Are SGLT2 Inhibitors New Hypertension Drugs?

Kazuomi Kario, MD, PhD; Keith C. Ferdinand, MD; Wanpen Vongpatanasin, MD


Circulation. 2021;143(18):1750-1753. 

Recent large, randomized, placebo-controlled clinical trials have demonstrated that treatment with SGLT2 (sodium-glucose cotransport 2) inhibitors (SGLT2i) significantly reduces the rate of cardiovascular events (including heart failure [HF]) and prevents the progression of renal dysfunction (and ultimately chronic kidney disease) in patients with or without diabetes who were already receiving optimal guideline-directed medical therapy, as recently reviewed.[1] One of these studies is the CREDENCE trial (Canagliflozin and Renal Events in Diabetes With Established Nephropathy Clinical Evaluation) of canagliflozin 100 mg/d versus placebo on cardiovascular and renal outcomes in patients with type 2 diabetes and nephropathy.

This issue of Circulation includes a post hoc analysis of data from the CREDENCE trial that examined the office blood pressure (BP)–lowering effects of canagliflozin, overall and in patient subgroups defined on the basis of baseline systolic BP (SBP), number of BP-lowering drug classes, and history of apparent treatment-resistant hypertension.[2] Canagliflozin recipients showed an early and sustained reduction in BP (–3.5 mm Hg versus baseline); this was consistent across all patient subgroups, but of a smaller magnitude than reductions in office BP reported in other SGLT2i studies (Table 1 and Table 2). Although statistically significant, the reduction in office BP in the CREDENCE trial can really be described as "modest" and certainly is not of a magnitude to fully account for the 30% reduction in primary end point events (composition of end-stage kidney disease, a doubling of serum creatinine, or death from renal or cardiovascular causes) and the 39% reduction in the rate of HF. For example, it has previously been reported that a 10-mm Hg reduction in BP during antihypertensive treatment with conventional pharmacotherapy led to a 20% reduction in the risk of major cardiovascular disease and a 28% reduction in HF in the largest meta-analysis of the clinical trials.[10] Indeed, the mediation analysis in the post hoc evaluation of CREDENCE trial data demonstrated that the contribution of office BP reduction to the reduction in event rates was relatively small.[2] Furthermore, the magnitude of BP reduction in the CREDENCE trial was not associated with outcomes. However, there was a significant interaction between baseline SBP and the effect of canagliflozin on HF (P=0.04). The hazard ratio (95% CI) for HF-related hospitalization was 0.52 (95% CI, 0.29–0.94) when baseline SBP was <130 mm Hg, 0.36 (95% CI, 0.19–0.66) when baseline SBP was 130 to <140 mm Hg, 0.58 (95% CI, 0.32–1.04) when baseline SBP was 140 to <150 mm Hg, and 1.00 (95% CI, 0.65–1.54) when baseline SBP was ≥150 mm Hg.[2] The fact that there was no apparent benefit of canagliflozin in those with the highest BP seems counterintuitive because patients with the highest BP had double the event rate of the other groups, and therefore might have been expected to benefit most from therapy. A potential explanation for this finding is unclear, but this population is more likely to be older, to be Black, and to be prescribed calcium channel blockers or β-blockers. It is possible that a lack of adjustment for these unmeasured confounding factors along with nonadherence to study medication or other guideline-directed medical therapy may play a role. In addition, the observation is post hoc and exploratory in nature.

Another important point is that change in BP seen in the analysis by Ye et al[2] might simply reflect regression to the mean. The fall in BP at 3 weeks during canagliflozin therapy in patients with the highest baseline BP was 11 mm Hg, but at the same time, the BP reduction in the placebo group was 7 mm Hg. In contrast, BP increased in both treatment groups for those with a baseline BP of <130 mm Hg. This highlights the limitations of a post hoc analysis, and a prespecified planned analysis with 24-hour ambulatory BP monitoring at baseline and during the trial may have been more informative.

It is also important to note that an analysis that is based on office BP is likely to underestimate the contribution of SGLT2i-induced BP reduction to the overall clinical benefit because numerous observational studies have established the superiority of home or 24-hour ambulatory BP monitoring as an independent predictor of cardiovascular events beyond office BP alone.[11,12] These data have led to the recommendations from many international guidelines to incorporate out-of-office BP monitoring in the effective management of hypertension.

There is a growing body of evidence for the effects of SGLT2i on out-of-office BP. In the SACRA study (SGLT2 Inhibitor and Angiotensin Receptor Blocker Combination Therapy in Patients With Diabetes and Uncontrolled Nocturnal Hypertension), 12 weeks of treatment with empagliflozin in patients with type 2 diabetes and uncontrolled nocturnal hypertension reduced nighttime ambulatory SBP,[3] regardless of age (by 7.9 mm Hg in patients <75 years of age and by 4.2 mm Hg in those ≥75 years of age; corresponding reductions in mean 24-hour SBP were –11.0 and –8.7 mm Hg).[13] Empagliflozin also significantly reduced 24-hour ambulatory systolic BP versus placebo after 12 and 24 weeks of therapy in Black patients with type 2 diabetes and hypertension (placebo-corrected difference, −5.2 mm Hg [95% CI, −9.2 to –1.2]; P=0.0117; and –8.4 mm Hg [95% CI, −13.7 to –3.0]; P=0.0025, respectively).[4] The EMPAREG 24-hour ambulatory BP monitoring substudy demonstrated that the placebo-subtracted change in 24-hour ambulatory systolic BP was 4.2 mm Hg with the dose of 25 mg after 12 weeks.[5] In addition, a recent ambulatory BP monitoring study of dapagliflozin showed the significant reduction of 24-hour brachial ambulatory systolic BP (mean±SD, −5.8±9.5 mm Hg; versus placebo, mean±SD, −0.1±8.7 mm Hg; P=0.005).[6] Significant reductions in home BP in patients with type 2 diabetes have also been reported with empagliflozin (SACRA study; placebo subtracted difference, −7.5 mm Hg),[3] canagliflozin (SHIFT-J study [Study of the beneficial effect of Canagliflozin on nocturnal home BP in Japanese T2DM patients]),[7] luseogliflozin (LUSCAR study [luseogliflozin on cardiovascular function]),[8] and dapagliflozin (Y-AIDA study [Yokohama Add-On Inhibitory Efficacy of Dapagliflozin on Albuminuria in Japanese Patients With Type 2 Diabetes]).[9] Future studies needed to focus on identifying predictors of response to the 24-hour BP-lowering effects of SGLT2i, and the differential effects of agents in this class on 24-hour BP. Currently available data for BP reductions with currently available SGLT2i are summarized in the Table, but head-to-head comparisons are needed to provide definitive data.

In this analysis of the CREDENCE trial, the BP-lowering effect of canagliflozin was also observed in patients with resistant hypertension.[2] However, patients treated with a mineralocorticoid receptor antagonist were excluded from the CREDENCE trial, and no participants in the SACRA study were treated with a mineralocorticoid receptor antagonist, which is now considered to be the cornerstone of treatment for resistant hypertension. Similar to SGLT2i, mineralocorticoid receptor antagonists possess not only diuretic action but also vascular protection and sympathoinhibitory effects.[14] Reductions in 24-hour ambulatory systolic BP with SGLT2i in the studies on the patients with higher baseline 24-hour BP levels[3,4,13] appear to be similar to those of spironolactone in the PATHWAY-2 study of resistant hypertension (home SBP difference of –8.7 mm Hg versus placebo).[15] Nevertheless, SGLT2i appears to have several advantages over mineralocorticoid receptor antagonists in this population, especially with respect to unwanted adverse effects.

A unique feature of the CREDENCE trial is the inclusion of patients with resistant hypertension. However, additional studies are needed to determine the role of SGLT2i in optimizing BP control and reducing cardiovascular outcomes in patients with resistant hypertension with or without diabetes.