IGFBP7 and left ventricular mass regression: a sub‐analysis of the EMPA‐HEART CardioLink‐6 randomized clinical trial

Abstract Aims Given recent suggestions that serum levels of insulin‐like growth factor‐binding protein 7 (IGFBP7) may identify patients who derive greater cardiorenal benefits from treatment with sodium‐glucose transport 2 inhibitors (SGLT2i), this exploratory sub‐analysis of the EMPA‐HEART CardioLink‐6 randomized controlled trial evaluated the association between serum levels of IGFBP7 and empagliflozin‐mediated left ventricular mass regression. Methods and results The EMPA‐HEART CardioLink‐6 trial used gold‐standard cardiac magnetic resonance imaging to detect change in left ventricular mass indexed to body surface area (LVMi) following 6 months of treatment with empagliflozin or matching placebo in 97 patients with type 2 diabetes and coronary artery disease. Serum samples were collected at baseline and analysed for IGFBP7 using an enzyme‐linked immunosorbent assay. A multivariate linear regression model was used to assess the association between IGFBP7 and baseline LVMi. A linear model adjusting for baseline differences in LVMi was used to test the relationship between baseline IGFBP7 level, change in LVMi over 6 months, and treatment arm. Of the 97 patients enrolled, 74 had complete covariate data and were included in our analysis. No association between baseline IGFBP7 and baseline LVMi was found [baseline LVMi: 0.14 g/m2 (95% CI: −0.29 g/m2 to 0.57 g/m2) per 1 ng/mL higher baseline IGFBP7]. In addition, no difference between patients treated with empagliflozin versus matching placebo was found when evaluating the association between serum IGFBP7, 6 month change in LVMi, and treatment arm [empagliflozin 6 month change in LVMi: 0.25 g/m2 (95% CI: −0.17 g/m2 to 0.67 g/m2) per 1 ng/mL higher IGFBP7 vs. matching placebo 6 month change in LVMi: 0.07 g/m2 (95% CI: −0.21 g/m2 to 0.35 g/m2) per 1 ng/mL higher IGFBP7; Pinteraction = 0.49]. Additional sensitivity analysis assessing IGFBP7 as a categorical variable (above/below the median) showed no significant association between IGFBP7, 6 month change in LVMi, and treatment arm. Conclusions Our study provides insight into the generalizability of IGFBP7 as a surrogate marker of cardiac remodelling in patients with type 2 diabetes and coronary artery disease. Our results suggest that SGLT2i‐mediated reverse cardiac remodelling may be independent of IGFBP7 levels. Further investigations evaluating the association between IGFBP7 and SGLT2i are suggested to understand if and how IGFBP7 levels may modulate benefits received from SLGT2i.