The association between anthropometric indicators of obesity and cardiac reverse remodelling with empagliflozin in patients with type 2 diabetes and coronary artery disease

Obesity is an established risk factor for heart failure and is associated with a high risk of cardiovascular (CV) disease.1, 2 While body mass index (BMI) is a commonly utilized metric of obesity, other anthropometric measures such as waist circumference and waist-to-hip ratio have been shown to be independent predictors of CV risk.3, 4 Sodium-glucose cotransporter-2 inhibitors (SGLT2i) offer CV benefits to people living with or without diabetes, irrespective of baseline BMI.5, 6 Although a recent meta-analysis of translational studies suggested that SGLT2i-associated left ventricular (LV) remodelling may be linked to the cardioprotective benefits of SGLT2i,7 whether SGLT2 inhibition uniformly promotes LV mass regression across baseline anthropometric variables is unknown. In the randomized, placebo-controlled Effects of Empagliflozin on Cardiac Structure in Patients with Type 2 Diabetes (EMPA-HEART) CardioLink-6 trial, empagliflozin assignment for 6 months resulted in significant regression in LV mass indexed to baseline body surface area (LVMi).8 This sub-analysis of the EMPA-HEART CardioLink-6 trial investigated the relationship between anthropometric indicators of obesity at baseline and empagliflozin-mediated LVMi regression over the 6-month treatment period. The design and primary results of the EMPA-HEART CardioLink-6 trial have been published.8 Briefly, 97 adult patients with type 2 diabetes and coronary artery disease were randomized to receive either empagliflozin (10 mg/daily) or matching placebo for 6 months. Cardiac parameters were measured according to a standardized image acquisition protocol at baseline and 6 months using cardiac magnetic resonance imaging. Waist circumference was measured using a standard technique at the umbilical level with a tape measure. All measurements were performed by trained staff with the participant standing and breathing at a normal minimal respiration. Our first analysis compared LVMi change after 6 months of treatment with empagliflozin in baseline BMI groups of <30 kg/m2 and ≥30 kg/m2. These BMI cut points were endorsed by the 2013 American College of Cardiology/American Heart Association/The Obesity Society guideline for the management of overweight and obesity in adults.9 Our second analysis evaluated the relationship of LVMi change after 6 months of treatment with empagliflozin according to waist circumferences. Here, we stratified the cohort using sex-specific waist circumference cut points recommended by the 2016 Canadian Cardiovascular Society Guidelines for the management of dyslipidaemia (<88 cm and ≥88 cm for women; <102 cm and ≥102 cm for men).10 Empagliflozin-associated 6-month change in LVMi in each subgroup was assessed using ANCOVA linear modelling. We adjusted for baseline differences in LVMi, duration of diabetes, random glucose levels and the presence/absence of a previous percutaneous coronary intervention because unpublished work from our group has suggested that these variables can influence the extent of LVM regression. Regression models are summarized as adjusted differences with two-sided 95% confidence intervals (CI). Values of p < .05 were considered statistically significant. Partial eta squared values (η2) were calculated as a measure of effect size with values 0.01, 0.06 and ≥0.14 deemed to define small, medium and large effect sizes, respectively. All statistical analyses were performed using the STATA statistical software version 17 (StataCorp LP). Of the 96 individuals included in this analysis, 23 had a BMI of at least 30 kg/m2 while the remaining 73 had a BMI <30 kg/m2. Waist circumference data were incomplete or missing for 20 participants and consequently, only data from the 39 participants who had a waist circumference <88 cm for women/102 cm for men and the 37 who had a waist circumference ≥88 cm for women/102 cm for men were included in the waist circumference analysis. Table S1 summarizes the baseline characteristics of the participants according to BMI ≥30 kg/m2 and BMI <30 kg/m2. Compared with individuals with a BMI <30 kg/m2, those with a BMI ≥30 kg/m2 were more likely to have experienced a previous stroke or transient ischaemic attack. Of note, median (interquartile range) baseline LVM was higher in the BMI ≥30 kg/m2 versus BMI <30 kg/m2 group [135 g (108 g, 168 g) vs. 107 g (96 g, 135 g); p = .001]. Table S2 summarizes the baseline characteristics of the participants according to waist circumference. Women with a waist circumference ≥88 cm and men with a waist circumference ≥102 cm were more likely to have a higher BMI, a smoking history, elevated diastolic blood pressure, higher total cholesterol, elevated haemoglobin and haematocrit levels, experienced a previous stroke/transient ischaemic attack, or undergone a percutaneous coronary intervention. This group of individuals also had a greater median (interquartile range) baseline LVM [126 g (109 g, 153 g)] than the group with women who had a waist circumference <88 cm and men who had waist circumferences <102 cm [103 g (88 g, 126 g); p < .001]. Figure 1 summarizes change in LVMi after 6 months of treatment with empagliflozin or placebo stratified by baseline BMI. The mean change in LVMi after 6 months of treatment with empagliflozin or placebo in those with BMI <30 kg/m2 was −3.42 (95% CI −6.23, −0.62) and −0.18 (95% CI −2.19, 1.83), respectively. The mean change in LVMi after 6 months of treatment with empagliflozin or placebo in those with BMI ≥30 kg/m2 was 0.77 (95% CI −5.13, 3.59) and 0.68 (95% CI −2.08, 3.44), respectively. In the group with BMI <30 kg/m2, change in LVMi over 6 months with empagliflozin just met the threshold for statistical significance compared with placebo (p = .049) whereas this was not seen for the group with BMI ≥30 kg/m2. However, subsequent analysis with an adjusted ANCOVA approach indicated a between-BMI group effect size of 0.0005 and a pinteraction value of .9 suggesting that LVMi regression in the two BMI groups were comparable. Figure 2 summarizes the change in LVMi after 6 months of treatment with empagliflozin or placebo stratified by baseline waist circumference. The mean change in LVMi after 6 months of treatment with empagliflozin or placebo in those with waist circumference <88 cm for women and <102 cm in men was −5.93 (95% CI −9.73, −2.14) and 1.45 (95% CI −0.81, 3.70), respectively. The mean change in LVMi after 6 months of treatment with empagliflozin or placebo in those with waist circumference ≥88 cm for women and ≥102 cm for men was −0.22 (95% CI −3.20, 2.87) and −0.36 (95% CI −4.30, 3.59), respectively. In the group with waist circumference <88 cm for women and <102 cm for men, change in LVMi over 6 months with empagliflozin was statistically different from that with placebo (p = .001). This was not true for the group with waist circumferences ≥88 cm for women and ≥102 cm for men. The follow-up adjusted ANCOVA analysis revealed a between-waist circumference group effect size of 0.098 and a pinteraction value of .15. The effects of empagliflozin on changes in LV end systolic volume indexed to baseline body surface area, LV end-diastolic volume indexed to baseline body surface area and LV ejection fraction are shown in Figure S1. There was no difference observed for these parameters in the two BMI strata and two waist circumference subgroups. We previously postulated that LV reverse remodelling may contribute to the CV benefits observed with SGLT2i therapy in patients with type 2 diabetes and coronary artery disease.8 As an extension, the current work suggests that empagliflozin-associated LVM regression occurs independent of anthropometric indicators of obesity. Notably, empagliflozin has been reported to reduce mortality and cardiorenal risk across BMI categories in adults with type 2 diabetes at high CV risk.6 Furthermore, dapagliflozin, another SGLT2i, improved the composite outcome of worsening heart failure or CV death across BMI groups in people with New York Heart Association class II, III or IV heart failure and an ejection fraction of ≤40%.5 Taken together, these studies support the notion that the CV benefits of SGLT2 inhibition occur independent of BMI status. Our observation that individuals with obesity had higher baseline LVM than those without obesity aligns with previous reports.11, 12 As obesity and greater LVM are independently associated with increased risk for poor CV outcomes, the collective results suggest that elevated LVM may contribute to the increased CV risk noted in patients with obesity.2, 13 The current study has several limitations that merit consideration in the interpretation of the data. First, the small overall sample size and imbalance in subgroup sizes for the BMI analysis may have contributed to the non-significant pinteraction values despite the statistically significant differences between the empagliflozin- and placebo-assigned groups for individuals in the BMI <30 kg/m2 and lower waist circumference strata. Second, this was a post hoc analysis. Third, that 93% of the cohort were men limits the generalizability of the findings to women. That said, although there are distinct differences in CV pathophysiology between the sexes, the bias enrolment of men is not unique to the EMPA-HEART CardioLink-6 trial and indeed reflective of the greater proportion of male individuals with type 2 diabetes and coronary artery disease being managed at the recruitment sites. Fourth, recommended BMI and waist circumference cut points for heightened CV risk vary across races so although the cut points applied in the current work are often used for anthropometric analyses, it will probably be that the stratification cut points used may not have been appropriate for all the individuals included in the study. Fifth, the short duration of exposure to empagliflozin may contribute to a greater variance in the data. Although the results of the current body of work can only be considered hypothesis-generating, they do support the need for future studies that are focused on determining whether the multiple benefits of SGLT2 inhibitors are present regardless of weight status particularly given the growing global prevalence of obesity. This work also underscores the importance of identifying events upstream of LV mass regression/remodelling with SGLT2 inhibition to understand better the mechanistic network(s) involved in mediating the downstream effects of SGLT2 inhibitors. CDM is supported by a Merit Award from the University of Toronto Department of Anaesthesiology and Pain Medicine and holds the Cara Phelan Chair in Critical Care at St Michael's Hospital-Unity Health Toronto. KAC holds the Keenan Chair in Research Leadership at St Michael's Hospital-Unity Health Toronto, University of Toronto. SV holds a Tier 1 Canada Research Chair in Cardiovascular Surgery. This work was supported by an unrestricted investigator-initiated study grant from Boehringer Ingelheim. KAC reports receiving research grants (to his institution), consultancy, and/or speaking honoraria from Abbott, AstraZeneca, Boehringer Ingelheim, Eli Lilly, Janssen, Merck, Novartis, Novo Nordisk, Sanofi and Servier; he is listed as an inventor on a patent application by Boehringer Ingelheim for the use of DPP-4 inhibitors in heart failure. HT reports personal fees from the Canadian Medical and Surgical Knowledge Translation Research Group. CDM reports advisory board honoraria/consulting fees from Amgen, AstraZeneca, BioAge, Boehringer Ingelheim and PhaseBio and DSMB stipends from Beth Israel Deaconess Medical Centre, Cerus and Takeda. SV reports receiving research grants and/or speaking honoraria from Amarin, Amgen, AstraZeneca, Bayer, Boehringer Ingelheim, Canadian Medical and Surgical Knowledge Translation Research Group, Eli Lilly, EOCI Pharmacomm Ltd, HLS Therapeutics, Janssen, Novartis, Novo Nordisk, Pfizer, PhaseBio, Sanofi, S&L Solutions, Sun Pharmaceuticals and the Toronto Knowledge Translation Working Group. SV is the President of the Canadian Medical and Surgical Knowledge Translation Research Group, a federally incorporated not-for-profit physician organization. All other authors have no conflicts of interest to declare. The datasets analysed during the current study are not publicly available but may be made available by the corresponding author on reasonable request. Table S1: Baseline characteristics of the EMPA-HEART CardioLink-6 participants stratified by baseline BMI. Table S2: Baseline characteristics the EMPA-HEART CardioLink-6 participants stratified by baseline waist circumference. Figure S1. Changes in LVESVi, LVEDVi and LVEF after 6 months of treatment with empagliflozin/placebo stratified by (A) baseline BMI and (B) baseline waist circumference. BMI, body mass index; LVESVi, left ventricular end-systolic volume indexed to baseline BMI; LVEDVi, left ventricular end-diastolic volume indexed to baseline BMI; LVEF, left ventricular ejection fraction. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.