Time-varying risk of microvascular complications in latent autoimmune diabetes of adulthood compared with type 2 diabetes in adults: a post-hoc analysis of the UK Prospective Diabetes Study 30-year follow-up data (UKPDS 86)

Background: Latent autoimmune diabetes of adulthood (LADA) differs in clinical features from type 2 diabetes. Whether this difference translates into different risks of complications remains controversial. We examined the long-term risk of microvascular complications in people enrolled in the UK Prospective Diabetes Study (UKPDS), according to their diabetes autoimmunity status. Methods: We did a post-hoc analysis of 30-year follow-up data from UKPDS (UKPDS 86). UKPDS participants with diabetes autoantibody measurements available and without previous microvascular events were included. Participants with at least one detectable autoantibody were identified as having latent autoimmune diabetes, and those who tested negative for all autoantibodies were identified as having type 2 diabetes. The incidence of the primary composite microvascular outcome (first occurrence of renal failure, renal death, blindness, vitreous haemorrhage, or retinal photocoagulation) was compared between adults with latent autoimmune diabetes and those with type 2 diabetes. The follow-up ended on Sept 30, 2007. Baseline and updated 9-year mean values of potential confounders were tested in Cox models to adjust hazard ratios (HRs). UKPDS is registered at the ISRCTN registry, 75451837. Findings: Among the 5028 participants included, 564 had latent autoimmune diabetes and 4464 had type 2 diabetes. After median 17·3 years (IQR 12·6–20·7) of follow-up, the composite microvascular outcome occurred in 1041 (21%) participants. The incidence for the composite microvascular outcome was 15·8 (95% CI 13·4–18·7) per 1000 person-years in latent autoimmune diabetes and 14·2 (13·3–15·2) per 1000 person-years in type 2 diabetes. Adults with latent autoimmune diabetes had a lower risk of the composite outcome during the first 9 years of follow-up than those with type 2 diabetes (adjusted HR 0·45 [95% CI 0·30–0·68], p<0·0001), whereas in subsequent years their risk was higher than for those with type 2 diabetes (1·25 [1·01–1·54], p=0·047). Correcting for the higher updated 9-year mean HbA1c seen in adults with latent autoimmune diabetes than in those with type 2 diabetes explained entirely their subsequent increased risk for the composite microvascular outcome (adjusted HR 0·99 [95% CI 0·80–1·23], p=0·93). Interpretation: At diabetes onset, adults with latent autoimmune diabetes have a lower risk of microvascular complications followed by a later higher risk of complications than do adults with type 2 diabetes, secondary to worse glycaemic control. Implementing strict glycaemic control from the time of diagnosis could reduce the later risk of microvascular complications in adults with latent autoimmune diabetes. Funding: European Foundation for the Study of Diabetes Mentorship Programme (AstraZeneca).