High percent body fat mass predicts lower risk of cardiac events in patients with heart failure: an explanation of the obesity paradox

Body mass index (BMI) is an easily measurable and quantitative anthropometric indicator of body mass and nutritional status, and it is widely used for the definition and classification of obesity. It has been established that obesity, generally defined as a BMI ≥ 25 kg/m², is an independent risk factor of incident chronic diseases including hypertension, type 2 diabetes mellitus, cancer, and heart failure (HF) [1‐3]. However, it has been shown that mortality is not proportionally increased with BMI-defined obesity. Results from epidemiological studies suggest that overweight (BMI, 25.0–29.9 kg/m²) or class I obesity (BMI, 30–34.9 kg/m²) is not associated with a worse clinical outcome or is associated even with a favorable outcome in the study population [4‐6]. That is the case in HF patients; HF patients with a higher BMI had a lower mortality rate than that in HF patients with normal or lower BMI, a phenomenon that has been termed the “obesity paradox” [2, 3, 7‐10]. In addition, a U-shaped relationship between BMI and mortality has been reported for both patients with heart failure with reduced ejection fraction (HFrEF) and patients with heart failure with preserved ejection fraction (HFpEF) [9, 10]. However, the reason for this complex relationship between BMI and mortality in HF patients has not been fully elucidated.

Cachexia is a hallmark in the advanced stage of chronic diseases including HF and is defined as involuntary loss of at least 5% of non-edematous body weight [11]. A pioneering study in this field by Anker et al. showed that the presence of cachexia is an independent predictor for mortality even after adjustment for age, exercise capacity, and severity of HF [12]. Body composition analyses by the use of dual-energy X-ray absorptiometry (DEXA) revealed that cachexic patients had reduced fat mass, reduced lean mass (muscle mass), and reduced bone mineral content [13]. Thus, a plausible explanation of the obesity paradox is that high BMI increases the risk of HF development, while HF-induced cachexia per se is associated with increased mortality. However, the natural history of each composition of the body (i.e., fat, muscle mass, bone mineral content) during progression of HF and the impact of the composition on prognosis of HF patients remain to be elucidated. In the present study, we used a DEXA scan, the best technique for analyzing body composition in research studies [14], to examine the relationships between fat and muscle masses and clinical outcome in HF patients.

This study was approved by the Clinical Investigation Ethics Committee of Sapporo Medical University Hospital (Number 302–104). This study was carried out by the opt-out method of our hospital website. Informed consent was obtained in the form of opt-out on the website.

The relationship between obesity and mortality is highly complex. The concept of the obesity paradox in a general population was derived from results of epidemiological studies showing a U- or J-shaped relationship between BMI-defined obesity and mortality [4‐6]. However, this concept has been questioned by the results of a recent meta-analysis [19]. The meta-analysis used data for study subjects who had no smoking history (never smokers) and no chronic diseases at baseline and were followed up for at least 5 years. The results showed that both overweight (BMI, 25.0–29.9 kg/m²) and obesity (BMI > 30 kg/m²) were associated with increased all-cause mortality, while no significant increase in mortality was observed in subjects with lower BMIs [19]. The reason for the apparent discrepancy between the results of the meta-analysis and the results of earlier epidemiological studies is unclear, but difference in baseline characteristics of study subjects is a reasonable explanation. In other words, concurrent chronic diseases and/or smoking habits might have been involved in the increase in mortality in subjects with a lowest range BMIs in the epidemiological studies. In fact, this notion is supported by recent findings that reduction in BMI between the first and second hospitalizations for HF was associated with increased risks of subsequent hospitalizations and cardiovascular mortality [20]. Taken together, the results of the earlier studies suggest that chronic diseases underlying BMI reduction as well as obesity have detrimental effects on clinical outcomes, resulting in a U-shaped relationship between BMI and mortality. Whether BMI reduction associated with chronic disease is just a surrogate marker of severity of the disease or whether change in the body mass composition that occurs together with BMI reduction contributes to the mortality remains an important question.

A recent study by Aimo et al. examined the association of estimated percent body fat calculated by prediction equation with prognosis in HF patients to unveil the underlying mechanism of obesity paradox [21]. Although HF patients with the lowest tertile of estimated percent body fat had worse prognosis in that study, the results should be confirmed by using sophisticated assessment of percent body fat. In the present study, we examined compositions of the body by DEXA scans and their relationships with cardiac events during a 180-day follow-up period. Study subjects in the present study were relatively old and mostly classified as non-obese subjects by BMI criteria (Fig. 1). Interestingly, patients with increased FM had a significantly lower cardiac event rate during the follow-up period than did patients without increased FM, though blood biochemistry data were consistent with obesity (i.e., low HDL-cholesterol levels, high triglyceride levels, hyperinsulinemia) (Table 2, Fig. 2a). In contrast, there was no significant difference between cardiac event rates in patients with and those without muscle wasting (Fig. 2b). These results are consistent with the results of a recent study by Thomas et al. using bioelectrical impedance analysis (BIA) for determination of body fat mass in HF outpatients (n = 359) [22]. Thomas et al. reported that survival rate was significantly better in patients with body fat mass index ≥ the median (i.e., 8.2 kg/m²) than in patients with body fat index < the median [22]. Cox regression multivariate analysis indicated that body fat mass index, but not lean body mass index, was associated with improved survival rate. Despite multiple differences including differences in methods for fat mass determination (DEXA vs. BIA), cutoff levels of fat mass index and presence or absence of gender-specific cutoff levels, the results of the present study and the study by Thomas et al. support the notion that higher percent body fat, but not muscle mass, predicts a lower cardiac event rate in HF patients.

Adipose tissue serves as a critical regulator of systemic energy control [23, 24]. Under a condition in which there is an energy supply, adipose tissue stores excess energy in the form of lipid droplets. Conversely, adipose tissue supplies energy via lipid breakdown in response to a starved condition such as anorexia in HF-induced cachexia. However, the cachexia-induced fat depletion cannot be explained solely by compensatory utilization of adipose tissue for energy production since an increase in energy supply by parenteral nutrition does not reverse the cachexic state [25]. Chronic diseases including chronic HF provoke systemic inflammation induced by innate immune signaling [26, 27], leading to inappropriate degradation of adipose tissue [11, 28]. Catecholamine excess and hormone imbalance are aggravative factors in this process [11, 29]. Importantly, innate immune signaling in HF is upregulated in an HF severity-dependent manner [30‐32], and recent studies have shown that inflammatory diseases as well as malnutrition and HF are closely associated with muscle wasting [33‐36]. Albeit there being shared signal pathways leading to fat depletion and muscle wasting, reduction in body fat is not necessarily concordant with reduction in skeletal muscle mass in the clinical course of HF [11]. In the present study subjects, the proportion of patients with muscle wasting was higher than that of patients with no increased FM (67% vs. 42%). When patients were divided into two groups by ASMI, the group with muscle wasting had slightly smaller body fat mass (median, 27.1% vs. 30.0%). On the other hand, when patients were divided into two groups by a cutoff level of increased FM, there was no significant difference in skeletal muscle mass (median, 5.8 vs. 5.5 kg/m²) between the two groups. These findings suggest that distinct mechanisms are operative for preservation of adipose tissue and preservation of skeletal muscle mass, while skeletal muscle preservation is partly dependent on the adipose tissue mass. In fact, starvation induces depletion of both skeletal muscle and body fat, and moderate to severe obesity potentially induces muscle wasting by fat-induced systemic inflammation [37, 38]. Nevertheless, a significant association between cardiac events and preserved body fat mass (Fig. 2) supports the notion that a favorable energy storage/supply balance in adipose tissue contributes to prevention of cardiac events.

An increase in fat mass induces systemic inflammation including cytokine release, leading to pathological/functional myocardial dysfunction [38, 39]. In addition to the absolute mass of body fat, fat distribution is also tightly linked to the development and progression of metabolic diseases and HF [40, 41]. In a clinical setting, waist circumference has been applied as a surrogate marker of visceral fat mass, which is one of the diagnostic criteria of metabolic syndrome [42, 43]. The contribution of an increase in abdominal fat mass to the progression of heart failure was supported by the results of a recent study showing that abdominal obesity, defined as waist circumferences of > 102 cm in men and > 88 cm in women, in patients with HFpEF was significantly associated with an increased risk of all-cause mortality [40]. In addition, obesity and diseases characterized by chronic inflammation lead to epicardial fat accumulation [41]. There is evidence indicating that epicardial fat plays a detrimental role in the pathophysiology of HF. Epicardial fat thickness is closely associated with the extent of myocardial fibrosis, leading to myocardial dysfunction [44, 45]. Furthermore, accumulation of epicardial fat induces systemic inflammation, possibly contributing to further fat accumulation. Unfortunately, we could not include data for abdominal fat and data for epicardial fat in the present analyses because of technical limitations and the retrospective nature of the study.

Circulating natriuretic peptide (NP) level is an established marker of HF prognosis [46, 47]. Venous blood level of NP is positively correlated with the extent of myocardial stretch, which is usually reflected by increased filling pressure of the ventricle. In addition to such a ventricular pressure-dependent release of NP into the circulation, several regulatory mechanisms of circulating levels of NP have been reported. Levels of NPs, especially NT-proBNP, are elevated by renal failure since NPs are mainly cleared from circulating blood by renal excretion [48]. Hyperinsulinemia, frequently seen in obese patients, is associated with lower NT-proBNP levels [49, 50]. Circulating levels of proinflammatory cytokines such as TNF-α and IL-1β, which are pronouncedly increased in a cachexic state, increase production of brain natriuretic peptide from cardiomyocytes [51]. Interestingly, NP has recently been reported to significantly stimulate lipolysis [52]. Thus, there is the possibility that marked elevation of NP level in cachexic HF patients exerts a detrimental effect on HF by acceleration of lipolysis, resulting in reduction of body fat mass, though protective effects of NPs on cardiovascular and renal functions have already been characterized. In fact, the prognostic impact of low body fat mass in the present study was lost by the inclusion of NT-proBNP into the Cox proportional hazard model (Table 3). Furthermore, a negative correlation between percent body fat and NT-proBNP levels was found (Fig. 4), suggesting an interaction of body fat mass level and level of NT-proBNP in HF patients. On the other hand, the prognostic impact of increased FM was also lost by the inclusion of NYHA functional class into the Cox proportional hazard model (Supplementary Table 1), suggesting that HF severity overcomes percent body fat in the prediction of short-term prognosis in HF patients. These findings need to be confirmed in a large population-based study of HF.

Fat mass, muscle mass, and bone mineral content are reduced as cachexic conditions progress [13]. Our recent study reported that presence of osteoporosis assessed during hospital stay is an independent predictor of adverse events after discharge and fat mass index was closely associated with the extent of osteoporosis [53]. Result of a recent study showed that muscle wasting is an independent predictor of mortality in stable ambulatory HF patients [54]. Taken together with findings of the present study, muscle wasting may be an early marker for HF progression, whereas reduction in fat mass and bone mineral content may serve as a marker of cachexia. Importantly, adipose tissue supplies energy via lipid breakdown in response to a starved condition such as anorexia in HF-induced cachexia, which might play a role in reduced cardiac events at the advanced stage of HF. A potential therapeutic approach targeting cachexia including fat depletion is a nutritional intervention since results of our very recent study showed that energy intake during hospital stay is a strong predictor of all-cause mortality even in elderly HF patients [55]. However, an increase in energy supply by parenteral nutrition did not reverse the cachexic state [25]. Therefore, further analyses are needed to demonstrate the complex relationship between fat mass and prognosis.

There are limitations in the present study. First, since this study was a retrospective observational study using a small number of patients in a single center, there might have been selection bias in study subjects. Importantly, the present study might have insufficient statistical power for detection of effects of fat/muscle mass on cardiac events among the groups with different etiologies of heart failure, e.g., HFrEF vs. HFpEF, though results of post-host analyses showed that prognostic impact of increased FM tended to be found in patients with HFrEF and with heart failure with mid-range ejection fraction (Supplementary Figure 1–2 and Supplementary Table 2). Second, sarcopenic obesity, i.e. coexistence of obesity and sarcopenia, is frequently observed in HF patients with HFpEF, which is a risk factor of hospitalization and death [56, 57]. Small number of patients with HFpEF in the present study (45/198) might be responsible for loss of prognostic effect of sarcopenic obesity as shown by the data: presence of muscle wasting had no effects on cardiac event rates also in HF patients with increased FM (Fig. 3). In addition, muscle strength, a criterion of sarcopenia, was not analyzed in the present study, which might contribute to underestimation of well-known prognostic impact of sarcopenia in this study subjects [18, 58]. Therefore, further analyses are needed to demonstrate the impact of sarcopenic obesity on cardiac event rates in our study population. Third, the patients enrolled in the present study were patients who were admitted to our institute for diagnosis and/or treatment of HF. Patients who were admitted for acute decompensation heart failure were also included. Although assessment of body composition was performed after the relief of worsening HF, the findings in the present study may not be extrapolated to ambulatory HF patients. Finally, previous studies repeatedly showed race/region-dependent variation in body composition [59, 60]. Thus, the results of the present study may not necessarily be applicable to other ethnicities.

Increased body fat mass, but not appendicular skeletal muscle mass, predicts a lower cardiac event rate after hospital discharge in non-obese HF patients.