A relative change in HbA1c over time is more predictive of type 1 diabetes (T1D) in the paediatric population than a single measure of HbA1c ≥6.5% (48 mmol/mol), the ADA 80th Scientific Sessions was told.
Professor Kendra Vehik, an epidemiologist at the University of South Florida and an investigator in The Environmental Determinants of Diabetes in the Young (TEDDY) study, said the use of HbA1c has been adapted from adult and T2D settings.
“But we have to remember that children are not small adults and any tool used for diagnosis should not be extrapolated from adult observational studies … without further studies in paediatric and adolescent populations,” she said.
Professor Vehik said earlier work using data from TEDDY, TrialNet, DPT-1 and TRIGR studies had shown a single HbA1c ≥6.5% (48 mmol/mol) was specific but not sensitive for diagnosing T1D in high-risk young people.
The study found the probability of an asymptomatic child with HbA1c above the threshold actually having T1D varied from 33% to 75%.
To revisit the question about how HbA1c could better predict progression to T1D, the TEDDY team looked at data from 720 children in the cohort who had developed multiple islet autoantibodies by the end of 2019 and for which two or more HbA1c measures were available.
The mean age of the children was 11 years and about a third of the children had progressed to T1D.
The study found children who had an increase in HbA1c levels over time had an increased risk of progressing to T1D. The number of autoantibodies, age and baseline HbA1c were all associated with a changing HbA1c.
Children with 10% relative change in HbA1c or a 0.5% absolute point change in HbA1c had a 6-fold increased risk of progression to T1D while children with a 20% relative change in HbA1c had a 23-fold increased risk of progression to T1D.
There was no difference between children who were asymptomatic at the time of diagnosis and those who presented with symptoms such as polyuria, polydipsia or weight loss.
The study also noted that the speed of change in HbA1c (<6 mths, 6-12 months, 1-2 years or >2 years) was not significantly associated with progression to T1D.
Validated in other studies
Professor Vehik said their finding – that a relative 10% change in HbA1c was a useful, non-invasive predictive measure of progression to T1D in a young population with islet autoimmunity – was also validated in the cross-sectional DPT-1 and TrialNet studies.
And its performance was similar to a 2-hour OGTT in children with multiple islet autoantibodies.
The findings when validated in TrialNet and DPT-1 cohorts were largely consistent although a rapid speed of change in HbA1c was found to modify the risk of progression to T1D in children from TrialNet and DPT-1 who were older than TEDDY’s cohort (10 v 5 years at baseline).
Professor Jennifer Couper, who leads Australia’s Environmental Determinants of Islet Autoimmunity (ENDIA) study, told the limbic that better understanding of the stages of progression to T1D was important to identify opportunities to intervene.
“Even if you can delay type 1 diabetes by 5-10 years, that is an enormous benefit,” she said.
“We are increasingly realising the difference over time and the difference between people; it’s quite heterogeneous. Age is very important but also their immune response.”
“So if they are close to type 1 diabetes, some interventions will work better. If they’re not close, other interventions work better. The more we understand what is happening at what speed, the better we can tailor future interventions for that child or adult. It won’t be one-size-fits-all.”
Professor Couper said in Australia children at risk of type 1 diabetes who are followed in research cohorts have glucose and HbA1c measurements regularly and soon will have access to continuous glucose monitoring.