figure b

Introduction

Sodium–glucose cotransporter 2 (SGLT2) inhibitors reduce the risk of cardiovascular (CV) events, including CV death or heart failure (HF), in patients with type 2 diabetes mellitus [1], and in patients with HF with reduced ejection fraction independent of type 2 diabetes status [2, 3]. However, since women remain underrepresented across clinical trials, it is important to define the efficacy and safety of SGLT2 inhibitors by participant sex. Supporting these concerns, sex disparities already exist in the management and treatment of CV risk factors in women with type 2 diabetes [4]. In the presence of a similar burden of risk factors, women are less likely than men to be treated with LDL-C-lowering therapies or to achieve adequate BP or glycaemic control [4]. As such, in a pre-specified analysis, we assessed in a large population with robust female representation (n = 6422, 37.4%) whether sex modifies the efficacy and safety of the SGLT2 inhibitor dapagliflozin in individuals with type 2 diabetes with or at increased risk of atherosclerotic disease in the Dapagliflozin Effect on Cardiovascular Events–Thrombolysis in Myocardial Infarction 58 (DECLARE-TIMI 58) trial [5].

Methods

Study population and procedures

The design and results of the DECLARE-TIMI 58 trial have been reported previously [5, 6]. In brief, DECLARE-TIMI 58 was a Phase III, multinational, double-blind, placebo-controlled trial that randomised 17,160 patients with type 2 diabetes with or at risk for atherosclerotic disease to dapagliflozin vs placebo. Eligible patients were 40 years or older with type 2 diabetes, had a creatinine clearance ≥60 ml/min and either multiple risk factors for atherosclerotic CV disease (ASCVD) or established ASCVD, coronary artery disease, cerebrovascular disease or peripheral artery disease. Eligible participants with multiple risk factors were men ≥55 years of age or women ≥60 years of age with at least one additional traditional ASCVD risk factor including hypertension, dyslipidaemia or current tobacco use. Following a single-blind placebo run-in period, patients who remained eligible were randomised in a double-blind fashion to dapagliflozin 10 mg/day vs matching placebo and followed up for a median of 4.2 years.

Outcomes

The dual efficacy outcomes were the composites of (1) CV death or HF hospitalisation (HHF) and (2) major adverse cardiovascular events (MACE; CV death, myocardial infarction or ischaemic stroke). The pre-specified cardiorenal outcome was the composite of a decrease of ≥40% in eGFR to <60 ml min−1 [1.73 m]−2, end-stage renal disease (ESRD) or CV or renal death. The pre-specified renal-specific outcome was the composite of a ≥40% drop in eGFR to <60 ml min−1 [1.73 m]−2, new ESRD or renal death. Safety events collected were adverse events leading to drug discontinuation, adverse events of special interest or serious adverse events (SAEs). An independent and blinded clinical events committee adjudicated all CV outcomes analysed.

Statistical analysis

Baseline characteristics are presented as medians (IQRs) for continuous variables and frequencies for categorical variables. Baseline characteristics were compared with the Wilcoxon rank sum tests for continuous variables and χ2 tests for categorical variables.

Mixed models for repeated measures in HbA1c, weight, systolic BP and diastolic BP were analysed to produce least-squares mean (LSM) estimates and 95% CIs by treatment and sex subgroup. Efficacy analyses were conducted with Cox proportional hazards models that included a treatment arm, two randomisation stratification factors (presence of established atherosclerotic disease and baseline haematuria) and run separately by participant sex as captured on the electronic case-report form. Effect modification was assessed by including interaction terms in the models. All efficacy analyses were conducted in the intention-to-treat study population and event rates are reported as Kaplan–Meier estimates at 4 years. Safety analyses were performed using the on-treatment analysis set, as previously described, except for amputation, fracture and malignancy outcomes, which included all events after first dose in all patients who were randomised and received at least one dose of the study drug [5, 7]. All tests were two-sided with a p value <0.05 considered to be significant. The TIMI study group conducted all analyses. Analyses were performed using Stata/SE version 16.1 (Stata, College Station, Texas) or SAS version 9.4 (SAS Institute, Cary, NC, USA).

Results

Of the 17,160 patients enrolled in the DECLARE-TIMI 58 trial, 6422 (37.4%) were women. The baseline characteristics of the study population by participant sex are summarised in Table 1. Women were treated with fewer non-insulin glucose-lowering medications than men were (Table 2), and these findings were largely consistent across regions (electronic supplementary material [ESM] Table 1) and by age and/or qualifying disease status (ESM Table 2). The use of metformin, dipeptidyl peptidase-4 (DPP-4) inhibitors and glucagon-like peptide 1 receptor agonists (GLP1-RA) was significantly lower in women than men (p < 0.001). Background use of insulin and sulfonylureas did not differ by sex (Table 2). Crude rates of study drug discontinuation were similar in both women and men (23.6% vs 22.8%, p = 0.23), including both the active (21.3% vs 21.0%, p = 0.72) and placebo arms (25.8% vs 24.6%, p = 0.21).

Table 1 Baseline characteristics for women and men in DECLARE-TIMI 58
Full size table
Table 2 Baseline use of glucose-lowering medication for women and men
Full size table

Effect of dapagliflozin on CV risk factors

Participants randomised to dapagliflozin had a lower HbA1c at month 12 than participants randomised to placebo did; this was true for both women (LSM absolute difference −0.49%, 95% CI −0.55, −0.43; or −3.59 mmol/mol, 95% CI −4.30, −2.87) and men (LSM absolute difference −0.55%, 95% CI −0.59, −0.51; or −3.81 mmol/mol, 95% CI −4.36, −3.25) (pinteraction = 0.07). Similarly, patients treated with dapagliflozin had a lower body weight at 12 months than placebo-treated patients did regardless of sex (women: 12-month LSM absolute difference −1.7 kg, 95% CI −1.9, −1.6; men: 12-month LSM absolute difference −1.8 kg, 95% CI −2.0, −1.7; pinteraction = 0.64]. At 12 months, patients treated with dapagliflozin had lower systolic BP than placebo-treated patients; this was true for both women (−2.7 mmHg, 95% CI −3.4, −2.0) and men (−3.0 mmHg, 95% CI −3.5, −2.5) (pinteraction = 0.52); similarly, the difference in diastolic BP between treatment groups was −0.8 mmHg (95% CI −1.2, −0.4) in women and −0.9 mmHg (95% CI −1.2, −0.6) in men (pinteraction = 0.87).

Efficacy outcomes

In the placebo arm, the crude incidence of CV death/HHF was 4.5% in women and 6.4% in men, and the incidence of MACE was 6.8% in women and 10.7% in men. All-cause mortality at 4 years was 5.4% in women and 6.3% in men. The incidence of MACE remained lower in women than men for those patients with (13.3% vs 16.2%) or without established ASCVD (4.0% vs 6.2%). Similarly, the incidence of CV death/HHF was lower in women than men in those with (15.2% vs 22.3%) or without a prior HF (3.5% vs 4.6%).

Dapagliflozin reduced the risk of CV death/HHF in women (HR 0.84, 95% CI 0.66, 1.07) and in men (HR 0.83, 95% CI 0.71, 0.96; pinteraction = 0.90; Fig. 1). The effects of dapagliflozin on risk of MACE did not differ between women (HR 0.93, 95% CI 0.77, 1.12) and men (HR 0.93, 95% CI 0.83, 1.05; pinteraction = 0.99; Table 3). Effects of dapagliflozin on risk of myocardial infarction also did not differ by sex (women: HR 0.89, 95% CI 0.67, 1.17; men: HR 0.88, 95% CI 0.75, 1.03; pinteraction = 0.99).

Fig. 1
figure 1

The cumulative incidence of CV death or HHF in women and men by randomised treatment arm in DECLARE-TIMI 58. Dapagliflozin similarly reduced the risk of CV death/HHF in women (HR 0.84, 95% CI 0.66, 1.07) and in men (HR 0.83, 95% CI 0.71, 0.96; pinteraction=0.90)

Full size image
Table 3 Efficacy of dapagliflozin vs placebo stratified by participant sex
Full size table

The cardiorenal composite outcome was reduced by dapagliflozin in women (HR 0.68, 95% CI 0.54, 0.86) and in men (HR 0.81, 95% CI 0.68, 0.96; pinteraction = 0.26; Table 3). For the renal-specific composite outcome, dapagliflozin reduced events in women (HR 0.50, 95% CI 0.35, 0.70) and in men (HR 0.55, 95% CI 0.42, 0.73; pinteraction = 0.64; Fig. 2; ESM Fig. 1).

Fig. 2
figure 2

The cumulative incidence of renal-specific events (a decrease of ≥40% in eGFR to <60 ml min˗1 [1.73 m]˗2, ESRD, or renal death) by participant sex and randomised treatment arm. Kaplan–Meier event rates at 4 years are displayed. Dapagliflozin reduced renal-specific events in women (HR 0.50, 95% CI 0.35, 0.70) and in men (HR 0.55, 95% CI 0.42, 0.73; pinteraction=0.64)

Full size image

In patients with established ASCVD, the HR for dapagliflozin vs placebo for risk of MACE was 0.85 (95% CI 0.66, 1.09) in women and 0.91 (95% CI 0.79, 1.05) in men (pinteraction = 0.63).

In patients with prior HF, dapagliflozin reduced the risk of CV death/HHF in women (0.78, 95% CI 0.51, 1.20) and in men (HR 0.81, 95% CI 0.62, 1.05; pinteraction = 0.89). In patients with prior myocardial infarction, the HR for dapagliflozin vs placebo for risk of MACE was 0.71 (95% CI 0.50, 1.02) in women and 0.88 (95% CI 0.74, 1.06) in men (pinteraction = 0.29). Similarly, the HR for dapagliflozin vs placebo for risk of recurrent myocardial infarction in patients with prior myocardial infarction were 0.70 (95% CI 0.45, 1.10) in women and 0.80 (95% CI 0.63, 1.00) in men (pinteraction = 0.65).

Safety outcomes

Treatment-emergent SAEs were less common in dapagliflozin-treated than placebo-treated women (29.3% vs 31.5%) and men (36.9% vs 39.0%; pinteraction = 0.78; Table 4). Urinary tract infections (SAEs or leading to drug discontinuation) were more frequent in women than men, but were not different in those randomised to dapagliflozin or placebo, irrespective of sex (women: 2.2% vs 2.1%; men: 1.0% vs 1.2%; pinteraction = 0.30); genital mycotic infections (SAEs or leading to drug discontinuation) were more common with dapagliflozin in both women (1.0% vs 0.1%) and men (0.8% vs 0.1%; pinteraction = 0.93). The incidence of diabetic ketoacidosis (DKA) with dapagliflozin vs placebo was 0.5% vs 0.2% in women and 0.2% vs 0.1% in men (pinteraction = 0.56). The incidence of amputation with dapagliflozin compared with placebo was not different between women (0.7% vs 0.6%) and men (1.9% vs 1.7%; pinteraction = 0.87; Table 4).

Table 4 The safety of dapagliflozin vs placebo stratified by participant sex
Full size table

Discussion

In patients with type 2 diabetes with or at high risk for ASCVD, the SGLT2 inhibitor dapagliflozin demonstrated comparable efficacy and safety in both women and men. Specifically, dapagliflozin significantly reduced the risk of CV death or HHF by 16–17%, irrespective of sex. Dapagliflozin also significantly reduced the risk of renal events by 45–50% irrespective of sex.

The current analysis uncovered notable differences at baseline in the management of type 2 diabetes in women and men. Although women had slightly higher baseline HbA1c and slightly longer duration of type 2 diabetes, women were less likely to be treated with non-insulin glucose-lowering medications including metformin, DPP-4 inhibitors and GLP-1 RAs. Although not previously well described for glucose-lowering medications in patients with type 2 diabetes, it is well established that women are less likely to be treated with evidence-based therapies across several disease states, including the management of CVD [8]. Although the reasons for these differences may be multifactorial and need to be elucidated, continued emphasis on the use of appropriate evidence-based therapies in the setting of CV risk factors in both women and men is of the utmost importance. In the current analysis, it cannot be determined whether the relative underuse of non-insulin glucose-lowering medications in women was warranted, but this would be an important avenue for future research.

To date, the efficacy and safety of SGLT2 inhibitors has not been compared between women and men. In the EMPA-REG OUTCOME trial (n = 2004 women), empagliflozin demonstrated comparable benefit toward reducing CV events and slowing nephropathy irrespective of sex, but suggested a possible absolute excess in the risk of genital infections with empagliflozin in women (10.0% vs 2.5%) compared with men (2.6% vs 1.5%). Other safety outcomes were not specifically reported by sex [9]. In the CANVAS programme (n = 3633 women) [10] and CREDENCE trial (n = 1494 women) [11], canagliflozin similarly had comparable CV and renal protective effects by sex, but safety data by sex were not published. Prior to the completion of DECLARE-TIMI 58, a pooled analysis of Phase IIb/III data for dapagliflozin demonstrated that women were more likely than men to experience urinary tract or genital infections irrespective of treatment with dapagliflozin, but did not specifically address the relative risk of these events for women and men treated with the drug owing to relatively fewer events (n = 667 women and 3296 men treated with dapagliflozin in a 24-week pool) [12].

In the present analyses of DECLARE-TIMI 58 (n = 6422 women with a median follow-up of 4.2 years), dapagliflozin demonstrated similar CV efficacy and renal protection in both women and men. In DECLARE-TIMI 58, in patients with prior myocardial infarction, dapagliflozin significantly reduced the risk of recurrent myocardial infarction by 22% (95% CI 5, 27) in the overall trial with directionally similar effects in women (30% relative risk reduction) and men (20% relative risk reduction), thereby supporting the concept that the CV benefits of SGLT2 inhibition toward reducing atherosclerotic events may be enhanced in patients with established coronary disease [1]. Although dapagliflozin increased the risk of genital mycotic infections (SAEs or those leading to drug discontinuation), the relative excess was similar in both women (1.0% vs 0.1%) and men (0.8% vs 0.1%), and urinary tract infections were not increased compared with placebo; however, individuals at highest risk of genitourinary infections may not have been enrolled in the trial. Although infrequent, a numerical excess in DKA cases was also observed with dapagliflozin vs placebo, as has been described with other SGLT2 inhibitors, in both women (0.5% vs 0.2%) and men (0.2% vs 0.1%). Symptoms of volume depletion and amputation risk were not increased with dapagliflozin in participants of either sex.

Limitations to the current analyses include that individual subgroups were underpowered for statistical significance; therefore, one cannot definitively exclude that a study with a larger population would detect differences in efficacy and safety by participant sex. Nonetheless, the DECLARE-TIMI 58 trial was the largest of the Phase III trials of an SGLT2 inhibitor in type 2 diabetes [10, 13, 14].

In summary, the use of and interest in SGLT2 inhibitors with regard to CV and kidney effects has continued to expand because randomised trials have demonstrated consistent CV and kidney benefit in patients with or without type 2 diabetes in the presence of chronic kidney disease or HF with reduced left ventricular ejection fraction. Therefore, the current results provide important reassurance that the efficacy and safety of dapagliflozin are consistent in both women and men.