Obesity, Heart Failure, and SGLT2 Inhibition

DECLARE-TIMI 58 Provides Insights

Naveed Sattar; Stefan D. Anker; Javed Butler; Subodh Verma


Eur Heart J. 2022;43(31):2968-2970. 

Graphical Abstract: Relationship between increasing BMI and heart failure hospitalizations in patients with diabetes in the DECLARE TIMI-58 study.

Mounting evidence supports obesity as an important risk factor for incident heart failure (HF) in people with and without diabetes.[1,2] In a recent Swedish report of patients with diabetes, while the association of glycaemia with incident myocardial infarction (MI) and incident HF was broadly similar, progressively higher body mass index (BMI) levels were strongly associated with incident HF whereas associations with incident MI were relatively flat across the BMI range.[3] In other Swedish data, elevated BMI in young adult men, measured at a time when there is minimal confounding by comorbidity, was also more strongly associated with incident HF than with incident MI.[4] The causal role of obesity in the pathobiology of HF has been well described in experimental and genetic studies (such as HERMES).[1,5–7] At the same time, there is increasing evidence for large-scale weight loss potentially reducing the risk for incident HF from multiple bariatric surgery cohort studies,[8,9] in people both with and without type 2 diabetes.

With this knowledge in mind, Oyama and colleagues, in their study published in this issue of the European Heart Journal,[10] examined whether the risks for cardiovascular (CV) death/hospitalization for HF and renal outcomes, as well as the benefits, of the sodium–glucose co-transporter protein 2 inhibitor (SGLT2i), dapagliflozin, in DECLARE-TIMI 58 differed by baseline BMI levels. The DECLARE-TIMI 58 trial recruited >17 000 patients with either multiple CV risk factors or existing vascular disease. Their findings for outcomes risk across the range of baseline BMI match prior research in that the hazard for incident hospitalization for HF (HHF) increased progressively across the BMI range (Graphical Abstract). Specifically, the excess risk in those with BMI levels >40 kg/m2 compared with 'normal' BMI was around four-fold; the same pattern was true for incident fibrillation (AF) or atrial flutter (AFL). There was also a gradient increase in risk by rising BMI for the renal-specific composite outcome, albeit much lower in steepness than seen for HHF or AF/AFL outcomes, whereas CV death risks appeared somewhat lower in all subjects who were overweight or above compared with normal weight.

The key finding of this analysis was that incident HHF and possibly AF/AFL appeared to be reduced more with dapagliflozin in absolute terms in patients with obesity compared with their non-obese counterparts (P for interaction = 0.02 and 0.09 in HHF and AF/AFL, respectively). That noted, there were no formal interactions seen when the data were interrogated in a continuous manner, probably due to less power in this latter analysis. In contrast, the absolute benefits for renal outcomes seemed broadly similar across BMI thresholds.

These new DECLARE-TIMI 58 data, whilst relatively simple in concept, were generally well analysed and contextualized, although it would have been helpful to have seen the data separated by presence or absence of established CV disease. Since the obesity paradox is more evident in people with established CV disease, analyses by these subgroups might have been valuable. It is also important to point out that the outcomes of AF/AFL were not pre-specified and were identified through a search of adverse event reporting. Also, the analyses did not evaluate outcomes by other anthropometric measures (such as waist circumference or waist-to-hip ratio), and did not have the ability to decipher outcomes by certain ethnic groups (such as South Asians, where BMI cut-offs to define overweight/obesity tend to be lower).[11,12] Last, details on the type of HF [i.e. HF with reduced ejection fraction (HFrEF) or HF with preserved ejection fraction (HFpEF)] were not available. Even so, the results do help the practising clinician in a number of ways. First, they remind us of the important and significant step-up of risk for HF with increasing body weight—in DECLARE-TIMI 58, the authors observed a hazard ratio (HR) of 1.30 per 5 kg/m2 increase in BMI. Second, these data underscore the important relationship between diabetes, obesity, and AF/AFL, even in a population of patients where the majority did not have established CV disease. While SGLT2is have been shown to exert greater absolute risk reduction in people with type 2 diabetes and atherosclerotic cardiovascular disease (ASCVD),[13] these data extend these observations across the spectrum of baseline ischaemic risk. Although the observation that SGLT2is can reduce AF/AFL is intriguing, prospective studies where AF/AFL are pre-specified will be required to be definitive about these conclusions. Mechanistically, it is possible that SGLT2is, through direct effects on myocardial ion channels or indirect effects on cardiac remodelling, reduce AF/AFL.[14,15]

Whilst healthcare professionals are good at preventing incident atherosclerotic disease, by not targeting weight gain or embracing newer evidence-based lifestyle interventions, we are missing an important opportunity to reduce risks of HF. Of note, obesity did appear to be a more common feature in patients with diabetes who were prescribed dapagliflozin and largely free from prior CV disease in a recent nationwide Scottish study; the average BMI in dapagliflozin users was 34.3 kg/m2 vs. 32.0 kg/m2 in never users.[16] It is likely that doctors prescribed more SGLT2is in heavier patients due to their weight loss properties rather than any considerations about future HHF risks. Though SGLT2is lessen weight and BMI, they do so modestly, with average weight losses of ~1.5–3 kg in most analyses, with slightly greater absolute losses at higher baseline BMIs as show in DECLARE-TIMI 58.[10] It is therefore unlikely that weight loss associated with SGLT2is accounts for their benefits in preventing HHF or AF/AFL. Indeed, a plethora of mechanisms have been proposed to mediate the HF benefits of SGLT2is.[15,17,18]

Next, turning to patients with prevalent HF, recent relevant analyses from DAPA-HF with dapagliflozin have just been reported.[19] These data showed no variation in the effect of dapagliflozin on the primary outcome and other outcomes by baseline BMI, e.g. the HR ratio for the primary outcome was in under/normal-weight 0.74, 95% confidence interval (CI) 0.58–0.94, overweight 0.81, 95% CI 0.65–1.02, obesity class I 0.68, 95% CI 0.50–0.92, and obesity class II/III 0.71, 95% CI 0.51–1.00 (P-value for interaction = 0.79).[20] Again, there is no reason to withhold SGLT2is on the basis of baseline BMI in patients with HFrEF. Overall, dapagliflozin-induced weight loss in HF patients was ~0.9 kg, implying that weight change could not account for the benefits observed in HFrEF. These data appear to be consistent in studies of empagliflozin and sotagliflozin in patients with HF.[21,22]

Finally, this brings us to the intriguing question on many people's minds. Namely, what could larger scale (~10 kg or more) intentional weight loss in patients with prevalent HF achieve? Large, randomized trials of low-calorie diets would be helpful to answer such questions, and some are currently underway in people with AF, such as LOSE-AF.[23] At the same time, two recent trials in patients with HFpEF with glucagon-like peptide-1 receptor agonists have been announced.[24,25] The results of these trials and others will help us better determine the value of large-scale weight loss to both prevent and treat HF and related conditions such as AF. Targeting obesity in HFpEF is of particular interest, given the large body of evidence suggesting a potential causal role for adiposity in the genesis of HFpEF.[26] However, for the time being, it is reassuring to know that SGLT2is work to reduce incident HF risks regardless of baseline BMI, but may do so better in those with higher levels of obesity. The ability of SGLT2is to prevent and treat HF and chronic kidney disease in people with and without diabetes truly represents one of the most formidable pharmacological advances of modern medicine.[27–29]