In this analysis of real-world clinical data, we demonstrated that dulaglutide significantly improved the glycemic control in Korean patients with T2DM. The glucose-lowering efficacy of dulaglutide was observed either in adding to OAD or adding to basal insulin, and the effect was sustained over a 12-month period. Moreover, dulaglutide induced significant reduction in body weight. Our findings are similar to those of previous RCTs and real-world clinical studies in patients with T2DM who received dulaglutide treatment.
In the present study, dulaglutide exhibited a significant reduction in HbA1c of − 1.13% at 6 months, which was consistent with the results from clinical trials [13–15], and similar or slightly greater than that from real-world clinical studies conducted in Western countries [16–18]. This may be associated with the pharmacokinetic and pharmacodynamic differences of GLP-1 RA in Caucasian and Asians, wherein Asian patients have higher postprandial glucose and postprandial excursion related to impaired β-cell function [19, 20]. We further analyzed the efficacy of dulaglutide separately as add-on to OAD or basal insulin because these two regimens are generally accepted in clinical practice. The glucose-lowering efficacy of dulaglutide therapy was observed in both groups. However, the magnitude of HbA1c reduction was greater in the add-on to OAD than in the add-on to insulin group (− 1.36% vs − 0.74%). Comparable glycemic efficacy of dulaglutide was reported in previous RCTs with dulaglutide in combination with OADs. The HbA1c-lowering effect of dulaglutide 1.5 mg was − 1.42% at 26 weeks with metformin in the Assessment of Weekly AdministRation of LY2189265 in Diabetes-6 (AWARD-6) trial and − 1.40% with sulfonylurea in the AWARD-8 trial [13, 21].
The observed glucose-lowering efficacy of dulaglutide as add-on to insulin (− 0.74%) was lower than those previously reported in the AWARD-9 trial [22]. In the AWARD-9 trial, the mean HbA1c changes from baseline was − 1.44% at 28 weeks in T2DM patients after adding to insulin glargine, and 66.7% of the patients achieved a target goal of HbA1c < 7.0%. Meanwhile, a real-world study in Korean patients revealed a − 0.97% mean change in HbA1c at 6 months after dulaglutide therapy as add-on to basal insulin [11]. These two studies similarly showed a decrease of − 11.7 IU in the total daily insulin dose, whereas our study showed a decrease of − 20.7 IU at 6 months. This disparity may be attributed to several differences in the study population between the studies as well as real-world factors. In our study, a considerable proportion (62.5%) of patients switched the treatment regimen from basal-bolus to basal insulin + dulaglutide in the add-on to insulin group in contrast to the abovementioned studies, and dulaglutide add-on decreased approximately 23.5 IU of insulin. Furthermore, as the insulin dose was self-titrated for the individual patient during dulaglutide treatment, there is a possibility that a marked glucose reduction or hypoglycemia might had led to a subsequently inadequate insulin dose, and the patient’s low adherence may have affected the glycemic control.
We observed a durable effect of dulaglutide over 12 months in the subgroup analysis. The observed reduction of HbA1c from baseline to 12 months was − 0.86%. HbA1c levels reached nadir at 3 months after dulaglutide initiation, and tended to increase until 12 months; however, HbA1c levels from 3 to 12 months were not statistically significant. This trend was consistent with the result from the AWARD trials and real-world studies [9, 23–25]. Several studies have evaluated the 12-month glucose-lowering efficacy of dulaglutide in the real-world scenario, and showed reductions in HbA1c of − 0.9 to − 1.0% that were comparable to our data [26, 27]. To the best of our knowledge, this is the first study to report the 12-month efficacy of dulaglutide in Korean patients with T2DM.
We analyzed the clinical predictive factors for a good response to dulaglutide. Higher glycemic parameters at baseline and no previous insulin therapy were significantly associated with a good dulaglutide response on multivariate regression analysis. A previous pilot study reported that low BMI and old age were linked to a good response to dulaglutide in Japanese patients with T2DM [28]. A real-world study in Korean patients demonstrated that a higher baseline HbA1c was the only predictable factor for a good response to dulaglutide [12].
Besides the HbA1c change, a significant improvement in body weight (− 2.9 kg) was observed at 6 months after dulaglutide initiation in overall population. The efficacy of weight reduction in our study is similar to the previous real-world data reported from Western countries, which reported − 2.7 to − 2.9 kg weight reduction at 6 months [16, 29]. Meanwhile, real-world data from Korea showed a weight reduction of − 2.1 kg during 6-month dulaglutide therapy in combination with various glucose-lowering treatments [12], and − 2.0 kg at 6 months with dulaglutide as an adjuvant to basal insulin therapy [11]. These differences may be attributed to the considerable reduction in insulin doses during dulaglutide therapy that leads to a synergic effect on weight loss, as shown in the present study. In addition, several combinations of OADs and selection bias may influence the body weight. In our study, if patients who stopped dulaglutide earlier due to complaints of no weight change, or took medications that conferred a possibility to gain weight, such as high-dose steroids, were excluded from the analysis.
In the present study, the incidence of gastrointestinal side effects, including mainly nausea, vomiting, and dyspepsia, was similar or slightly higher than those reported in previous RCTs and real-world data [25, 30–32]. Self-reported hypoglycemia occurred more frequently in our study (13.4%) than in previous studies and meta-analyses [13, 33, 34], and all of them received sulfonylurea and/or basal insulin. Similar to previous studies, all reported cases of hypoglycemia were mild, and no severe hypoglycemia was documented in this study. In this study cohort, 3.6% of patients had an injection-site reaction; however, the incidence was gradually decreased with repeated injections. Our data supported the safety profiles of dulaglutide in a clinical management plan that comprised various glucose-lowering regimens.
There are several limitations of the present study due to its retrospective design. The total cohort sample was small, in particular, for the analysis of glucose-lowering efficacy with regard to the background treatment regimens. Moreover, several parameters, including FPG, lipid profiles, blood pressure, and insulin sensitivity or resistance, were unavailable for an evaluation of the changes after dulaglutide therapy. Furthermore, safety profiles were not collected systematically; therefore, the self-report system might have resulted in over- and under-estimation of adverse events. Nevertheless, this study provided meaningful data on the real-world efficacy of dulaglutide in combination with various background glucose-lowering regimen and its long-term efficacy over 12 months.