This is the first analysis designed to explore the efficacy and safety of once weekly dulaglutide 1.5 mg and 0.75 mg compared with glimepiride or insulin glargine by subgroups of HbA1c (< 8.5% or ≥ 8.5%) in Chinese patients with T2DM. In this subgroup analysis, dulaglutide 1.5 mg and 0.75 mg treatments resulted in clinically meaningful reductions in HbA1c compared with GLIM/GLAR, with greater HbA1c reductions in patients with higher baseline HbA1c (≥ 8.5%) compared with patients with lower baseline HbA1c (< 8.5%). Our observation that poorly glycemic-controlled patients with HbA1c ≥ 8.5% had greater HbA1c reductions compared with patients with HbA1c < 8.5% is consistent with published data that patients with lower HbA1c tend to experience smaller treatment-induced changes in HbA1c than those with higher HbA1c at baseline [13,14,15]. This was also reported for liraglutide by Henry et al. . In that post hoc analysis of phase III randomized controlled trials, patients with T2DM were stratified by HbA1c values into five categories (≤ 5%, > 7.5–8.0%, > 8.0–8.5%, > 8.5–9.0%, and > 9.0%), and reductions in HbA1c levels with liraglutide were generally greater in groups with higher baseline HbA1c . Similar findings were also reported for lixisenatide, with greater reductions in HbA1c in patients with higher baseline HbA1c levels, as shown in a pooled analysis of the lixisenatide GetGoal studies . The observed effect is due to the greater potential for improvement in glycemic control in patients with higher baseline HbA1c. Similar relationships between HbA1c levels and improvements in glycemic control have also been identified in previous reports of meta-analyses of various glucose-lowering therapies , meta-analyses of GLP-1RAs other than dulaglutide , and global phase III studies of dulaglutide . Similarly, the FBG reductions were consistent with HbA1c reductions in both dulaglutide doses.
Post hoc analyses of the dulaglutide clinical development program/global AWARD studies (AWARD-1 to -6 and -8 clinical trials), which included mainly Caucasian patients with T2DM, demonstrated significant improvements in glycemic control irrespective of HbA1c levels, with greater HbA1c and FBG reductions in patients with a higher baseline HbA1c . The dulaglutide clinical development program showed that patients with HbA1c ≥ 8.5% had greater HbA1c reductions than patients with baseline HbA1c < 8.5%, (≥ 8.5%: LS mean − 1.86% [95% CI − 1.97, − 1.75]; < 8.5%: LS mean − 1.02% [95% CI − 1.12, − 0.93]) . Also, global AWARD studies showed that reductions in FBG were consistent with HbA1c changes. The findings of this post hoc analysis of Chinese data are consistent with the global AWARD studies. In AWARD-CHN1, the HbA1c reductions at 26 weeks with baseline HbA1c < 8.5% and ≥ 8.5%, respectively, were dulaglutide 1.5 mg: − 1.1% and − 2.2%; dulaglutide 0.75 mg: − 0.9% and − 2.0%; glimepiride: − 0.7% and − 1.4%. In AWARD-CHN2, the HbA1c reductions at 26 weeks with baseline HbA1c < 8.5% and ≥ 8.5%, respectively, were dulaglutide 1.5 mg: − 1.2% and − 2.3%; dulaglutide 0.75 mg: − 1.0% and − 1.7%; insulin glargine: − 0.6% and − 1.7%. In the dulaglutide and active comparator treatment groups (GLIM/GLAR), the patients with HbA1c ≥ 8.5% experienced greater reductions in HbA1c than patients with HbA1c < 8.5%. Moreover, the present analysis showed dulaglutide 1.5 and 0.75 mg treatments resulted in a greater proportion of patients achieving HbA1c values of less than both 6.5% and 7% at 26 weeks compared with active comparators (GLIM/GLAR). Also, in this post hoc analysis, the HbA1c < 8.5% subgroup had a greater proportion of patients achieving HbA1c values of less than both 6.5% and 7% at 26 weeks compared with the HbA1c ≥ 8.5% subgroup. This indicates that early initiation of dulaglutide treatment when the HbA1c target is not being met may lead to a better patient outcome.
In this post hoc analysis, dulaglutide-treated patients experienced GI AEs more frequently than active comparator-treated patients; however, all the observed GI AEs were reported as mild to moderate intensity and well tolerated. The incidences of total hypoglycemia, asymptomatic, and nocturnal hypoglycemia were lower with dulaglutide than GLIM/GLAR in both subgroups, with low incidence of hypoglycemia in patients with baseline HbA1c ≥ 8.5% compared with the patients with baseline HbA1c < 8.5%. Moreover, incidences of hypoglycemia were higher in insulin glargine-treated patients with lower HbA1c than in patients with higher HbA1c. Overall, in this post hoc analysis, both doses of dulaglutide were well tolerated and the safety profile of dulaglutide was similar to the GLP-1RA class of drugs, suggesting a favorable benefit-to-risk profile for dulaglutide. The findings of the present post hoc analysis are consistent with the findings from global studies (AWARD program)  with dulaglutide and with those from other studies with published data for other GLP-1RAs [20, 21]. The safety and tolerability profile of dulaglutide is similar to that of other agents in the GLP-1RA class [11, 12, 20, 21]. The most common side effects are GI related and include nausea, vomiting, diarrhea, and abdominal distension [11, 12]. Also consistent with the GLP-1RA class, GI side effects are mostly mild to moderate, occur early in the course of treatment, and are transient.
The present post hoc analysis has some limitations. Pooled analyses of data of both the included AWARD-CHN studies were not conducted to prevent the confounding effect of the various concomitant background medications used in each study. Additional limitations include: an imbalanced sample size and smaller number of patients with HbA1c ≥ 8.5% compared with HbA1c of ≤ 8.5%.