Abstract Body (Do not enter title and authors here): Introduction
Numerous studies have shown a U-shaped relationship between ambient temperature and cardiovascular mortality. However, few studies examined a temperature- heart failure (HF) hospitalization relationship or its regional variations in the United States (US).

Research Question
How does ambient temperature impact HF hospitalizations in older US adults? Does the temperature-HF relationship vary by region?

Methods
We conducted a place and time-stratified case-crossover study of older adults (age ≥ 66 years) hospitalized for HF between 2007-2022 using 20% national Medicare data. Daily mean temperatures (T_mean) were identified from the Parameter-elevation Relationships on Independent Slopes Model linked through residential zip code. HF cases were defined as hospitalization with a primary HF diagnosis. Up to 5 controls were selected as same day of week, month, and year as cases. We used conditional logit for distributed lag non-linear models in overall population and subgroups by region (Northeast, West, South, Midwest). We used 14-day lag for the primary analysis, which was varied from 2-28 days.

Results
Among 1,517,756 HF hospitalization cases, the effect of cold T_mean on HF hospitalizations is delayed and peaked ~4-8 days after exposure; the risk ratio (RR) of temperatures < 50°F vs. 65°F ranged from 1.01-1.07 (Figure 1). The cumulative risk of colder T_mean were only significant in longer lag models (24°F vs. 65°F –lag 7: RR 0.98 [0.96–1.00]; lag 14: RR 1.06 [1.03–1.08]; lag 21: RR 1.1[1.07–1.13]). Overall, heat had minimum and non-significant effects that were near immediate (lag 0-2) (Figure 1).

In regional analyses (Figure 2), the magnitude of colder T_mean effect varied (23°F vs. 65°F – Northeast: RR 0.96[0.92–1.01]; West: RR 1.34[1.21–1.50]; South: RR 1.01[0.96–1.06]; Midwest: RR 1.08[1.03–1.13]) The effect of high T_mean also varied; heat had non-significant increase in HF hospitalization in the Northeast region only with max RR of 1.42 (100°F vs. 65°F).

Conclusion
Colder T_mean was associated with increased HF hospitalizations. The elevated risk was apparent approximately 4 days after exposure; the magnitude varied by region. Extreme heat’s association with HF hospitalization was indicated in Northeast region and confined within 0-2 days after exposure. The regional variation may reflect acclimatization to regional climate, available local resources, and varied behavioral responses to temperature.