The study was motivated by a gap in the evidence. Acute pulmonary embolism is a common cardiovascular disorder with substantial morbidity and mortality, and prognosis is largely determined by hemodynamic status, the extent of right ventricular involvement, and biomarkers of cardiac injury and wall stress such as troponin and N-terminal pro-B-type natriuretic peptide (NT-proBNP). The sudden increase in systolic pulmonary artery pressure in pulmonary embolism leads to right ventricular dilatation and dysfunction, which may in turn dictate the patient’s hemodynamic condition. SGLT2 inhibitors are considered a cornerstone therapy in heart failure and chronic kidney disease, but previous research has mainly targeted their effects on the left ventricle, and the limited data on the right ventricle come from echocardiographic and clinical studies conducted in chronic conditions such as chronic heart failure and adult congenital heart disease. According to the authors, no prior study had evaluated these drugs in an acute condition affecting the right ventricle.

The research team, led by Gassan Moady of Galilee Medical Center and the Azrieli Faculty of Medicine at Bar-Ilan University, screened 632 patients hospitalized with acute pulmonary embolism over a four-year period. Thirty of them, or 4.7 percent, had been taking an SGLT2 inhibitor for at least two months before admission, with a median duration of four months. The investigators deliberately excluded high-risk patients and intermediate-high-risk patients who deteriorated and required intravenous thrombolysis or endovascular ultrasound-facilitated intervention, as well as anyone with a prior history of heart failure or left ventricular dysfunction and anyone with missing laboratory or echocardiographic data. That left a cohort of low-risk and intermediate-low-risk patients treated with oral anticoagulation, a population in which continuation of chronic outpatient medication is clinically plausible, since these drugs are generally withheld during critical illness.

To compensate for the obvious baseline differences between the groups, the team used propensity-score matching. Before matching, 86.7 percent of the SGLT2 inhibitor group had diabetes compared with 34.4 percent of controls, and the treated group also had better renal function, with a mean creatinine of 0.83 milligrams per deciliter versus 1.12. Using a logistic regression model with SGLT2 inhibitor use as the dependent variable, each treated patient was matched to three controls on age, sex, body mass index, background disorders, laboratory parameters, and baseline medical therapy, yielding a final cohort of 30 treated patients and 90 controls. A caliper-based sensitivity approach was explored but retained too few treated patients to serve as the primary analysis. After matching, absolute standardized mean differences fell below the conventional 0.1 threshold for most variables, although some residual imbalance remained, including in insulin use and creatinine.

The echocardiographic signal was the study’s central finding. Tricuspid annular plane systolic excursion, or TAPSE, an M-mode measure of how far the lateral tricuspid annulus moves toward the apex during systole, averaged 19.4 millimeters in the SGLT2 inhibitor group versus 17.3 millimeters in controls, a mean difference of 2.1 millimeters with a 95 percent confidence interval of 0.7 to 3.5 and a p-value of 0.004. Clinical congestion, assessed through peripheral edema, jugular venous distension, and the need for intravenous diuretics, was documented in 6.7 percent of treated patients versus 25.6 percent of controls, an odds ratio of 0.21 with a 95 percent confidence interval of 0.05 to 0.94 and a p-value of 0.03. Interestingly, systolic pulmonary artery pressure, which averaged 36.5 millimeters of mercury in the treated group versus 39.2 in controls, and the rate of right ventricular dilatation, 13.3 percent versus 27.8 percent, did not differ significantly between the groups, suggesting that the association manifested in ventricular performance rather than in measured pulmonary pressures themselves.

The biomarker differences were striking. Median troponin, which reflects the extent of myocardial injury, was 14.5 nanograms per liter in the SGLT2 inhibitor group, with an interquartile range of 5.3 to 51, compared with 64.5 in controls, while median NT-proBNP, a marker of ventricular wall stress, was 158 picograms per milliliter versus 1,000, with a 95 percent confidence interval for the difference of 467 to 1,153 picograms per milliliter and p-values below 0.001 for both comparisons. Because these biomarkers respond rapidly to the acute hemodynamic insult and track the clinical course, with persistent NT-proBNP elevation indicating ongoing right ventricular dysfunction and worse outcomes, the authors interpret the lower levels as consistent with attenuated right ventricular strain. The simplified pulmonary embolism severity index was comparable between groups, and systolic blood pressure and heart rate on admission were similar, arguing against a simple difference in presenting severity.

Mechanistically, the authors propose several non-mutually-exclusive pathways, while stressing that direct protection against embolism-induced right ventricular dysfunction remains unproven and that lower left ventricular filling pressures alone cannot explain the pattern of findings. Reduced preload and afterload through natriuresis, diuresis, and vasodilation could lower right ventricular stress even if measured pulmonary artery pressures were similar at the time of imaging. In animal models of pulmonary hypertension, SGLT2 inhibitors have been shown to attenuate pulmonary vascular remodeling, right ventricular hypertrophy, and fibrosis, effects attributed to anti-inflammatory mechanisms including inhibition of NLRP3 inflammasome activation, downregulation of interleukin-mediated signaling, modulation of mitochondrial autophagy and nutrient-sensing pathways, and inhibition of pulmonary artery smooth muscle cell proliferation through activation of AMP-activated protein kinase. These are preclinical observations, and their relevance to human pulmonary embolism has not been established.

The long-term outcome data tell a more sobering story. One-year readmissions for dyspnea, chest pain, or recurrent thromboembolic events occurred at similar rates in both groups, with an odds ratio of 0.34, a 95 percent confidence interval of 0.1 to 1.2, and a p-value of 0.13. The authors attribute this null result partly to the small sample size and to the inclusion of only hemodynamically stable patients, who had relatively low readmission rates overall. They also note that the readmission analysis assumed continued SGLT2 inhibitor use after discharge and no crossover between the groups, without systematically verifying longitudinal adherence or medication changes. Readmissions after pulmonary embolism are common in the broader literature, with almost half of patients readmitted within the first year, mainly for recurrent venous thromboembolism, anticoagulation-related bleeding, or persistent dyspnea and chest pain.

The study’s limitations place firm boundaries on interpretation. As a retrospective, single-center, exploratory analysis with only 30 treated patients, it cannot establish causation, and unmeasured factors such as thrombotic burden and baseline right ventricular function may have influenced the results, since PE-specific severity variables could not be included in the propensity model because they represented study outcomes. Baseline NT-proBNP was unavailable, so pre-existing differences in volume status cannot be entirely excluded, although patients with a history of heart failure were excluded and significant baseline volume overload was considered unlikely. There were no serial echocardiographic assessments after the index event to track recovery or persistent dysfunction. No episodes of diabetic ketoacidosis were documented in either group, a finding the authors say is unsurprising given the established safety profile of the drugs, the small sample size, and the hemodynamically stable population, but which should be interpreted cautiously.

The authors call for prospective, ideally randomized studies to test whether continuing or initiating SGLT2 inhibitors after clinical stabilization in appropriately selected patients with pulmonary embolism preserves right ventricular function and improves clinical outcomes. They note that because many patients with acute pulmonary embolism also have comorbidities such as diabetes and chronic kidney disease, initiation after stabilization may be reasonable in selected patients given the favorable safety profile, but that this approach requires further evaluation. Until such trials are done, the association reported here should be read as an intriguing signal about the right ventricle, an organ long overshadowed by its left-sided counterpart in cardiovascular research, rather than as a new treatment guideline.

Subject of Research: Association of chronic SGLT2 inhibitor use with right ventricular function in hemodynamically stable patients with acute pulmonary embolism

Article Title: Association between chronic sodium-glucose cotransporter 2 inhibitors use and right ventricular function in hemodynamically stable patients with acute pulmonary embolism

Article References: Moady, G., Maman, S., Dragu, R., Levi-Gofman, L., & Atar, S. (2026). Association between chronic sodium-glucose cotransporter 2 inhibitors use and right ventricular function in hemodynamically stable patients with acute pulmonary embolism. Frontiers in Cardiovascular Medicine, 13, Article 1975531. https://doi.org/10.3389/fcvm.2026.1975531

Image Credits: AI Generated

DOI: 10.3389/fcvm.2026.1975531

Keywords: SGLT2 inhibitors, pulmonary embolism, right ventricle, TAPSE, NT-proBNP, troponin, echocardiography, propensity-score matching, retrospective study, clinical congestion, heart failure, cardiometabolic