Prediabetes, Atrial Fibrillation, and Heart Failure: An Intersection of Comorbidities

Prediabetes for each of several definitions (1) is increasingly common and has been associated with an increased risk for progression to type 2 diabetes (T2D) (2,3) and for incident atherosclerotic cardiovascular disease (ASCVD) (4–6) and atrial fibrillation (AF) and heart failure (HF) (7). Prediabetes may progress to T2D or persist as prediabetes as well as regress to normoglycemia. Progression to diabetes is generally associated with increased risk for both macrovascular and microvascular complications. Investigations of whether “diabetes prevention” reduces the risk for T2D complications based on studies of lifestyle and medications have shown inconsistent results (8). Effects of prediabetes regression to normoglycemia on clinical outcomes has not been extensively studied.

In this issue of Diabetes Care, Huang et al. (9) report associations of normoglycemia and dysglycemia (prediabetes and T2D) with HF hospitalization in a group of patients who had new-onset AF. They further report the risk for HF in the group with prediabetes who reverted to normoglycemia versus those who persisted in the prediabetes category or who progressed to diabetes over a period of 2 years. Incident risks for HF were nominally higher in each of three models with prediabetes than those for normoglycemia (point estimates of subdistribution hazard ratios [SHRs] [10] ranged from 1.07 to 1.12), and SHRs were statistically higher with T2D (SHR range 1.23–1.30) than with normoglycemia. They also divided the prediabetes cohort into three groups: regression to normoglycemia, persistent prediabetes, and progression to T2D. With the persistent prediabetes group as the referent, the SHR for reversion to normoglycemia was lower (range 0.72–0.61) and the SHR for those who progressed to T2D was higher (range 1.57–1.50). These are interesting observations, but they raise the following question: Is the reduced risk for AF-related HF a direct consequence of reversion from prediabetes to normoglycemia or a result of modifications of risk factors for each observation, such as changes in weight, blood pressure (BP), or lipid profiles?

The study by Huang et al. does not address these considerations. There is a paucity of weight, BP, and lipid (beyond “dyslipidemia”) data, and there are no data on changes in these variables. These variables are associated with insulin resistance in prediabetes and are also commonly associated with AF and HF risks. Data from selected studies that inform the relationships among dysglycemia, obesity, hypertension, AF, and HF may be helpful to put the current study into context. A brief review of published data is the focus of the remainder of this commentary.

In the Valsartan Antihypertensive Long-term Use Evaluation (VALUE) trial, Aksnes et al. (11) reported that hypertensive patients with new-onset diabetes and AF had a greater than threefold risk of HF than those with new-onset diabetes without AF. The results of this study confirm that diabetes and AF together are associated with increased HF risk. Krisai et al. (12) analyzed risks for HF from three large trials of antithrombotic therapy in which all subjects had AF. The presence of diabetes increased risk for HF by threefold. Associated risk factors for HF included weight, BP, and serum creatinine. In the multivariable analyses of these data, the most significant risk factor for HF was presence of diabetes with a hazard ratio (HR) of 1.8 (95% CI 1.56–2.09). Two observational studies deserve comment. Schnabel et al. (13) evaluated risks for HF in individuals with AF from the Framingham and Framingham Offspring cohorts. In a fully adjusted model for HF prediction, both obesity (BMI) and diabetes contributed to HF risk. Investigators of a large observational study (n = 294,057; n = 28,233 for AF; n = 25,604 for HF) from Sweden with almost two decades of follow-up reported on the relationships of AF not only with HF with diabetes but also with dysglycemia, defined as impaired fasting glucose (IFG) (7). IFG and diabetes were both associated with increased risk for AF (IFG HR 1.19, 95% CI 1.13–1.26; diabetes HR 1.30, 95% CI 1.21–1.41) and for HF (IFG HR 1.4, 95% CI 1.13–1.26; diabetes HR 2.22, 95% CI 2.08–2.36). This Swedish study included BMI data for only 19% of the cohort; however, in a limited multivariable analysis, a high BMI attenuated the association of dysglycemia with AF and remained significant only for diabetes (HR 1.25, 95% CI 1.02–1.53). These authors concluded the following: “Fasting glucose at prediabetes levels is associated with development of atrial fibrillation and heart failure. To some extent increased BMI may drive this association.”

Few studies have evaluated the impact of prediabetes reversion to normoglycemia on cardiovascular disease or HF outcomes. Mando et al. (14) analyzed major adverse cardiovascular events (MACE) in an observational cohort (n = 119,271) of patients from Michigan including patients with prediabetes that progressed to diabetes (n = 6,324), patients with prediabetes that stayed in the prediabetes range (n = 13,520), and patients with prediabetes who reverted to normoglycemia (n = 1,585). Of note, one of the components of their combined end point was HF. HF was not examined separately, but the observations are informative. There were 102,087 MACE in 43,303 (36.3%) patients, and this included 40,104 HF admissions. After adjustment for multiple risk factors including BMI, risk for MACE was highest with diabetes and prediabetes conversion to diabetes. There were more MACE events among subjects for prediabetes to prediabetes and prediabetes to normoglycemia than for normoglycemia. Subjects who went from prediabetes status to normoglycemia status had nominally fewer events compared with those who had persistent prediabetes. Finally, data from the Whitehall study (15) for cardiovascular disease and mortality were similar to data in the current study by Huang et al. (9).

Weight is an important variable in several studies above; thus, an obvious question is, what is the impact of weight loss on AF and HF outcomes? Middeldorp et al. (16) reported on the effects of weight loss using a structured diet/exercise intervention on various forms of AF (e.g., from paroxysmal to persistent) in a post hoc analysis from an earlier trial. Slightly less than 30% of individuals had diabetes, and only ∼10% had IFG, so these groups were too small to be analyzed separately. The magnitude of weight reduction was associated with slower progression of AF. Weight reduction was also associated with significant reductions in BP. The European Society for Cardiology (ESC) recently published management guidelines for patients with AF (17). This document cites several studies in which weight loss favorably affected AF. The ESC class IIa level B evidence states, “In obese patients with AF, weight loss together with management of other risk factors should be considered to reduce AF incidence, AF progression, AF recurrences, and symptoms.”

Proposed mechanisms for associations or intersection of dysglycemia, obesity, ASCVD, AF, and HF include insulin resistance, inflammation, and fat distribution including epicardial fat. These mechanisms obviously deserve study but currently do not drive clinical practice. The accumulating evidence of the intersection of glucose-lowering agents, ASCVD risk, and HF has been incorporated into guidelines for practicing physicians including rationale and recommendations for weight loss strategies (18) and sodium–glucose cotransporter 2 inhibitor use in diabetes and HF (19). These guidelines are likely applicable to prediabetes.

In conclusion, results of the study by Huang et al. and some of the studies above suggest that reversion of prediabetes may have favorable effects on AF and HF. Whether the observation of Huang et al. is related to weight loss is still uncertain. Nevertheless, other evidence strongly suggests that weight reduction may have favorable effects in AF and HF across the spectrum of dysglycemia, and pharmacotherapy proposed for diabetes (19) may also be applicable to prediabetes.