The Association Between Metformin Treatment and Outcomes in Type 2 Diabetes Mellitus Patients With Heart Failure With Preserved Ejection Fraction: A Retrospective Study

Background: Metformin is the first-line antidiabetic medication for type 2 diabetes mellitus (T2DM). However, the association between metformin and outcomes in T2DM patients with heart failure with preserved ejection fraction (HFpEF) is still unknown. We aimed to explore the association between metformin and adverse outcome in T2DM patients with HFpEF. Methods: A total of 372 T2DM patients with HFpEF hospitalized from January 1, 2013, to December 31, 2017, were included in this retrospective cohort study. There were 113 and 259 subjects in metformin and non-metformin group, respectively. Subjects were followed up for all-cause mortality, cardiovascular death, all-cause hospitalization, and heart failure hospitalization. Results: The median follow-up period was 47 months. Eleven patients (2.49% per patient-year) in the metformin group and 56 patients (5.52% per patient-year) in the non-metformin group deceased during follow-up (P = 0.031). However, a multivariable Cox regression failed to show that metformin was an independent factor of all-cause mortality [HR (95% CI) = 0.682 (0.346-1.345); P = 0.269]. A subgroup analysis revealed a significant association between metformin and all-cause mortality in patients with a higher hemoglobin Al c (HbA1 c) level (HbA1c >= 7%) [HR (95% CI) = 0.339 (0.117-0.997); P = 0.045]. The 4-year estimated number needed to treat (NNT) with metformin compared with non-metformin for all-cause mortality was 12 in all populations and 8 in the HbAlc >= 7% subgroup. Conclusions: Metformin was not independently associated with clinical outcomes in patients with T2DM and HFpEF, but was associated with lower all-cause mortality in the subgroup of patients with poor glycemic control. Prospective, randomized controlled trials are needed to further verify these findings.

Automated summary

In patients with T2DM and HFpEF, metformin did not show an independent association with clinical outcomes, but was linked to lower all-cause mortality in the subgroup of patients with poor glycemic control. Further prospective, randomized controlled trials are needed for confirmation.