Abstract
BACKGROUND: Hospitalization indicates the presence of severe disease and constitutes a leading cost in health care. We aimed to prospectively assess if prevalent diabetes mellitus contributes to excess all-cause and cause-specific hospital admissions and mortality at the population level. METHODS: We used a Swedish prospective population-based cohort, including 25,642 individuals of whom 4.2% had prevalent diabetes at baseline (mean age 61.2 ± 6.8 years, age range 44.8-73.4 years). We compared the number of hospitalizations and mortality classified according to the main chapters of the 10th revision of the International Classification of Diseases (ICD-10) during follow-up using nationwide inpatient registries, comparing individuals with and without prevalent diabetes, using multivariate adjusted negative binomial regression (incidence rate ratio, IRR) and Cox regression, respectively. RESULTS: During a median follow-up of 19 years, 18,904 subjects were hospitalized at least once [median 3 (IQR 2-6)] and 6767 (26.4%) individuals died. Overall, subjects with diabetes were hospitalized (IRR 1.83, p < 0.001) more often, and had a higher incidence rate of hospital admissions due to endocrine diseases (IRR 14.6, p < 0.001), dermatological diseases (IRR 3.7, p < 0.001), injuries and poisoning (IRR 2.7, p < 0.001), infectious diseases (IRR 2.5, p < 0.001), psychiatric diseases (IRR 2.0, p < 0.001), but also cardiovascular, hematological, genitourinary, neurologic and respiratory diseases compared with non-diabetic individuals. No difference was observed for hospital admissions due to cancer or musculoskeletal disorders. All-cause mortality was higher (HR 1.77, p < 0.001) in individuals with diabetes, with disease-specific mortality being significant only for cardiovascular and endocrine disease-related death. CONCLUSIONS: At the population level, prevalent diabetes increased the hospitalization burden longitudinally due to diseases of most of the ICD-10 main chapters, except for cancer and musculoskeletal disorders. These novel findings challenge the current view on the spectrum of prevalent diabetes-related conditions and may have implications for screening and treatment strategies in diabetes. PERMISSION OF GRAPHICAL ILLUSTRATIONS: Source: Pixabay.com. No permission or acknowledgement is required.