Cardiopulmonary changes and its association with clinical features in noncirrhotic portal hypertension

Abstract

BACKGROUND

Cardiopulmonary changes in noncirrhotic portal hypertension (NCPH) are poorly understood.

AIM

To investigate cardiopulmonary changes using transthoracic echocardiography (TTE) in NCPH and their correlation with clinical features.

METHODS

Prospective cohort including 10 preclinical NCPH [without portal hypertension (PH)] and 32 NCPH subjects who underwent TTE with agitated saline injection and comprehensive clinical evaluation were assessed. PH was defined by presence of either varices, ascites or portosystemic shunting. Intrapulmonary vascular dilatation (IPVD) is defined as appearance of microbubbles in the left atrium after three heartbeats. Right ventricular systolic pressure (RVSP) > 38 mmHg was used to identify possible porto-pulmonary hypertension. Cardiomyopathy is defined using cirrhotic cardiomyopathy consortium criteria.

RESULTS

Among 42 subjects, 17 (40%) had IPVD, 4 (9.5%) had RVSP > 38 mmHg, and 6 (14%) had cardiomyopathy. Aspartate aminotransferase to alanine aminotransferase (AST/ALT) (1.3 vs 1, P = 0.04) and liver stiffness measurement (LSM) (12.4 kPa vs 7.1 kPa, P = 0.03) were higher in those with IPVD. Presence of either LSM > 10 or AST/ALT > 1.2 aided in identifying subjects with IPVD-sensitivity, specificity, and accuracy of 76%. RVSP correlated with oxygen saturation (r = -0.33), and free right hepatic vein pressure (r = 0.43). Those with PH had higher left atrial volume (LAV) (62 mL vs 48 mL, P < 0.01), and LAV index (LAVI) (35 m²vs 23 m², P < 0.01) compared to those without PH. Total bile acids, especially primary bile acids positively correlated with LAV (r = 0.36), and LAVI (r = 0.41).

CONCLUSION

Similar to cirrhotic patients, cardiopulmonary changes are prevalent in NCPH, especially among those with PH. In NCPH, cardiopulmonary changes occur despite preserved synthetic function, suggesting the NCPH model's value in understanding cardiopulmonary dysfunction in liver disease.

Key Words: Cardiomyopathy; Hepatopulmonary syndrome; Portopulmonary hypertension; Porto-sinusoidal vascular disease; Echocardiography; Nodular regenerative hyperplasia; Intrapulmonary vascular dilatation; Bile acids; Angiopoietin 2; Vascular cell adhesion molecule 1

Core Tip: In this study, cardiopulmonary changes in noncirrhotic portal hypertension (NCPH) were investigated using echocardiography and its association with clinical features was explored. Cardiopulmonary changes are prevalent in NCPH patients as in those with cirrhosis. Unlike cirrhosis, cardiopulmonary complications occur despite preserved synthetic function of liver. Liver stiffness measurement and aspartate aminotransferase to alanine aminotransferase ratio help identify patients at risk of hepatopulmonary syndrome.

INTRODUCTION

Noncirrhotic portal hypertension (NCPH) is caused by a diverse group of disorders that primarily affect hepatic vasculature resulting in portal hypertension (PH) in the absence of cirrhosis[1,2]. NCPH can be classified based on the location of resistance to blood flow, as pre-hepatic, hepatic, and post-hepatic causes. Hepatic causes in turn can be divided into presinusoidal, sinusoidal, and post-sinusoidal causes[1]. Porto-sinusoidal vascular disease (PSVD), one of the presinusoidal causes of NCPH is the most common cause of NCPH in the Western world[3]. Individuals with NCPH often maintain normal liver synthetic function and lack hepatic fibrosis, yet they remain susceptible to complications associated with PH, similar to cirrhotic patients[4]. One such group of complications is cardiopulmonary dysfunction, including hepatopulmonary syndrome (HPS), porto-pulmonary hypertension (PoPH), and cardiomyopathy, which are well-described in cirrhosis. In cirrhotic patients, these entities are attributed to a combination of hepatic dysfunction, immune dysregulation, and PH including a hyperdynamic circulatory system[5]. However, cardiopulmonary changes in NCPH patients remains poorly understood.

Among patients with cirrhosis, HPS, PoPH, and cardiomyopathy are relatively common and are initially assessed by transthoracic echocardiography (TTE)[6]. HPS, which results from microvascular remodeling in the alveolar circulation, occurs in 10-30% of patients and is detected by contrast injection and delayed shunting. PoPH occurs in 5% to 6% of patients and is screened by calculating right ventricular systolic pressure (RVSP) using tricuspid regurgitant velocity (TRV)[7,8]. Cardiomyopathy appears to be present in about 30%-35% of patients and is detected by evaluating for the presence of diastolic or systolic dysfunction[9]. Despite the substantial impact of these disorders on prognosis, they are underdiagnosed as most patients remain asymptomatic in the early stages[5]. Although sporadic case series and studies have described the occurrence of cardiopulmonary dysfunction in NCPH, our understanding of frequency, and associations with clinical features remains limited[10-13].

We aimed to explore the frequency of echocardiography findings of HPS, PoPH, and cardiomyopathy in a prospective cohort of NCPH patients and their potential association with PH. An additional exploratory focus was to analyze associations of cardiopulmonary changes with hepatic factors such as bile acids and vascular factors such as those related to angiogenesis.

MATERIALS AND METHODS

Data sources/cohort

A cross-sectional analysis of consecutive subjects with biopsy proven NCPH followed in a single center prospective cohort from July 2015 through July 2023. Prospective natural history study included both preclinical NCPH (without PH) and NCPH subjects (NCT02417740)[14]. PH was defined by the presence of features, namely gastric, esophageal, or ectopic varices, portal hypertensive gastropathy, or the presence of ascites or portosystemic shunting in imaging. Subjects were classified as preclinical NCPH if they had histological features of early nodular regenerative hyperplasia or obliterative portal venopathy[15] and had underlying etiology known to cause NCPH[1], but did not have any above mentioned features of PH. PSVD, one of hepatic causes of NCPH, was defined based on the previously published criteria[16]. Preclinical NCPH subjects were included in this study, since this is an exploratory study and also as it remains unclear whether cardiopulmonary changes can be observed prior to development of features of PH. Exclusion criteria included all known causes of chronic liver disease, known coronary artery disease, heart failure, valvular heart disease, intrinsic lung disease, chronic kidney disease, active infection, malignancy, human immunodeficiency virus, and sarcoidosis. All participants with varices seen after the publication of the Baveno VII guidelines in 2022 were given beta blockers after discussing with participants about potential risks and benefits[16]. Use of beta blockers and diuretics were documented but not suspended during study procedures as this was considered unethical.

All participants underwent evaluation including TTE (Supplementary material), detailed clinical history, anthropometric measurements, vital signs (recorded after sitting in an upright posture for 5 minutes), liver magnetic resonance imaging, laboratory testing, liver stiffness measurement (LSM), esophago-gastroduodenoscopy and liver biopsy. Liver biopsy was considered adequate if it was ≥ 2 cm length or was otherwise considered adequate for interpretation by the expert pathologist (DEK)[16]. LSM examinations using FibroScan^® (Echosens, Paris, France) were performed by trained operators after a minimum 4-hour fasting period. Results, measured in kilopascals (kPa), required at least 10 valid individual measurements interquartile range/median ≤ 30%. Trans-jugular liver biopsy with portal pressure measurements was performed in 30 (71%) participants. Hepatic venous pressure gradient (HVPG) expressed in millimetres of mercury (mmHg), was calculated by subtracting the free hepatic venous pressure from the wedged hepatic venous pressure. At least two readings, from middle or left and right hepatic veins were obtained to ensure reproducibility and accuracy. All evaluations were performed within 7 days of the TTE. Laboratory assays including bile acid measurements and vascular factors measurements [human vascular cell adhesion molecule 1 (VCAM1), angiopoietin 2 (ANG2) and von Willebrand factor (vWF)] were performed for all patients (Supplementary material).

Definition

The presence of intrapulmonary vascular dilatation (IPVD) or shunting was diagnosed by the appearance of microbubbles in the left atrium (LA) 3 or more cardiac cycles after right atrial opacification[17]. RVSP > 38 mmHg was used to identify possible PoPH[18].

Left ventricular diastolic dysfunction (LVDD) in individuals with normal ejection fraction (EF) was diagnosed using the 2020 cirrhotic cardiomyopathy consortium (CCC) criteria which used abnormality in 3 out of 4 parameters - septal early diastolic mitral annular flow velocity (e′) < 7 cm/s, early diastolic transmitral flow to early diastolic mitral annular velocity (E/e′) ≥ 15, left atrial volume index (LAVI) > 34 mL/m², and TRV > 2.8 m/s. Left ventricular systolic dysfunction (LVSD) was defined by an EF ≤ 50% at rest, or an absolute global longitudinal strain by speckle-tracking echocardiography < 18% at rest[9].

Statistical analysis

Statistical analysis was performed using R statistical software (version 4.1.2) and PRISM Data are presented as mean (standard deviation), or median (interquartile range) respectively. Categorical variables were expressed as frequencies and percentages. Quantitative variables were compared using the independent student’s t-test or the Mann-Whitney U test for normally and non-normally distributed variables, respectively. Qualitative variables were compared using the corrected χ² test or a 2-sided Fisher’s exact test, as appropriate. Based on variables with statistical significance a multivariable model was constructed using binary logistic regression. Test characteristics-sensitivity, specificity, negative predictive value, positive predictive value, and accuracy, were calculated. The relationship between parameters was estimated using Spearman correlation. All tests were 2-sided and P-values < 0.05 were considered to be significant.

Ethical approval

This prospective study was approved by the Institutional Review Board of the National Institutes of Health (NCT02417740). Written informed consent was obtained from all participants.

RESULTS

Baseline characteristics of the cohort

Among 50 subjects enrolled, 8 were excluded (one did not undergo TTE, and seven had extrahepatic factors that could affect the cardiopulmonary function, Figure 1). Baseline characteristics of the overall cohort, and comparison among those with preclinical NCPH and NCPH are summarized in Table 1. The median age of the cohort was 52 years, with a slight male (62%) predominance. A detailed description of the cohort was recently published[14]. Vital signs, including heart rate, blood pressure, and oxygen saturation, were similar between the groups. Only participants with NCPH were on nonselective beta-blockers for variceal prophylaxis. In the overall cohort, median fibrosis 4 (FIB-4) score, LSM, and HVPG were 3.4, 8.6 kPa, and 6 mmHg respectively. These were significantly higher in the group with NCPH compared to preclinical NCPH. The cause of NCPH is mentioned in Supplementary Table 1.

Figure 1 Consort diagram of the cohort. Portal hypertension was defined by presence of gastric, esophageal, or ectopic varices, portal hypertensive gastropathy, or the presence of ascites or portosystemic shunting in imaging. NCPH: Noncirrhotic portal hypertension; RVSP: Right ventricular systolic pressure; LVDD: Left ventricular diastolic volume; LVSD: Left ventricular systolic volume.

Table 1 Baseline characteristics of the cohort.

Baseline characteristics	Overall cohort (n = 42)	NCPH (n = 32)	Preclinical NCPH (n = 10)	P value1
Age (years)	52 (41-61)	52 (43 – 62)	49 (40-59)	0.5
Gender, male (%)	26 (62)	20 (63)	6 (60)	0.9
Body mass index (kg/m2)	24.1 (22.1-26.8)	24.0 (22.6-26.8)	25.3 (19.0-26.6)	0.2
Heart rate (bpm)	69 (62-74)	69 (63-74)	69 (60-74)	0.5
Systolic blood pressure (mmHg)	116 (104-127)	114 (103-125)	126 (115-132)	0.2
Diastolic blood pressure (mmHg)	67 (60-71)	62 (59-72)	68 (68-70)	0.3
Oxygen saturation (%)	97 (96-98)	98 (96-99)	97 (96-98)	0.6
Smoking status				0.05
Current or previous smoker, n (%)	15 (36)	14 (44)	1 (10)
Non-smoker, n (%)	27 (64)	18 (56)	9 (90)
Underlying disorder				< 0.01
Immunodeficiency, n (%)	17 (40)	9 (28)	8 (80)
Autoimmune disorders, n (%)	2 (5)	2 (6)
Hematological disorders, n (%)	4 (10)	3 (9)	1 (10)
Exposure to medication, n (%)	4 (10)	4 (13)
Genetic disorders, n (%)	3 (7)	2 (6)	1 (10)
CVID and exposure to drugs, n (%)	1 (2)	1 (3)
Undiagnosed, n (%)	11 (26)	11 (35)
Medication use
Beta blockers, n (%)	17 (40)	17 (53)
Diuretics, n (%)	13 (31)	11 (34)	2 (20)	0.4
Antihypertensives, n (%)	6 (14)	5 (16)	1 (10)	0.4
Antidiabetics including insulin, n (%)	6 (14)	5 (16)	1 (10)	0.4
Immunomodulators, n (%)	12 (29)	7 (22)	5 (50)	0.08
Haemoglobin (g/dL)	12.9 (11.1-14.4)	12.7 (11.0-14.1)	13.3 (12.8-15.4)	0.06
Sodium (mEq/L)	138 (138-140)	138 (138-141)	139 (138-139)	0.6
Creatinine (mg/dL)	0.8 (0.7-0.9)	0.8 (0.7-0.9)	0.8 (0.7 – 1.0)	0.9
Albumin (g/dL)	3.9 (3.7-4.2)	3.9 (3.7-4.2)	4.1 (3.6-4.2)	0.7
C-reactive protein (mg/dL)	2.2 (1.0-4.5)	2.3 (1.2-4.9)	1.7 (1.0 – 3.0)	0.09
Prothrombin time (seconds)	14.2 (13.1-15.0)	14.4 (13.6-15.3)	12.9 (12.7-13.8)	< 0.01
International normalized ratio	1.1 (0.9-1.2)	1.1 (1.0-1.2)	1.0 (0.9-1.0)	< 0.01
Platelet count (109/L)	94 (62-153)	77.5 (53.8-134)	171 (128-246)	0.02
Alanine aminotransferase, AST (U/L)	45 (25-60)	40 (26-59)	45 (25-77)	0.6
Aspartate aminotransferase, ALT (U/L)	46 (31-62)	44 (31-60)	46.5 (31.3-75.3)	0.6
AST/ALT ratio	1.1 (0.9-1.3)	1.1 (0.9-1.3)	1.1 (1.0-1.2)	0.9
Fibrosis-4 index	3.4 (1.9-6.3)	4.1 (2.7-7.4)	1.5 (1.4-2.6)	< 0.01
MELD-Na	9 (8-10)	9 (8-10)	8 (7-10)	0.06
Liver stiffness measurement (kPa)	8.6 (6.0-13.5)	11.6 (7.7-14.5)	5.3 (4.3-6.1)	< 0.01
Hepatic venous pressure gradient (mmHg)2	6 (4-11)	10 (5-15)	4 (3-5)	< 0.01

¹P values show a comparison among those with and without portal hypertension cohort using either Student’s t-test or the Mann-Whitney U test for continuous variables and corrected χ² test or a 2-sided Fisher’s exact test for categorical variables.

²30 patients (23 with portal hypertension and 7 without portal hypertension) underwent transjugular portal pressure measurement.

Data presented as median (IQR), mean (SD) or n (%). MELD-Na: Model for end-stage liver disease-sodium; CVID: Common variable immunodeficiency; NCPH: Noncirrhotic portal hypertension; AST/ALT: Aspartate aminotransferase to alanine aminotransferase.

Prevalence of cardiopulmonary changes

Among 42 subjects, four subjects had intracardiac shunting. Among 38 subjects without intracardiac shunting, 17 (45%) had IPVD, of which 15 had PH and 2 did not. Four (13%) subjects with PH had RVSP > 38 mmHg, suggesting the possible PoPH. Using the CCC criteria, 6 (14%) had cardiomyopathy, 3 subjects each with LVDD and LVSD (Figure 1). Supplementary Figure 1 shows presence of cardiopulmonary changes based on their PH status.

Factors associated with intrapulmonary shunting

Among those with shunting, only 4 (24%) subjects complained of dyspnea at rest or with exertion, and none had a history of smoking. Only 2 (12%) subjects with shunting had oxygen saturation < 96%, but neither required supplemental oxygen. Interestingly, oxygen saturation among those with or without shunting did not differ. Smoking status and medication use were similar between the groups. The liver related factors aspartate aminotransferase to alanine aminotransferase (AST/ALT) ratio (1.3 vs 1.1, P = 0.04) and LSM (12.4 kPa vs 7.1 kPa, P = 0.03) were significantly higher in the group with shunting compared to those without (Table 2). TTE parameters were not different among those with and without shunting (Supplementary Table 2).

Table 2 Factors associated with shunting.

Characteristics	Shunt present (n = 17)	Shunt absent (n = 21)	P value1
Age (years)	51 (39-61)	49 (41-57)	0.7
Male gender, n (%)	8 (47)	16 (76)	0.09
Body mass index (kg/m2)	24.1 (23.4-28.7)	23.5 (21.9-25.7)	0.04
Heart rate (bpm)	70 (63-78)	68 (60-70)	0.1
Systolic blood pressure (mmHg)	116 (101-124)	116 (105-127)	0.6
Diastolic blood pressure (mmHg)	61 (60-69)	68 (62-70)	0.2
Oxygen saturation (%)	97 (96-98)	98 (97-99)	0.08
Had features of portal hypertension (%)2	15 (88)	13 (62)	0.1
Smoking status			0.6
Current or previous smoker, n (%)	5 (29)	8 (38)
Non-smoker, n (%)	12 (71)	13 (62)
Underlying disorder			< 0.01
Immunodeficiency, n (%)	3 (18)	13 (62)
Autoimmune disorders, n (%)	1 (5)	1 (5)
Hematological disorders, n (%)	2 (12)	1 (5)
Exposure to medication, n (%)	2 (12)	2 (9)
Genetic disorders, n (%)	3 (18)
CVID and exposure to drugs, n (%)		1 (5)
Undiagnosed, n (%)	6 (35)	3 (14)
Medication use
Beta blockers, n (%)	8 (47)	8 (38)	0.7
Diuretics, n (%)	5 (29)	5 (24)	0.7
Antihypertensives, n (%)	2 (12)	3 (14)	0.9
Antidiabetics including insulin, n (%)	4 (24)	1 (5)	0.2
Immunomodulators, n (%)	3 (18)	9 (43)	0.2
Sodium (mEq/L)	138 (138-140)	138 (138-139)	0.3
Creatinine (mg/dL)	0.8 (0.7-0.9)	0.8 (0.7-0.9)	0.6
Albumin (g/dL)	3.9 (3.8-4.2)	3.9 (3.6-4.2)	0.6
Haemoglobin (g/dL)	13.5 (10.8-14.2)	12.9 (11.8-14.4)	0.6
C-reactive protein (mg/dL)	2.3 (1.4-3.6)	1.9 (1.0-5.2)	0.1
Prothrombin time (seconds)	14.2 (13.4-15.4)	14.1 (13.0-14.5)	0.2
International normalized ratio	1.1 (1.0-1.2)	1.1 (0.9-1.1)	0.2
Platelet count (109/L)	85 (65-146)	97 (60-144)	0.8
Alanine aminotransferase, ALT (U/L)	32 (25-53)	45 (26-76)	0.2
Aspartate aminotransferase, AST (U/L)	42 (32-49)	46 (30-71)	0.7
Chenodeoxycholic acid (nmol/mL)	6.1 (2.0-14.5)	4.1 (1.2-1.0)	0.1
Cholic acid (nmol/mL)	4.8 (1.8-11.1)	1.3 (0.7-10.4)	0.7
Deoxycholic acid (nmol/mL)	1.7 (0.8-9.1)	0.9 (0.1-2.4)	0.09
Ursodeoxycholic acid (nmol/mL)	0.4 (0.2-1.7)	0.3 (0.1-0.7)	0.2
Primary bile acid (nmol/mL)	14.2 (3.8-26.0)	5.1 (2.0-18.7)	0.3
Secondary bile acid (nmol/mL)	1.7 (1.2-11.6)	1.3 (0.3-2.4)	0.08
Total Bile acid (nmol/mL)	15.9 (5.2-33.0)	7.5 (4.4-22.2)	0.2
Spleen length (cm)	16.2 (12.7-20.0)	15.8 (12.6-18.7)	0.8
Wedged hepatic venous pressure (mmHg)3	23 (14-27)	19 (14-26)	0.6
Free hepatic vein pressure (mmHg)3	13 (10-15)	11 (9-15)	0.4
Hepatic venous pressure gradient (mmHg)3	8 (5-11)	6 (3-14)	0.6
AST/ALT ratio	1.3 (1.0-1.4)	1.0 (0.8-1.1)	0.04
FIB-4 index	3.7 (2.0-5.0)	3.3 (1.6-6.6)	0.9
MELD-Na	10.0 (8.0-10.0)	8.0 (7.0-10.0)	0.7
Liver stiffness measurement (kPa)	12.4 (7.2-16.8)	7.1 (5.4-10.3)	0.03

¹P values show a comparison among the groups using either Student’s t-test or the Mann-Whitney U test for continuous variables and corrected χ² test or a 2-sided Fisher’s exact test for categorical variables.

²Features of portal hypertension were defined by presence of gastric, esophageal, or ectopic varices, portal hypertensive gastropathy, or the presence of ascites or portosystemic shunting in imaging.

³30 patients (14 with shunt and 16 without shunt) underwent transjugular portal pressure measurement.

Data presented as median (IQR), mean (SD) or n (%). MELD-Na: Model for end-stage liver disease-sodium; CVID: Common variable immunodeficiency; AST/ALT: Aspartate aminotransferase to alanine aminotransferase.

Liver-related criteria to predict shunting

To aid in deciding the need to screen for shunting, a criterion was developed using liver related factors, which were significantly different among the groups in univariate analysis. On multivariable logistic regression analysis, a combination of AST/ALT ratio and LSM performed with an area under receiver operating curve (area under receiver operating characteristic = 0.75, 95%CI: 0.59-0.91, P < 0.01) were used to identify shunting. To simplify the screening criteria, presence of either LSM > 10 kPa or AST/ALT ratio > 1.2 were identified as cut off in deciding the need for TTE shunt screening. This criterion had more than 70% performance in all test characteristics (Figure 2). Patients who met the criteria had 10.4 (95%CI: 2.3-38.9, P < 0.01) times higher odds of having shunting.

Figure 2 Performance of liver-related criteria to screen for shunting. Wilson-Brown method was used to calculate the confidence interval of test characteristics. PPV: Positive predictive value; NPV: Negative predictive value; AUROC: Area under receiver operating characteristic; IPVD: Intrapulmonary vascular dilatation; AST/ALT: Aspartate aminotransferase to alanine aminotransferase.

Features associated with elevated RSVP

Four patients had RVSP > 38 mmHg. There was no difference in oxygen saturation (97% vs 97%, P = 0.09), LSM (11.9 kPa vs 8.4 kPa, P = 0.5), and AST/ALT ratio (1.13 vs 1.07, P = 0.4) among those with and without elevated RVSP. However, the FIB-4 index was significantly higher in the group with possible PoPH (8.9 vs 3.2, P = 0.03). Multiple factors either positively or negatively correlated with RVSP as shown in Table 3. Notably, despite oxygen saturation being within the normal range, it negatively correlated with elevation in RVSP (r = -0.33, P = 0.04). Interestingly, AST/ALT ratio positively correlated with RVSP, but LSM did not. Free right hepatic vein pressure and IVC diameter measured by TTE positively correlated with RVSP (Supplementary Figure 2). Similar correlation findings were observed when the analysis was limited to the PH group alone (Supplementary Table 3).

Table 3 Factors associated with right ventricular systolic pressure.

Variables	Correlation1, r	P value
Age	0.31	0.04
Oxygen saturation (%)	-0.33	0.03
Cardiac output (L/min)	0.33	0.04
Cardiac index (L/min/m2)	0.29	0.07
LV mass index (gm/m2)	0.31	0.05
LA volume (mL)	0.51	< 0.01
LA volume index (mL/m2)	0.5	0.01
RA volume (mL)	0.37	0.02
RA volume index (mL/m2)	0.42	< 0.01
Septal MV E/e’	0.28	0.08
Lateral MV E/e’	0.39	0.01
TAPSE (mm)	0.35	0.04
IVC diameter (mm)	0.32	0.04
AST/ALT ratio	0.33	0.03
Liver stiffness measurement (kPa)	0.04	0.7
Wedged hepatic venous pressure (mmHg)2	0.09	0.6
Free hepatic vein pressure (mmHg)2	0.43	0.02
Hepatic venous pressure gradient (mmHg)2	-0.08	0.6

¹Spearman correlation.

²30 patients underwent transjugular portal pressure measurement.

LV: Left ventricle; TAPSE: Tricuspid annular plane systolic excursion; IVC: Inferior vena cava; AST/ALT: Aspartate aminotransferase to alanine aminotransferase; LA: Left atrial; RA: Right atrial.

Features associated with cardiomyopathy

All six subjects with cardiomyopathy had PH and 5 had esophageal varices. There was no difference in oxygen saturation (97% vs 97%, P = 0.3), FIB-4 index (3.3 vs 3.5, P = 0.5), LSM (14.5 kPa vs 8.3 kPa, P = 0.1) and AST/ALT ratio (1.2 vs 1.1, P = 0.3) among those with and without cardiomyopathy.

Echocardiographic features among those with preclinical NCPH and NCPH

A comparison of echocardiographic findings between the groups is shown in Supplementary Table 4. As expected, cardiac chamber volumes were significantly increased in those with PH: Left ventricle (LV) diastolic volume index (66 mL/m²vs 56 mL/m², P < 0.01), LA volume index (35 mL/m²vs 23 mL/m², P < 0.01). Since smoking status and beta blocker use were significantly different between the groups, TTE parameters were compared after removing participants with exposure to beta- blockers and a history of smoking. In patients with no history of beta-blocker use, chamber volumes remained increased in the group with PH (Table 4).

Table 4 Features among those with and without portal hypertension (excluding smokers and beta blocker users).

Baseline characteristics	NCPH (n = 7)	Preclinical NCPH (n = 9)	P value1
Age (years)	51 (42-63)	46 (39-57)	0.6
Male gender, n (%)	4 (57)	5 (56)	0.8
Body mass index (kg/m2)	25.8 (23.3-31.6)	25.1 (18.6-25.7)	0.09
Heart rate (bpm)	70 (67-80)	70 (65-75)	0.2
Systolic blood pressure (mmHg)	116 (113-128)	124 (114-129)	0.7
Diastolic blood pressure (mmHg)	67 (55-71)	68 (68-69)	0.5
Oxygen saturation (%)	96 (96-98)	97 (96-97)	0.7
Sodium (mEq/L)	139 (138-140)	139 (138-139)	0.9
Creatinine (mg/dL)	0.8 (0.7-0.8)	0.8 (0.7-1.0)	0.7
Albumin (g/dL)	3.9 (3.8-3.9)	4.1 (3.5-4.2)	0.9
Haemoglobin (g/dL)	12.7 (11.9-14.1)	13 (12.8-15.7)	0.3
C-reactive protein (mg/dL)	1.2 (0.6-3.9)	1.5 (1.0-2.6)	0.4
Left ventricular ejection fraction (%)	60 (59-62)	62 (62-65)	0.1
Cardiac output (L/min)	5.6 (5.2-7.8)	5.2 (4.4-6.0)	0.3
Cardiac index (L/min/m2)	2.9 (2.7-3.8)	2.8 (2.7-3.0)	0.5
LV diastolic volume (mL)	146 (133-164)	97 (89-121)	0.02
LV diastolic volume index (mL/m2)	69 (64-86)	57 (48-59)	0.04
LV mass index (gm/m2)	84.9 (80.5-98.3)	68.4 (64.1-86.5)	0.3
LV global longitudinal strain (%)	-22.7 (-24.4--19.0)	-21.6 (-22.3--19.9)	0.9
RV free wall strain (%)	-24.6 (-26.9--23.5)	-25.9 (-28.9--24.3)	0.6
RVSP (mmHg)	26 (24-38)	27 (25-30)	0.3
LA volume (ml)	80 (73-85)	46 (34-59)	0.03
LA volume index (mL/m2)	37 (35-44)	23 (22-29)	0.04
RA volume (mL)	57 (52-64)	28 (26-43)	0.09
RA volume index (mL/m2)	26 (26-30)	16 (13-20)	0.1
Septal MV E/e’	10.0 (8.1-11.7)	7.4 (6.5-8.8)	0.1
Lateral MV E/e’	7.8 (6.5-9.8)	4.9 (4.5-6.9)	0.1
Septal e’ (cm/s)	9 (8-11)	11 (10-11)	0.4
Lateral e’ (cm/s)	12 (12-15)	15 (10-15)	0.8
TAPSE (mm)	25 (22-26)	24 (23-25)	0.9
Mitral E to A ratio	1.2 (0.8-1.2)	1.4 (1.1-1.5)	0.9
IVC diameter (mm)	23.0 (16.8-30.8)	20.0 (12.0-21.0)	0.1

¹P values show a comparison among the groups using either student’s t-test or the Mann-Whitney U test for continuous variables and corrected χ² test or a 2-sided Fisher’s exact test for categorical variables.

Data presented as median (IQR), mean (SD) or n (%). LVID: Left ventricular internal diameter; LV: Left ventricle; RV: Right ventricular; RVSP: Right ventricular systolic pressure; LA: Left atrial; RA: Right atrial; TAPSE: Tricuspid annular plane systolic excursion; IVC: Inferior vena cava; NCPH: Noncirrhotic portal hypertension.

Bile acids and cardiac parameters

There was no difference in bile acids levels among those with and without shunting (Table 2). Bile acids also did not correlate with RVSP. Bile acids levels were not different among those with/without cardiomyopathy and elevated RVSP (data not shown). However, both total and primary bile acid levels positively correlated with left atrial and left ventricular volume and index (Supplementary Table 5).

Angiogenesis factors and cardiac dysfunction

Factors associated with angiogenesis, notably ANG2 and VCAM1 are shown to be elevated in patients with HPS due to cirrhosis[19]. In this NCPH cohort, ANG2 (3.5 ng/mL vs 1.5 ng/mL, P = 0.3), and VCAM1 (6929 ng/mL vs 6308 ng/mL, P = 0.7) were not different among those with and without shunting (Supplementary Figure 3). Also, vWF levels were not different (24939 ng/mL vs 19659 ng/mL, P = 0.2) among those with and without shunting. Angiogenesis factors were not different among those with elevated RVSP and cardiomyopathy.

DISCUSSION

This exploratory analysis demonstrates that subjects with NCPH, despite their preserved synthetic function experience a spectrum of cardiopulmonary changes like IPVD, and elevated RVSP similar to what is seen in cirrhotic patients. In this NCPH cohort, cardiomyopathy and features suggestive of possible HPS and PoPH are prevalent at rates comparable to those in patients with cirrhosis. Furthermore, a noninvasive assessment combining LSM and AST/ALT ratio can aid in identifying which subjects could be considered for HPS screening.

In this study, the prevalence of IPVD in PH group was 47%, higher than prior reports, which was approximately 12%[11,12]. This rate was also higher than what is seen in cirrhotics[20]. This difference might be due to relatively older age of our cohort, the difference in underlying disorders that led to NCPH, and variations in the severity of PH. However, similar to what has been observed previously, pulse oximetry has proven to be a less effective screening test for HPS[21]. Among NCPH participants with IPVD, only 12% had oxygen saturation below 96%. A study by DuBrock et al[22] found significant differences in CI and CO among patients with HPS compared to those who do not. However, in this cohort, there was no difference in CI and CO among those with and without IPVD. Among liver related factors, AST/ALT ratio and LSM were higher in the group with shunting compared to those who did not, indicating a potential association between advanced liver disease and the development of IPVD. Current recommendation advice screening for cardiopulmonary dysfunction in individuals with symptoms or advanced liver disease undergoing liver transplant evaluation[17]. However, symptoms and pulse oximetry are not sensitive enough to identify those with underlying HPS. Instead, using a non-invasive test which reflect disease severity, either AST/ALT ratio > 1.2 or LSM > 10 kPa can guide in deciding the need for IPVD screening in non-transplant candidates. In our cohort, these criteria performed with more than 70% test performance for all the test characteristics. However, further validation in an independent cohort is needed.

Echocardiography is a valuable screening tool for identifying PoPH when other potential etiologies for pulmonary hypertension have been excluded[18]. In this cohort, 4 subjects with PH were identified to possibly have PoPH. However, lack of cardiac catheterization to confirm PoPH diagnosis is a limitation. Among factors associated with RVSP, it was observed that oxygen saturation was negatively correlated with RVSP. All patients with RVSP > 38 mmHg had oxygen saturation < 98% in sitting position. The pathophysiology of PoPH remains incompletely understood, but one widely accepted theory suggests that PH facilitates the formation of portosystemic shunts. This, in turn, leads to the shunting of vasoactive metabolites into the pulmonary arteries, resulting in the obliteration of pulmonary arteries and an increase in pulmonary vascular resistance[23]. In our NCPH cohort, there was a positive correlation between RVSP and free hepatic pressure, which could either reflect the increased downstream effect of PoPH or increased portal pressure leading to the development of PoPH. However, there was no correlation between HVPG and RVSP. This could be because HVPG is underestimated in presinusoidal/prehepatic condition like NCPH and hence not reliably corelate with severity of PH in these patients[2].

In the overall cohort, 14% had cardiomyopathy. In cirrhosis, the prevalence of cardiomyopathy varies depending on the population being studied. Most prevalence reports focus on patients being assessed for liver transplantation[5,24]. However, a European study that examined the prevalence of cardiomyopathy in all patients with cirrhosis, following the 2020 CCM guidelines, reported that diastolic dysfunction was observed in 7.4% of cases, while systolic dysfunction was identified in 12.3%[25,26]. These results are comparable to what was seen in consecutive NCPH participants enrolled in this study. Detecting cardiomyopathy in patients with liver disease has shown to be important as it is associated with poor outcomes[27].

Studies involving cirrhotic patients have shown that higher LAVI are associated with poor outcomes[27,28]. A study by Merli et al[28] showed among cirrhotic patients, during a median follow up of 2 years, a LAVI > 35.23 mL/m² was associated with decreased survival compared to those with lower LAVI. In our cohort, those with PH had higher LAVI compared to those without. Prognostic importance of these findings in NCPH need to be explored in future studies.

Excess bile acids are known to be cardiotoxic. One of the possible mechanisms suggested is that an increase in bile acid levels causes activation of stress kinases, inhibiting fatty acid oxidation, which is the primary energy source for the heart, resulting in metabolic derangement. Elevated bile acids levels have been shown to be associated with enlarged left atrial volume[29]. In a study of pediatric patients with biliary atresia, it was shown that bile acids levels correlated with left ventricular diameter[30]. In this NCPH cohort, a positive correlation with total bile acids and LAVI as well as LV diastolic volume index were found. These associations were mostly with primary bile acids, suggesting a potential role of bile acids in cardiac changes, but future studies are required to establish causality.

Dysregulation in angiogenesis is thought to be one of the factors associated with HPS. A study by Kawut et al[19] found significantly higher levels of ANG2 in those with HPS compared to those without. They also found VCAM1, which plays an important role in immune response and inflammation, and vWF, which has a primary role in blood clotting and platelet function, are higher in cirrhotic participants with HPS. In this study, there was no difference in any of these factors among those with and without IPVD. This could be because the underlying disorder that led to the development of NCPH, including immune deficiency and hematological disorders, potentially affects the levels of these factors.

This study has some limitations. The relatively small cohort size restricted our capacity to perform multivariate analyses, underscoring the need for caution when generalizing the findings. Echocardiography has some inherent limitations in tomographic assessments of the left ventricular outflow tract and often underestimates these values and subsequent stroke volume and cardiac output calculations. It is also unclear whether use of beta blockers during the TTE evaluation could have influenced some of the findings. Since betablockers reduce heart rate, decrease contractility, and alter chamber filling, their use may have influenced some of our echocardiographic findings. In addition, some essential features critical for diagnosing HPS and risk stratification, such as arterial blood gas and shunt quantification, were not available. Similarly, the lack of confirmatory right heart catheterization made it difficult to fully substantiate the diagnosis of PoPH. Lastly, given that the study was conducted in a referral centre, there is the possibility of selection bias, as patients referred to such centres may have unique characteristics compared to the broader population. Hence these findings need to be validated in an independent cohort.

CONCLUSION

In summary, patients with NCPH despite preserved hepatic synthetic function can develop cardiopulmonary changes, especially those with PH. Future studies are needed to explore the underlying pathophysiology, importance of cardiopulmonary changes and prognostic implications in these patients.

ACKNOWLEDGEMENTS

We acknowledge the assistance of Stanislav Sidenko, Wen Li, Lisbet Odou-Lorenzo for their assistance with data collection.