Lessons Learned From Major Clinical Outcomes Trials Involving Sodium–Glucose Cotransporter 2 Inhibitors

To date, a total of nine large, randomized clinical trials have been completed with SGLT2 inhibitors. Seven have been conducted in subjects with type 2 diabetes, six of which were designed as noninferiority cardiovascular outcomes trials (CVOTs), each additionally powered to assess for cardiovascular effectiveness. Two have been dedicated HF trials, and one was primarily a renal outcomes trial. Below, we briefly describe the findings of each.

EMPA-REG OUTCOME was the first major trial of SGLT2 inhibitors to assess cardiovascular outcomes. The study included adult subjects with type 2 diabetes and established CVD. Additional inclusion criteria were a BMI ≤45 kg/m², estimated glomerular filtration rate (eGFR) ≥30 mL/min/1.73 m², and A1C between 7.0 and 9.0% (for drug-naive participants) or between 7.0 and 10.0% for those already using pharmacologic glucose-lowering therapy. A total of 7,020 subjects were randomized to empagliflozin 10 or 25 mg daily or placebo and followed for a median of 3.1 years (8).

The primary outcome was the time to a composite three-point end point of major adverse cardiovascular events (MACE), including cardiovascular death, nonfatal myocardial infarction (MI), or nonfatal stroke. The relative risk for this MACE composite was reduced by 14% in those assigned to empagliflozin versus placebo (hazard ratio [HR] 0.86, 95% CI 0.74–0.99), not only meeting the statistical threshold for noninferiority (P <0.001) but also for superiority (P = 0.04) (8).

With regard to secondary cardiovascular outcomes, there were dramatic relative risk reductions (RRRs) in cardiovascular death (38%, HR 0.32, 95% CI 0.49–0.77), all-cause death (32%, HR 0.68, 95% CI 0.57–0.82), and hospitalization for heart failure (HHF) (35%, HR 0.65, 95% CI 0.50–0.85). Renal function was also tracked as a secondary outcome, with an equally large 39% RRR in incident or worsening nephropathy, defined as urinary albumin-to-creatinine ratio (UACR) >300 mg/g, doubling of serum creatinine with an eGFR <45 mL/min/1.73 m², initiation of renal replacement therapy, or death from renal disease (HR 0.61, 95% CI 0.53–0.70). No significant effects on other vascular outcomes such as MI or stroke were observed (8).

These benefits occurred with only modest improvements in the active therapy group in terms of A1C, blood pressure, and body weight, suggesting that the benefits of empagliflozin were independent of these intermediate parameters. In fact, in a follow-up mediation analysis from the trial, it was the absolute increase in hematocrit (∼3%) that appeared to be the most significant mediator (52%) of the reduction in cardiovascular death, suggesting that plasma volume reduction from this medication with known diuretic actions may be important.

With regard to adverse events, subjects on active therapy experienced an increase in genital mycotic infections, with rare cases of diabetic ketoacidosis (DKA) (8).

The second major CVOT of an SGLT2 inhibitor was the CANVAS research program, which combined 10,142 subjects from two separate trials (10). CANVAS trial subjects were randomized to placebo or 100 or 300 mg canagliflozin daily. In CANVAS-R, subjects were randomized to placebo or 100 mg canagliflozin with an optional increase to 300 mg after 3 months based on glycemic control. The study included adults with type 2 diabetes who were ≥30 years of age with a history of atherosclerotic cardiovascular disease (ASCVD) or ≥50 years of age with two or more cardiovascular risk factors and an A1C between 7.0 and 10.5%. Nearly two-thirds of the study population had a history of CVD. The mean follow-up time was 3.6 years (10).

The main result was a 40% RRR in a renal composite outcome including 40% reduction in eGFR, initiation of renal replacement therapy, or death from a renal cause (HR 0.60, 95% CI 0.47–0.77). There was a 27% reduction in progression of albuminuria (HR 0.73, 95% CI 0.67–0.79), a 14% reduction in three-point MACE (HR 0.86, 95% CI 0.75–0.97), and a 33% reduction in HHF (HR 0.67, 95% CI 0.52–0.87). There were no significant reductions in cardiovascular death (HR 0.87, 95% CI 0.72–1.06) or all-cause mortality (HR 0.87, 95% CI 0.74–1.01) (10).

There was an increase not only in genital mycotic infections, but also in bone fractures and lower-extremity amputations in the canagliflozin group (10).

CREDENCE was a double-blind, randomized controlled trial designed to evaluate the effects of canagliflozin 100 mg daily on renal outcomes (9). A total of 4,401 subjects underwent randomization to placebo or canagliflozin 100 mg daily and were followed for a median of 2.6 years. The study included subjects with type 2 diabetes who were ≥30 years of age with underlying kidney disease as manifested by an eGFR of 30 to <90 mL/min/1.73 m² and a UACR >300–5,000 mg/g. About half of the study participants also had a history of CVD. All participants were treated with renin-angiotensin system blockade for ≥4 weeks before randomization (9).

The main result was a 30% improvement in a primary composite outcome of development of end-stage renal disease (ESRD), including initiation of dialysis, renal transplantation, or a sustained eGFR <15 mL/min/1.73 m² (HR 0.70, 95% CI 0.59–0.82). There was a 34% improvement in the renal-specific composite outcome comprised of ESRD, doubling of serum creatinine, renal death, or cardiovascular death (HR 0.66, 95% CI 0.53–0.81); a 32% improvement in ESRD (HR 0.68, 95% CI 0.54–0.86); and a 28% improvement in the more restrictive renal composite of dialysis, kidney transplantation, or renal death (HR 0.72, 95% CI 0.54–0.97). There was also a 20% reduction in three-point MACE (HR 0.80, 95% CI 0.67–0.95). There was no significant effect on all-cause mortality (HR 0.83, 95% CI 0.68–1.02) (9).

The side effect profile was similar to that in the EMPA-REG OUTCOME trial. Neither the amputation nor the fracture signal found in the CANVAS Program was confirmed in the CREDENCE trial (9).

This trial randomized 17,160 subjects with type 2 diabetes to either dapagliflozin or placebo and followed them for a median of 4.2 years. Subjects either had established CVD or risk factors defined by age (males ≥55 years, females ≥60 years), hypertension, hyperlipidemia, or tobacco use. One-third (33%) of the study participants had CVD at baseline (11).

The study had two coprimary outcomes—three-point MACE and a composite of cardiovascular death and HHF. The first proved neutral, whereas the risk for the second was significantly reduced by dapagliflozin by 17% (HR 0.83, 95% CI 0.73–0.95), driven by a 27% reduction in HHF (HR 0.73, 95% CI 0.61–0.88). There was also a 24% RRR in a renal composite consisting of ≥40% decrease in eGFR, ESRD, or death from renal or cardiovascular cause (HR 0.76, 95 % CI 0.67–0.87) (11).

Adverse events were similar to those observed in prior studies, but no fracture or amputation risk was seen (11).

The DECLARE-TIMI 58 trial supported what prior studies had shown; however, it did not have as robust results, possibly related to a mainly primary prevention population and a lack of subjects with significantly reduced kidney function at baseline.

The VERTIS CV trial consisted of 8,246 subjects who were ≥40 years of age and had type 2 diabetes and known CVD. Subjects were randomized to ertugliflozin 5 or 15 mg or placebo daily (12).

Here, there was no significant effect on three-point MACE or its components, including cardiovascular death. There was, however, a 30% reduction (HR 0.70, 95%CI 0.54–0.90) in HHF, but the renal outcome was numerically but not significantly reduced (HR 0.81, 95% CI 0.63–1.04) (12).

With regard to adverse effects, subjects were more likely to have genital infections, with an increase in DKA and also a small increase in amputations (12).

DAPA-HF was the first trial evaluating the effect of SGLT2 inhibitors on HF outcomes independent of preexisting type 2 diabetes. The trial included 4,744 patients with HF with reduced ejection fraction (HFrEF) who underwent randomization to dapagliflozin 10 mg daily or placebo and were followed for 2 years. Subjects were included if they were ≥18 years of age, had New York Heart Association (NYHA) class II–IV HF with ejection fraction ≤40%, and had an N-terminal-pro hormone B-type natriuretic peptide (NT-proBNP) level ≥600 pg/mL (≥400 pg/mL if recently hospitalized and ≥900 pg/mL with atrial fibrillation). Subjects were optimized from an HF perspective with standard therapy, including ACE inhibitor, β-blocker, implantable cardioverter-defibrillator (ICD), and cardiac resynchronization therapy (CRT), as indicated (13).

There was a 26% reduction in the primary outcome, which was a composite of hospitalization or urgent visit for HF and cardiovascular death (HR 0.74, 95% CI 0.65–0.85). There was also a 30% reduction in HHF (HR 0.70, 95% CI 0.59–0.83), an 18% reduction in cardiovascular death (HR 0.82, 95% CI 0.69–0.98), and a 17% reduction in all-cause mortality (HR 0.83, 95% CI 0.71–0.97). Subjects also had improvements in NT-proBNP, HF symptoms, physical function, and quality of life as measured by the Kansas City Cardiomyopathy questionnaire (13,14). There was no significant heterogeneity in the effect of dapagliflozin across subgroups; importantly, the point estimates for major outcomes were essentially the same for those with and without diabetes (13).

The adverse effect profile of dapagliflozin was surprisingly bland in this trial. This was the first study showing efficacy of SGLT2 inhibitors on HHF and cardiovascular death irrespective of preexisting diabetes (13).

This was the second trial evaluating the effect of SGLT2 inhibitors on cardiovascular and HF outcomes independent of diabetes. It evaluated HF outcomes in 3,730 subjects randomized to placebo or 10 mg of empagliflozin daily. Subjects were ≥18 years of age with chronic HF NYHA class II–IV, an ejection fraction ≤40%, HHF within the past 12 months, and NT-proBNP ≥2,500 pg/mL (≥1,000 pg/mL with ejection fraction ≤31–35% and ≥600 pg/mL with ejection fraction ≤30%). NT-proBNP thresholds were doubled in subjects with atrial fibrillation. Subjects were optimized from an HF standpoint, including diuretics, renin-angiotensin inhibitors, neprilysin inhibitors, β-blockers, mineralocorticoid receptor antagonists, ICD, and CRT, as indicated. Median follow-up time was 16 months (15).

There was a 25% reduction of the primary composite outcome of cardiovascular death or HHF (HR 0.75, 95% CI 0.65–0.86). There was also a 30% reduction in HHF (HR 0.70, 95% CI 0.58–0.85) and a 50% reduction in a renal composite defined as chronic dialysis, renal transplant, or sustained reduction in eGFR (HR 0.50, 95% CI 0.32–0.77). In contrast to the DAPA-HF trial, there were no significant reductions in cardiovascular death (HR 0.92, 95% CI 0.75–1.12) or all-cause mortality (HR 0.92, 95% CI 0.77–1.10). Patients with or without diabetes at baseline experienced similar benefits (15).

With regard to adverse events, there were increases in hypotension, volume depletion, and genital infections and very modest increases in amputations and fractures (15).

DAPA-CKD was the first trial evaluating the effects SGLT2 inhibitors on CKD in patients with or without diabetes. It evaluated CKD outcomes in 4,034 subjects randomized to placebo or 10 mg dapagliflozin daily. Inclusion criteria were age ≥18 years, eGFR 25–75 mL/min/1.73 m², and UACR 200–5,000 mg/g. All trial participants were previously optimized with an ACE inhibitor or angiotensin receptor blocker. Median follow-up was 2.4 years (16).

Risk of the primary composite outcome of sustained ≥50% decline in eGFR, ESRD, or death from renal or cardiovascular causes was reduced by 39% by dapagliflozin (HR 0.61, 95% CI 0.51–0.72). There was also a 44% reduction in a renal-specific composite outcome of ≥50% decline in eGFR, ESRD, and death from a renal cause (HR 0.56, 95% CI 0.45–0.68), a 29% reduction in a composite of cardiovascular death or HHF (HR 0.71, 95% CI 0.55–0.92), and a 31% reduction in all-cause mortality (HR 0.69, 95% CI 0.53–0.88). As with the HF studies, there was no statistical interaction between the presence or absence of diabetes (16).

No new safety concerns were raised (16). DAPA-CKD was the first study showing efficacy of SGLT2 inhibitors on CKD regardless of preexisting diabetes.

Overall, SGLT2 inhibitors substantially increase the risk of genital mycotic infections, although these typically minor infections are usually easily treated with antifungal medications. Drugs in this class also increase the risk of DKA, but this is very uncommon in type 2 diabetes. (DKA is more commonly encountered when these medications are used in trials or off label in practice in patients with type 1 diabetes—something we do not advise [8–12].)

There have been post-marketing reports to regulatory agencies of episodes of necrotizing fasciitis of the perineum, known as Fournier’s gangrene, in patients treated with SGLT2 inhibitors (17). These events have not been detected in clinical trials, but, as they are rare, this is not surprising (8–13,15,16). Based on the nature of these reports, it is not known whether SGLT2 inhibitors actually increase this risk.

These and other adverse effects are summarized in Table 1. The fracture and amputation signals with canagliflozin in the CANVAS Program have not been seen, generally speaking, with the SGLT2 inhibitors as a class (8–12). Caution remains advisable, however, in patients with severe peripheral arterial disease and those with prior amputations. Other concerns such as the initial small drop in eGFR, polyuria, urinary frequency, and propensity for dehydration should be considered at an individual patient level. As with any therapeutic, the risks and benefits of these agents must be weighed with each prescription. Importantly, users must have a reasonable eGFR for these medications to work, although their cardiovascular and renal effects may persist at levels of CKD that do not allow for significant glucose lowering. The specific dosing and eGFR prescribing requirements for each SGLT2 inhibitor are detailed in Table 2 (18–21).

Several meta-analyses have compiled and systematically evaluated data from the major SGLT2 trials.

Zelniker et al. (22) conducted a pooled analysis of the EMPA-REG OUTCOME, CANVAS Program, and DECLARE-TIMI 58 studies, for a total of 34,322 subjects. The investigators found that SGLT2 inhibitors led to RRRs of the three-point MACE of 14% (HR 0.86, 95% CI 0.80–0.93), cardiovascular mortality of 16% (HR 0.84, 95% CI 0.75–0.94), HHF of 31% (HR 0.69, 95% CI 0.61–0.79), a composite of cardiovascular death or HHF of 23% (HR 0.77, 95% CI 0.71–0.84), worsening kidney disease of 45% (HR 0.55, 95% CI 0.48–0.64), and all-cause mortality of 15% (HR 0.85, 95% CI 0.78–0.93). When stratified by the presence or absence of baseline ASCVD, there was a 14% reduction in three-point MACE in subjects with preexisting disease (HR 0.86, 95% CI 0.80–0.93) but no significant reduction in subjects with only cardiovascular risk factors (HR 1.00, 95% CI 0.87–1.16). Subjects also had a significant decrease in renal composite outcomes irrespective of the presence of ASCVD (with ASCVD: HR 0.56, 95% CI 0.47–0.67; with risk factors only: HR 0.54, 95% CI 0.42–0.71). Subjects randomized to SGLT2 inhibitors experienced a decrease in HHF irrespective of ASCVD at baseline (with ASCVD: HR 0.71, 95% CI 0.62–0.82; with risk factors only: HR 0.64, 95% CI 0.48–0.84) or HF at baseline (with HF: HR 0.71, 95% CI 0.61–0.84; without HF: HR 0.79, 95% CI 0.71–0.88) (22). A graphic summary of the main findings is present in Figure 1.

Another meta-analysis, by Arnott et al. (23), pooled data from these three trials plus the CREDENCE trial. The investigators found RRRs of 12% in three-point MACE (HR 0.88, 95% CI 0.82–0.94), 17% in cardiovascular death (HR 0.83, 95% CI 0.75–0.92), 15% in all-cause mortality (HR 0.85, 95% CI 0.79–0.92), and 12% in MI (HR 0.88, 95% CI 0.80–0.97). There was no significant effect on stroke. For HHF, the RRR was 32% (HR 0.68, 95% CI 0.60–0.76). This meta-analysis also assessed serious adverse events and did not find any significant effect with regard to severe hypoglycemia or bone fracture. There was, however, a significant increase in DKA (relative risk [RR 2.46, 95% CI 1.43–4.24) and amputations (RR 1.23, 95% CI 1.05–1.44). The heterogeneity on the finding of DKA was low (I² = 0%); however, heterogeneity was high for the finding of amputations (I² = 70.0%), likely driven by the higher amputation rate in the CANVAS Program that was not confirmed in other studies (23). The primary findings of this safety analysis are summarized in Figure 2.

Taken together, these pooled analyses demonstrate substantive benefits in HF and CKD that are highly consistent among the individual trials. In contrast, the benefit for MACE is not as definitive, with small reductions in three of the four studies included. The MACE benefit appears to be contingent on preexisting ASCVD, whereas improvements in HF and renal outcomes do not. The VERTIS CV trial (12), which has not yet been incorporated into a published meta-analysis, has shown neutrality for MACE.

The outcomes of the CVOTs have had a major impact on labeled indications for these compounds and for medical society guidelines and recommendations for clinical practice.

All marketed SGLT2 inhibitors in the United States are labeled for use as adjuncts to diet and exercise to improve glycemic control in adults with type 2 diabetes. Empagliflozin also has a labeled indication to reduce the risk of cardiovascular death in adults with type 2 diabetes and established CVD (18). Canagliflozin is labeled to reduce the risk of MACE in adults with type 2 diabetes and established CVD and to reduce the risk of ESRD, doubling of serum creatinine, cardiovascular death, HHF, and diabetic nephropathy with albuminuria (20). Dapagliflozin has a labeled indication to reduce the risk of HHF in adults with type 2 diabetes and established CVD or multiple cardiovascular risk factors, as well as the risk of cardiovascular death and HHF in adults with HFrEF, irrespective of the presence of diabetes (19).

With regard to guidelines, an American Diabetes Association/European Association for the Study of Diabetes 2018 consensus statement recommended that patients be stratified according to risk for ASCVD, HF, and CKD initially. By these guidelines, patients with ASCVD who are already taking metformin should be started on either an evidence-based glucagon-like peptide 1 (GLP-1) receptor agonist or an SGLT2 inhibitor as a second-line agent (24). For those with HF or mild-to-moderate CKD, the preferred choice would be an SGLT2 inhibitor. A 2019 update to this consensus statement included two modifications: first, the decision to add these medications should be considered independent of current A1C or A1C target, and second, patients without established ASCVD but at high risk should also be considered for such a strategy (25). Patients without heart or kidney disease could still be considered for SGLT2 inhibitor therapy if weight or hypoglycemia are concerns. However, both SGLT2 inhibitors and GLP-1 receptor agonists are currently very expensive and therefore may be cost-prohibitive for many patients (26).

The European Society of Cardiology in 2019 released a guideline recommending that patients with ASCVD or high risk of ASCVD be prescribed an SGLT2 inhibitor or GLP-1 receptor agonist irrespective of baseline metformin use. It also recommended that patients with diabetes be treated with an SGLT2 inhibitor to reduce HF risk and optimize management of CKD (eGFR ≥30 mL/min/1.73 m²) (27).

The American College of Cardiology in 2020 updated its expert consensus decision pathway for use of novel antihyperglycemic agents. This document recommends that patients who are ≥18 years of age and have type 2 diabetes, HF, diabetic kidney disease, high risk for ASCVD, or a history of ASCVD should be considered for initiation of an SGLT2 inhibitor as long as there are no contraindications to these medications. A GLP-1 receptor agonist was recommended for patients ≥18 years of age with type 2 diabetes, high risk for ASCVD, or a history of ASCVD (28).

The KDIGO (Kidney Disease Improving Global Outcomes) consortium recently updated its clinical practice guidelines on diabetes management in CKD. This group now recommends that most patients with type 2 diabetes, CKD, and an eGFR ≥30 mL/min/1.73 m² would benefit from treatment with an SGLT2 inhibitor. Patients who initiate SGLT2 inhibitor therapy at an eGFR ≥30 mL/min/1.73 m² and then decline to an eGFR <30 mL/min/1.73 m² may be continued on the medication. However, treatment with an SGLT2 inhibitor should not be initiated in patients with an eGFR <30 mL/min/1.73 m². Also, individuals with type 1 diabetes and kidney transplant recipients are excluded from these recommendations (29).

We generally concur with these recommendations, although we still prefer the initial or concurrent use of metformin given its low cost, glycemic efficacy, long safety record, and possible cardiovascular benefits (the latter admittedly from older and smaller trials).

SGLT2 inhibitors are newer glucose-lowering agents approved for use in individuals with type 2 diabetes. Large cardiovascular and renal outcome trials indicate a mostly consistent overall benefit from these drugs on HF and kidney disease, with more variable effects on MACE and cardiovascular death. These trials have expanded the prescribing indications of drugs in this class to improve both cardiovascular and renal outcomes. Two dedicated trials involving subjects with HFrEF have also demonstrated that SGLT2 inhibitors reduce HF outcomes regardless of the presence of diabetes, resulting in new indications for the class beyond diabetes. These recent findings and regulatory feedback have now resulted in major shifts in treatment guidelines, which now broadly encourage the use of SGLT2 inhibitors in at-risk patients.

S.E.I. has participated on clinical trial executive/steering/publications committees and/or served as an advisor for AstraZeneca, Boehringer Ingelheim, Esperion, and Novo Nordisk and has delivered lectures supported by AstraZeneca, Boehringer Ingelheim, and Merck. No other potential conflicts of interest relevant to this article were reported.

B.S.W. researched data and wrote the manuscript, and S.E.I. reviewed and edited the manuscript. B.S.W. is the guarantor of this work and, as such, takes responsibility for the integrity of this review.