This study compared the reduction of glycated haemoglobin (HbA1c) with sodium-glucose cotransporter type-2 inhibitors (SGLT2is) vs. dipeptidyl peptidase-4 inhibitors (DPP-4is) as add-ons to metformin in patients with type 2 diabetes mellitus (T2DM), with a specific focus on HbA1c changes according to baseline HbA1c.

Electronic databases were scrutinized for randomized controlled trials (RCTs) evaluating the reduction of HbA1c from baseline (Δ HbA1c) with an SGLT2i or DPP-4i in patients with T2DM not well controlled by metformin monotherapy. The endpoint was Δ HbA1c using both indirect and direct comparisons.

Overall, Δ HbA1c was slightly greater with SGLT2is (−0.80±0.20% from 8.03±0.35%; 44 analyses, 29 RCTs, 15 with two doses, n=9321) than with DPP-4is (−0.71±0.23% from 8.05±0.43%; 61 analyses, 59 RCTs, n=17,914; P=0.0354). When the mean baseline HbA1c was<8% ([64mmol/mol] 7.79±0.15% vs. 7.71±0.23%), Δ HbA1c averaged −0.735±0.17% vs. −0.62±0.16% (P=0.0117) with SGLT2is vs. DPP-4is, respectively. However, this difference vanished when the mean baseline HbA1c was≥8% (−0.87±0.22% from 8.27±0.32% with SGLT2is vs. −0.80±0.24% from 8.35±0.33% with DPP-4is; P=0.2756). The relationship between Δ HbA1c and baseline HbA1c was only slightly stronger with SGLT2is (slope: −0.39, r²=−0.43; P<0.0001) than with DPP-4is (slope: −0.26, r²=−0.25; P<0.0001).

Because of the small difference in Δ HbA1c whatever the baseline HbA1c level with SGLT2is vs. DPP-4is as add-ons to metformin, choosing between these glucose-lowering agents in clinical practice should be based on other efficacy criteria (such as weight and blood pressure changes, cardiovascular and renal protection) or on safety profiles rather than on HbA1c levels.