Introduction: Prescribing DPP-4i medications versus metformin as a first-line treatment is common, although large-scale safety evidence favoring DPP-4is monotherapy is scarce. The goal of this study is to compare the incidence of hypoglycemia in patients with a DPP-4i monotherapy versus metformin monotherapy

Methods: US patients in Komodo's Healthcare Map with diabetes were studied. 2,016,156 patients (59.3% female) on a single-gradient metformin monotherapy and 45,826 patients (52.6% female) on a DPP-4i monotherapy were found from 2016 - 2020. No antidiabetic treatment was prescribed in the previous year and patients were continuously enrolled 1 year before/6 months after their first medication prescription. Propensity score matching (PSM) was performed using age, gender, US regions, chronic kidney disease (CKD), obesity, emergency room visits, Charlson Comorbidity Index (CCI), and proportion of Days Covered (PDC). The PSM sample comprised 45,400 1:1 matched users of each group. The incidence rate of hypoglycemia was calculated using Poisson regression.

Results: Patients in the DPP-4i group were older, 64.3 (SD14.0) versus 50.5 (15.8) years in the metformin group. CKD was recorded in 28.1% of the DPP-4i users and 4.3% of metformin users. A CCI score ≥5 was seen in 53.7% of DPP-4i users and 15.9% of Metformin users. PDC was ≥ 80% in 37.0% and 42.7% of DPP-4I and metformin users. The estimated pre-matched incidence of hypoglycemia was 19.04 and 16.88 per 1000 person-years in DPP-4i and metformin users. The post-matching incidence rates were 19.17 for DPP-4i and 16.82 for metformin users. The DPP-4i to metformin incidence rate ratio was 1.14 (95% CI 0.99 - 1.31, p = 0.062).

Conclusion: The incidence rate of hypoglycemia in a large cohort of patients treated with DPP-4i was comparable to those treated with metformin. The sizable risk of hypoglycemia after DPP-4is should be considered in prescription decision-making.


U. Periyanayagam: None. Y. Wei: None.

Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at