Early changes in left ventricular mass index after empagliflozin initiation in type 2 diabetes: a retrospective cohort study

Sodium-glucose cotransporter-2 inhibitors (SGLT2i) promote reverse cardiac remodeling; however, evidence regarding very early changes in left ventricular mass index (LVMI) in patients with type 2 diabetes without established cardiovascular disease remains limited. We evaluated 3-month changes in LVMI after empagliflozin initiation and explored whether the echocardiographic response differed by obesity status, insulin use, and baseline left ventricular hypertrophy (LVH). This retrospective cohort study included 233 patients with type 2 diabetes mellitus who had paired baseline and follow-up echocardiographic examinations performed 75–105 days apart. Of these, 119 received empagliflozin and 114 constituted the control group. Echocardiographic, clinical, and laboratory data were extracted retrospectively. The primary outcome was the between-group difference in LVMI change (ΔLVMI). Multivariable analysis adjusted for baseline LVMI, age, sex, body mass index, systolic blood pressure, HbA1c, estimated glomerular filtration rate, diabetes duration, and insulin use. Inverse probability of treatment weighting (IPTW) was performed as a sensitivity analysis. Secondary analyses evaluated changes in other echocardiographic parameters. Exploratory subgroup analyses were performed according to obesity status, insulin use, and guideline-defined LVH. LVMI decreased by 1.87 ± 3.96 g/m² in the empagliflozin group and by 0.07 ± 3.92 g/m² in the control group. The between-group difference in change was − 1.80 g/m² (95% confidence interval, − 2.82 to − 0.78; p < 0.001). After multivariable adjustment, empagliflozin initiation remained associated with a lower follow-up LVMI (adjusted difference, − 1.98 g/m²; 95% CI, − 3.02 to − 0.94; p < 0.001), with consistent findings in the IPTW sensitivity analysis. Empagliflozin was also associated with favorable changes in left atrial volume index, E/A ratio, septal and lateral e′ velocities, tricuspid annular plane systolic excursion, and pulmonary artery systolic pressure, whereas changes in left ventricular ejection fraction and E/e′ did not differ significantly between groups. No significant treatment-by-subgroup interactions were observed for obesity, insulin use, or baseline LVH. Empagliflozin treatment was associated with a modest but statistically significant reduction in LVMI within 3 months. Given the short follow-up period, these early changes may partly reflect favorable alterations in loading conditions rather than definitive structural reverse remodeling. Longer-term prospective studies are needed to determine whether these findings persist and translate into sustained cardiac remodeling.