with effect sizes much more similar to our findings.[19] Importantly these studies used different endpoints, and some did not adjust for many of the covariates that we were able to include in our analyses.

The biological mechanism underlying the association of low body temperature and outcomes remains unclear. Excessive levels of angiotensin II may be a potential link, as elevated angiotensin II can result in hypothermia[20] via its effects on its central nervous system receptors.[21] Decreased thermogenesis that results from wasting seen in severe HF is another possible mechanism that can explain the association of low body temperature and worse outcomes. Wasting results in decreased skeletal muscle activity and adipose tissue mass, both of which play a significant role in thermogenesis. The association of low body temperature and low body weight found in this study may support this possible link.

The independent association of low body temperature and outcome in HF patients, if corroborated in further studies, may have therapeutic implications as well. Low body temperature may be part of a vicious circle that contributes to HF progression. Artificially induced core body cooling increases plasma levels of norepinephrine;[22] and increased norepinephrine in HF has been associated with worse outcomes. Similarly, a colder ambient environment seems independently to compromise HF patients.[23] Interestingly, warming HF patients appears to improve their symptoms, and their neuroendocrine and haemodynamic status.[24]


Our findings should be interpreted in the context of the following potential limitations. In this study, it was not possible to evaluate the association of low body temperature and other biomarkers that are elevated due to neurohormonal activation, such as BNP. BNP was measured inconsistently in EVEREST, with the use of both BNP and pro-BNP during different phases of the study, and was not included in our analyses. Additionally, temperature measurements were obtained at randomization, which was within 48 h. This may have affected the association, because treatments provided in the first 48 h of admission may have affected body temperature. Moreover, the results may not be generalizable to temperature measurements obtained at other time points, such as hospital admission and discharge.