The effect of empagliflozin on arterial stiffness and heart rate variability in subjects with uncomplicated type 1 diabetes mellitus

In this 8-week, open-label, prospective clinical trial, 42 participants with T1D were treated with empagliflozin 25 mg once daily (NCT01392560). Inclusion and exclusion criteria at screening have been described elsewhere [24] (http://​www.​clinicaltrials.​gov/​ct2/​show/​NCT01392560). In brief, men and women ≥18 years with T1D ≥ 12 months with estimated GFR ≥60 ml/min/1.73 m² and no hypertension or macroalbuminuria were included. Patients taking medication known to interfere with RAAS activity and/or renal function based on investigator judgment were excluded [24]. Of note, none of the patients were taking medications that can affect arterial stiffness or blood pressure, such as renin angiotensin aldosterone system blockers, beta blockers, calcium channel blockers, diuretics or other antihypertensive agents. Two cases of diabetic ketoacidosis leading to discontinuation occurred during the study: one in the context of insulin pump failure and the other in the setting of acute gastroenteritis. These patients were withdrawn within the first 3 days of drug exposure and were included only in the safety analyses. Detailed data regarding other adverse effects of empagliflozin in this cohort have been reported elsewhere [24]. 40 subjects completed the study (Table 1) and the primary endpoint of change in renal function in response to empagliflozin has been reported elsewhere [24]. In this manuscript we report the pre-specified exploratory vascular endpoint results. The Research Ethics Board at the University Health Network (Toronto, Canada) approved the protocol and all subjects gave informed consent prior to start of study procedures. The study was conducted according to the International Conference on Harmonization on Good Clinical Practice.

Subjects adhered to a sodium-replete and moderate protein diet during the 7-day period before each experiment, as described previously [27]. After admission to the Renal Physiology Laboratory, euglycemic (4–6 mmol/L) conditions on the first day, followed by hyperglycemic (9–11 mmol/L) conditions on the second day, were maintained by a modified glucose clamp technique. Studies were performed on two consecutive days at baseline after maintenance of the clamp within the desired range for approximately 5 hours preceding and during all investigations, as described previously [24].

Following the glucose clamp, right radial artery and carotid waveforms were recorded with a high-fidelity micromanometer and using the validated transfer function, corresponding central aortic pressure waveform data were generated (SPC-301, Millar Instruments SphygmoCor, AtCor Medical Systems Inc., Sydney, Australia). Augmentation index, an estimate of systemic arterial stiffness was calculated as the difference between the second systolic peak and inflection point, expressed as a percentage of the central pulse pressure corrected to an average heart rate of 75 beats per minute. The aortic pulse wave velocity (PWV) was measured using the same device by sequentially recording ECG-gated right carotid and radial artery waveforms. Two vascular measurements were obtained for each parameter and the average value reported. The use of the SphygmoCor device to assess arterial stiffness parameters has been previously published by our group [28]. Automated blood pressure measurements were obtained using a DINAMAP machine (Critikon, Tampa, Florida) and the average of two values obtained immediately prior to the arterial stiffness assessments are reported.

After completion of arterial stiffness testing, HRV testing was performed using AtCor software (Atcor Medical Systems Inc., Sydney, Australia). In brief, two 10-minute segments were recorded. Vagal tone (Root Mean Square Successive Difference - RMSSD) and sympathetic activity (Standard Deviation of Normal-to-Normal interval - SDNN) measures were obtained at each of the two periods and the results were then averaged. Plasma noradrenalin and adrenaline concentrations were obtained on each of the study days according to standard, previously described methods [29].

The primary endpoint of this study was change in GFR after treatment with empagliflozin for 8 weeks [24]. Sample size calculations were based on anticipated changes in GFR [24]. Accordingly, results reported in this manuscript were pre-specified, exploratory endpoints. Paired t-tests were performed to evaluate differences in vascular measurements and neurohormonal outcomes before and after treatment with empagliflozin. A repeated measures model was used to evaluate differences in response to empagliflozin between euglycemic and hyperglycemic conditions. All statistical analyses were performed using the statistical package SAS (Version 9.2).

The baseline characteristics of this T1D study cohort are described in Table 1. This cohort comprised a mainly young group of participants with uncomplicated T1D that has been described in more detail elsewhere [22, 24]. All except for 4 patients (ages 31, 33, 34 and 44 years old) were between the ages of 18–30.

During clamped euglycaemia, empagliflozin significantly reduced systolic blood pressure, radial augmentation index, carotid augmentation index and aortic augmentation index (Table 2, Figure 2). Carotid-radial pulse wave velocity decreased significantly (p = 0.0001), while similar trends for carotid-femoral pulse wave velocity were not significant. Effects on vagal tone (RMSSD) and SNS activity (SDNN, plasma noradrenalin and adrenaline) were not significant.

During clamped hyperglycemic conditions, treatment with empagliflozin did not reduce blood pressure significantly, despite significant effects on radial augmentation index, carotid augmentation index and aortic augmentation index (Table 2, Figure 2). Empagliflozin decreased both carotid-radial and carotid-femoral pulse wave velocity under clamped hyperglycemic conditions (Table 2). Similar to effects during clamped euglycemic conditions, effects on SNS function including RMSSD, SDNN, plasma noradrenalin and adrenaline, were not significant (Table 2, Figure 2). In the repeated measures model used to evaluate differences in response to empagliflozin between glycemic conditions, there were no significant differences in any of the hemodynamic parameters (p > 0.05, data not shown).

Previous studies in T2D have suggested that oral hypoglycemic agents such as metformin reduce blood pressure, primarily through improving glycemic control [12‐14, 43]. In contrast, other oral hypoglycemic agents such as sulfonylureas that increase the risk of severe hypoglycemic events and lead to body weight gain may promote the development of hypertension [44]. Perhaps as a result of increasing weight, these traditional oral hypoglycemic agents do not improve arterial stiffness parameters and a significant proportion of patients develop hypertension with increasing diabetes duration [45, 46]. Since existing oral agents are not approved nor appropriate in T1D, the options for improving glycemic control in patients with T1D are limited to intensifying insulin regimens, which can have the same adverse effects on weight, hypoglycemic risk and blood pressure as in T2D patients [30‐32]. This study shows for the first time that an oral hypoglycemic agent that improves glycemic control over an 8-week period in patients with T1D can also reduce arterial stiffness, which may in part be responsible for the antihypertensive effects associated with SGLT2 inhibition.

Arterial stiffness is influenced by a variety of factors, including neurohormonal mediators and changes in arterial structure [4]. We have previously demonstrated in this cohort that empagliflozin induces a modest increase in plasma RAAS mediators, including aldosterone and angiotensin II, which is likely a compensatory response to effective circulating volume contraction, similar to expected effects of a thiazide diuretic [24, 47]. The mild increase in plasma RAAS levels with empagliflozin was insufficient to counteract the effect of effective circulating volume contraction, resulting in decreases in blood pressure and arterial stiffness. A similar observation was made for vasodilatory NO, for which we observed a decline in plasma levels in our previous work, which was also unlikely to have accounted for decreases in blood pressure or arterial stiffness in this cohort [24]. We therefore concluded that the beneficial hemodynamic changes observed with empagliflozin were unlikely to be influenced by concomitant changes in systemic RAAS and NO activity. Consequently we hypothesized that improved arterial compliance after empagliflozin could be related to changes in autonomic nervous system function, reflected by increased HRV and lower circulating levels of adrenaline and noradrenalin. However, empagliflozin did not affect HRV or plasma adrenergic system markers, suggesting that the mechanisms responsible for blood pressure and arterial stiffness lowering are independent of effects on autonomic nervous system activity. Nevertheless, our results do not exclude the possibility that other neurohormonal factors related to hyperglycaemia may have contributed to changes in blood pressure and arterial stiffness, such as decreased reactive oxygen species generation [48].

While neither RAAS, NO nor SNS activity can explain our results, we postulate that changes in arterial stiffness with empagliflozin are due to several other factors (Figure 1). First, weight loss independently decreases blood pressure and arterial stiffness [49]. Empagliflozin was associated with a significant reduction in weight of 2.7 kg in our cohort and this may have contributed to benefits on blood pressure and arterial stiffness, as previously reported [22, 24]. This weight loss was likely in part due to loss of fat, since waist circumference declined. Second, decreases in daily insulin doses have been correlated with improved arterial compliance in T2D, and a similar interaction may have occurred in our T1D cohort since total daily insulin doses decreased significantly by the end of treatment [22, 50]. Third, other diuretic agents such as thiazides lower arterial stiffness and these effects are likely in part due to direct effects on vascular smooth muscle relaxation after induction of a negative sodium balance [51, 52]. Similar diuretic effects with SGLT2 inhibition may have contributed to the decline in arterial stiffness in our cohort. Next, empagliflozin-induced weight loss and improved glycemic control may exert anti-inflammatory changes, which favour improvements in blood pressure and arterial stiffness [49]. For example, in experimental models of diabetes, SGLT2 inhibition reduces oxidative stress and suppresses markers of inflammation and fibrosis, including nuclear factor κβ and collagen IV expression [53, 54]. Due to the strong relationship between inflammation, cardiovascular complications and renal disease, future studies should clarify the ability of SGLT2 inhibitors to suppress inflammation in humans [55].

Due to their insulin-independent mechanism of action and as shown in animal models [56], SGLT2 inhibitors have the potential to be used in T1D [22] in addition to T2D [33]. Furthermore, SGLT2 inhibition may avoid some of the most common treatment-related side effects in T2D that have limited the benefits of traditional oral hypoglycemic agents, including significant weight gain, development and worsening of hypertension and severe hypoglycemic events [33, 39, 57]. Our observations are, to our knowledge, the first to suggest possible functional benefits of an oral hypoglycemic agent on large vessel function in T1D patients. The improvement in arterial compliance is important because previous studies have demonstrated strong, independent associations between macrovascular complications and arterial stiffness in T2D [58, 59]. Increased arterial stiffness is also independently associated with clinically relevant outcomes in T1D patients, including cardiovascular, renal, retinal and autonomic complications [42]. Importantly, the beneficial effects of conventional cardiovascular protective agents such as ACE inhibitors have been in part attributed to improvements in arterial compliance [60]. Ultimately, large ongoing clinical outcomes trials in T2D will determine if changes in arterial stiffness or blood pressure with SGLT2 inhibition translate into long-term renal or cardiovascular protection.

This trial has some limitations. First, the duration of therapy was limited to 8 weeks. Therefore, conclusions regarding longer term clinical benefits cannot be made from this study. Second, further work is needed to explore the interdependence among beneficial effects on arterial stiffness, including weight loss, glycemic control and blood pressure seen with empagliflozin. In particular, we could not determine if changes in blood pressure were based on declines in arterial stiffness, or whether reduced arterial stiffness was responsible for the fall in systolic blood pressure. Nevertheless, consistent changes in these vascular parameters are encouraging and support the rationale for ongoing long-term cardiovascular outcome trials. Third, these results in young patients with uncomplicated T1D cannot be generalized to older T1D patients, T1D patients with hypertension or to those with T2D. Given the important relationship between age and arterial stiffness, future work should determine if SGLT2 inhibition exerts similar effects in older individuals. Finally, measures of arterial stiffness have a high variability. As a consequence, our results should be viewed as exploratory and should be confirmed in future studies.