Are different phenotypes of type 2 diabetes associated with different risks of vascular outcomes? A cohort study in east London

Problem

Background: Recent studies have identified distinct subtypes of type 2 diabetes (T2DM). Characterizing diabetes phenotypes in a young, ethnically diverse inner-city population and their associations with vascular risk is an essential step towards tailoring diabetes management for high-risk groups. The aim of this study was to identify and characterize subtypes of type 2 diabetes based on clinical characteristics at diagnosis and determine associations with vascular outcomes.

Approach

An observational cohort study using data from the east London Database was conducted using 30,000 adults aged 18 and over diagnosed with T2DM between 2008 and 2018. Latent class analysis was used to identify distinct subgroups of type 2 diabetes using baseline data on HbA1c, age at diagnosis and BMI. The ideal number of classes was selected based on model fit criteria and clinical relevance. Multinomial logistic regression was used to identify demographic and baseline characteristics associated with class membership. Amongst those free from prevalent vascular disease, multivariable cox-proportional hazards regression was used to compare risk of incident cardio- and micro-vascular disease between classes.

Findings

Three clinically meaningful T2DM classes were identified; 70% of the population were classified as having ‘Typical diabetes’; 22.8% were classed as having ‘Age-related diabetes’ (mean age at diagnosis 72.7), while 8.1% were classed as having ‘Severe hyperglycaemia related diabetes’ (mean HbA1c at diagnosis 12.1%/108.7 mmol/L. The majority of individuals in the age-related group were of white ethnicity (42.9%), while the other two classes were largely south Asian (43.2% severe hyperglycaemia, 56.8% typical onset). Compared to those with typical onset, risk of macrovascular disease was 2.75 times higher in the age-related diabetes group (95%CI 2.42-3.14). Risk of microvascular disease was 29% higher in the age-related diabetes group (95%CI 1.21-1.36) and 41% higher in the severe hyperglycaemia group (95%CI 1.30-1.52) relative to those with typical onset.

Consequences

We characterized differences between individuals belonging to three distinct classes of type 2 diabetes and identified strong associations between some diabetes subtypes and vascular risk. Further exploration into the clustering and trajectories of cardio-metabolic risk factors, both around the time of diagnosis and over the longer term will be key to developing targeted strategies to identify and manage those at highest risk of poor outcomes.