Introduction

Obesity is a strong risk factor for the development of type 2 diabetes, hypertension, coronary artery disease (CAD), and both systolic and diastolic heart failure (HF). Diet, exercise, and weight-loss pharmacotherapies have limited potential to achieve significant and sustainable weight loss. These non-operative strategies may have particularly limited efficacy for patients with symptomatic systolic HF who cannot exercise. Bariatric surgeries, such as laparoscopic sleeve gastrectomy (LSG) and Roux-en-Y gastric bypass (RYGB), have emerged as the most successful long-term strategy both in achieving weight loss and in promoting diabetes, hypertension, and hyperlipidemia remission for patients with obesity. These benefits translate to reductions in cardiovascular events and overall mortality. There are also data supporting improvements in obesity-associated abnormalities in myocardial structure and function, such as ventricular hypertrophy, diastolic dysfunction, and subclinical systolic dysfunction, with a limited literature now supporting improved clinical outcomes in HF patients after bariatric surgery. However, substantial clinical uncertainty persists regarding the risk-benefit ratio of bariatric surgery in obese systolic HF patients, especially given the observed obesity survival paradox in HF epidemiological studies. This review seeks to determine the role of bariatric surgery for patients with systolic HF and obesity and to define the research required to inform future clinical practice in this field.

Epidemiology of Obesity in Heart Failure

It is widely accepted that obesity and low physical activity have contributed extensively to the modern epidemic of cardiovascular disease, including the development of heart failure (HF) with preserved or reduced ejection fraction [1,2,3]. The overweight and obese states are commonly defined by body mass index (BMI), with 25–30 kg/m2 indicating overweightness and ≥ 30 kg/m2 indicating obesity. Within the Framingham study, the risk of clinically symptomatic HF increased by 5% for men and 7% for women per unit BMI increase, despite adjustments for known confounders [1]. In a prospective study of 4080 older males, the adjusted hazard ratios incident HF associated with a 1 standard deviation (SD) increase in BMI were 1.37 (95% confidence interval 1.09–1.72) and 1.18 (1.00–1.39) in men with and without CAD, respectively. The prevalence of obesity has grown rapidly since the 1960s, with over a third of adults in the USA (39.8% crude rate) currently classified as obese by BMI criteria [4]. The prevalence of obesity is high among those with established HF: 57.6% of 795 patients who participated in contemporary HF clinical trials were classified as obese at enrollment, with 19.7% having a BMI ≥ 40 kg/m2 [5•]. The common coexistence of obesity and HF becomes particularly clinically challenging when evaluating the most advanced HF patients for potential cardiac transplantation, because International Society of Heart And Lung Transplantation (ISHLT) guidance recommends that patients should achieve a BMI of ≤ 35 kg/m2 prior to transplantation listing (class IIa, level of evidence C), and transplant candidates with elevated BMIs often wait longer for a suitable weight-matched organ donor [6].

However, it is also recognized that BMI alone is an imperfect metric for obesity assessment, especially among patients with HF. BMI does not capture the complexities of the location of excess adiposity, with visceral adiposity being more harmful than subcutaneous, and does not reflect the degree of systemic metabolic dysfunction or the ratio of white to brown fat. Excess fluid from decompensated HF cannot be distinguished from excess adiposity in the calculation of BMI, and bioelectrical impedance scales for the determination of body composition have not performed well in HF [7]. Likewise, obesity complicates the diagnosis of HF. Dyspnea, lower extremity edema, orthopnea, and reduced exercise capacity are features of both conditions. The key diagnostic biomarkers of HF, brain natriuretic peptide (BNP) and N-terminal pro-B natriuretic peptide (NT-proBNP), are inversely associated with BMI, and the performance of echocardiography is lower in obese individuals, all further complicating diagnosis [8].

To compound the diagnostic complexities in obese HF patients, there are persistent uncertainties about the impact of obesity on clinical outcomes in HF. Even though excess adiposity has been strongly associated with incident HF, the relationship between obesity and survival in patients with established systolic HF is less clear. Multiple epidemiology studies have suggested that obese chronic HF subjects have either have an equivalent or higher survival probability compared to their normal weight peers [7, 9,10,11,12,13,14,15,16]. For example, all-cause mortality in the Digitalis Investigation Group (DIG) trial was lower in subjects with high BMIs, ranging from 45% mortality in the underweight group to 28% mortality in the obese group (p for trend < 0.001) [9]. A meta-analysis of 28,209 patients compared normal-weight HF patients to those classified as overweight or obese. The overweight and obese groups had significantly lower mortality (adjusted HR 0.88, 95% confidence interval (CI) 0.83–0.93 and adjusted HR 0.93, 95% CI 0.89–0.97, respectively) compared to the normal weight group, with a mean follow-up of 2.7 years. Conversely, underweight and low-normal weight (< 23 kg/m2) subjects had the highest hazard of mortality (adjusted HR 1.11, 95% CI 1.01–1.23) [10]. There are several potential explanations of this so-called “obesity survival paradox” as an epidemiological artifact, including lead-time bias, collider stratification bias, a healthy survivor effect, or inadequate risk adjustment between obese and non-obese cohorts. However, it is also possible that obese HF patients derive benefit from a greater metabolic reserve provided by their excess adiposity and are relatively protected from the adverse effects of cardiac cachexia development. Clinical trial participants with ≥ 5% weight loss over 6-month period had a greater than 50% excess mortality hazard as compared to subjects with weight stability [11]. It is also possible that components of the adipokine and gut hormone milieu seen in obese subjects are beneficial for the myocardium or other components of the cardiovascular system.

The Impact of Obesity

A number of pathways may explain the association between excess adiposity and HF development, including increased circulating blood volume or chronically elevated intrathoracic pressures leading to chamber dilatation, hypertension causing left ventricular hypertrophy, the presence of epicardial and myocardial fat, myocardial signals from gut hormones (such as glucagon-like peptide-1 (GLP-1)), or excess adipose tissue causing cardiotoxicity via adipokines (such as adiponectin and leptin) and inflammatory proteins [12]. The degree of metabolic dysfunction is an important contributor to the risk of HF development [13]. The pediatric literature provides some of the clearest evidence for a direct association between obesity and myocardial dysfunction because obese children have a low prevalence of intermediaries such as diabetes and CAD. Excess adiposity in childhood is strongly associated with increased left ventricular mass, which may be independent of the effect of hypertension [14,15,16]. Abnormalities of myocardial deformation and diastolic function have also been reported in children with obesity, supporting a hypothesis of a period of left ventricular hypertrophy, diastolic dysfunction, and possibly subclinical systolic dysfunction that precede the development of symptomatic HF [17,18,19]. Fortunately, these pediatric abnormalities of cardiac structure and function appear to be reversible with successful weight loss [20, 21]. Speckle tracking-derived strain and strain rate imaging has also identified subclinical systolic dysfunction associated with obesity in adults [22, 23].

There has been increasing recognition of adipose tissue as an endocrine organ with widespread homeostatic influences and the potential to affect cardiovascular health. The anatomical and functional abnormalities of adipose tissue in obese individuals result in adverse endocrine and immune responses that can directly and indirectly contribute to HF [24]. The adipokines of particular relevance to the cardiovascular system are adiponectin, resistin, leptin, chemerin, visatin, apelin, and omentin, as well as the inflammatory mediators tumor necrosis factor alpha (TNF-α), interleukin-6 (IL-6), angiotensinogen, and plasminogen activator inhibitor-1 (PAI-1) [25, 26].

The net effect on the cardiovascular system is thought to be determined by the balance of adipokines involved. For example adiponectin, which is produced both by adipocytes and myocytes, promotes insulin sensitivity, normal endothelial function, and is anti-inflammatory and anti-atherogenic [27, 28]. It can also ameliorate the negative consequences of myocardial pressure overload and ischemia-reperfusion injury. Adiponectin levels are usually reduced in obesity. In patients with cardiovascular disease, levels may be elevated with evidence of functional adiponectin resistance [29]. Adiponectin is also secreted as from epicardial adipose tissue (EAT) and has a paracrine effect on the heart—adiponectin levels are significantly lower in EAT from patients with severe CAD [30]. Conversely, leptin can mediate pro-inflammatory, atherogenic, thrombotic, and angiogenic activity and may contribute to the left ventricular hypertrophy and HF development [31,32,33,34]. Circulating leptin levels are usually high in obesity due to leptin resistance [35]. In a prospective study of 4080 older men, higher circulating leptin was significantly associated with the risk of incident HF in men without pre-existing coronary heart disease, independent of BMI and potential mediators, although no association was seen in those with pre-existing coronary heart disease [36]. Despite the predominantly negative cardiovascular effects of leptin, it may also afford some protection against ischemia-reperfusion injury [37].

The gut hormones may also contribute the relationship between obesity and myocardial dysfunction. Ghrelin is a potent simulator of appetite and has positive effects on left ventricular function [38, 39]. The incretins, including glucagon-like peptide 1 (GLP-1) and gastric inhibitory polypeptide (GIP), are a family of gut hormones that stimulate post-prandial insulin release, inhibit glucagon, slow gastric empting rate, and promote weight loss. GLP-1 analogs such as liraglutide are in widespread clinical use for type 2 diabetes management. GLP-1 receptors are found on cardiomyocytes, and initial pilot studies suggested a positive effect of GLP-1 on left ventricular function [40,41,42]. Recognizing that the myocardium becomes increasingly insulin-resistant with progression of systolic HF, the Functional Impact of GLP-1 for Heart Failure Treatment (FIGHT) study randomized subjects to liraglutide vs placebo after a systolic HF hospitalization, but the study result was neutral [43]. However, GLP-1 and the related gut hormones may be pathophysiologically important in the myocardial response to weight gain and weight loss.

Bariatric Surgery and Cardiovascular Effects

Bariatric surgery is now established as the most effective and durable strategy for managing severe obesity. Restrictive procedures such as gastric banding and LSG promote satiety and reductions in food intake, whereas malabsorptive operations such as biliopancreatic diversion decrease nutrition absorption. Roux-en-Y gastric bypass (RYBG) combines restrictive and malabsorptive features [44]. Worldwide, laparoscopic RYGB has been the most common bariatric surgery, but in the USA, LSG is currently the leading procedure [45]. Candidacy for bariatric surgery continues to evolve, but typical qualifications are outlined in Table 1 [46, 47]. Bariatric surgery should always be considered within the context of a broader strategy of dietary and exercise management, as described in the 2011 American Heart Association Scientific Statement on Bariatric Surgery and Cardiovascular Risk Factors [48].

Table 1 Standard qualifications for bariatric surgery
Full size table

Bariatric procedures typically result in an excess weight loss of around 50%, meaning that half of the initial weight that exceeded ideal weight for height has been lost [49]. A recent report of 418 patients who underwent RYGB, vs 417 who did not receive surgery and 321 who did not seek surgery, showed an adjusted mean change from baseline body weight of − 45.0 kg (95% CI − 47.2 to − 42.9) in the surgery group at 2 years, − 36.3 kg (95% CI − 39.0 to − 33.5) at 6 years, and − 35.0 kg (95% CI − 38.4 to − 31.7) at 12 years. The mean change at 12 years of the two non-surgery groups were − 2.9 and 0 kg, respectively [50•]. Profound effects on cardiovascular risk factors and metabolic dysfunction are also observed after bariatric surgery [51]. A 22,000-patient meta-analysis demonstrated that an average post-operative excess weight loss of 61% was accompanied by significant improvements in type 2 diabetes mellitus, hypertension, dyslipidemia, and obstructive sleep apnea [49].

The Surgical Treatment and Medications Potentially Eradicate Diabetes Efficiently (STAMPEDE) trial subsequently underlined the scope of bariatric surgery in the management of patients with type 2 diabetes and BMI range 27 to 43 kg/m2 [52]. Five years after randomization, the primary end point of hemoglobin A1c < 6.0% was achieved by 2 of 38 patients (5%) who received medical therapy alone, 14 of 49 patients (29%) who underwent RYGB (unadjusted p = 0.01, adjusted p = 0.03, p = 0.08 in the intention-to-treat analysis), and 11 of 47 patients (23%) who underwent LSG (unadjusted p = 0.03, adjusted p = 0.07, p = 0.17 in the intention-to-treat analysis) [53•]. In addition, bariatric surgery can prevent the onset of type II diabetes [54], and is associated with reduced cardiovascular morbidity and mortality over follow-up periods of 4.4 to 14.7 years [50•, 51, 52, 53•].

These longer-term cardiovascular and mortality benefits should be balanced against the early risks of undergoing a surgical procedure. Early studies described a 30-day post-operative mortality of 0.1% for the restrictive procedures (2297 patients undergoing banding and 749 patients with gastroplasty), 0.5% in 5644 patients undergoing RYGB and 1.1% in 3030 patients undergoing a biliopancreatic diversion or duodenal switch. Mortality is highest in male patients, those with age ≥ 45 years, BMI ≥ 50 kg/m2, patients with hypertension or thromboembolic risk factors, and those requiring an open (rather than laparoscopic) procedure. The 30-day mortality rate was 0.3% in the Longitudinal Assessment of Bariatric Surgery (LABS) cohort (USA 2005–2007, n = 4776) [55], and 0.22% among 136,903 observational study subjects in a recent meta-analysis (international 2003–2012) [56]. Most high-volume bariatric programs have a contemporary post-operative mortality rate of approximately 1% at 1 year, which is in line with a laparoscopic cholecystectomy, with limited data suggesting that similar mortality can be anticipated among appropriately selected surgical candidates with LV systolic dysfunction [57••].

Despite the low post-operative mortality, there are several post-operative complications that require careful consideration for a surgical candidate with HF and prompt management should they occur. Venous thromboembolic events represent the leading cause of death after bariatric surgery [58]. Anatomical surgical complications vary by the procedure performed. Laparoscopic adjustable banding has the lowest weight loss efficacy but also a lower peri-operative risk than RYGB or LSG; early complications can include gastroesophageal perforation, band slippage, acute stomal obstruction, band infection, and bleeding. Late complications include band erosion, esophageal dilatation, esophagitis, and failure of weight loss, with reoperation rates ranging from 10 to 50% [59].

Serious complications early post-RYGB include anastomotic leak (up to 5%), ileus, obstruction, and hemorrhage (1%), with longer-term complications including anastomotic strictures, hernias, intussusception, gallstones, and marginal ulcers (due to exposure of the unprotected jejunal mucosa to gastric acidity, 4%). LSG can be complicated by proximal leaks (3.5%), hemorrhage (3.5%), stricture (2.3%), trocar site hernia (1.2%), and recuts abdominal hematomas (1.2%) [60]. Later complications can include gastroesophageal reflux, vomiting, stenosis, leak, gastrocutaneous fistula, and weight regain. Both RYGB and LSG are associated with the dumping syndrome, characterized by diaphoresis, dizziness, palpitations, abdominal pain, and nausea, often related to ingestion of a high-carbohydrate meal. Nutritional deficiencies are more commonly seen after RYGB than LSG, with the commoner deficiencies being iron, folic acid, calcium, copper, and vitamins D, B1, and B12. Overall, laparoscopic RYGB has superior weight loss efficacy compared to LSG, but is associated with a higher rate of surgical complications due to the number of anastomoses [53•, 61].

Limited data suggest that the post-operative complication profile is only modestly increased in patients with LV systolic dysfunction in comparison to the general bariatric surgery cohort. Among 42 patients with pre-operative left ventricular ejection fraction (LVEF) < 50%, the frequencies of thromboembolism, stroke, acute kidney injury, pneumonia, acute blood loss anemia, sepsis, wound infection, endoscopy, and reoperation were equivalent to the remainder of the population without known LV systolic dysfunction (n = 2588) [57••]. However post-operative decompensated HF/volume overload occurred more frequently in the LV systolic dysfunction group (10 vs 0.2%, p < 0.001), as did myocardial infarction (2 vs 0.04%, p = 0.032). Additional studies are required to define the rates of post-operative complications in symptomatic HF patients who have been cardiovascularly optimized in preparation for bariatric surgery. However, clinical consensus favors the LSG for candidates with cardiac comorbidities due to good weight loss and diabetes regression efficacy, but lower post-operative complications as compared to RYGB. There are some additional anesthetic considerations relevant to the intraoperative management of bariatric surgery patients with HF, including anticipation of the potential for venous thromboembolic disease, the impact of the pneumoperitoneum on cardiorespiratory function, and the potential role of radial artery and/or pulmonary artery catheterization in hemodynamically challenging cases [62].

There are marked changes in the adipokine and gut hormone profile that occur in the weeks and months after RYGB or LSG. Circulating levels of CRP and IL-6 fall in parallel with the improved insulin sensitivity almost immediately post-operatively [63], although it has been argued that it is the decreased calorie intake rather than the surgery itself that induces these changes. Circulating levels of adiponectin, leptin, and GLP-1 also rapidly improve, long preceding the nadir of weight loss and seemingly independent of caloric restriction [64, 65]. Changes in cardiac structure and function are also seen in the months and years after either dietary or surgical weight loss, principally with reductions in left ventricular mass, which may be independent of blood pressure decreases [66,67,68,69,70,71]. Many of these observations are from echocardiography studies, but cardiac magnetic resonance imaging (MRI) is also a useful tool for studying the structural and functional changes after bariatric surgery and offers superior volumetric assessments. Thirty obese subjects without cardiac risk factors underwent MRIs at baseline and 1-year post-weight loss (bariatric surgery or diet) [67]. There was a 10% mean reduction in left ventricular mass and a 40% reduction in right ventricular mass. Left ventricular end-systolic volume, stroke volume, and cardiac output also fell with weight loss. Leptin has been proposed as a mediator of LVH regression [31]. Post-bariatric surgery improvements in diastolic function, as reflected by isovolumic relaxation time, tissue Doppler velocities, mitral inflow patterns, and left atrial volume, as well as subclinical systolic dysfunction, as reflected by myocardial deformation, have also been widely reported [72,73,74,75, 76•]. Thirteen obese patients with LVEFs above 40% demonstrated regression of these subclinical abnormalities of myocardial deformability in the 6 to 24 months after bariatric surgery [74].

Therefore, it is not surprising that successful weight loss has recently been shown to prevent the development of new onset HF. Three recently published studies demonstrated a continuum of HF prevention efficacy across non-surgical and surgical strategies for weight loss. Benotti and colleagues studied a single-center cohort of 1724 RYGB surgery patients who were closely matched to a contemporaneous cohort of non-surgical controls. Each group was followed for a median of 6.3 years (maximum 12 years) for cardiovascular events and HF incidence. Seventy-nine patients developed HF, with an unadjusted hazard ratio of 0.53 (95% confidence interval 0.33–0.85, p = 0.0089) and adjusted hazard ratio of 0.38 (0.22–0.64, p = 0.0003) for the development of HF in the RYGB group, as compared to the control group [77••]. Similarly, Persson and colleagues studied 47,859 Swedish adults with a primary diagnosis of obesity, of whom 46.6% underwent bariatric surgery at a mean age of 41 years [78••]. There were a total of 1033 incident HF cases over a mean follow-up period of 3.7 years. Patients who underwent bariatric surgery had a markedly reduced risk of HF development compared with non-surgical obese patients, with an adjusted hazard ratio of 0.37 (0.30–0.46). There was also a numerically lower mortality rate in surgical patients, as compared to non-surgical patients. Furthermore, Sundstrom and colleagues compared incident HF in a Swedish registry of people treated with a structured intensive lifestyle program vs the Scandinavian Obesity Surgery Registry. The 25,804 RYGB patients lost an average of 18.8 kg more weight after 1 year, and 22.6 kg more after 2 years, than the 13,701 lifestyle modification patients [79••]. The RYGB group had a significantly lower HF incidence than the lifestyle group, with a propensity-matched hazard ratio 0.54 (0.36–0.82), at a median of 4.1-year follow-up.

Weight Loss Strategies in Systolic Heart Failure

Despite the epidemiological uncertainties regarding the obesity survival paradox in HF, there are a host of potentially beneficial effects of weight loss for the obese HF patient, including improved glycemia or even remission from diabetes, improved sleep-disordered breathing, greater functional capacity, reductions in coronary and cerebrovascular events, increased freedom from atrial fibrillation, improved depression and quality of life, and access to heart transplantation listing where indicated [44, 53•,80]. The sustained loss of as little as 5–10% of body weight is thought to positively impact blood pressure, insulin resistance, lipid profile, and left ventricular hypertrophy [67, 81, 82].

Dietary modifications and exercise are routinely recommended for patients with excess adiposity, but meaningful weight loss may be challenging for patients with limited exercise capacity due to symptomatic HF. Therefore, it is important to personalize weight loss recommendations within the context of the individual patient’s abilities and limitations. There are few studies of diet and exercise interventions specific to systolic HF patients, but those that have been published suggest that weight loss efficacy may be limited. One such pilot study randomized 20 subjects with systolic HF and the metabolic syndrome to a walking program and a reduced calorie diet with two Slim Fast® meal replacements daily vs standard medical therapy [83]. Five patients in each group had lost weight at 3 months, although the change in weight was marginal at − 0.50 ± 3.64 and − 0.84 ± 3.82 kg (p = 0.85) in the intervention vs control groups, respectively, and quality of life and 6-min walk were unchanged. Another small study of a high-protein diet, vs a standard protein diet, vs a conventional diet, was more promising. Fourteen subjects with NYHA class II-III symptoms and BMI > 27 kg/m2 were enrolled for 3 months [84]. The high-protein diet was associated with greater weight loss (− 9.9 vs − 5.6 vs − 1.5 kg, p = 0.005) and body fat reduction (− 2.5 vs − 1.1 vs − 1.2%, p = 0.036) than the standard protein and conventional diets, respectively. The most comprehensive exercise intervention in systolic HF was HF-ACTION, where 49% of subjects were classified as obese [85]. The structured aerobic exercise program was associated with improved quality of life [86], with the greatest benefits seen in subjects with higher BMIs [87]. Exercise was also associated with non-significant trends towards lower all-cause mortality and hospitalization in each BMI category. However, weight loss was marginal—for example, subjects in the 35.0–39.9-kg/m2 initial BMI group had a − 0.5-kg median weight change (interquartile range − 2.8, 1.5 kg) when randomized to exercise, vs + 0.3 kg (− 1.4, 2.9 kg) with usual care.

Pharmacotherapy has a clear role in supporting weight loss in appropriately selected obese patients, although placebo-adjusted weight losses remain modest, lying in the 2–6-kg range [88]. The anorectic sympathomimetic drugs are usually inappropriate for HF patients due to the potential for hypertension and tachycardia, and caution is also advised in HF regarding lorcaserin, a 5-HT2C receptor agonist that promotes satiety. Orlistat is an inhibitor of pancreatic lipase and arguably has the most appropriate side-effect profile of the currently approved weight loss drugs for HF patients, which was demonstrated in a small pilot study of obese HF patients [89]. The GLP-1 agonist liraglutide 3 mg once daily has an indication for weight loss, although of note the FIGHT trial showed no beneficial effects of this drug when used as an adjunctive HF therapy. The newer combination therapies of phentermine/topiramate and naltrexone/bupropion are currently unstudied in symptomatic systolic HF, thus presenting limited pharmacological weight-loss options for the HF patient. The option of bariatric surgery for obese systolic HF patient should thus be evaluated within the context of this limited dietary, exercise, and pharmacology weight loss efficacy.

Until recently, the majority of the literature surrounds bariatric surgery in patients with systolic HF consistent of case reports. Many of these reports focused on recovery of LV function after surgical weight loss and featured very obese individuals who were young and had predominantly non-ischemic cardiomyopathies [90,91,92]. There are also several small case series in this field. The first reports fractional shortening pre- and post-vertical band gastroplasty and included 13 subjects with low pre-operative systolic function [93]. There were modest improvements in fractional shortening (22 ± 2 to 31 ± 2%, p < 0.01) at a mean of 4.3 months after weight loss plateaued, accompanied by reductions in left ventricular end-diastolic diameter and blood pressure. The same group published a study of fractional shortening pre- and post-vertical band gastroplasty in 14 subjects with clinical diagnoses of HF and mild systolic dysfunction [94]. This cohort showed improvements in symptom status, but no statistically significant improvements in systolic function. An overlapping cohort of HF patients that underwent bariatric surgery generated two publications [95, 96]. In Ramani et al., 12 patients with mean age 41 years, BMI 53 kg/m2, and LVEF of 22 ± 7% underwent RYGB (n = 9), sleeve gastrectomy (n = 2), or adjustable gastric banding (n = 1). Surgical subjects were matched to 10 controls who received diet and exercise management only. At 1 year, hospital readmission was lower among the surgical subjects than controls (0.4 ± 0.8 vs 2.5 ± 2.6, p = 0.04). There was a LVEF improvement for the surgical group (22 ± 7 to 35 ± 15%, p = 0.005), but not for controls, and the NYHA class improved in the surgical group but deteriorated in controls. Another case series features a subset of 9 patients with LVEF ≤ 50%, within a 57-patient cohort of obese subjects with mean BMI 49 kg/m2, who underwent RYGB. Although there was a trend towards increased LVEF in these nine patients (pre-operative LVEF 44.8 ± 7 to post-operative LVEF 59.5 ± 10.1), a similar rise was seen in non-surgical controls with initial LVEF ≤ 50% (44.9 ± 7.9 to 58.6 ± 14.1) [70].

More recently, a retrospective cohort study compared 42 patients with LVEF < 50% who underwent bariatric surgery to 2588 patients without known systolic dysfunction [57••]. Weight loss efficacy was slightly inferior at 1 year in the LV systolic dysfunction, with 22.6% post-operative weight loss as compared to 28.1% in the general cohort (p = 0.011). There was an excess of post-operative decompensated HF and MI in the LV systolic dysfunction group, but despite the higher rate of cardiac complications, there was no excess of reoperation, tracheostomy, intensive care unit admission, or hospital readmission within the 30 days post-operatively, and no between-group differences in 30-day or 1-year mortality. Blinded echocardiographic readers reported on 38 systolic dysfunction patients both pre- and post-operative echocardiographic images available. There was an LVEF improvement post-operatively on the cusp of clinical significance in the surgical group of + 5.1 ± 8.3% (pre/post-comparison, p < 0.001) for surgical subjects, but not for controls. Subsequently, a self-controlled case series of 524 obese patients with HF who underwent bariatric surgery in three US states indicated that the rate of HF hospitalizations/emergency department visits was significantly lower in the 13–24 months after bariatric surgery, compared to pre-operatively (adjusted odds ratio 0.57, p = 0.003) [97••]. However, the proportion of systolic vs diastolic HF cases was not defined and operative mortality was not reported. Together, these two studies indicate that bariatric surgery may be a safe and symptomatically beneficial procedure for carefully selected obese HF patients at experienced centers. However, the optimal HF selection criteria and the impact of malabsorptive bariatric surgery on HF medication absorption both require further investigation.

There are also several case reports of patients with end-stage systolic HF requiring left ventricular assist device (LVAD) implantation who have received a bariatric surgery to help attain the BMI threshold for cardiac transplantation eligibility. LSG is the most commonly utilized bariatric procedure in the more recent case reports [98, 99]. Some centers have performed LVAD implantation and LSG in a single combined procedure, although the patient education and lifestyle adjustments required for each component operation should not be underestimated and may be overwhelming when performed simultaneously. Anti-coagulation is a critical component of successful bariatric surgery for patients with LVAD support, and a potential decrease in warfarin dose requirement should be anticipated post-gastrectomy.

Conclusion

As many as half of HF patients in the USA are obese, with imaging and clinical outcomes data suggesting that significant weight loss can improve cardiac structure and function, and reduce HF symptom burden. Recent literature and clinical experience at high-volume bariatric surgery centers suggests that clinicians should be more aggressive in considering surgical weight loss for selected HF patients with acceptable peri-operative risk profiles. Surgical weight loss is probably most appropriate for HF patients with BMI > 40 kg/m2 and age less than 50–60 years, with severely depressed systolic function and NYHA II-III symptoms, who have failed non-surgical strategies and have a high likelihood of future cardiac transplantation candidacy after weight loss. Symptomatic patients with HFpEF and BMI > 40 kg/m2 who have failed non-invasive weight loss strategies may also be good candidates for bariatric surgery, and it is possible that successful weight loss may reduce the frequency of future HF hospitalizations among obese HF patients. LSG is currently the procedure of choice. However, optimal patient selection remains incompletely defined and prospective studies of surgical vs non-surgical weight loss are urgently needed in HF patients with both reduced and preserved ejection fractions. Rates of post-operative complications in HF patients are also poorly defined, and further personalization of the multidisciplinary approach will be required to ensure equivalent weight loss efficacy in surgical candidates with HF as the general surgical population. In summary, bariatric surgery is the most effective strategy for attaining significant and sustained weight loss for patients with obesity and the current operative safety profile offers considerable opportunities to offer this intervention to select patients with systolic HF at experienced centers.