Cardiac index and hepatopulmonary syndrome in liver transplantation candidates: The Pulmonary Vascular Complications of Liver Disease Study

Abstract

Background and Aims:

Hepatopulmonary syndrome (HPS) and a hyperdynamic circulation are common complications of advanced liver disease, but the relationship between HPS and cardiac index (CI) is poorly understood. We sought to compare CI in patients with and without HPS and to assess the relationship between CI and symptoms, quality of life (QOL), gas exchange, and exercise capacity among liver transplantation (LT) candidates.

Approach and Results:

We performed a cross-sectional analysis within the Pulmonary Vascular Complications of Liver Disease (PVCLD2) study, a multicenter prospective cohort study of patients being evaluated for LT. We excluded patients with obstructive or restrictive lung disease, intracardiac shunting and portopulmonary hypertension. We included 214 patients (81 with HPS and 133 controls without HPS). Compared to controls, patients with HPS had a higher CI (least square mean 3.2 L/min/m², 95% CI 3.1–3.4 versus 2.8 L/min/m², 95% CI 2.7–3.0, p<0.001) after adjustment for age, sex, Model for End stage Liver Disease-Sodium (MELD-Na) score and beta blocker use and a lower systemic vascular resistance. Among all LT candidates, CI was correlated with oxygenation (Alveolar-arterial oxygen gradient r=0.27, p<0.001), intrapulmonary vasodilatation severity (p<0.001) and biomarkers of angiogenesis. Higher CI was independently associated with dyspnea and worse functional class and physical QOL after adjusting for age, sex, MELD-Na, beta blocker use and HPS status.

Conclusions:

HPS was associated with higher CI among LT candidates. Independent of HPS, higher CI was associated with increased dyspnea and worse functional class, QOL, and arterial oxygenation.

Hepatopulmonary syndrome (HPS) is characterized by the clinical triad of liver disease, intrapulmonary vascular dilatation (IPVD), and abnormal arterial oxygenation(1). HPS affects up to 30% of patients evaluated for liver transplantation (LT) and is associated with worse quality of life (QOL) and an increased risk of death regardless of the degree of hypoxemia (1, 2). Although HPS is characterized by IPVD and a biomarker profile of angiogenesis, the systemic hemodynamic manifestations of HPS are poorly understood(3). A hyperdynamic circulation is another vasodilatory complication of cirrhosis(4). Although there is no established threshold to define a hyperdynamic state, it is generally characterized by an elevated cardiac index (CI) and reduced systemic vascular resistance. Nitric oxide (NO) is hypothesized to play a role in the pathogenesis of both HPS and the hyperdynamic state of cirrhosis as well as other complications of liver disease, such as ascites and spontaneous bacterial peritonitis(5–7). Both HPS and the hyperdynamic circulation also resolve following LT. Despite these similarities, few studies have investigated the association between HPS and the hyperdynamic state of cirrhosis, and little is known regarding the relationship between these two vasodilatory complications of liver disease.

The hyperdynamic state of cirrhosis may potentially worsen gas exchange mediated by diffusion limitation or ventilation-perfusion (V-Q) mismatch in HPS, but the relationship between oxygenation and cardiac function among LT candidates is poorly understood. Dyspnea, impaired QOL and reduced exercise capacity are common in patients with cirrhosis and are often attributed to pulmonary complications, such as HPS or portopulmonary hypertension (8–10). The independent effect of CI on symptoms, QOL and exercise capacity, however, is not known. In this study, we sought to characterize the systemic hemodynamic profile of HPS and to identify whether CI was associated with symptoms, QOL, gas exchange, and exercise capacity.

Experimental Procedures

Study Sample:

The Pulmonary Vascular Complications of Liver Disease 2 (PVCLD2) Study was a multicenter, prospective cohort study of adult patients with portal hypertension undergoing evaluation for LT and has been previously described(11). The inclusion criteria for cohort assembly were the presence of portal hypertension with or without intrinsic liver disease and undergoing an initial evaluation for LT. Patients with active infection, recent gastrointestinal bleeding (<2 weeks from date of evaluation), or a history of prior liver or lung transplantation were excluded. The study sample for this analysis was drawn from 425 patients undergoing initial LT evaluation enrolled at the University of Pennsylvania, Mayo Clinic, and University of Texas at Houston between 2013 and 2017. HPS was defined by current criteria as the presence of IPVD and an Alveolar-arterial (A-a) gradient ≥15 or >20mmHg if age >64 in the setting of liver disease and in the absence of significant obstructive or restrictive lung disease(1). For this analysis, we excluded patients with missing CI (assessed by transthoracic echocardiography, see below), arterial blood gas or spirometry measurements and patients with significant obstructive or restrictive ventilatory defects [forced expiratory volume in 1 second (FEV1)/forced vital capacity (FVC)<0.7 with FEV1<80% predicted or FVC<70% predicted, respectively], intracardiac shunting, and portopulmonary hypertension.

Study procedures:

Patients being seen in LT evaluation clinic were screened for eligibility at each study site. Informed consent was obtained from eligible patients, who were then scheduled for research assessment, which included a history and physical examination (including assessment of dyspnea), completion of the Medical Outcome Study Short Form-36 (SF-36), anthropometrics, pulse oximetry, blood pressure measurement, contrast enhanced (CE) transthoracic echocardiogram, arterial blood gas (ABG) sampling, spirometry and six-minute walk testing(12, 13). For the purposes of this analysis, mean arterial pressure (MAP) was calculated as [diastolic blood pressure + 2/3 (pulse pressure)] and systemic vascular resistance was calculated as 80*[(MAP-right atrial pressure)/cardiac output].

CE was performed by the injection of agitated saline via a peripheral vein during echocardiographic imaging. The apical four-chamber view was the preferred window for image acquisition, although other views were utilized if the four-chamber view was suboptimal or unavailable. At least 10 continuous cardiac cycles were captured, beginning immediately prior to contrast injection, to allow accurate assessment of cardiac cycles to determine delay from injection of agitated saline until visualization of contrast entering the left heart. Identification of microbubbles in either the left atrium or left ventricle after ≥ 3 cardiac cycles was considered to indicate the presence of IPVD. IPVD severity was categorized as Grade I (mild) if few microbubbles were visualized in the left heart without appreciable change in density of the left ventricular cavity, Grade II (moderate) if microbubbles were visualized in left heart with less than 50% of comparable density in the right heart, or Grade III (severe) if microbubbles were visualized in the left heart with ≥50% of comparable density in the right heart. Patients with immediate (< 3 cycles) opacification of the left atrium or ventricle were presumed to have an intra-cardiac shunt. Doppler flow signal across the atrial septum was presumed to indicate an intra-cardiac shunt. Stroke volume (SV) and cardiac output (CO) were measured using the doppler velocity time integral method as recommended by the American Society of Echocardiography guidelines(14). According to this method, SV is estimated as the left ventricular outflow tract cross-sectional area multiplied by the velocity-time integral. CO is then calculated according to the following equation: CO=SV X heart rate. CO is divided by the body surface area in order to obtain the CI. Body surface area was calculated according to the Mosteller method as √[height (cm) x weight (kg)/3600](15). Similar to other recent studies of the hyperdynamic state in cirrhosis, we chose to focus on CI rather than CO in order to account for differences in body size(16). The Echocardiography Core Laboratory at the Mayo Clinic evaluated all CEs performed at individual study sites and readers interpreted the studies offline and were blinded to clinical information.

Clinical data were collected from the medical record and formal patient interviews. Clinical laboratory results obtained closest to the date of the study visit were recorded. Model for End Stage Liver Disease Sodium (MELD-Na) scores were calculated using the following equations: MELD Score = 10 * ((0.957 * ln(Creatinine)) + (0.378 * ln(Bilirubin)) + (1.12 * ln(INR))) + 6.43 and MELD-Na Score = MELD score – Serum Na – (0.025*MELD Score * (140 - Serum Na)) + 140.

Pulse oximetry was performed using a standard professional grade oximeter (Datascope Accutorr Plus Vital Signs Monitor, Nonin Pulse Oximeter, Welch Allyn MasimSET, or Protocol Systems Inc. Quik Signs) after the study participant maintained an upright posture for five minutes and then was repositioned supine for 5 minutes. ABG sampling was performed on ambient air in a seated position after 10 minutes rest. The samples were processed in a blood gas analyzer after a one-point calibration. The A-a gradient was calculated using the following formula: AaPO₂=[(FiO2*[P_atm – P_H20])-(PaCO₂ /R)]] –PaO₂ where R is assumed to be 0.8 and P_atm was the barometric pressure measured in the city on the date of the study visit.(17) Spirometry was performed according to American Thoracic Society (ATS)-European Respiratory Society (ERS) recommendations as previously described(18).

Statistical analysis:

Categorical variables were summarized by frequencies and proportions, and continuous variables were summarized by means and standard deviations (SD) for normally distributed data and median (interquartile range) for non-parametric data. Categorical variables were compared using a χ² test or Fisher’s exact test. Continuous variables were compared using a Student’s t-test or Wilcoxon rank-sum test as appropriate. Relationships between continuous variables were assessed using Pearson’s correlation. One-way analysis of variance was used to assess the relationship between IPVD severity and CI. Multivariate logistic and linear regression models were used to assess the relationship between CI and HPS status, symptoms, QOL and exercise capacity after adjusting for covariates. A p-value<0.05 was considered significant. All data analysis was performed in SAS, Version 9.4. The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki as reflected in a priori approval by institutional review boards at all study sites (University of Pennsylvania, Office of Regulatory Affairs Protocol #816361, Mayo Clinic Institutional Review Boards Protocol #12-007715 University of Texas Committee for the Protection of Human Subjects Protocol #HSC-MS-12-0481). All authors had access to the study data and reviewed and approved all statistical analyses as well as the resulting final manuscript.

Results

The cohort included 454 patients; 81 patients with HPS and 133 controls without HPS fulfilled the inclusion criteria for this analysis. We excluded 43 patients with portopulmonary hypertension, 26 with missing ABG or pulmonary function tests data, 105 with obstructive or restrictive lung disease, 49 with intracardiac shunting, and 17 with missing cardiac output measurements, leaving 214 in the study sample for this analysis (Figure 1).

Figure 1. Study flow chart.

.

HPS versus non-HPS

A detailed comparison of HPS and non-HPS patients within the PVCLD2 cohort has been previously published(3). In this analysis which additionally excluded patients who did not have CO/CI measurements, there were no differences between patients with and without HPS in regard to age or gender, but patients with HPS had higher MELD-Na scores (15.9 ± 4.9 versus 13.7 ± 5.1, p=0.003) and were more likely to have ascites and encephalopathy and less likely to have a history of hepatocellular carcinoma (Table 1).

Table 1.

Patient characteristics

Characteristic	n	Controls	n	HPS	P-value
Age, years	133	57.4 ± 9.0	81	55.0 ± 9.5	0.07
Male Gender	133	95 (71.4)	81	51 (63.0)	0.20
Vital Signs
Weight	133	88.6 ± 20.1	81	92.5 ± 24.6	0.23
Height	133	171.1 ± 9.2	81	170.9 ± 9.5	0.88
Systolic blood pressure, mm Hg	133	124.0 ± 17.8	81	121.1 ± 16.2	0.24
Diastolic blood pressure, mm Hg	133	70.5 ± 11.6	81	66.1 ± 9.3	0.003
Pulse, beats per minute	133	70.8 ± 13.6	81	73.8 ± 14.5	0.13
Mean arterial pressure, mm Hg	133	106.3 ± 14.7	81	103.0 ± 12.8	0.09
Pulse pressure, mm Hg	133	53.4 ± 13.3	81	55.0 ± 13.6	0.40
Systemic vascular resistance, dynes/sec/cm5	133	1505.6 ± 495.3	81	1198.9 ± 335.3	<0.001
Model for End Stage Liver Disease-Sodium Score	131	13.7 ± 5.1	79	15.9 ± 4.9	0.003
Complications of Liver Disease
Ascites	133	82 (61.7)	81	61 (75.3)	0.04
Esophageal varices	133	85 (63.9)	81	62 (76.5)	0.05
Encephalopathy	133	63 (47.4)	81	55 (67.9)	0.003
Hepatocellular Carcinoma	133	56 (42.1)	81	21 (25.9)	0.02
Beta-Blocker Use	132	70 (53.0)	81	42 (51.9)	0.87
Spontaneous bacterial peritonitis prophylaxis/antibiotics	132	51 (38.6)	81	41 (50.6)	0.09

Data expressed as mean ± standard deviation or n (percent).

CO and CI were strongly correlated (r=0.87, p<0.001) and we chose to focus on CI to account for differences in body size. Compared to controls, patients with HPS had a significantly higher CO, CI and SV with similar heart rates (Table 1). Left ventricular ejection fraction and tricuspid annular plane systolic excursion were also higher in HPS patients. Beta blocker use was similar in patients with (n=47, 50.5%) and without HPS (n=79, 52.3%) (p=0.79). Beta blocker use was associated with a lower heart rate (64.5 ± 12.2 beats/min versus 71.2 ± 14.1 beats/min, p<0.001), CO (5.9 ± 1.6 L/min versus 6.5 ± 2.1 L/min, p=0.02) and CI (2.9 ± 0.8 L/min/m² versus 3.1 ± 1.0 L/min/m², p=0.049) but no differences in dyspnea, oxygenation, PCS scores, or 6-minute walk distance (p>0.05 for all). In multivariate analysis, patients with HPS had a significantly higher CI (least square means 3.2 L/min/m², 95% CI 3.1–3.4 versus 2.8 L/min/m², 95% CI 2.7–3.0, p<0.001) (Figure 2) and CO (least square means 6.7 L/min, 95% CI 6.3–7.0 versus 5.8 L/min, 95% CI 5.5–6.1, p<0.001) after adjustment for age, sex, beta blocker use and MELD-Na score.

Figure 2. Patients with HPS have a significantly higher cardiac index.

Patients with HPS had significantly higher least square means for cardiac index (3.2 L/min/m², 95% CI 3.1–3.4 versus 2.8 L/min/m², 95% CI 2.7–3.0, p<0.001) after adjustment for age, sex, beta-blocker use and MELD-Na score.

Additionally, CI was significantly associated with IPVD presence and severity among all LT candidates regardless of HPS diagnosis. Patients with IPVD (n=124) had a significantly higher CI compared to patients without IPVD (n=86) (3.2 ± 0.9 versus 2.7 ± 0.7 L/min/m², p<0.001). Additionally, higher CI was associated with IPVD severity (No IPVD: 2.8 L/min/m², 95% CI 2.6–2.9, Mild IPVD: 3.1 L/min/m², 95% CI 3.0–3.2, Moderate to severe IPVD: 3.3 L/min/m², 95% CI 3.1–3.5, p<0.001) (Figure 3). Among patients with HPS, CI was higher in patients with increased disease severity (Mild HPS: 3.3, 95% CI 3.0–3.5 L/min/m², Moderate HPS: 3.4, 95% CI 3.0–3.7 L/min/m², Severe and very severe: 3.6, 95% CI 2.8–4.4 L/min/m²) but the relationship between CI and HPS severity was not statistically significant (p=0.67).

Figure 3. Cardiac index and intrapulmonary vasodilatation.

Least square means and 95% confidence intervals for cardiac index in patients with no intrapulmonary vasodilatation (IPVD), mild IPVD and moderate to severe IPVD. Cardiac index was significantly associated with intrapulmonary vasodilatation (IPVD) severity; patients with more severe IPVD had higher cardiac index (p<0.001).

Consistent with a hyperdynamic state, diastolic blood pressure as well as systemic vascular resistance (SVR) were also significantly lower in patients with HPS versus controls (Table 1). In multivariate analysis, SVR was lower in patients with HPS (least square mean 1252.7, 95% CI 1160.1–1345.2 dynes/sec/cm⁵) versus non-HPS (least square mean 1474.1, 95% CI 1402.5–1545.7 dynes/sec/cm⁵) (p<0.001) after adjustment for age, gender, MELD-Na and beta blocker use.

Association of cardiac index with clinical characteristics

Higher CI was independently associated with an increased prevalence of dyspnea, worse functional class and worse physical QOL. CI was independently associated with an increased risk of dyspnea after adjustment for age, sex, MELD-Na scores, beta blocker use and HPS status (Odds ratio (OR) 1.55, 95% CI 1.04–2.32, p=0.03). Higher CI was also associated with worse functional class, even after adjustment for age, sex, MELD-Na, beta blocker use and HPS status (Figure 4). CI was correlated with worse physical QOL, as measured by SF-36 PCS scores (r=−0.27, p<0.001) (Figure 5) but not mental component summary scores (r=−0.07, p=0.31). CI was independently associated with worse physical component summary (PCS) scores after adjustment for age, sex, MELD-Na scores, beta blocker use and HPS status (p=0.02). There was no significant effect modification between CI and HPS on PCS scores (p-value for interaction term=0.95).

Figure 4. Cardiac index and World Health Organization (WHO) functional class.

Least square means for cardiac index by WHO functional class after adjustment for age, sex, MELD-Na score, beta blocker use and HPS status. Higher cardiac index was associated with worse functional class (Class 1: 2.8 L/min/m², 95% CI 2.6–3.0, Class 2: 2.9 L/min/m², 95% CI 2.8–3.1, WHO Class 3/4: 3.4 L/min/m², 95% CI 3.2–3.7, p<0.001).

Figure 5. Cardiac index and oxygenation, quality of life, exercise capacity and Angiopoietin-2 levels.

Cardiac index was significantly correlated with Alveolar-arterial oxygen gradient (A), SF-36 physical component summary (PCS) scores (B), 6-minute walk distance (C) and Angiopoietin-2 levels as well as partial pressure of oxygen in arterial blood (PaO2) (data not shown), partial pressure of carbon dioxide in arterial blood (PaCO2) (data not shown), and pH (data not shown).

Among all patients, higher CI was significantly correlated with worse oxygenation [A-a gradient (r=0.28, p<0.001), PaO2 (r=−0.16, p=0.02)], lower PaCO2 (r=−0.26, P<0.001) and higher pH (r=0.24, p<0.001) (Figure 5). CI was inversely correlated with 6-minute walk distance (r=−0.23, p<0.001) (Figure 5) but the relationship between CI and 6-minute walk distance no longer persisted in multivariate analysis after adjusting for age, sex, MELD-Na scores, beta blocker use and HPS (p=0.19).

Cardiac index was also higher in patients with ascites compared to patients without ascites (3.1 ± 0.8 versus 2.8 ± 0.8 L/min/m², p=0.003) and patients treated with antibiotics for spontaneous bacterial peritonitis prophylaxis (3.1 ± 0.9 versus 2.9 ± 0.8 L/min/m², p=0.008). There was no association between cardiac index and renal function as measured by serum creatinine (r=−0.03, p=0.54).

Lastly, CI was associated with elevations in biomarkers of angiogenesis associated with HPS. CI was strongly correlated with angiopoietin-2 levels (r=0.39, p<0.001) which persisted after adjustment for HPS status (β=4.7, 95% CI 2.9–6.6, p<0.001). CI was also correlated with vWF percent (r=0.29, p<0.001) and this relationship also persisted after adjustment for HPS diagnosis (β=53.7, 95% CI 21.2–86.1, p=0.001).

Discussion

In this study, we found that HPS is characterized by a relative hyperdynamic state with a higher CI and lower SVR compared to liver disease controls. Additionally, higher CI is associated with increased dyspnea and worse functional class, physical QOL and oxygenation independent of HPS status. The results of our study characterize HPS as a hyperdynamic condition and highlight the association between CI and clinical characteristics.

In this multicenter prospective cohort study, we found that HPS was characterized by a relative hyperdynamic state with a CI that was 0.4L/min/m² higher than non-HPS patients. These findings highlight the important association between IPVD and HPS and the hyperdynamic state of cirrhosis. A prior single-center study similarly found that elevations in CI were associated with worse oxygenation among patients with cirrhosis, but did not specifically evaluate the hemodynamics of patients with HPS or systematically exclude patients with alternative etiologies of hypoxemia, such as obstructive or restrictive lung disease(19). Our study was a larger, prospective multi-center study with blinded and systematic CE echocardiography as part of the study protocol, resulting in greater generalizability. Elevations in CI were also associated with IPVD severity and worse oxygenation among all LT candidates regardless of HPS diagnosis. We previously found that patients with IPVD alone in the absence of abnormal gas exchange had a higher CO compared to controls(20). This suggests the possibility that elevations in CO may precede the onset of hypoxemia associated with HPS and could represent an HPS precursor state. Further supporting this, we observed an independent association between cardiac index and biomarkers of angiogenesis that have been implicated in the pathogenesis and prognosis of HPS.

A hyperdynamic state is characterized by an elevated CI and a reduced SVR, two features that were significantly different in HPS patients versus controls. CI is determined by both heart rate and SV, and we found that the increased CI in HPS was mediated primarily by an increase in SV rather than heart rate. We also found that other echocardiographic parameters, such as left ventricular ejection fraction, were higher in patients with HPS. This finding is also consistent with a hyperdynamic state and increased stroke volume. The mechanistic link between HPS and CI is not known but may be related to splanchnic vasodilatation driven by circulating vasodilators, such as nitric oxide. Splanchnic vasodilatation and elevations in nitric oxide has also been implicated in the pathogenesis of ascites and spontaneous bacterial peritonitis, and we found that cardiac index was higher in patients with ascites and those treated with antibiotics for spontaneous bacterial peritonitis prophylaxis(5, 6, 21, 22). Alternatively, increased angiogenesis may play a role as suggested by elevations in angiopoietin-2. Increased angiogenesis has been implicated in HPS disease pathogenesis and could potentially contribute to increased blood vessel formation which could lead to increased blood volume and flow(23).

Higher CI was associated with increased dyspnea, worse functional class and worse physical QOL, and these relationships persisted in multivariate analysis even after adjustment for HPS. Our group has previously reported that HPS was associated with dyspnea, worse physical QOL and worse survival(2, 3). Interestingly, our current study suggests that elevations in CI are strongly and independently associated with these clinical manifestations of HPS.

Similar to prior studies, we found that CI was strongly correlated with gas exchange parameters(19, 24). Higher CI was associated with a higher A-a gradient, lower PaO2, higher pH and lower PaCO2 among LT candidates but was not significantly associated with HPS disease severity. Hypoxemia develops in HPS via several mechanisms, including ventilation-perfusion (V-Q) mismatch, diffusion limitation, and anatomic shunting(25, 26). Elevations in CI could potentially worsen hypoxemia mediated by diffusion limitation or V-Q mismatch and thus may serve as a novel therapeutic target to improve oxygenation in HPS. Respiratory alkalosis, characterized by an elevated pH and a reduced PaCO2, is common in cirrhosis and is typically attributed to hyperventilation mediated by hormones, such as estrogen and progesterone(27). Interestingly, these hormones have also been postulated to play a role in the hyperdynamic state of cirrhosis as well and have been shown to decrease following LT in parallel to improvements in CI(24). Similar to this prior small study, the results of our study suggest that systemic hemodynamic consequences of cirrhosis and gas exchange abnormalities are closely linked.

There are currently no effective medical therapies for HPS(1). Two prior studies found no significant effect of beta blockade on arterial oxygenation in patients with cirrhosis although one of these studies reported a trend of improved A-a gradient(28, 29). We found that higher CI in HPS was primarily driven by increases in SV rather than heart rate, suggesting that SV rather than heart rate would be a more appropriate therapeutic target. Additionally, we found that beta blocker use was not associated with differences in oxygenation or other clinical characteristics. A prior study reported improvement in hypoxemia and a decrease in CO after administration of intravenous methylene blue, a systemic vasoconstrictor (30). It is not known, however, whether decreases in CO mediated improvements in oxygenation since the correlation between these changes was not described. Although there are currently no approved medications that target the hyperdynamic circulation of cirrhosis, bevacizumab, an anti-angiogenic agent, has been shown to decrease CI and symptoms of high output heart failure associated with hepatic arteriovenous malformations in hereditary hemorrhagic telangiectasia(31). Sorafenib, an anti-angiogenic agent, demonstrated efficacy in animal models of HPS but not in clinical trials(32). Future prospective studies to determine whether treatment targeting the elevations in SV and CI or reduced SVR in cirrhosis is associated with clinical improvements in symptoms, QOL and oxygenation are warranted.

Our study had several limitations. First of all, CI was not directly measured with right heart catheterization, but was estimated non-invasively with echocardiography. Although echocardiogram is not the gold standard for measurement of cardiac output, it does provide a non-invasive assessment and our measurements were performed in accordance with ASE guidelines. Additionally, a large percentage of the cohort was excluded which may have resulted in selection bias. Since a diagnosis of HPS requires excluding other causes of hypoxemia, such as significant obstructive or restrictive lung disease, we opted to exclude patients with abnormal spirometry to minimize confounding. Lastly, it is unclear from our study whether the hyperdynamic state of cirrhosis represents a modifiable variable that can improve with interventions.

In conclusion, HPS is associated with an elevated CI which is independently associated with increased dyspnea and worse functional class, QOL and oxygenation. Future studies to better understand the mechanistic link between the hyperdynamic circulation and HPS and to determine whether treatment targeting the hyperdynamic state of cirrhosis can improve symptoms, QOL and gas exchange are warranted.

Table 2.

Study test results

Test result	n	Controls	n	HPS	P-value
Echocardiogram
Cardiac output, L/min	133	5.7 ± 1.6	81	6.9 ± 2.1	<0.001
Cardiac index, L/min/m2	133	2.8 ± 0.7	81	3.3 ± 1.0	<0.001
Stroke volume, ml	133	87.4 ± 24.2	81	98.8 ± 24.5	0.001
Heart rate, beats per min	133	66.4 ± 13.4	81	69.9 ± 13.4	0.06
LV Ejection fraction, %	133	62.4 ± 6.3	81	64.2 ± 4.4	0.02
TAPSE, mm	100	26.0 ± 5.2	62	28.6 ± 5.4	0.002
IPVD Severity	129		81		<0.001
None		86 (66.7)		0 (0)
Mild		39 (30.2)		46 (56.8)
Moderate to Severe		4 (3.1)		35 (43.2)
Laboratory Results
Sodium	133	138.2 ± 3.9	80	137.5 ± 4.4	0.25
Creatinine, mg/dL	133	1.09 ± 0.50	80	0.98 ± 0.44	0.11
Platelets, 1000/uL	132	109.8 ± 67.0	80	89.1 ± 39.9	0.005
Hemoglobin, g/dl	133	12.3 ± 1.8	81	12.2 ± 2.3	0.67
INR	131	1.3 ± 0.3	80	1.5 ± 0.3	0.002
Albumin, g/dL	132	3.3 ± 0.6	81	3.0 ± 0.7	0.004
Total bilirubin, mg/dL	132	2.4 ± 2.8	81	3.0 ± 2.0	0.05
Alanine aminotransferase, units/L	132	79.4 ± 77.0	81	80.2 ± 49.2	0.92
Aspartate aminotransferase, units/L	132	61.4 ± 50.8	81	52.0 ± 31.5	0.10
Arterial Blood Gas
pH	133	7.4 ± 0.0	81	7.4 ± 0.0	0.23
PaCO2, mmHg	133	35.2 ± 4.7	81	33.2 ± 4.5	0.002
PaCO2, mmHg	133	92.4 ± 13.5	81	78.8 ± 12.9	<0.001
A-a gradient, mmHg	133	14.3 ± 10.4	81	29.6 ± 14.0	<0.001
≤ 64 years	109	14.0 ± 10.9	70	28.5 ± 13.8	<0.001
>64 years	24	15.5 ± 7.4	11	36.7 ± 13.4	<0.001
Exercise Testing
6-minute walk distance, m	113	426.8 ± 91.9	70	387.5 ± 99.1	0.007
Biomarkers of Angiogenesis
Angiopoietin, ng/ml	130	14.4 ± 9.6	75	21.3 ± 14.9	<0.001

Data expressed as mean ± standard deviation or n (percent). Abbreviations: A-a: Alveolar arterial, FVC: Forced vital capacity, FEV1: Forced expiratory volume in 1 second, INR: International normalized ratio, LV: Left ventricular, PaCO2: Partial pressure of carbon dioxide in arterial blood, PaO2: partial pressure of oxygen in arterial blood, TAPSE: Tricuspid annular systolic plane excursion

Abbreviations

6MWD

6-minute walk distance

A-a

Alveolar-arterial

ABG

Arterial blood gas

ALT

Alanine aminotransferase

AST

Aspartate aminotransferase

ATS

American Thoracic Society

BMI

Body mass index

BP

Blood pressure

CI

Cardiac index

CO

Cardiac output

ERS

European Respiratory Society

FEV1

Forced expiratory volume in 1 second

FVC

Forced vital capacity

HPS

Hepatopulmonary syndrome

HR

Hazard ratio

INR

International normalized ratio

IPVD

Intrapulmonary vasodilatation

LT

Liver transplantation

LV

Left ventricular

MELD-Na

Model for End Stage Liver Disease-Sodium

NAFLD

Non-alcoholic fatty liver disease

NO

Nitric oxide

OR

Odds ratio

PaCO2

Partial pressure of carbon dioxide in arterial blood

PaO2

Partial pressure of oxygen in arterial blood

PCS

Physical component summary

PVCLD2

Pulmonary vascular complications of liver disease 2

SD

Standard deviation

SF-36

Short form-36

SpO2

Oxygen saturation

TAPSE

Tricuspid annular systolic plane excursion

V-Q

Ventilation-perfusion

WHO

World Health Organization