Efficacy and Safety of Initial Combination Therapy in Treatment-Naïve Type 2 Diabetes Patients: A Systematic Review and Meta-analysis

Initial hypoglycemic monotherapy is usually used in newly diagnosed type 2 diabetes patients, as currently recommended by the guidelines of the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD) [1, 2]. However, initial monotherapy is frequently insufficient to enable patients to achieve or sustain glycemic targets [3, 4]. Thus, initial combination therapy has emerged as an alternative approach. The latest position statement from the ADA/EASD [2] called for an initial combination of two non-insulin agents in patients with a high baseline glycated hemoglobin (HbA1c) level (≥ 9.0%). Additionally, the latest American Association of Clinical Endocrinologists (AACE) treatment algorithm [5] recommended that patients with a HbA1c level of > 7.5% should receive combination therapy with metformin plus an additional drug.

However, we asked the question of whether initial combination therapy is actually more efficacious than monotherapy in terms of glucose control and confirmed safety. To search for the answer, we identified two published systematic reviews and meta-analyses. In one meta-analysis [6] that included 15 randomized controlled trials (RCTs), the authors found that compared to metformin alone, combination therapy with metformin plus another anti-diabetes drug provided statistically significant reductions of 0.43% in HbA1c level and of 14.30 mg/dl in fasting plasma glucose (FPG) level. In another meta-analysis [7] that included eight RCTs, the authors reported that compared with metformin monotherapy, initial combination therapy with dipeptidyl peptidase-4 (DPP-4) inhibitors plus metformin was associated with a higher reduction of 0.49% in HbA1c level, a higher reduction of 0.80 mmol/l in FPG level and a lower weight loss of 0.44 kg. However, the authors of both of these meta-analyses did not present any further analysis with regard to the different types of hypoglycemic agent tested. Therefore, the aims of this study reported here were to comprehensively evaluate the efficacy and safety of initial combination therapies versus monotherapy using updated trial data in type 2 diabetes patients.

Montherapy is unlike to achieve glycemic targets in patients with a high baseline HbA1c level (≥ 9%) [2], and in such cases the guidelines of the ADA/EASD recommend that the patient receive initial combination therapy [2]. In terms of “high” baseline HbA1c level, the AACE recommends initial pharmacologic combination treatment in patients with a HbA1c level of > 7.5% [5], and the Canadian Diabetes Association recommends initial combination therapy in patients with a HbA1c level of > 8.5% [49]. Among all sets of guidelines, the justification for initiating combination therapy is that patient would be unlikely to reach the glycemic target with monotherapy. The results of our meta-analysis supports that rationale, with most initial combination therapies—compared with monotherapy—showing superior glucose control in type 2 diabetes patients with an initial HbA1c level of > 7.5% at a similar risk of hypoglycemia.

As previously indicated [50, 51], there are a number of rationales for initial combination therapy in patients with type 2 diabetes. First, such therapy may lead to early robust lowering of HbA1c levels; as demonstrated by our meta-analysis, most initial combination therapies showed superior glucose control compared to monotherapy. Second, initial combination therapy may avoid the clinical inertia associated with a stepwise approach to therapy. The authors of one study suggested that the time to receive additional anti-hyperglycemic medication exceeded 1 year for patients who failed metformin monotherapy and that this delay was associated with clinical inertia [52]. Consequently, initial combination therapy may one of the best options to directly address the causes of clinical inertia [52]. Third, initial combination therapy may improve ß-cell function [50, 51]. However, this finding was not clearly evident in our meta-analysis due to the lack of data. Fourth, the complementary mechanisms of action provided by initial combination therapy may require comparatively lower doses of individual agents and therefore may cause fewer AEs. This benefit was indicated by the results of our meta-analysis which showed that most initial combination therapies exhibited better glucose control with comparable risks of hypoglycemia, SAEs, discontinuation due to AEs and GI side effects. Fifth, initial combination therapy may avoid the long-term consequences of metabolic memory, as the initial use of combination therapy could lead to greater HbA1c reduction, enabling more individuals to achieve their glycemic goals while avoiding AEs stemming from multiple metabolic defects [51, 53, 54]. However, this latter potential benefit may not be concluded from the present meta-analysis because most of the studies included were of short-term duration.

The evidence is compelling that type 2 diabetes is a progressive, physiologically and genetically complex heterogeneous disease. Achieving glycemic control is necessary to prevent or delay the progression of vascular complications. As current treatment approaches do not adequately acknowledge the complexity of diabetes, a compelling case may be made for combination treatment [51]. Initial combination therapy may be required to address the complex pathophysiology of type 2 diabetes, which includes improving insulin secretion and insulin sensitivity, inhibiting hepatic glucose production and addressing delayed gastric emptying or glucose absorption, while focusing on satiety and renal glucosuria. Among the mechanisms of hypoglycemic agents [55], metformin inhibits hepatic gluconeogenesis and improves peripheral insulin sensitivity, SUs/glinides stimulate insulin secretion by β-cells, DPP-4 inhibitors stimulate insulin secretion and suppress glucagon secretion, SGLT2 inhibitors reduce renal glucose reabsorption and induce urinary glucose excretion, TZDs activate peroxisome proliferator-activated receptor gamma (PPAR-γ) and increase insulin sensitivity. Therefore, choices for initial combinations of the above agents should also be supported by the pathophysiology of type 2 diabetes.

However, a number of unresolved issues associated with initial combination therapy in type 2 diabetes patients remain. One of these is whether initial combination therapy improve adherence. To date, there is no evidence suggesting that initial combination therapy versus monotherapy or sequential titration therapy would result in a greater adherence of patients to the therapeutic regimen. However, published studies do show that the more complex the drug regimen, the lower the adherence to that regimen [56]. In our meta-analysis, we did not collect any data on a possible improvement in adherence. Another issue is cost; is initial combination therapy less costly? The relatively high cost of including novel agents, such as DPP-4 inhibitors or SGLT2 inhibitors, in an initial combination with metformin remains a significant barrier to their use in many regions of the world [51]. Several studies have estimated the cost-effectiveness associated with monotherapy compared to combination therapy with oral anti-diabetes agents, but a number of these these were derived from non-RCT data and had multiple confounders [57, 58]. Moreover, the authors of another study indicated that it was difficult to quantify the cost-effectiveness of softer outcomes such as fewer hypoglycemic events or improved quality of life [59]. We did not collect any data on the costs of initial combination therapy in our meta-analysis, but there are other economic models which could be used to answer this question. Moreover, the association between initial combination therapy and cardiovascular risk has not been fully examined in the literature. Gaps still exist in the evidence on treatment paradigms utilizing sequential versus initial combination therapy. Therefore, carefully designed, pragmatic, prospective real-world studies to assess the clinical effectiveness of initial combinations versus sequential treatment in patients with newly diagnosed or poorly controlled type 2 diabetes should be performed to provide more evidence.

There were several limitations to our meta-analysis. First, data from the separate studies covered different durations of the study. As previously indicated, RRs are sensitive to the length of the follow-up; consequently, the pooling of results from studies with different durations of follow-up might lead to an artificial heterogeneity and discrepancy in the meta-analyses [60]. We therefore explored the outcomes in subgroup analyses by pooling all of the studies with a study period of 24 weeks to conduct a sensitivity analysis, which showed similar results with the total results. Second, the definitions of treatment-naïve patients varied depending on the protocols of the trials included in our meta-analysis, and these differences may also be associated with the high heterogeneity of this study and also lower the ability of the authors of this study to propose solid conclusions. Therefore, we also conducted a sensitivity analysis to minimize the bias and found the similar results to the efficacy and safety evaluations. The large differences in the number of studies for several combinations is another limitation. For those treatment groups with only one trial included [41‐48], no further meta-analysis was done for evaluation purposes. Another problem may be the variations in dosages used in the different studies. Therefore, the standard doses recommended and approved in the clinical practice were used in this meta-analysis to minimize the bias. Since baseline characteristics were variable across studies, we used the random-effects model for analysis when the level of heterogeneity was high. Given these factors, we suggest that our results be interpreted cautiously.

In conclusion, compared with monotherapy, all initial combination therapies resulted in significantly reduced HbA1c levels in treatment-naïve type 2 diabetes patients. Compared with metformin monotherapy, the initial combination therapies of DPP-4 inhibitors plus metformin and SGLT2 inhibitors plus metformin exhibited similar risks of hypoglycemia, but the initial combination therapies of SU plus metformin and TZD plus metformin exhibited higher risks of hypoglycemia.