Eligibility and Response to Sodium-glucose Cotransporter-2 (SGLT2) Inhibitors Therapy: How Useful are Cardiac Biomarkers?

Despite considerable progress in the management of cardiovascular diseases (CVD), these remain the leading cause of mortality and morbidity in persons with diabetes (1). Sodium glucose co-transporter 2 inhibitors (SGLT2i) were initially developed as anti-diabetic drugs but have emerged as powerful cardioprotective medications. Clinical trials of SGLT2is (2, 3), including the CANVAS trial (4), have demonstrated that this class of medications significantly lowers major adverse cardiovascular events (MACE), including heart failure, among individuals with diabetes, although the absolute benefits appear largest among those with established cardiovascular disease (2, 3). This has raised the question of whether circulating cardiac biomarkers can provide information about which patients will derive the greatest benefit from SGLT2is. Evidence suggests SGLT2is have more pronounced absolute benefits for MACE among persons with elevated biomarkers levels such as high sensitivity troponin T (hs-TnT) or NT-proB-type natriuretic peptide (NT-proBNP) (5, 6). This confirms the promise held by biomarkers as essential tools in clinical patient management for screening, diagnosis, and treatment. However, few studies have specifically explored the response of these biomarkers to treatment with SGLT2is.

In this issue of JACC, the CANVAS biomarker sub-study by Vaduganathan et al. (7) illustrates how a combination of biomarkers can inform CVD risk stratification in the context of SGLT2i therapy. The study included 4,330 participants with type 2 diabetes (mean age 63 years, 82% White adults, 67% with a history of CVD, 52% treated with insulin, 2% on glucagon-like peptide agonists [GLP-1 RAs]) among whom three biomarkers (high-sensitivity cardiac troponin T [hs-cTnT], soluble ST2 [sST2], and insulin-like growth factor–binding protein 7 [IGFBP7]) were assessed at baseline and after six years. Vaduganathan et al. made a number of key observations (7). First, an important fraction of the CANVAS participants with diabetes had elevated levels of circulating levels of the various markers (39% had hs-cTnT ≥14 pg/mL, 6% had sST2 >35 ng/mL, and 49% had IGFBP7 >96.5 ng/mL), suggesting a high presence of subclinical CVD. Second, over a 6-year period, canagliflozin significantly slowed increases of hs-cTnT and sST2, but not IGBP7. Third, each of these three biomarkers was independently associated with cardiovascular and kidney outcomes. The effects of canagliflozin on the heart failure and kidney risks were independent of baseline biomarkers concentrations, but patients with elevated hs-cTnT (≥14 ng/L) and sST2 (>35 ng/mL) derived greater relative for MACE. Fourth, a combination of all the three biomarkers in a score predicted incident cardiac and kidney outcomes above and beyond the traditional risk factors. The participants with the highest number of elevated circulating biomarkers experienced the highest absolute and relative reductions in MACE from canagliflozin treatment.

The analyses conducted by Vaduganathan et al. were robust in several ways (7). They evaluated the effects of SGLT2i treatment on 6-year change in the cardiac biomarkers—few trials have assessed these measures at more than one timepoint. In their evaluation of the associations of cardiac biomarkers with MACE and kidney outcomes, they accounted for relevant confounders including ongoing cardioprotective therapies other than canagliflozin and therapies such as insulin use that could affect IGFBP7 levels. The authors also used a composite score of several biomarkers (hs-cTnT, sST2, and IGBP7) measured at the same time point. Using a combination of biomarkers can help overcome the modest performance of individual biomarkers for risk prediction. When judged by measures of risk prediction and discrimination, the authors showed that a multi-marker approach modestly outperformed the use of any single biomarker to provide information on CVD risk above and beyond traditional risks factors. The findings from this study demonstrate the potential of using biomarkers for clinical risk stratification among high risk individuals-- the vast majority of CANVAS participants had a history of CVD. There has been a paucity of multi-marker studies among individuals with a high burden of conventional risk factors and/or existing CVD (8, 9).

The study by Vaduganathan et al has some shortcomings that should be considered in the interpretation of these results (7). First, NT-proBNP, a biomarker well-known to be associated with cardiovascular outcomes and shown by the authors be affected by the treatment with canagliflozin (5), was not included in the biomarker panel. The extent to which NT-proBNP may have added information to the biomarker panel is unknown. Second, a significant proportion of the CANVAS participants had a history of CVD (>60%). In current clinical practice, these high-risk individuals are those who would benefit the most from preventive and therapeutic measures, thus are typically managed aggressively. Hence, additional biomarker-guided risk stratification may only marginally influence management decisions among such individuals. Indeed, current guidelines do not recommend routine biomarker assessment to guide management decisions in patients already known to be at high risk. As per the contemporary guidelines (10), patients with prior CVD would be directly eligible for therapy using cardioprotective agents including SGLT2i and GLP-1RAs (2, 3, 11).Third, a limited proportion of the CANVAS biomarkers study participants were from racial/ethnic groups other than Whites (i.e., only 9% of the study population was Asian American and 2.5% were Black adults), which limits the generalizability of the findings. CVD risk and some biomarker levels vary by race/ethnicity.

From a mechanistic perspective, that canagliflozin slows the long-term rise of blood biomarkers of myocardial injury, inflammation, and fibrosis supports the concept that CVD is a complex phenotype involving multiple biological pathways. These results also support that SGLT2is may affect more than one pathway, including among others cardiac stress, inflammation, atherosclerosis, vascular structure and function, and metabolism. From a clinical perspective, these results support the need to scale up contemporary practice to improve the uptake of SGTL2i among eligible patients, as well as highlight the potential heterogeneity of CVD risk among individuals with type 2 diabetes.

The burden of CVD disease among individuals with diabetes and the demonstrated efficacy of SGLT2i creates an onus to design more effective primary and secondary CVD prevention strategies. The elegant study by Vaduganathan et al. highlights the prognostic utility of biomarkers for CVD risk estimation, especially when used in combination, among individuals with diabetes changes in the setting of SGLT2i therapy. Advances in molecular biology and high-throughput assay technologies have allowed the development of genetic, transcriptomic, metabolomic and proteomic biomarkers; hence, ‘omics may help extend the findings of this study (12). Indeed, ‘omics approaches can help refine the identification of biomarkers in different pathways from those represented by existing biomarkers (12). Additional studies in multiethnic cohorts, preferably of individuals without known CVD, will extend these findings and help identify populations most likely to benefit from SGLT2i therapy. Ultimately, this study adds to the growing body of evidence that circulating biomarkers hold great promise to improve CVD prevention among patients with diabetes without a clinical history of CVD.