Summary

• The CANVAS Program^1,2 (N=10,142) comprised the 2 large INVOKANA cardiovascular (CV) outcome studies, CANVAS,^3,4 and CANVAS-R,^5,6 and included a prespecified integrated analysis to meet Food and Drug Administration (FDA) postmarketing requirement (PMR) for CV safety, and evaluated potential for CV protection in type 2 diabetes mellitus (T2DM) patients with a prior history of cardiovascular disease (CVD) or ≥2 CV risk factors.^1,2 The integrated analysis also evaluated effects of INVOKANA on renal and safety outcomes.¹
• In the CANVAS Program, INVOKANA met the primary outcome, significantly reducing rates of the composite of major adverse cardiovascular events (MACE): CV mortality, nonfatal myocardial infarction (MI), or nonfatal stroke (26.9 vs 31.5/1000 patient-years [PY]; hazard ratio [HR]: 0.86; 95% confidence interval [CI]: 0.75-0.97; P<0.0001 for noninferiority; P=0.0158 for superiority) compared with placebo (PBO). All 3 components of MACE had point estimates suggesting benefit with INVOKANA.¹
• The adverse events (AEs) reported in the CANVAS Program are generally consistent with the known safety profile of INVOKANA. Serious AEs were reported less frequently with INVOKANA vs PBO (104.25 vs 120.02/1000 PY; HR 0.93; 95% CI: 0.87-1.00; P<0.05). The incidence of AEs leading to discontinuation was higher for INVOKANA vs PBO (35.49 vs 32.76/1000 PY; HR 1.13; 95% CI: 0.99-1.28; P=0.07).¹
• In the CANVAS Program, an increased incidence of amputation was reported with INVOKANA vs PBO at 6.3 vs 3.4/1000 PY; HR 1.97; 95% CI: 1.41-2.75; P<0.001.¹
• In the CANVAS Program, incidence of all fractures was elevated with INVOKANA vs PBO (15.4 vs 11.9/1000 PY; HR 1.26; 95% CI: 1.04-1.52; P=0.02). However, the difference was not statistically significant for low-trauma fractures (11.6 vs 9.2/1000 PY; HR 1.23; 95% CI: 0.99-1.52; P=0.06). There was heterogeneity (P≤0.005) in bone fracture between trials, with elevated risks observed in CANVAS, but not CANVAS-R.¹
• Several analyses of different outcomes and subpopulations in the CANVAS Program have also been published.^7-16

REGULATORY REQUIREMENTS TO ESTABLISH CV SAFETY

FDA provides guidance to show that all new antihyperglycemic agents (AHAs) are not associated with unacceptable increase in CV risk. Clinical development for a new T2DM therapy should include an independent, blinded CV endpoints committee to prospectively adjudicate events (eg, CV mortality, MI, and stroke, etc.) during all phase 2 and 3 trials. A meta-analysis should be performed upon completion, prior to marketing to establish a CV safety profile supporting approval (upper bound of 2-sided 95% CI for estimated risk ratio [RR] <1.8). A postmarketing trial should definitively show the upper bound of the 2-sided 95% CI for estimated RR <1.3.^2,17 This can be achieved through a single trial or combining results from a premarketing safety trial with a similarly designed postmarketing trial.^2,17

CANVAS Program included prespecified integrated analysis of CANVAS^3,4 and CANVAS-R^5,6 (similar in design, patient population, procedures, and assessments). The integrated analysis, designed to meet the FDA PMR, evaluated CV safety and potential for CV protection of INVOKANA in T2DM patients with a prior history of CVD or ≥2 CV risk factors.^1,2

The CANVAS Program

In order to meet requirements of the FDA Guidance¹⁷ and Approval Letter,¹⁸ CV safety of INVOKANA was evaluated in 2 CV outcome studies, CANVAS^3,4 and CANVAS-R^5,6.^1,2 Efficacy and safety data from CANVAS^3,4 was combined with CANVAS-R^5,6 in a prespecified integrated analysis of CV safety outcomes.^1,2 This analysis met the FDA PMR to determine CV safety, and evaluated potential for CV protection of INVOKANA in T2DM patients who had either a prior history of CVD or ≥2 CV risk factors. The integrated analysis of CANVAS and CANVAS-R also evaluated the effects of INVOKANA on renal and safety outcomes.^1,2

All MACE, hospitalization for heart failure (HHF), renal outcomes, deaths, and selected safety outcomes (ie, diabetic ketoacidosis, pancreatitis, and bone fractures) were adjudicated by an independent Endpoint Adjudication Committee shared across trials.^1,3,5 Likewise, CANVAS and CANVAS-R shared an Independent Data Monitoring Committee, whose role was to implement prespecified interim monitoring that would see the trial stopped prematurely if there was early evidence of significant benefit or harm.^2,3,5

• The CANVAS study was a randomized, double-blind, PBO-controlled, parallel-group, phase 3, multicenter, event-driven study evaluating CV risk for MACE, including CV death, nonfatal MI, and nonfatal stroke with INVOKANA in patients with prior history of CVD or ≥2 CV risk factors. Effects of INVOKANA vs PBO on risk of CVD were assessed in patients on background standard of care. Safety and tolerability were also evaluated.^2-4
  • Patients were randomized 1:1:1 to INVOKANA 100 mg, INVOKANA 300 mg, or PBO.
• The CANVAS-R study was a randomized, double-blind, PBO-controlled, parallel group, multicentered, phase 4 study that assessed effect on progression of albuminuria in T2DM patients with prior history of CVD or ≥2 CV risk factors. CANVAS-R was conducted in response to the FDA requirement of a “CANVAS-like” study, data from which could be combined with data from CANVAS to satisfy PMR for incidence of MACE.^2,5,6
  • Participants were randomized 1:1 to initial dose of INVOKANA 100 mg or PBO. Optional up-titration to INVOKANA 300 mg (or PBO) was encouraged at week 13 (or thereafter) if additional glycemic control was required and patient was tolerating the initial dose.

Endpoints

• The primary endpoint was time to first occurrence of MACE: CV mortality, nonfatal MI, or nonfatal stroke. In the prespecified integrated analysis, CV safety is shown if the upper bound of the 2-sided 95% CI of the HR is <1.3, suggesting noninferiority for INVOKANA vs PBO, using the full integrated CANVAS program dataset and intent-to-treat (ITT) approach. Additionally, superiority is demonstrated if the null hypothesis H₀ is rejected and upper bound of the 2-sided 95% CI of the HR is <1.0.¹
• Prespecified secondary endpoints planned for sequential hypothesis testing were all-cause mortality (ACM), CV mortality, progression of albuminuria, and composite of CV mortality and HHF based on a truncated integrated dataset or CANVAS-R alone.¹
  • Progression of albuminuria was defined as a ≥30% increase from baseline in urinary albumin:creatinine ratio (UACR), accompanied by a change from either normo- to micro- or macroalbuminuria, or from micro- to macroalbuminuria.
• If sequential hypothesis testing was not significant for all outcomes, then the remaining outcomes were assessed as exploratory using the full integrated CANVAS Program dataset (all randomized patients). See Figure: The CANVAS Program: Sequential Hypothesis Testing Plan.
• Exploratory prespecified CV outcomes were nonfatal MI, nonfatal stroke, and HHF, and the composite of CV death or HHF.¹
• Key prespecified exploratory renal outcomes were based on ITT analysis, including regression of albuminuria and renal composite: 40% reduction in estimated glomerular filtration rate (eGFR) sustained for ≥2 consecutive measures, need for renal replacement therapy (RRT; dialysis or transplantation), or renal death.¹
  • Albuminuria regression was defined as development of normoalbuminuria in a patient with baseline micro- or macroalbuminuria or microalbuminuria in a patient with baseline macroalbuminuria, with a decrease in UACR of ≥30% from baseline.
  • A 40% eGFR reduction had to occur from baseline and persist for ≥30 days, not thought to be due to reversible causes.
  • eGFR was calculated using the Modification of Diet in Renal Disease (MDRD) formula.³
• Analysis of these exploratory CV outcomes, renal outcomes, death, and hospitalizations was prespecified based on the full integrated dataset (all randomized patients).¹

Statistical Analysis

• The primary CV safety hypothesis was noninferiority for HR <1.3 for INVOKANA vs PBO for the primary outcome using the full integrated CANVAS Program dataset and ITT population. Superiority INVOKANA vs PBO would be demonstrated if the null hypothesis H₀ is rejected and the upper bound of the 2-sided 95% CI of the HR is <1.0.^1,2
• To minimize risk of bias resulting from interim unblinding of CANVAS data at the request of a health authority, the efficacy endpoint addressing the effect of INVOKANA on the secondary mortality endpoints in CANVAS Program data were to be carried out on a modified integrated dataset. Specifically, data from CANVAS that contributed to the dataset used for testing of the secondary mortality endpoint were left-truncated, such that study time and mortality events prior to November 20, 2012, were excluded.²
• Sequential hypothesis testing was utilized, conditional on the primary safety hypothesis and each subsequent test for superiority being met in the full integrated dataset, a truncated dataset, or the CANVAS-R dataset.¹ P-values for efficacy were reported only where hypothesis was proven. Hypothesis testing in the sequential hypothesis testing plan was not to be done beyond the first nonsignificant result. For all subsequent and exploratory outcomes, reporting consisted of HR estimates and nominal 95% CIs.¹ See Figure: The CANVAS Program: Sequential Hypothesis Testing Plan.

Inclusion Criteria

• Adults with uncontrolled T2DM (hemoglobin A1C [A1C] ≥7.0 and ≤10.5%) and at an elevated risk of CVD who were either^1,2:
  • Aged ≥30 years with symptomatic atherosclerotic CVD: stroke, MI, hospital admission for unstable angina, coronary artery bypass graft (CABG), percutaneous coronary intervention (with or without stenting), peripheral revascularization (angioplasty or surgery), symptomatic with hemodynamically significant carotid or peripheral vascular disease, or amputation due to vascular disease (secondary prevention cohort).^1,2
  • OR ≥50 years with no prior CVD, but with ≥2 CV risk factors: T2DM duration ≥10 years, systolic blood pressure (SBP) >140 mmHg on ≥1 antihypertensives, current daily smoker, micro- or macroalbuminuria, or high-density lipoprotein cholesterol (HDL-C) <1 mmol/L (<39 mg/dL) (primary prevention cohort).^1,2
• All participants were required to have a baseline eGFR of >30 mL/min/1.73 m².^1,2

Baseline Characteristics

• Baseline characteristics were similar for INVOKANA compared with PBO in CANVAS Program. INVOKANA was uptitrated to 300 mg in ~71% of CANVAS-R patients. History of atherosclerotic vascular disease was reported in ~72.2% of patients, and 65.6% reported history of CVD in CANVAS Program. Use of background therapy for glycemic and other risk factor management was according to best practice within local guidelines. In CANVAS Program, 77.2% of patients at baseline were on metformin, 50.2% on insulin, 43% on sulfonylurea, and 12.4% on dipeptidyl peptidase-4 inhibitor.¹
• For more information related to the CANVAS Program baseline characteristics, see Table: The CANVAS Program: Baseline Characteristics.

Results

CV Safety and Efficacy Outcomes

• In CANVAS Program, INVOKANA significantly reduced rates of the composite of MACE: CV mortality, nonfatal MI, or nonfatal stroke (26.9 vs 31.5/1000 PY; HR 0.86; 95% CI: 0.75-0.97; P<0.0001 for noninferiority; P=0.0158 for superiority), compared with PBO.¹
• All 3 components of MACE, including CV death, nonfatal MI, and nonfatal stroke, reported point estimates of effect, suggesting benefit with INVOKANA.¹
  • CV death event rates were 11.6 and 12.8/1000 PY and nominal effect estimates were HR 0.87; 95% CI: 0.72-1.06 in INVOKANA and PBO groups, respectively.¹
  • Nonfatal MI event rates were 9.7 and 11.6/1000 PY and nominal effect estimates were HR 0.85; 95% CI: 0.69-1.05 in INVOKANA and PBO groups, respectively.¹
  • Nonfatal stroke event rates were 7.1 and 8.4/1000 PY and nominal effect estimates were HR 0.90; 95% CI: 0.71-1.15 in INVOKANA and PBO groups, respectively.¹ The direction of stroke effect observed in CANVAS Program differs from previous reports^19-22 of a possible adverse effect on stroke risk.¹
• Overall, consistent effects for MACE across a broad range of prespecified subgroups were reported (P<0.001 for homogeneity between studies; see Figure 4 in published paper).¹
• In CANVAS Program, INVOKANA was associated with protection against HHF at 5.5 and 8.7/1000 PY with INVOKANA and PBO, respectively, with nominal effect estimates reported as HR 0.67; 95% CI: 0.52-0.87.¹
• In CANVAS Program, event rates for the composite of CV death or HHF were 16.3 and 20.8/1000 PY in INVOKANA and PBO groups, respectively, and nominal effect estimate was HR 0.78; 95% CI: 0.67-0.91.¹
• In the full integrated dataset, nominal effect estimates for ACM were HR 0.87; 95% CI: 0.74-1.01.¹ See Table: The CANVAS Program: Effect of INVOKANA vs PBO for CV Outcomes, Death, and Hospitalization. Additionally, see Figure: The CANVAS Program: Effects of INVOKANA vs PBO on Cardiovascular, Renal, Hospitalization, and Death Events.

• In the hypothesis testing sequence, superiority was not demonstrated with INVOKANA for the first secondary outcome of reducing ACM (using the truncated dataset) at 19.05 and 20.12/1000 PY in INVOKANA and PBO, respectively; HR 0.90; 95% CI: 0.76-1.07; P=0.245. As such, there was no hypothesis testing of other outcomes; nominal effect estimates were calculated for subsequent and exploratory outcomes.¹ For additional information on results for the prespecified sequential hypothesis testing plan, see Table: The CANVAS Program Results for the Prespecified Sequential Hypothesis Testing Plan.

Renal Outcomes

• In CANVAS Program, there were nominal findings for progression of albuminuria, which occurred less frequently with INVOKANA vs PBO (89.4 vs 128.7/1000 PY; HR 0.73; 95% CI: 0.67-0.79; P=0.0184). Greater effects on progression of albuminuria were reported in CANVAS-R (HR 0.64; 95% CI: 0.57-0.73) compared to CANVAS (HR 0.80; 95% CI: 0.72-0.90); P=0.02 for homogeneity between CANVAS and CANVAS-R trials.¹

Key prespecified exploratory renal outcomes were based on ITT integrated analysis and included regression of albuminuria, and renal composite: 40% reduction in eGFR sustained for ≥2 consecutive measures, need for RRT (dialysis or transplantation), or renal death.¹

• In the CANVAS Program:
  • Regression of albuminuria was more frequent in those receiving INVOKANA vs PBO (293.4 vs 187.5/1000 PY; HR 1.70; 95% CI: 1.51-1.91).¹
  • Nominal findings for the composite of sustained ≥40% eGFR reduction, need for RRT, or renal death, occurred less frequently with INVOKANA vs PBO (5.5 vs 9.0/1000 PY; HR 0.60; 95% CI: 0.47-0.77), with no difference between trials.¹
• For more information, see Table: CANVAS Program and CANVAS-R: Effects of INVOKANA vs PBO on Renal Outcomes. Additionally, see Figure: The CANVAS Program: Effects of INVOKANA vs PBO on Cardiovascular, Renal, Hospitalization, and Death Events.

The CANVAS Program: Patients with CVD and Patients at Risk for CVD at Baseline

Mahaffey et al (2017)²³ conducted a prespecified subgroup analysis of CANVAS Program, comparing INVOKANA vs PBO on CV, renal, and safety outcomes among T2DM patients who had either history of CVD (secondary prevention: N=6656; 66%) or ≥2 risk factors for CVD (primary prevention: N=3486; 34%) at baseline. At baseline, patients at risk for CVD and patients with CVD were 63 and 64 years, 45% and 31% were female, and duration of T2DM was 14 years and 13 years, respectively. Patients with CVD at baseline had increased use of common cardioprotective medications, including statins, beta-blockers, and antiplatelet agents. For more information, see Table: Select Baseline Characteristics of Patients With CVD vs Patients at Risk for CVD in CANVAS Program.

• INVOKANA reduced rates of 3 point-MACE: CV mortality, nonfatal MI, or nonfatal stroke, in patients with CVD (34.1 vs 41.3/1000 PY; HR 0.82; 95% CI: 0.72-0.95) and patients at risk for CVD (15.8 vs 15.5/1000 PY; HR 0.98; 95% CI: 0.74-1.30; P=0.18 for homogeneity) compared with PBO, respectively.²³
• INVOKANA was associated with protection against HHF in patients with CVD (7.3 vs 11.3/1000 PY; HR 0.68; 95% CI: 0.51-0.90) and at risk for CVD (2.6 vs 4.2/1000 PY; HR 0.64; 95% CI: 0.35-1.15; P=0.91 for homogeneity) compared with PBO.²³
• Additionally, INVOKANA reduced the renal composite of sustained ≥40% eGFR reduction, need for RRT, or renal death in patients with CVD (6.4 vs 10.5/1000 PY; HR 0.59; 95% CI: 0.44-0.79) and in patients at risk for CVD (4.1 vs 6.6/1000 PY; HR 0.63; 95% CI: 0.39-1.02; P=0.73 for homogeneity) compared with PBO, respectively.²³

Matthews et al (2017)²⁴ also evaluated CANVAS Program CV outcomes in patients with CVD and at risk for CVD. In CANVAS Program, 8% of patients at risk for CVD and 18% of patients with CVD had New York Heart Association (NYHA) Class I-III heart failure (HF) at baseline.²⁴ For the effect of INVOKANA on the primary outcome, as well as other endpoints, including CV death and ACM, see Figure: The CANVAS Program: Patients With CV Disease and Patients at Risk for CV Disease.

Effect on A1C, Body Weight, and SBP

• The mean difference in A1c was -0.58% (95% CI: -0.61 to -0.56), mean difference in body weight (BW) was -1.60 kg (95% CI: -1.70 to -1.51), and mean difference in SBP was -3.93 mmHg (95% CI: -4.30 to -3.56) for INVOKANA vs PBO; all P<0.001.¹ See Figures: The CANVAS Program: Effect of INVOKANA on A1C, The CANVAS Program: Effect of INVOKANA on Body Weight, and The CANVAS Program: Effect of INVOKANA on Systolic Blood Pressure.
• Use of other AHAs during follow-up was 9.3% lower (95% CI: -11.0% to -7.6%) with INVOKANA vs PBO.¹

Overall Safety

• AEs reported in CANVAS Program were generally consistent with the known safety profile of INVOKANA. Serious AEs were reported less frequently for INVOKANA than PBO (104.3 vs 120.0/1000 PY; HR 0.93; 95% CI: 0.87-1.00; P=0.04).¹ Incidence of AEs leading to discontinuation was higher for INVOKANA vs PBO (35.5 vs 32.8/1000 PY; HR 1.13; 95% CI: 0.99-1.28; P=0.07).¹
• Since the safety profile of INVOKANA was well-established during phase 3 trials, AE collection in CANVAS-R was streamlined to include: serious AEs, AEs resulting in study drug discontinuation, and all AEs for selected AEs of interest. Subsequently, AE data collection in CANVAS was streamlined in the same fashion.¹
• In CANVAS Program, increased incidence of amputation was reported with INVOKANA vs PBO at 6.3 vs 3.4/1000 PY; HR 1.97; 95% CI: 1.41-2.75; P<0.001.¹
• Across the integrated dataset, incidence of all fractures (adjudicated) was higher with INVOKANA vs PBO (15.4 vs 11.9/1000 PY; HR 1.26; 95% CI: 1.04-1.52; P=0.02). However, the difference was not statistically significant for low-trauma fractures (11.6 vs 9.2/ 1000 PY; HR 1.23; 95% CI: 0.99-1.52; P=0.06) and there was heterogeneity (P≤0.005) in findings between trials, with higher risks observed in CANVAS, but not CANVAS-R.¹
• For detailed safety information, see Table: The CANVAS Program and CANVAS: AEs.¹

Additional Analyses of the CANVAS Program

• Rasmussen et al (2024)²⁵ conducted a post hoc analysis to explore the association of type III collagen turnover with CV outcomes and impact of INVOKANA on this turnover in participants of the CANVAS Program.
• Khan et al (2023)²⁶ conducted a post hoc analysis of the CANVAS trial to examine the varying treatment effects of INVOKANA on HHF based on the baseline HF risk using the following diabetes-specific HF risk scores: WATCH DM (weight, age, hypertension, creatinine, HDL-C, diabetes control, QRS duration, MI, and CABG) and TRS-HF DM (TIMI Risk Score for HF in Diabetes).
• Borisov et al (2023)²⁷ conducted a secondary post hoc analysis of the effects of INVOKANA on liver-related outcomes in patients with advanced T2DM and a high CV risk.
• Ohkuma et al (2020)¹⁴ conducted a post hoc analysis of the effects of INVOKANA on CV, renal, safety, and BW outcomes for participants based on baseline body mass index: <25, 25-<30, and ≥30 kg/m².
• Oshima et al (2020)¹³ conducted a post hoc analysis of the effects of INVOKANA on 57%, 50%, 40%, and 30% reductions in eGFR sustained for ≥2 consecutive measurements and ≥30 days apart.
• Matthews et al (2020)¹⁵ evaluated the effects of INVOKANA on initiation of insulin and other AHAs.
• Matthews et al (2019)⁸ conducted an analysis of the effects of INVOKANA on amputation risk by baseline characteristics and proximate etiologies.
• Figtree et al (2019)⁷ explored the effects of INVOKANA on HF outcomes, specifically, those associated with preserved and reduced ejection fraction.
• Zhou et al (2019)¹² further explored the effects of INVOKANA on fatal and nonfatal stroke, along with stroke subtypes and additional vascular outcomes.
• Perkovic et al (2018)¹¹ completed an analysis of prespecified renal outcomes, including a composite of doubling of serum creatinine, end-stage kidney disease, or death from renal causes; the individual components of this composite outcome; annual reductions in eGFR; and changes in UACR.
• Radholm et al (2018)¹⁰ further explored the effects of INVOKANA on HF and CV death overall, and evaluated effects of INVOKANA on a range of efficacy and safety outcomes among participants with and without baseline history of HF.
• Neuen et al (2018)⁹ conducted a subgroup analysis to determine effects of INVOKANA on renal and safety outcomes in patients who had a baseline eGFR of ≥60 mL/min/1.73 m² or <60 mL/min/1.73 m².

LITERATURE SEARCH

A literature search of MEDLINE^®, Embase^®, BIOSIS Previews^®, and Derwent Drug File (and/or other resources, including internal/external databases) pertaining to this topic was conducted on 23 December 2024.