Network meta-analysis of responses to DPP-4 and SGLT 2 inhibitors: What next after metformin with a baseline HbA1c ≤8%?

W.D. Strain1, M.C. Thomas2, J.E. Foley 3, P.M. Paldánius4

1University of Exeter Medical School, Exeter, UK
2Baker IDI Heart and Diabetes Institute, Melbourne, Australia
3Novartis Pharmaceuticals Corporation, East Hanover, NJ, USA
4Novartis Pharma AG, Basel, Switzerland

Background: Treatment decisions for timely delivery of care for optimised management of type 2 diabetes mellitus (T2DM) should be supported by evidence-based data and guidelines, which are tailored to individual patient preferences, projected long-term plans and presence or anticipated future risk of comorbidities. Guidelines indicate early and more aggressive intensification of failing metformin monotherapy for prevention of hyperglycaemia-driven complications of T2DM especially in younger patients with shorter duration of T2DM. Dipeptidyl peptidase-4 inhibitors (DPP4i) and sodium-glucose co-transporter 2 inhibitors (SGLT2i) lower glucose levels in patients with T2DM proportional to the baseline HbA1c albeit with different mechanisms of action. This relationship, in those with normal renal function at least, appears to be steeper with SGLT2i, with greater glycosuria and efficacy in patients with sustained high HbA1c levels, but attenuating efficacy closer to target levels for glycaemic control. 

Aim: To explore the relative benefits of DPP4i and SGLT2i in people inadequately controlled with monotherapy at the level of HbA1c indicated for therapy escalation by most guidelines.

Methods: A network meta-analyses based on a systematic literature review was undertaken to identify all studies that provide information on the comparative efficacy of DPP4i and SGLT2i in patients with T2DM, as add-on to metformin or to metformin and sulphonylurea, exploring the HbA1c reduction from a baseline ≤8.0%. A base-case analysis of 12–26 weeks and a sensitivity analysis of 12–52 weeks inclusive were performed with a fixed and random effects model. The mean difference in change from baseline (ΔHbA1c) between intervention and control were used as model inputs.

Results: Outcome data on patients with baseline ≤8% HbA1c was found in 19 phase III/IV randomised controlled trials, 13 of which reported data on patients receiving only metformin as background therapy. The only DPP4i with information reviewed was vildagliptin. For the background metformin monotherapy network, the fixed effect model indicated that vildagliptin (daily dose of 100 mg) and SGLT2i were more effective than placebo and (additional) metformin. For the background metformin and sulphonylurea network, the fixed effect model indicates that vildagliptin 100 mg is more effective than placebo (ΔHbA1c=−0.90; 95% CI −1.25, −0.55), dapagliflozin 10 mg (ΔHbA1c=−0.54; −0.96, −0.12), canagliflozin 100 mg (ΔHbA1c=−0.45; −0.88, −0.03) and comparable to canagliflozin 300mg (ΔHbA1c=−0.26; −0.69, +0.16). There were no reports using empagliflozin (10 or 25 mg), nor of weight change within the population of interest. 

Discussion: The limited literature supports the hypothesis that in patients with a HbA1c ≤8%, SGLT2i may be less effective for lowering glucose than other treatment strategies such as DPP4i. For delivery of individualised care, decisions must therefore be made balancing the optimisation of glucose control versus the potential benefits of SGLT2i in those with pre-existing cardiovascular disease.

Network meta-analysis of responses to DPP-4 and SGLT2 inhibitors: What next after metformin with a baseline HbA1c ≤8%?