Aims: Clinical data indicating which patients initiating GLP-1 receptor agonists are likely to achieve desired treatment responses are limited, but baseline (BL) A1C is the strongest predictor of A1C lowering. We aimed to identify factors beyond BL A1C that predicted response.

Methods: In this post-hoc analysis of pooled data from 1414 patients who received exenatide twice daily in 8 randomized controlled trials, we divided the population into higher, average, and lower A1C responders at 6 months, based upon tertiles of linear regression residuals that were corrected for individual participants’ BL A1C (mean A1C 8.14%, 7.71%, and 8.08%, respectively). We examined BL variables including sex, ethnicity, concomitant medications, renal function, age, weight, diabetes duration, fasting glucose (FG), self-monitored blood glucose profile, HOMA-B, HOMA-S and C-peptide by response tertile.

Results: A1C mean changes (95% CI) from BL at 6 months by tertile were: −1.92% (−1.98, −1.86), −0.98% (−1.02, −0.94), and −0.23% (−0.29, −0.17), respectively. Most variables did not differ by tertile. Potentially important individual BL characteristics of higher vs lower responders were older age, Asian ethnicity, dual therapy, lower weight, lower FG, and higher postprandial glucose (PPG) excursion (Table). Multivariate analysis to further explore the observed ethnicity effect indicated that this finding could potentially be explained by the BL pre-breakfast blood glucose (mean: Asians 9.4 mmol/l; non-Asians 9.1 mmol/l) and BL post-breakfast glucose excursion (4.3 and 2.1 mmol/l, respectively).

Conclusions: Asian ethnicity was associated with a greater BL A1C-adjusted response to exenatide twice daily that appeared to be clinically meaningful. The baseline pre-breakfast glucose and the post-breakfast glucose excursion accounted for most of this finding.

Disclosure statement: This study was supported by AstraZeneca.