(DCP061) STABILITY OF HEMOGLOBIN A1C VALUES OVER TIME AMONG ADULTS WITH TYPE 1 DIABETES: A MULTISTATE MODEL FROM A RETROSPECTIVE POPULATION-BASED COHORT STUDY
Thursday, October 26, 2023
15:45 – 16:00 EST
Location: ePoster Screen 12
Disclosure(s):
Alanna Weisman, MD, PhD: No financial relationships to disclose
Background: A Hemoglobin A1c (HbA1c) ≤7.0% is recommended by multiple international societies for adults with type 1 diabetes (T1D) to reduce the risk of long-term complications. However, HbA1c variability within individuals over time has not been well-studied. Whether HbA1c values remain relatively constant or fluctuate substantially has implications for clinical practice and potentially for risk of complications. Therefore, we evaluated changes between HbA1c categories (‘states’) among adults with T1D over time.
METHODS AND RESULTS: We performed a secondary analysis of a population-based retrospective cohort study of adults with T1D using administrative health data in Ontario from April 1, 2012 to March 31, 2019. T1D was identified using a previously validated algorithm and HbA1c values were obtained from the Ontario Laboratory Information System (OLIS), which captures 99% of HbA1c results. Using homogeneous Markov multistate models, we estimated transition intensity rates between four HbA1c categories (≤7.0%, 7.1-8.0%, 8.1-9.0%, >9.0%), and determined probabilities of future HbA1c categories at 1, 3, 5 and 7 years, given the initial HbA1c category.
24,315 adults with T1D [12,062 (50.1%) female] had a median age of 27 years [interquartile range 19-40] and median diabetes duration of 13 years [7-19]. There were 273,432 HbA1c tests (median 2 per individual per year (IQR 1-3)) and the mean HbA1c value was 8.07% (95% CI 8.05 to 8.08%) in 2013 (fiscal year), rising to 8.11% (95% CI 8.09-8.12%) in 2019. The model-predicted probability of being in a given HbA1c category in one-year was similar regardless of the baseline HbA1c category: 17% probability for HbA1c ≤7.0%, 36% probability for HbA1c 7.1-8.0%, 33% probability for HbA1c 8.1-9.0%, and 14% probability for HbA1c >9.0%. Probabilities were similar at 3, 5 and 7 years of follow-up. Individuals with an HbA1c 7.1-8.0% were 23% more likely for their subsequent HbA1c to be ≤7.0% than 8.1-9%, and those with an HbA1c 8.1-9.0% were 28% more likely for their subsequent HbA1c to be 7.1-8.0% than >9.0%.
Conclusion: There was a high probability of future HbA1c values being 7.1-8.0% or 8.1-9.0%, as opposed to ≤7.0% or >9.0%. For individuals whose initial HbA1c met the guideline-recommended target of ≤7.0%, the probability of future HbA1c values continuing to be ≤7.0% was only 17%. While low attainment of glycemic targets has been previously demonstrated, our results suggest that few individuals meeting the guideline-recommended HbA1c target are able to sustain optimal glycemic levels over the short- and long-term.