Author + information
- Received March 3, 2016
- Revision received April 18, 2016
- Accepted May 3, 2016
- Published online August 2, 2016.
- Yogesh N.V. Reddy, MD,
- Vojtech Melenovsky, MD, PhD,
- Margaret M. Redfield, MD,
- Rick A. Nishimura, MD and
- Barry A. Borlaug, MD∗ ()
- ↵∗Reprint requests and correspondence:
Dr. Barry A. Borlaug, Mayo Clinic and Foundation, 200 First Street SW, Rochester, Minnesota 55905.
Background High-output heart failure (HF) is an unusual cause of cardiac failure that has not been well-characterized.
Objectives This study sought to characterize the etiologies, pathophysiology, clinical and hemodynamic characteristics, and outcomes of high-output HF in the modern era.
Methods We performed a retrospective analysis of all consecutive patients referred to the Mayo Clinic catheterization laboratory for hemodynamic assessment between 2000 and 2014. Subjects with definite HF, as defined by the Framingham criteria, were compared to controls of similar age and sex.
Results The most common etiologies of high-output HF (n = 120) were obesity (31%), liver disease (23%), arteriovenous shunts (23%), lung disease (16%), and myeloproliferative disorders (8%). Compared with controls (n = 24), subjects with high-output HF displayed eccentric left ventricular remodeling, greater natriuretic peptide activation, higher filling pressures, pulmonary hypertension, and increased cardiac output, despite similar ejection fraction. Elevated cardiac output in high-output HF patients was related to both lower arterial afterload (decreased systemic vascular resistance) and higher metabolic rate. Mortality was increased in high-output HF as compared with controls (hazard ratio: 3.4; 95% confidence interval: 1.6 to 7.6). Hemodynamics and outcomes were poorest amongst patients with the lowest systemic vascular resistance.
Conclusions High-output HF is an important cause of clinical HF in the modern era that is related to excessive vasodilation, and most frequently caused by obesity, arteriovenous shunts, and liver disease. Given the high mortality and increasing prevalence of these comorbidities in Western countries, high-output HF must be considered in the differential diagnosis of patients presenting with dyspnea, congestion, and a normal ejection fraction.
Cardiac output is usually normal or low in patients with heart failure (HF), but a minority of patients present in a high-output state, which has historically been referred to as high-output HF (1,2). The pathophysiology is believed to be related to decreased systemic vascular resistance (1–3), but this understanding is based upon reviews and limited case reports. Standard therapies for HF, such as vasodilators and inotropes, are potentially detrimental in high-output HF, and there is no proven treatment. This makes improved understanding of the pathophysiology, causes, and clinical course of high-output HF in the modern era an important unmet clinical need, forming the basis for the current study.
Consecutive patients with elevated cardiac index (≥4 l/min/m2) referred to the Mayo Clinic for right heart catheterization between January 1, 2000, and August 20, 2014, were identified. A cardiologist evaluated all patients before and after catheterization. Patients were identified based upon cardiologist diagnosis of congestive HF, and demonstrated elevated cardiac filling pressures and/or pulmonary hypertension by direct invasive assessment. High-output HF cases were adjudicated using the Framingham criteria to confirm the diagnosis (Online Table 1) (4). Patients were categorized as having left-sided HF if an elevation in pulmonary capillary wedge pressure was identified at the time of catheterization (≥15 mm Hg). Patients with clinical HF, elevated mean pulmonary artery pressure (≥25 mm Hg), but normal pulmonary wedge pressure (<15 mm Hg) were defined as having right-sided HF.
Subjects with alternative causes of high cardiac output, either physiological (pregnancy, fever, infection), congenital, or iatrogenic (pulmonary vasodilators, inotropes) were excluded, as were patients with severe anemia (hemoglobin <8 mg/dl), thyrotoxicosis, valvular heart disease (> mild stenosis, > moderate regurgitation), constrictive pericarditis, left ventricular systolic dysfunction (ejection fraction [EF] <45%), cardiomyopathies, and heart transplantation.
Control subjects free of HF enrolled in a separate prospective study were included as the comparator population (5,6). Control subjects displayed no evidence of cardiac pathology after detailed noninvasive and invasive assessment, including history, physical, imaging, and normal rest and exercise pulmonary wedge pressures and mean pulmonary artery pressures (5,6).
The etiology of high-output HF was determined from detailed chart review with patients categorized into the following 5 cohorts: 1) systemic arteriovenous shunts (including congenital, traumatic, or hemodialysis fistulas); 2) cirrhotic liver disease; 3) pulmonary disease; 4) myeloproliferative hematologic disorders; and 5) obesity (body mass index >35 kg/m2). Obesity was only considered as the etiology if no other identifiable cause of HF or a high-output state could be identified. Mortality was assessed from the medical records, the Mayo Clinic registration database, the Rochester Epidemiology Project death database, and the Social Security Death Index to obtain 100% verification of vital status.
Invasive hemodynamic assessment
Subjects were studied on their chronic medications in the fasted state and supine position after minimal sedation, as described previously (5–7). Pressures in the right atrium, pulmonary artery, and pulmonary capillary wedge position were measured at end-expiration from electronically stored continuous recordings of pressure tracings digitized at 240 Hz. Systemic and mixed venous oxygen contents were determined by blood sampling (= saturation × hemoglobin × 1.34) to determine the arterial-venous oxygen content difference. Cardiac output was determined by the direct Fick method (oxygen consumption/[arterial-venous oxygen difference]) (n = 53) or by thermodilution (n = 91). Oxygen consumption was measured directly by expired gas analysis (MedGraphics, St. Paul, Minnesota), or calculated indirectly by multiplying the thermodilution cardiac output by the directly measured arterial-venous oxygen content difference according to the Fick principle.
Stroke volume was calculated as cardiac output divided by heart rate. Systemic vascular resistance was calculated from the pressure drop across the systemic circulations (mean arterial pressure-right atrial pressure) multiplied by 80, divided by cardiac output (5–7). Plasma volume at the time of catheterization was estimated by: (1 − hematocrit) × (a + [b × weight in kg]), where a = 1,530 in men and 864 in women, and b = 41 in men and 47.9 in women (8).
Experienced sonographers and echocardiologists performed 2-dimensional and Doppler echocardiography within 4 weeks of catheterization according to American Society of Echocardiography guidelines (9).
Data are reported as mean ± standard deviation or median (25th to 75th interquartile range). Between-group differences for categorical variables were compared by chi-square or Wilcoxon rank-sum tests, as appropriate. Analysis of variance or Student t tests were used to examine differences between the groups for continuous variables. To correct for multiple comparisons between groups, the Tukey test or Steel-Dwass test were used. Survival was assessed by the Kaplan-Meier method, with adjustment for baseline differences by Cox regression analysis. Receiver operating characteristic curve analysis was performed to determine optimal cutoff points according to the maximum Youden index to noninvasively diagnose high-output failure. Linear regression was used to determine the correlation between continuous variables of interest. All tests were 2-sided, with a p value <0.05 considered significant. Analyses were performed using JMP version 10.0.0 (SAS Institute, Cary, North Carolina).
Of 16,462 consecutive patients undergoing right heart catheterization between January 1, 2000, and August 20, 2014, 525 displayed an elevated cardiac index. From this group, 120 cases of definite high-output HF were identified (Figure 1A). Compared with controls, patients with high-output HF displayed higher body mass index, more natriuretic peptide activation, greater estimated plasma volume, and a higher likelihood of being treated with diuretic agents (Table 1). Age, sex, comorbid conditions, and body surface area were similar in high-output HF cases and controls.
Most of the high-output patients presented with left-sided HF (n = 91, 76%), although a minority presented with right HF (n = 29, 24%). The clinical, echocardiographic, and hemodynamic differences between left- and right-sided high-output HF are provided in Online Table 2. Left-sided HF patients demonstrated more systemic vasodilation, with a resultant higher plasma volume and wedge pressure. Right-sided HF patients had higher pulmonary vascular resistance at baseline, but their degree of natriuretic peptide activation was similar.
Ventricular Structure-Function, Hemodynamics, and Metabolism
Compared to controls, high-output HF patients displayed hyperdynamic hearts with higher EF, and eccentric ventricular remodeling with increased ventricular chamber size and mass (Table 1). High-output HF subjects also displayed higher echocardiographic estimates of ventricular filling pressures (E/e’ ratio), greater estimated right ventricular systolic pressures, and higher estimated cardiac indexes on echocardiography.
At cardiac catheterization, patients with high-output HF displayed 2-fold higher cardiac filling pressures (right atrial and pulmonary wedge pressures), coupled with markedly higher pulmonary artery pressures (all p <0.0001) (Table 1). Elevated cardiac output was related to both a greater stroke volume and heart rate in patients compared with controls. A higher ventricular preload (increased left ventricular diastolic dimension), more complete emptying (higher EF), and lower arterial afterload (decreased systemic vascular resistance) drove the increased stroke volume.
A high cardiac output can be caused by an increased metabolic demand (reflected by higher oxygen consumption). Compared with controls, total and body mass−indexed oxygen consumption were higher in high-output HF, indicating a relatively hypermetabolic state. Oxygen carrying capacity was lower in high-output HF (lower hemoglobin), with resultant decreased oxygen content of arterial and mixed venous blood samples. However, arterial-venous oxygen content difference (reflecting tissue oxygen extraction) was lower in high-output HF compared with controls, suggesting that, on average, this increased tissue perfusion was in excess of metabolic demand in patients with high-output HF.
Cardiac output remained higher in high-output HF after adjusting for oxygen consumption (p < 0.0001), but not after adjusting for systemic vascular resistance (p = 0.10), suggesting that excessive vasodilation was the stronger pathophysiologic contributor to the high-output state. The pathophysiologic importance of excessive vasodilation was further supported by greater elevation in filling pressures with lower systemic vascular resistance, as well as with increasing cardiac output (Figure 2).
Causes of high-output HF
Morbid obesity was the most common cause of high-output HF (n = 37, 31%), followed by arteriovenous shunts and liver disease (Figure 1B). Within the shunt group (n = 27, 22%), 18 patients had systemic arteriovenous fistula, of which 17 were hemodialysis fistulas, and 8 had hereditary hemorrhagic telangiectasia. All patients with liver disease−associated high-output HF (n = 27, 23%) displayed advanced hepatic cirrhosis.
Lung disease−associated high-output HF (n = 19, 16%) included patients with chronic obstructive pulmonary disease (n = 7), connective tissue disorders (n = 5), bronchiolitis obliterans (n = 1), interstitial lung disease (n = 5), and bronchiectasis (n = 1). Hematologic disorders were the only identifiable cause in 8% of cases (n = 10). Each of these were myeloproliferative disorders characterized by hepatosplenomegaly and extramedullary hematopoiesis, including primary and secondary myelofibrosis (n = 8), and chronic myelomonocytic leukemia (n = 2).
Characteristics of high-output HF etiologies
Most hemodynamic and ventricular function findings were fairly uniform among the different etiologies of high-output HF, although there were some notable differences (Table 2). Obesity and liver-associated patients had greater body size, weight, and higher estimated plasma volume. The hematologic and liver disease−associated high-output HF groups had marginally lower hemoglobin levels than the remaining groups. As expected, the shunt group had worse renal function, given that 67% had arteriovenous fistulas for the purpose of hemodialysis access.
Absolute oxygen consumption was highest in obesity-related high-output HF, but this was explained solely by greater body mass, as the difference was eliminated after adjustment for weight (Table 2). Oxygen consumption corrected to weight was highest and mixed venous oxygen content was the lowest in the myeloproliferative disorder group, suggesting a hypermetabolic state with relatively inadequate tissue perfusion relative to demand. Low arterial-venous oxygen content difference and depressed systemic vascular resistance were common to all of the different etiologies, but tended to be lowest in the liver-associated high-output HF group. Patients with shunt-related high-output HF had the most ventricular remodeling whereas ventricular dimensions were smallest in lung disease−associated high-output HF.
Patients with high-output HF displayed increased 3-year mortality compared with controls (38% vs. 0%; hazard ratio [HR]: 3.4; 95% confidence interval [CI]: 1.6 to 7.6; p = 0.002) (Figure 3A). Among the individual causes, obesity-related high-output HF had the lowest 5-year mortality (19%), whereas liver disease (59%) and shunt-associated high-output HF (58%) had the highest 5-year mortality (p = 0.01) (Figure 3B).
Among all high-output HF etiologies, excessive vasodilation was associated with the poorest prognosis: patients with a very low systemic vascular resistance (bottom quartile, <1,030 dyne·m2/s·cm5) displayed increased mortality compared to the remainder of patients with mildly depressed or normal systemic vascular resistance (61% vs. 36%; HR: 2.5; 95% CI: 1.2 to 5.1; p = 0.01) (Figure 3C).
Noninvasive identification of high-output HF
A cardiac index, estimated by echocardiography, of 3.54 l/min/m2 or greater identified high-output HF with 62% sensitivity and 96% specificity (area under the curve: 0.85, p < 0.0001). Most patients with high-output HF also displayed an elevated Doppler-estimated right ventricular systolic pressure (≥42 mm Hg; 92% sensitivity; 100% specificity [area under the curve: 0.97; p < 0.0001]).
This study represents the first large-scale, systematic characterization of patients with invasively proven high-output HF. We show that compared to controls without HF, patients with high-output HF demonstrate a hyperdynamic state characterized by natriuretic peptide activation, plasma volume expansion, elevated cardiac filling pressures, and pulmonary hypertension. Although ventricular dimensions were higher in cases than controls, marked eccentric remodeling was not present. The most common causes of high-output HF in this contemporary series were obesity, cirrhosis, and arteriovenous shunts, although pulmonary and myeloproliferative disorders were also important, previously under-recognized causes of this syndrome (Central Illustration). Excessive vasodilation and, to a lesser extent, an increased metabolic rate, were common to all etiologies and were associated with more deranged hemodynamics. Survival was depressed in high-output HF patients as compared with controls, emphasizing the clinical importance of this disorder, and patients with more dramatic vasodilation displayed the highest mortality, supporting the pathophysiologic importance of excessive afterload reduction in this cause of HF. Given the increasing prevalence of obesity and liver and kidney disease in Western countries, high-output HF must be considered in the differential diagnosis of patients presenting with the clinical syndrome of cardiac failure with preserved EF.
Pathophysiology of high-output HF
All other things being equal, cardiac output increases with isolated reductions in vascular resistance (cardiac load), increases in oxygen consumption (cardiac demand), or both. Excessive systemic vasodilation has been considered to be a unifying mechanism of high-output HF (1–3), and this study confirms this hypothesis in a large series of patients with varying etiologies. Reduction in arterial afterload may be related to shunting from macrovascular causes, such as arteriovenous fistulas, or secondary to a systemic process producing peripheral vasodilation, such as cirrhosis, hypercapnia from advanced lung disease, or morbid obesity.
We further show for the first time that resting oxygen consumption is increased in high-output HF, contributing to the high-output state. Increases in flow due to physiological causes, such as exercise or fever, are considered to be normal, but we show that pathological conditions, such as obesity and myelofibrosis, may be associated with increased metabolic demand for tissue perfusion. This may be caused by excess quantity of tissue to perfuse (obesity) or potentially by qualitative changes in tissue metabolism related to underlying noncardiac diseases (extramedullary hematopoiesis).
Importantly, cardiac output was much higher in high-output HF patients even after adjusting for VO2 (p < 0.0001), indicating that heightened metabolic demand alone does not explain the elevated flow. Adjusting for systemic vascular resistance eliminated group differences in output. This indicates that excessive reduction in afterload indeed plays the dominant role in the pathophysiology of high-output HF. The pathophysiologic importance of excessive afterload reduction is underscored by the correlations between lower systemic vascular resistance (arterial underfilling), plasma volume expansion, and higher filling pressures (Figure 2), and by the greater mortality observed in patients with the greatest reduction in systemic vascular resistance (Figure 3C). The fluid retention and congestion in patients with high-output failure is similar to what has previously been described in elegant pathophysiologic studies of low-output HF (10) and severe anemia (11). Further study is needed to determine whether interventions to mitigate excessive vasodilation may be effective in treating the various forms of high-output failure.
Traditionally described causes of high-output HF, such as thiamine deficiency (beriberi) and Paget’s disease were not observed in this series, and only 1 case of thyrotoxicosis was identified (not included in the analysis according to study design). We also excluded patients with severe anemia, as this is a well-described and more easily reversible cause of high-output−related congestion (11). Although this may, in part, be related to the referral bias of a cardiac catheterization laboratory population, it appears that in the modern era, obesity, cirrhosis, and arteriovenous shunts may be the most common causes of this form of HF.
Obesity-related high-output HF
Two-thirds of American adults are currently overweight or obese, and the prevalence is growing (12). As such, the problem of obesity-related heart diseases will only be expected to increase in the coming years (13). Obesity was the most common cause of high-output HF in this contemporary series. Reductions in arterial resistance are well-described in obesity (13–15), caused by increased capillary perfusion to adipose tissue and elaboration in fat cells of vasoactive adipokines (16). The resultant renal hypoperfusion caused by low vascular resistance is believed to promote sodium retention and increase cardiac work (1–3). Total body oxygen consumption was increased in obesity, but this was driven largely by higher total body mass, in comparison to the other etiologies of high-output HF.
In addition to peripheral effects on vascular resistance and metabolism, obesity affects the heart directly. Left ventricular efficiency is 50% lower when the heart utilizes fatty acids as opposed to glucose as its primary substrate (17). In obese women without HF, myocardial oxygen consumption, fatty acid uptake, and stroke work vary directly with body mass, whereas ventricular efficiency varies inversely with body mass (18). In patients without HF, weight loss decreases blood volume, cardiac work, ventricular size, mass, and filling pressures, while improving diastolic and systolic ventricular function, insulin sensitivity, and glucose uptake (19). Further study is required to evaluate the effects of weight loss in obesity-related high-output HF.
Other etiologies of high-output HF
Arteriovenous shunts and cirrhosis were the next most common causes of high-output HF. Shunts are important to consider because surgical modification or banding may improve or even eliminate congestive symptoms in this cohort, although data is limited in this regard. Patients with cirrhosis develop splanchnic vasodilation, decreasing systemic vascular resistance (20). This is believed to be caused by bacterial translocation from the intestinal lumen causing cytokine release, as well as increased concentrations of vasoactive substances that are not cleared by the failing liver (21). The liver-induced high-output HF group had the lowest arterial resistance and highest mortality, although it is likely that the poor survival was related, in part, to the impact of hepatic dysfunction.
Two other major causes of high-output HF that are less well-described in published reports were identified: chronic pulmonary and myeloproliferative diseases. Hypoxia and hypercapnia are associated with a high output, reduced arterial resistance, salt and water retention, and impaired renal blood flow in patients with chronic obstructive pulmonary disease (22,23). One-half of the patients in this series with lung disease−associated high-output HF presented with isolated right HF in the absence of high left heart filling pressures. This is likely explained by the combination of a high-flow state from systemic vasodilation in the setting of vascular remodeling and hypoxic pulmonary vasoconstriction.
Patients with hematologic disorders presented with some form of myeloproliferative disease characterized by extramedullary hematopoiesis, which may contribute to both an increase in oxygen consumption and reduction in arterial resistance. One-third of myelofibrotic patients presented with isolated right HF, which may be related to the adverse effects of high circulating progenitor cells on right ventricular and pulmonary vascular function, superimposed on a high-flow state (24).
Gross indicators of systolic performance, including EF, were normal in the patients studied, indicating that significant cardiac pump failure was not present; yet all patients fulfilled rigorous clinical and hemodynamic definitions of HF, as applied in practice. This emphasizes that the clinical syndrome of HF does not necessarily require significant abnormalities in pump function, and may be caused by abnormalities external to the heart affecting vascular load and metabolism.
Patients in the high-output group displayed an elevated E/e’ ratio (often used as a noninvasive measure of diastolic dysfunction) and normal EF, suggesting that many of these patients might have been erroneously diagnosed as having HF with preserved EF (HFpEF) if there had been no direct assessment of cardiac output. This observation emphasizes the importance of considering high-output HF in the differential diagnosis. There is increasing evidence that patients with HFpEF are a heterogeneous group comprising many phenotypes (25). The current data show that high-output HF must also be considered in the differential diagnosis, and probably should be categorized separately from more garden-variety patients with HFpEF. The current data show that the presence of an increased echocardiographic Doppler-derived cardiac index >3.5 l/min/m2 should prompt clinicians to consider further evaluation to clarify the diagnosis.
This study was performed in a catheterization laboratory referral population, thus some causes of high-output HF might have been underrepresented if patients were less likely to be referred for invasive study. Data were not acquired prospectively, but from chart review. Very few patients received specific treatments directed toward high-output HF, other than diuretic agents, so we cannot make any conclusions regarding treatment from these data. Cause of death was not available. Oxygen consumption was not directly measured in all subjects, but rather solved for by using the Fick equation in patients where thermodilution outputs were obtained.
High-output HF is an important cause of cardiac failure that is associated with increased mortality. Despite the wide variety of underlying etiologies, the causes share a common pathophysiology of excessively depressed arterial resistance and heightened metabolic demand that causes congestion, despite a hyperdynamic circulation. Given the high mortality in high-output HF, and the increasing prevalence of obesity and liver and kidney disease worldwide, this disease must be considered in the differential diagnosis of patients presenting with clinical HF in the setting of a normal EF.
COMPETENCY IN MEDICAL KNOWLEDGE: The most frequent causes of high-output HF have changed over time, and today include obesity, liver disease, and arteriovenous fistulas, although pulmonary and myeloproliferative diseases are additional, underappreciated causes. Outcomes differ little from those in patients with low-output HF due to systolic dysfunction.
TRANSLATIONAL OUTLOOK: Further studies are needed to develop criteria that clinicians can use to distinguish patients with high-output HF from those with HFpEF.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- confidence interval
- ejection fraction
- heart failure
- heart failure with preserved ejection fraction
- Received March 3, 2016.
- Revision received April 18, 2016.
- Accepted May 3, 2016.
- 2016 American College of Cardiology Foundation
- Ho K.K.,
- Pinsky J.L.,
- Kannel W.B.,
- et al.
- Andersen M.J.,
- Olson T.P.,
- Melenovsky V.,
- et al.
- Andersen M.J.,
- Hwang S.J.,
- Kane G.C.,
- et al.
- Borlaug B.A.,
- Nishimura R.A.,
- Sorajja P.,
- et al.
- Ling H.Z.,
- Flint J.,
- Damgaard M.,
- et al.
- Lang R.M.,
- Badano L.P.,
- Mor-Avi V.,
- et al.
- Anand I.S.,
- Ferrari R.,
- Kalra G.S.,
- et al.
- Anand I.S.,
- Chandrashekhar Y.,
- Ferrari R.,
- et al.
- Go A.S.,
- Mozaffarian D.,
- Roger V.L.,
- et al.,
- American Heart Association Statistics Committee and Stroke Statistics Subcommittee
- Lavie C.J.,
- Alpert M.A.,
- Arena R.,
- et al.
- Peterson L.R.,
- Herrero P.,
- Schechtman K.B.,
- et al.
- Fowler N.O.,
- Westcott R.N.,
- Scott R.C.,
- et al.
- Anand I.S.,
- Chandrashekhar Y.,
- Ferrari R.,
- et al.
- Shah S.J.,
- Katz D.H.,
- Selvaraj S.,
- et al.