Benefits of Various SGLT2 Inhibitors Mostly Consistent in Observational Study

The SGLT2 inhibitors, as a class, lead to similar treatment benefits at comparable clinical doses, although there are some discrepancies between them, particularly when dapagliflozin (Farxiga; AstraZeneca) is used at low doses, a new observational study suggests.

Compared with empagliflozin (Jardiance; Boehringer Ingelheim/Eli Lilly), dapagliflozin produced a similar impact on MI/stroke risk but was associated with an increased risk of heart failure (HF) hospitalization, a risk that was elevated when the 5-mg dose was used.

“These findings reinforce the use of dapagliflozin 10 mg, or other SGLT2 inhibitors, when the therapeutic goal is to reduce the risk of heart failure,” write HoJin Shin, PhD (Brigham and Women’s Hospital, Boston, MA), and colleagues in the paper published this week in JAMA Internal Medicine.

Stephen Wiviott, MD (Brigham and Women’s Hospital), who was not involved in the analysis but who led the DECLARE-TIMI 58 trial with dapagliflozin, urged caution in making too much of observational data. “This is true even when using advanced statistical techniques like propensity matching and trial emulation employed in this study,” he told TCTMD. “[I] would not clinically choose one agent over another based on these types of data.”

The “magic” of trial randomization is that it accounts for both measured and unmeasured confounders, he stressed, which observational studies do not. That’s particularly true when physicians choose to select a lower dose because the treatment decision “may be based on unmeasured factors, such as perceived frailty or concern about tolerating a full dose medication,” said Wiviott.  

No Head-to-Head Studies

SGLT2 inhibitors, which were initially approved for glycemic control, have been shown to protect the heart and kidneys in patients with type 2 diabetes, as well as reduce the risk of HF hospitalizations in patients at risk for cardiovascular disease.

However, as the researchers point out, the effect on atherosclerotic outcomes has varied between trials. Specifically, canagliflozin (Invokana; Janssen) and empagliflozin both reduced the risk of major adverse cardiovascular events compared with placebo in the CANVAS and EMPA-REG OUTCOME studies, respectively. In DECLARE-TIMI 58, dapagliflozin reduced the risk of study’s primary endpoint—all-cause mortality and HF hospitalization—but didn’t reduce the risk of MACE.

In the absence of head-to-head studies, researchers sought to compare the different SGLT2 inhibitors in adults with type 2 diabetes treated in clinical practice. Using data from three large US health insurance claims databases, they compared canagliflozin and dapagliflozin to empagliflozin using inclusion and exclusion criteria, treatment doses, and outcomes that emulated those of clinical trials. The dosages included two approved strengths of each of the SGLT2 inhibitors: canagliflozin 100 or 300 mg, dapagliflozin 5 or 10 mg, and empagliflozin 10 or 25 mg. Ertugliflozin (Steglatro; Merck Sharp & Dohme), the newest of the SGLT2 inhibitors, approved in late 2017, was excluded from the analysis.

Overall, 232,890 patients were treated with canagliflozin, 129,881 with dapagliflozin, and 295,043 with empagliflozin across the three datasets. Before propensity-score weighting, dapagliflozin-treated patients were younger than those treated with empagliflozin, both at the low and high doses, while empagliflozin-treated patients were more likely to have diabetes-related conditions, cardiovascular disease, or chronic kidney disease. Follow-up in the analysis was up to 8 years.

After propensity-score weighting, there was no difference in the risk of MI/stroke or HF hospitalization between canagliflozin- and empagliflozin-treated patients. When dapagliflozin and empagliflozin were compared, the risk of MI/stroke was similar, but dapagliflozin was associated with a higher risk of HF hospitalization (HR 1.19; 95% CI 1.02-1.39). Canagliflozin-treated patients had a borderline lower risk of all-cause mortality compared with empagliflozin (HR 0.87; 95% CI 0.76-1.00), but mortality was similar in the dapagliflozin and empagliflozin users.

The results were consistent across the three databases and when patients were stratified by cardiovascular disease history. In an analysis focused on dose, the pattern seen in the primary analysis was amplified, with low-dose dapagliflozin associated with a higher risk of HF hospitalization compared with empagliflozin (HR 1.30; 95% CI 1.12-1.50); no such difference was seen with the higher dose.

Consistent Results Seen in Other Studies

From a safety standpoint, use of canagliflozin or dapagliflozin was associated with a slightly lower risk of genital infections than empagliflozin. Risk of severe urinary tract infections was higher with canagliflozin, while dapagliflozin was associated with a lower risk of diabetic ketoacidosis.  

The researchers note that there is a risk of residual confounding by unmeasured factors, such as duration and severity of diabetes and cardiovascular disease. They even point that while they observed a higher risk of HF hospitalization with dapagliflozin, that finding was not seen in other observational studies, including one from Denmark and another from Japan. In 2024, they add, a network meta-analysis of 21 placebo-controlled, randomized trials testing multiple SGLT2 inhibitors found no differences in the risk of all-cause mortality, cardiovascular mortality, HF hospitalizations, and diabetic disease progression between the different drugs.

To TCTMD, Wiviott said there is a large body of evidence supporting the SGLT2 inhibitor drug class, noting the medications have been a “remarkable advance in diabetes management and the prevention and treatment of cardiovascular and renal outcomes in people with and without diabetes.”

In his opinion, the results across the trials have been consistent, with the various agents reducing HF and kidney disease progression and demonstrating a modest effect on atherosclerotic outcomes, such as MI.

“There have been consistent safety observations such as a small increase in [diabetic ketoacidosis] and genital infections that do not seem to offset the overall benefit,” said Wiviott. “The meta-analyses have largely supported consistency of effect among agents by showing a lack of statistical heterogeneity between trial results.”