Introduction
Heart failure (HF) is one of the leading causes of intensive care unit (ICU) admission in Germany, affecting over four million individuals, with incidence expected to rise due to demographic change. Beyond established cardiovascular risk factors, atmospheric conditions such as temperature extremes, pressure fluctuations and air pollutants are recognized short-term triggers of acute cardiac decompensation. However, data on multiple simultaneously acting atmospheric exposures and their sex- and age-specific effects on ICU-level HF admissions remain limited.
Methods
This analysis forms part of the ALERT-ITS project, which evaluates atmospheric constellations as predictors of short-term ICU demand. Anonymized nationwide claims data from Germany's largest statutory health insurer (Allgemeine Ortskrankenkasse, AOK) for 2009 to 2023 were screened for inpatient stays involving HF (ICD-10-GM I11.0, I13.0, I13.2 and I50) that required intensive care. Daily case counts were aggregated for the Southeast German biometeorological region and stratified by sex and age (dichotomised at 60 years). For each day, 31 regionally averaged atmospheric parameters were compiled, combining meteorological fields from the ERA5 reanalysis with pollutant concentrations from the Copernicus Atmosphere Monitoring Service. The relationship between these exposures and daily admissions was estimated using quasi-Poisson generalized additive models, in which lag terms captured delayed effects (up to 10 days for meteorological and 7 days for air quality factors), while smooth terms accounted for seasonality and long-term trend.
Results
Over the 15-year observation period, a total of 54,198 ICU admissions for HF were identified in Southeast Germany (50.6% female, 91.5% aged 60 years or older). Mean annual admission rates were 78 cases per 100,000 insured individuals per year. Nine disease-specific predictors were identified across three exposure domains: pressure, heat, cold, and air quality. In males ≥60 years, summer day (SumDay) (p=0.0001, lag 1d) and heat wave duration (HW-Dur) (p=0.009, lag 1d) were significantly associated. In females ≥60 years, cold stress wave day reached significance (p=0.012, lag 8d). Among younger patients, SumDday and HW-Dur were significant in females <60 years (p=0.011, lag 8d; p=0.019, lag 0d), while tropical night and sea level pressure were significant in males <60 years (p=0.001, lag 9d; p=0.030, lag 2d). Notably, SumDay was associated with a 9.4% reduced admission risk in males ≥60 years (RR 0.906, 95% CI 0.861–0.952).
Discussion
This study provides region-specific evidence on the impact of atmospheric constellations on ICU-relevant HF admissions in Southeast Germany. The predominant vulnerability of older patients, particularly those aged 60 years or older, is consistent with known pathophysiological mechanisms including impaired thermoregulation, diuretic therapy, and reduced cardiovascular reserve. The extended lag periods observed may reflect harvesting effects, consistent with the selective inclusion of ICU-level admissions. These findings underline the potential of integrating atmospheric forecasting into short-term ICU capacity management and contribute to the broader ALERT-ITS framework of weather-informed cardiovascular care.
