Seasonal Patterns and Severe Cold Exposure in Heart Failure With Mildly Reduced Ejection Fraction

By
Henning Johann Steffen
Michael Behnes
Daniel Duerschmied
Ibrahim Akin
Tobias Schupp
October 13, 2026

Clinical Research In Cardiology

Overview

In this retrospective study of 2,184 patients hospitalized with heart failure with mildly reduced ejection fraction (HFmrEF), admission season was not associated with 30-day mortality. Admission during extreme cold, however, was associated with approximately twice the adjusted hazard of 30-day death; extreme heat and large day-to-day temperature fluctuations were not associated with mortality.

Background

Seasonal variation in heart failure outcomes has been reported, with winter mortality peaks in studies from France and the United States and a summer peak in a Japanese study. Much of the prior evidence concerns unselected heart failure populations and broad seasonal categories rather than measured temperatures. HFmrEF is defined in the article as LVEF 41–49%, and its seasonal profile and short-term prognosis have been less studied. The article also notes that the 2026 ESC guidelines incorporated HFmrEF into an enlarged HFrEF spectrum with LVEF below 50%.

Data Highlights

Measure	Result
Study population and period	2,184 consecutive patients hospitalized with HFmrEF at University Medical Centre Mannheim, January 2016–December 2022
Admissions by season	Spring: 556 (25%); summer: 548 (25%); autumn: 518 (24%); winter: 562 (26%)
Thirty-day deaths	92 (4.2%)
Season and 30-day mortality	No association identified; log-rank p = 0.809
Extreme cold on admission (≤0.8 °C)	9/105 deaths (8.6%) versus 79/1,972 (4.0%) in the reference group; adjusted HR 2.012 (95% CI 1.003–4.033; p = 0.049)
Extreme cold, preceding 3-day mean (≤1.33 °C)	Adjusted HR 2.058 (95% CI 1.028–4.118; p = 0.042)
Extreme heat on admission (≥23.3 °C)	4/107 deaths (3.7%); adjusted HR 0.849 (95% CI 0.310–2.325; p = 0.750)
Top 5th percentile of day-to-day temperature fluctuation	Adjusted HR 0.854 (95% CI 0.313–2.328; p = 0.757)

Key Findings

Admissions were distributed relatively evenly across the four seasons, and monthly counts showed no discernible winter predominance.
Most measured clinical characteristics and discharge heart failure pharmacotherapies were similar across seasons; diabetes and current smoking differed statistically but showed no consistent seasonal gradient.
Season of admission was not associated with 30-day all-cause mortality (log-rank p = 0.809).
Admission during extreme cold (≤0.8 °C) was associated with higher 30-day mortality after adjustment (HR 2.012; 95% CI 1.003–4.033; p = 0.049).
Extreme cold defined by the preceding 3-day mean temperature (≤1.33 °C) showed a comparable adjusted association with 30-day mortality (HR 2.058; 95% CI 1.028–4.118; p = 0.042).
Extreme high temperature and top-5th-percentile day-to-day temperature fluctuation were not associated with 30-day mortality in the reported analyses.

Clinical Implications

The study reports an association between extreme cold exposure and 30-day mortality among hospitalized patients with HFmrEF, while broad seasonal categories were not associated with mortality. These observational findings do not establish causality; the source does not report a tested intervention or make a specific clinical management recommendation.

Conclusion

In this single-centre retrospective cohort, season of admission did not predict 30-day mortality in HFmrEF, whereas extreme cold at admission and in the preceding 3-day mean was associated with higher adjusted mortality. The findings distinguish measured cold exposure from broad seasonal classification.