Identifying insulin treatment responders with a composite measure: beyond Hba1c < 7% in patients with type 2 diabetes.

Objectives: Many insulin-treated patients with type 2 diabetes (T2D) do not reach hemoglobin A1c (HbA1c) < 7%, but have clinically relevant HbA1c reductions. Using an integrated database (IDB) of 53 insulin lispro clinical trials and a real-world evidence (RWE) database of T2D patients initiating insulin therapy, an expanded HbA1c measure was used to identify responders to insulin therapy.

Methods: Analysis included 4,908 patients (IDB) and 1,134 patients (RWE) with T2D treated with any insulin regimen with a baseline and ≥1 post-baseline HbA1c. Responders were defined as patients with endpoint HbA1c < 7% (cut point [CP]) and/or either ≥1% absolute (ABS) decrease, and/or ≥10% relative (REL) decrease in HbA1c from baseline. The percentage of responders with CP vs ABS and concordance between ABS and REL were calculated. As the ABS and REL measures were highly correlated (94%), the ABS measure was used to compare characteristics of responders and non-responders by age, diabetes duration, race/ethnicity, baseline HbA1c, and insulin regimen at 24 weeks.

Results: In both databases, more responders were identified with ABS or REL (>62%) than CP (<41%). More ABS responders had a baseline HbA1c ≥ 9% and a shorter diabetes duration than non-responders. Basal insulin-treated patients in the IDB had 78.2% responders at 24 weeks, compared to 69.7% with basal/bolus or pre-mixed insulin (75.4%). Results were similar in the IDB and RWE.

Conclusion: Composite HbA1c measures identified more patients with clinically meaningful responses to therapy than the broadly accepted HbA1c < 7% and may be useful in assessing clinical trials, clinical care, and quality measures.