HbA_1c cutpoint predicts pediatric T1DM within a year

ORLANDO – Among children with genetic risks for type 1 diabetes and autoantibodies against pancreatic islet cells, a hemoglobin A_1c at or above 5.6% strongly predicts the onset of type 1 diabetes within a year, according to investigators from The Environmental Determinants of Diabetes in the Young (TEDDY) study.

^{M. Alexander Otto/MDedge News}

Michael Killian (left) and Dr. William Hagopian

The risk is even greater if other factors – age at A_1c test, month of A_1c test, continent of residence, and the z-score for one antibody in particular, IA2A, a tyrosine phosphatase antigen protein – are taken into account, with an area under the curve (AUC) of about 0.97. The study team is developing an online risk calculator for clinicians so they can simply plug in the numbers.

But the 5.6% cutpoint works well even by itself, with an AUC of about 0.86. Among the children with genetic risk factors and islet cell autoantibodies who hit that mark in the study, the median time to diagnosis of type 1 diabetes mellitus (T1DM) was 7.1 months.

The point of the work is to catch disease onset early to prevent children from going into diabetic ketoacidosis.

The ultimate goal of TEDDY is to identify infectious agents, dietary factors, and other environmental agents that either trigger or protect against T1DM in genetically susceptible children, to develop strategies to prevent, delay, or even reverse T1DM. For more than a decade, the consortium has been following more than 8,000 children who screened positive at birth for genetic anomalies in the human leukocyte antigen region on their 6th chromosome, a major risk factor for T1DM. The team is following them to find out what puts them over the edge; A_1c seems to be key.

Catching the disease early is a goal. Until now, the TEDDY team has used the development of pancreatic islet cell autoantibodies to trigger oral glucose tolerance tests (OGTTs) every 6 months.

The problem is that islet antibodies “are great at saying that you are going to get diabetes but not when. We have kids who have been multiple antibody positive now for 6, 7, 8 years” but who still haven’t developed T1DM. In the meantime, their parents have been dragging them in for OGTTs every 6 months for years, said lead investigator and TEDDY study coordinator Michael Killian, of the Pacific Northwest Research Institute, Seattle.

With the A_1c cutpoint, it should be safe to hold off on direct glycemic surveillance until A_1c levels reach 5.6%, and even safer when the full risk prediction model with 1A2A z-scores and other factors are available. Once children hit the parameters, it’s near certain they will develop T1DM within a year, so the findings should reduce the number of OGTT tests to just two or three. In the end, “we’ll get better compliance” with testing and still “catch these kids early,” Mr. Killian said at the annual scientific sessions of the American Diabetes Association.

Rare is the newborn who is screened for diabetes risk at birth, and even rarer is the child who is followed for autoantibodies. If the findings hold out, however, such early action could be the future.

“We envision a time” when newborns are screened for T1DM risk and followed for autoantibodies. Glycemic surveillance will kick in if they hit A_1c levels of 5.6%, said William Hagopian, MD, PhD, clinical associate professor at the University of Washington, Seattle, and a TEDDY principal investigator.

Among the 8,000-plus children followed, 456 had persistent islet autoantibodies and at least three A_1c tests in the prior 12 months; 104 progressed to T1DM and 352 did not. The mean age to islet autoantibody seroconversion was 43 months, and mean age to T1DM diagnosis was 73 months. The team ran a receiver operating curve analysis to identify the optimum A_1c cutpoint, 5.6%.

The month of testing mattered, because “glycemia is higher in the winter, lower in the summer, and insulin sensitivity is lower in the winter, and higher in the summer,” Dr. Hagopian said. It probably has something to do with what the stresses of early human evolution bred into the genes.

The IA2A autoantibody z-score is the only number that will be cumbersome to plug into the upcoming online risk calculator. It’s the number of standard deviations off your reference lab’s mean. “You have to talk to the lab that does your IA2As to know what the standard deviation is,” Dr. Hagopian said.

The investigators had no disclosures. TEDDY is supported by the National Institutes of Health, among other entities.

aotto@mdedge.com

SOURCE: Killian M et al. ADA 2018, Abstract 162-LB