Endothelial function in patients with type 2 diabetes was assessed using FMD following treatment with empagliflozin and glimepiride, a sulfonylurea. We found that although FPG, HbA1c, and GA were equally improved in both empagliflozin and glimepiride treated patient groups, the ΔFMD did not show a significant difference. Empagliflozin had no effect toward improving endothelial function irrespective of glucose level improvement in this study. To observe changes in endothelial function induced by drugs that lacked the ability to improve the glucose level, this study was conducted with patients in a steady glucose–controlled state. Although baseline HbA1c was targeted in this study, severe hypoglycemia did not occur after reducing glargine if the fasting plasma glucose was maintained under 90 mg/dL. Additionally, GA improved significantly in both groups. It is possible that any glucose variation was improved by the additional treatment. The insulin dose was also decreased significantly for both groups and this may have reduced any increase in body weight observed. Although ΔHbA1c, body weight, HDL–C, and TG had an effect on the change of FMD and at least HbA1c was improved significantly in the empagliflozin group, FMD showed no significant change in this study. These findings suggested that empagliflozin had no effect toward improving FMD although empagliflozin was effective for improving HbA1c.
Renal function was found to be worse in the empagliflozin group. However, our study was conducted over a relatively short observation period, whereas empagliflozin was previously shown to stably maintain the eGFR during long term administration . In a recent study, canagliflozin reduced the risk of kidney failure at a median follow up of 2.62 years , whereas an SGLT–2 inhibitor decreased hyperfiltration by diuresis . Therefore, empagliflozin would be expected to prevent the progression of kidney disease over an extended period of time.
UA was significantly decreased in the empagliflozin compared to the glimepiride treated group. Hyperuricemia is a known risk factor for developing end stage renal disease  and the serum UA level constitutes an independent predictor of future cardiovascular mortality . A reduced serum UA may affect kidney function and cardiovascular events in the future. Notably, increased fractional excretion of UA was shown to be related to reduced serum UA following canagliflozin treatment .
In previous studies, empagliflozin was found to decrease body weight [11, 26]. We also observed a change in body weight of –0.59 ± 2.5 kg in the present study, which was less than that described in previous reports; this discrepancy may be due to a lack of patient compliance with the diet. However, for the subgroup showing decreased body weight, FMD was found not to be significantly different between the two groups. This suggested that a change in body weight only may not affect the FMD.
In another study, empagliflozin together with basal insulin for 18 weeks led to a significant decrease in body weight (–1.7 ± 0.6 kg; P = 0.035) . Additionally, empagliflozin combined with multiple titrated daily injections of insulin for 18 weeks also led to a significantly decreased body weight (–0.97 ± 0.18 kg) . Our study examined the effect of empagliflozin together with basal insulin and metformin. The dose of glargine used was not related to a change in body weight and FMD, and was found to be significantly decreased for both groups; however, the Δdose of glargine showed no significant difference. In comparison, insulin therapy was found by others to induce weight gain in the absence of a well–controlled diet ; however, in the present study, the use of insulin was not related to the lack of a significantly decreased body weight with empagliflozin treatment.
Abdominal adiposity is linked to a risk of cardiovascular disease, as shown by assessments made noninvasively using waist measurements . Empagliflozin was found to have decreased the waist circumference of patients, as also observed in a previous report ; however, glimepiride increased it significantly. This suggested that empagliflozin may have decreased abdominal adiposity compared to glimepiride and is therefore expected to contribute to the prevention of cardiovascular disease. To confirm this effect, it will be necessary to follow up with abdominal computed tomography studies to assess visceral fat with more accuracy.
The body fluid volume was also found to have significantly decreased following 12 weeks of empagliflozin treatment. However, the two groups did not show a significant difference in Δbody fluid volume. Additionally, it was unclear whether the effect of empagliflozin on decreasing the body fluid volume was greater than that of glimepiride. Empagliflozin has been shown to improve hospitalization rates after heart failure  and is thought to ameliorate this disease by decreasing body fluid volume. In the observational period, heart failure was not observed for patients in either treatment group. This suggested that empagliflozin may have had a coronary protective effect that is not derived from any impact on endothelial function.
The FMD may be improved when the baseline FMD is low. However, in the subgroup showing a lower than median baseline FMD, the FMD was not significantly changed by additional treatments and the ΔFMD did not significantly differ between the two groups in both subgroup (empagliflozin; P = 0.59, glimepiride; P = 0.64) and overall analyses. Notably, SGLT2 inhibitors have a secondary preventive role in adverse cardiovascular events but lack a primary preventive role . Although SGLT2 inhibitors have a greater effect on the improvement of FMD in patients with a history of cardiovascular events than on those without such a history, no patients in the present study had such a background; it therefore would be necessary to undertake a secondary intervention.
In particular, the effects of other oral hypoglycemic agents along with glucagon like peptide 1 (GLP–1) analogs on endothelial function have been previously reported; e.g., pioglitazone improved endothelial function [18, 33]. In comparison, a dipeptidyl peptidase (DPP) 4 inhibitor improved , had no effect , or worsened  endothelial function but did not affect cardiovascular events [37, 38]. Moreover, GLP–1 treatment itself enhanced [39, 40] or had no effect on  endothelial function. However, in patients with type 2 diabetes, liraglutide, a GLP–1 analog, was successful in preventing nonfatal myocardial infarction or stroke along with death from cardiovascular causes . Consistent with these inconsistent findings, significant heterogeneity existed between a meta–analysis study of a DPP4 inhibitor and GLP–1 . The size of the study, the duration of intervention, and the age or sex of participants enrolled did not affect the mean difference in FMD . Rather, a change in FMD was found to be dependent on the baseline FMD. As the baseline FMD of our study was less than that in reports of improved FMD by additional treatment [19, 33, 34], this may have affected our findings. In our study, the baseline blood glucose was well controlled. However, empagliflozin may not improve endothelial function in patients with moderately controlled blood glucose and no history of cardiovascular events.
In the empagliflozin group, blood glucose, serum UA, body weight, and waist measurement improved significantly, whereas the lipid profile, blood pressure, and insulin resistance remained unchanged for the observation period. Changes in these risk factors are required to improve endothelial function as assessed by FMD. However, empagliflozin would be expected to help prevent heart failure if metabolic risk factors remained unchanged. With regard to body composition, empagliflozin significantly decreased body fluid, which supports the idea of the potential of empagliflozin to reduce heart failure. However, any change in FMD is required to be assessed for longer periods than evaluated in the present study to more fully evaluate any coronary protective effect by empagliflozin.
Several limitations were evident in this study. First, as the study participants were outpatients, it may not have been possible to completely exclude patients who smoked or had a meal before the FMD was examined. Some patients may also not have been very compliant with their diet, which may have affected the change in body weight observed. Second, the number of patients examined was small and the study observation period was relatively short. A larger patient cohort and longer study period to monitor any adverse events are both required in any future studies.
Overall, we found that empagliflozin did not improve endothelial function compared with glimepiride in patients with type 2 diabetes. However, empagliflozin significantly reduced body fluid volume. Thus, the coronary protective effect of empagliflozin may be derived not from preventing endothelial dysfunction but rather from reducing heart failure.