Heart Failure Burden in Diabetes: Can Fenofibrate Provide Additional Hope?

Heart failure (HF) prevalence is increasing globally (1), and its health care burden is expected to increase further given the rising prevalence of diabetes (2)—a major and independent risk factor for HF. As a sign of this, the number of hospitalizations for HF (HHF), as well as the proportion of HHF related to diabetes, is progressively rising (3,4). For instance, among ∼1,200,000 HHF recorded in 2018 in the U.S., ∼49% occurred in patients with diabetes, as compared with ∼41% in 2008 (4). In addition to the higher risk of developing HF, patients with diabetes experience an ∼3.0-fold increase in HHF risk and an ∼1.5-fold increases in mortality from HF (5). The introduction of sodium–glucose cotransporter 2 inhibitors (SGLT2i), which consistently reduce HHF risk by 32% (95% CI 24–39), has raised hope for improvement of these outcomes (6). Yet, given the epidemiological picture described above, novel treatments to further reduce the HF burden among subjects with diabetes are direly needed.

In this context, the important study by Ferreira et al. (7) in this issue of Diabetes Care provides new information as to whether fenofibrate could be used to reduce the HF burden among patients with type 2 diabetes. Approved by the U.S. Food and Drug Administration almost 30 years ago, fenofibrate is well known for reducing triglycerides and increasing HDL cholesterol levels but, similar to other fibrates, has several other effects. As an agonist of peroxisome proliferator–activated receptor-α (PPAR-α) (a transcription factor expressed ubiquitously), it can regulate vascular inflammation, cell differentiation, and cardiac energy metabolism (8–11). Indeed, PPAR-α is a master regulator of fatty acid β-oxidation and is known to modulate nutrient-sensing pathways toward a cardioprotective, fasting-like state (11).

Despite this wide array of actions, cardiovascular trials of fibrates have mainly focused on major adverse cardiovascular events (MACE) (i.e., cardiovascular death, nonfatal myocardial infarction, and stroke). Findings of meta-analyses of these studies have shown a nominally significant 10% risk reduction for MACE, with a consistently larger benefit (∼35% risk reduction [95% CI 22–64]) among subjects with atherogenic dyslipidemia (defined by low HDL cholesterol and high triglycerides levels) (12,13) or in those carrying a common PPARA polymorphism (∼49% risk reduction [28–66]) (14). However, data have been provided for only two of these studies on the effect of fenofibrate or gemfibrozil on HHF (Action to Control Cardiovascular Risk in Diabetes lipid trial [ACCORD Lipid] and Veterans Affairs HDL Intervention Trial [VA-HIT]), and, when combined, they showed a significant 18% reduction in the risk of HHF (95% CI 4–30; P = 0.012) (12). Surprisingly, despite these findings and the known role of PPAR-α in regulating cardiac energy homeostasis, the potential benefit of fenofibrate on HF has not been further investigated until the current study by Ferreira et al. (7).

Leveraging data from ACCORD Lipid (15), this study has two important findings: 1) treatment with fenofibrate versus placebo in addition to statin significantly reduced risk of HHF and cardiovascular death, by 18% (95% C.I. 0–32; P = 0.048), and 2) this beneficial effect was influenced by the intensive glycemic control intervention tested in the Action to Control Cardiovascular Risk in Diabetes (ACCORD) clinical trial. Indeed, after accounting for the factorial design of ACCORD study, the fenofibrate benefit was significantly larger among subjects on standard glycemic control (target HbA_1c of 7.0%–7.9%), reducing the risk of HHF or cardiovascular death by 36% (25–52) and that of HHF alone by 40% (15–58), while no benefits at all were observed among those on intensive glycemic control (target HbA_1c < 6%).

The nominal statistical significance of the above interactions (P = 0.017–0.025) and the lack of replication or validation warrant great caution in the interpretation of these hypothesis-generating data. However, these findings could have a major clinical impact. Guidelines currently recommend, for individuals with clinical characteristics like those included in ACCORD Lipid, glycemic targets such as those tested in the ACCORD standard glycemic arm (16), which are indeed similar to HbA_1c levels observed in SGLT2i trials (6). In this context, applying due caution to the comparison of results between trials conducted in different populations and with different background therapies, it is remarkable that fenofibrate may have a beneficial effect on HHF that is similar in magnitude to that reported for SGLT2i (40% risk reduction [95% CI 15–58] and 32% [24–39], respectively).

While these results are quite promising, the study has several limitations, which are summarized in Fig. 1A. First, beside the lack of replication, no mechanisms were demonstrated for the lack of fenofibrate efficacy for HHF in the intensive glycemic group. The authors elegantly showed that the on-trial use of insulin or thiazolidinediones (more frequently prescribed in intensive glycemic arm and known to increase HHF risk [17,18]) did not influence fenofibrate effectiveness. This, however, indirectly suggests that the fenofibrate effectiveness could be blunted in the intensive glycemic arm by the low HbA_1c levels per se or by the occurrence of hypoglycemic events (which increased HHF risk in ACCORD [19]). For instance, low glycemic levels might modify substrate availability for cardiac energy homeostasis, influencing the same pathways that might determine fenofibrate efficacy for HF (10,11). Further analyses will there be important to see whether fenofibrate benefit for HHF is present among subjects with low HbA_1c levels, regardless of how these were obtained.

Second, no data supporting the mechanism of actions of fenofibrate benefit on HF were presented. Remarkably, the authors showed that the fenofibrate-induced changes in lipid profile and the reduction of long-term estimated glomerular filtration rate decline, which follows the fenofibrate-induced initial decrease in creatinine/estimated glomerular filtration rate, did not explain the fenofibrate effectiveness for HF outcomes. However, these analyses were performed with the entire data set, whereas they would have been more informative if restricted to the data of subjects on standard glycemic control, among whom the fenofibrate benefit was prominent, specifically with regard to HHF. Nonetheless, as previously reported also from studies on MACE and progression of diabetic retinopathy (14,20,21), these findings reinforce the concept that fenofibrate has several possible beneficial mechanisms of action beyond its influence on serum lipids. The authors suggested that fenofibrate benefit for HF could have therefore been driven by regulation of the “nutrient deprivation” cascade (e.g., SIRT-1/PGC-1α/FGF-21). However, while these hypotheses are biologically plausible, further studies are needed to validate them.

Finally, an intrinsic limitation of this study is that ACCORD was completed in 2009, when SGLT2i were not yet in use. Thus, one of the key questions is whether fenofibrate has beneficial effects for HF also among patients on SGLT2i, which are currently strongly recommended for HHF prevention (22,23). This question is especially important if one considers that SGLT2i regulate ketone body metabolism through the PPAR-α/SIRT-1/FGF-21 pathway (24–26). Therefore, if the same pathway is already “saturated” by SGLT2i, the addition of PPAR-α agonists might provide less benefit than that reported by Ferreira et al. Nonetheless, in this scenario, fenofibrate might still be useful for individuals intolerant of SGLT2i. Conversely, if the pathway is not saturated or if different mechanisms are involved, addition of fenofibrate to SGLT2i might provide additive benefit for HF outcomes.

The role of PPAR-α modulation in HHF risk will soon be explored further in the Pemafibrate to Reduce Cardiovascular OutcoMes by Reducing Triglycerides IN patiENts With diabeTes (PROMINENT) trial, with evaluation of the effect of a novel selective PPAR-α modulator—pemafibrate—among subjects with atherogenic dyslipidemia and type 2 diabetes (27). The study, after enrollment of >10,000 subjects, was recently stopped due to interim analyses showing low probability of achieving the primary outcome (i.e., a significant reduction in risk of MACE plus coronary revascularization) (28). However, HHF was a prespecified secondary outcome of the study (27), and when complete results are reported they will likely provide novel information on the effect of PPAR-α modulation on HF.

Altogether, the study by Ferreira et al. provides novel and important insights into the possible beneficial effect of fenofibrate on HF. It also adds new data to the accumulating evidence of interindividual variability in the response to fenofibrate (Fig. 1B) and reinforces the need to re-evaluate this relatively inexpensive but rarely used drug, which may provide multiple benefits to individuals with type 2 diabetes if prescribed to the right patients.