Incidence of diabetes mellitus following hospitalisation with influenza: a population-based cohort study in England

OBJECTIVES: To establish the incidence of developing diabetes mellitus (DM) post hospitalisation with influenza. DESIGN: Retrospective cohort study. SETTING: Electronic healthcare records from Clinical Practice Research Datalink (CPRD) linked to Hospital Episode Statistics in England. PARTICIPANTS: 13 710 adults with a first episode of hospitalised influenza as the primary cause for admission between 1 July 2004 and 1 March 2021 based on ICD-10 codes without pre-existing DM were included. A randomly selected group (a) from CPRD records matched for age, sex and General Practitioner (GP) practice and (b) an unmatched group of hospitalised sepsis patients were used as comparator groups. OUTCOME MEASURES: Patients were followed from 1 day after discharge till either DM diagnosis, death or end of GP record. HRs for incidence of DM were calculated using adjusted Cox regression models. RESULTS: Incidence of DM was 12.5 per 1000 person years. Adjusted HRs (aHR) for developing DM after hospitalised influenza compared with matched controls was 1.54 (95% CI 1.39 to 1.70, p<0.001) and to hospitalised sepsis comparators 1.14 (95% CI 1.03 to 1.26, p=0.013). The greatest risk for developing DM in hospitalised influenza patients was within 90 days of discharge (aHR 2.71 (95% CI 1.94 to 3.77, p<0.001)) compared with matched controls. Risk factors for DM after influenza hospitalisation included being male, pre-existing DM risk, obesity and critical care admission during acute illness. CONCLUSION: Patients' post hospitalisation with influenza had a greater incidence of DM when compared with both matched controls and patients following hospitalisation with sepsis.