Effect modification of diabetes as a comorbidity on temperature variability and hospitalization for cardiovascular disease

Background and aim: Evidence has accumulated that large temperature variability (TV) is associated with an increased risk for hospitalization or death. However, no previous studies examined whether the association for TV is modified by diabetes comorbidity, which is particularly vulnerable to cardiovascular disease (CVD). We aimed to explore the short-term association between TV and CVD hospitalization, stratified by presence of diabetes as a comorbidity in 47 prefectures of Japan from 2011 to 2018. Methods: A two-stage meta-analysis was used. First, we used a time-stratified case-crossover design to estimate the association between TV and CVD admissions with and without diabetes comorbidity for each prefecture. We then pooled the estimates using a multivariate meta-regression model across the prefectures. The associations were also stratified by sex, age, smoking status, and season. Results: A total of 4,482,583 hospitalizations for CVD were included in this study. Each 1°C increase in 0–7 days TV exposure was associated with a 0.58% (95% confidence interval, CI= 0.20% to 0.96%) increase in risk for total CVD admission and a 2.07% (95% CI= 1.16% to 2.99%) increase in risk for heart failure admission in subjects with diabetes, while the respective increases in subjects without diabetes were 0.37% (95% CI= 0.12% to 0.63%) and 0.61% (95% CI= -0.02% to 1.23%). These patterns were broadly similar to those stratified by age, sex, and smoking status, while the association for TV was greater in spring in the season-stratified analysis. Conclusions: Susceptibility to TV in relation to CVD onset may be stronger in subjects with diabetes than those without. Our findings may suggest that subjects with diabetes should be more cautious to sudden temperature changes. Keywords: Temperature variability; Comorbidity.