Identifying and visualising temporal trajectories of hospitalisations for traditional and non-traditional complications in people with type 2 diabetes: a population-based study.

Background: People with type 2 diabetes are increasingly susceptible to complications that are not specific to diabetes. We aimed to examine the temporal trajectories of hospitalisations for traditional and non-traditional complications in people with type 2 diabetes.

Methods: We included 758,254 people with incident type 2 diabetes between 2002 and 2018 in Hong Kong, followed up until 2019. We included hospitalisations for 72 selected diseases and all-cause deaths. We derived the temporal trajectories of hospitalisations based on pairs of disease associations and identified trajectory clusters using Markov Cluster Algorithm.

Findings: During a median follow-up of 7.8 (IQR: 4-12) years, 57.6% of people experienced a hospitalisation for any of the 72 selected diseases and 22.6% of people died. Among the 5184 directional disease pairs, 95 were identified as having a significant and directional association. The three most common disease pairs were hospitalisations for urinary tract infection followed by pneumonia, ischemic heart disease followed by heart failure, and ischemic stroke followed by pneumonia. Cardiovascular and kidney diseases were predominant in the hospitalisation trajectories. However, these traditional complications had complex associations both among themselves and with various non-traditional complications across multiple systems. Three distinct trajectory clusters were identified, with heart failure/chronic kidney disease, pneumonia, and urinary tract infection as central diseases.

Interpretation: Cardiovascular and kidney diseases interacted with a broad set of non-traditional complications to influence the overall patterns of hospitalisation progression in people with diabetes, highlighting the need to broaden diabetes care to consider complications beyond the traditional focus.

Funding: Direct Grant for Research from The Chinese University of Hong Kong.