Clinical Question

Are dipeptidyl-peptidase-4 (DPP-4) inhibitors, glucagon-like peptide-1 (GLP-1) receptor agonists, and sodium-glucose cotransporter-2 (SGLT-2) inhibitors safe and effective at reducing cardiovascular mortality, all-cause mortality, and other cardiovascular outcomes (e.g., myocardial infarction, stroke, hospitalization from heart failure) in people with cardiovascular disease (CVD)?

Evidence-Based Answer

DPP-4 inhibitors do not reduce mortality or cardiovascular outcomes in people with CVD, and they increase the risk of pancreatitis. (Strength of Recommendation [SOR]: A, consistent, good-quality patient-oriented evidence.)

GLP-1 receptor agonists reduce cardiovascular mortality, all-cause mortality, and stroke in people with CVD. (SOR: A, consistent, good-quality patient-oriented evidence.)

SGLT-2 inhibitors reduce cardiovascular mortality, all-cause mortality, and hospitalization from heart failure in people with CVD.¹ (SOR: A, consistent, good-quality patient-oriented evidence.)

Practice Pointers

In the United States, CVD is the top cause of death, responsible for 24% of all deaths in 2019.² Diabetes mellitus is a leading cause of CVD.³ Although metformin remains the first-line therapy for diabetes, with demonstrated cardiovascular benefit in people with or without diabetes, three newer medication classes were approved recently for diabetes management.⁴ The authors of this review sought to determine if DPP-4 inhibitors, GLP-1 receptor agonists, and SGLT-2 inhibitors improve outcomes in people with CVD regardless of whether they have diabetes.

This Cochrane review included 20 studies and 129,465 participants in several separate meta-analyses.¹ Participants 18 years and older who had baseline CVD with or without diabetes were recruited from across the world. This review included only placebo-controlled trials, and patients could be taking other medications to treat diabetes, including metformin, sulfonylureas, thiazolidinediones, and insulin. Treatment lasted 24 weeks or longer. Primary outcomes were cardiovascular mortality, myocardial infarction, and stroke. Secondary outcomes were all-cause mortality, hospitalization from heart failure, and adverse effects (including worsening renal function, hypoglycemia, pancreatitis, and fractures).

DPP-4 inhibitors (10 trials) did not improve any primary or secondary outcomes compared with placebo over three years.

GLP-1 receptor agonists were studied up to 3.8 years and reduced cardiovascular mortality (six trials; n = 46,093; number needed to treat [NNT] = 181; 95% CI, 109 to 531), all-cause mortality (seven trials; n = 46,393; NNT = 129; 95% CI, 82 to 307), and fatal and nonfatal stroke (five trials; n = 42,910; NNT = 268; 95% CI, 147 to 1,565), but they did not reduce myocardial infarctions or hospitalizations from heart failure, compared with placebo.

SGLT-2 inhibitors (five trials; n = 24,962) were studied for up to 3.5 years and reduced cardiovascular mortality (NNT = 48; 95% CI, 36 to 71), all-cause mortality (NNT = 43; 95% CI, 32 to 64), and hospitalization from heart failure (NNT = 19; 95% CI, 16 to 22), but they did not reduce myocardial infarctions or stroke, compared with placebo.

Regarding adverse effects, DPP-4 inhibitors increased the risk of pancreatitis (five trials; n = 47,684; number needed to harm = 860; 95% CI, 489 to 3,584) compared with placebo. GLP-1 receptor agonists (one trial; n = 3,297; NNT = 43; 95% CI, 26 to 121) and SGLT-2 inhibitors (two trials; n = 8,474; NNT = 109; 95% CI, 67 to 286) provided renal protection vs. placebo. Neither GLP-1 receptor agonists nor SGLT-2 inhibitors increased the risk of pancreatitis compared with placebo.

Limitations of this review include limited follow-up duration (metformin and its diabetes and cardiovascular outcomes have been under investigation since the 1950s) and the fact that some trials involved populations in which metformin was not used as a first-line therapy. Although all participants had baseline CVD, only three of the 20 pooled studies included people with CVD and without diabetes. It was unclear if the benefits of these medications resulted from their effect on improving glucose control or occurred through an independent mechanism.

Despite a lack of head-to-head trials investigating these medications, the authors performed a network meta-analysis, a methodology that allows for comparisons of different interventions, provided the interventions have a common comparator (in this case, the medications were compared with placebo). The network meta-analysis in this review largely indicated that SGLT-2 inhibitors were likely best for reducing CVD and all-cause mortality.

The American Diabetes Association and the European Association for the Study of Diabetes recommend GLP-1 receptor agonists for people with diabetes and CVD. They also recommend SGLT-2 inhibitors for people with diabetes and heart failure, regardless of their atherosclerotic CVD risk score.⁵ The American College of Cardiology recommends initiating a patient-physician discussion about SGLT-2 inhibitors and GLP-1 receptor agonists for people with diabetes and CVD (atherosclerotic CVD or heart failure).⁶ Although there is robust evidence for the use of these medications in people with diabetes, the evidence is insufficient to support the use of these new medication classes (i.e., DPP-4 inhibitors, GLP-1 receptor agonists, and SGLT-2 inhibitors) to improve cardiovascular outcomes in people without diabetes who have CVD.

Editor’s Note: The NNTs and number needed to harm and their corresponding CIs reported in this Cochrane for Clinicians were calculated by the authors based on raw data provided in the original Cochrane review.

The practice recommendations in this activity are available at https://www.cochrane.org/CD013650.

The opinions and assertions expressed herein are those of the authors and do not reflect the official policy or position of the Uniformed Services University of the Health Sciences, the U.S. Army, or the Department of Defense.