Portopulmonary hypertension: Results from a 10-year screening algorithm


  • Michael J. Krowka,

    Corresponding author
    1. Divisions of Pulmonary and Critical Care Medicine, Mayo Clinic College of Medicine, Rochester, MN
    2. Gastroenterology and Hepatology, Mayo Clinic College of Medicine, Rochester, MN
    3. Liver Transplant Center, Mayo Clinic College of Medicine, Rochester, MN
    • Professor of Medicine, Mayo Clinic College of Medicine, 200 1st Street SW, Rochester, MN 55905
    Search for more papers by this author
    • fax: 507-266-4372

  • Karen L. Swanson,

    1. Divisions of Pulmonary and Critical Care Medicine, Mayo Clinic College of Medicine, Rochester, MN
    Search for more papers by this author
  • Robert P. Frantz,

    1. Cardiovascular Disease, Mayo Clinic College of Medicine, Rochester, MN
    Search for more papers by this author
  • Michael D. McGoon,

    1. Cardiovascular Disease, Mayo Clinic College of Medicine, Rochester, MN
    Search for more papers by this author
  • Russell H. Wiesner

    1. Gastroenterology and Hepatology, Mayo Clinic College of Medicine, Rochester, MN
    2. Liver Transplant Center, Mayo Clinic College of Medicine, Rochester, MN
    Search for more papers by this author

  • Presented in abstract form at the 2005 International Liver Transplant Society Meeting, Los Angeles, CA.

  • Potential conflict of interest: Dr. Krowka is a consultant for and received grants from Cotherix. Dr. McGoon is a consultant for and received grants from Cotherix, Medtronic, and Myogen. He is a consultant for Actelion.


Portopulmonary hypertension (POPH) is the elevation of pulmonary artery pressure due to increased resistance to pulmonary blood flow in the setting of portal hypertension. Increased mortality has occurred with attempted liver transplantation in such patients and thus, screening for POPH is advised. We examined the relationship between screening echocardiography and right heart catheterization determinations of pressure, flow, volume, and resistance. A prospective, echocardiography–catheterization algorithm was followed from 1996 to 2005. Consecutive transplantation candidates underwent Doppler echocardiography to determine right ventricular systolic pressure (RVSP). Of 1,235 patients, 101 with RVSP >50 mm Hg underwent catheterization to measure mean pulmonary artery pressure (MPAP), flow via cardiac output (CO), central volume via pulmonary artery occlusion pressure (PAOP), and resistance via calculated pulmonary vascular resistance (PVR). Bland-Altman analysis suggested marked discordance between echocardiography-derived RVSP and catheterization results. All-cause pulmonary hypertension (MPAP >25 mm Hg) was documented in 90/101 (90%) patients. Using current pressure and resistance diagnostic guidelines (MPAP >25 mm Hg, PVR ≥240 dynes/s/cm−5), POPH was documented in 66/101 (65%) patients. Elevated MPAP was due to increased CO and/or PAOP in 35/101 (35%) patients with normal resistance (PVR <240 dynes/s/cm−5). The transpulmonary gradient (MPAP–PAOP) further characterized POPH in the presence of increased volume. Model for end stage liver disease (MELD) scores correlated poorly with MPAP and PVR. In conclusion, right heart catheterization is necessary to confirm POPH and frequently identifies other reasons for pulmonary hypertension (e.g., high flow and increased central volume) in liver transplantation candidates. Severity of POPH correlates poorly with MELD scores. (HEPATOLOGY 2006;44:1502–1510.)

Pulmonary hypertension associated with advanced liver disease has variable etiologies and prognostic implication.1–4 First, the hyperdynamic, high-flow circulatory state (as a consequence of splanchnic vasodilatation caused by portal hypertension) results in high cardiac output (CO) and an increase in mean pulmonary artery pressure (MPAP), but the pulmonary vascular resistance (PVR) remains normal. Second, elevated MPAP with increased central blood volume estimate as measured by the pulmonary arterial occlusion pressure (PAOP) results in variable effect on PVR. Third, for reasons yet to be characterized, and regardless of liver disease severity, additional hemodynamic change can occur due to development of pulmonary artery endothelial/smooth muscle proliferation, vasoconstriction, in situ thrombosis, and plexogenic arteriopathy. Such pathology results in progressive obstruction to pulmonary arterial flow from the right ventricle to the lungs, very high pulmonary artery pressure, and markedly increased PVR.1, 2, 5–7 This pathophysiology characterizes the concept of portopulmonary hypertension (POPH), an uncommon hemodynamic problem associated with intraoperative liver transplantation death and posttransplant hospitalization mortality.2, 8 POPH is now recognized as a unique entity within the modified World Health Organization Group I category of pulmonary arterial hypertension.9

Noninvasive screening for POPH using Doppler echocardiography in liver transplantation candidates has been proposed,10–12 but reports of echocardiography–right heart catheterization relationships are few with limited interpretation of right heart catheterization results.1, 10, 11, 13, 14 We report the results of our prospective screening echocardiography–right heart catheterization algorithm conducted in consecutive liver transplantation candidates at the Mayo Clinic from 1996 to 2005. We categorized pulmonary hemodynamic results based on PVR, the key concept that characterizes obstruction to pulmonary arterial flow. We described the existence of distinct hemodynamic subgroups (that may prove to have prognostic significance); and hypothesized that no relationship would exist between MELD scores and the severity of portopulmonary hypertension.


POPH, portopulmonary hypertension; RVSP, right ventricular systolic pressure; MPAP, mean pulmonary artery pressure; CO, cardiac output; PAOP, pulmonary artery occlusion pressure; PVR, pulmonary vascular resistance; MELD, model for end-stage liver disease; ERS-EASL, European Respiratory Society–European Association for the Study of the Liver; TPG, transpulmonary pressure gradient; CTP, Child-Turcotte-Pugh.

Patients and Methods

Patient Selection.

From July 1996 to December 2005, consecutive patients met minimal listing criteria (Child-Turcotte-Pugh score ≥7) for liver transplantation in the Mayo Clinic Liver Transplant Clinic. Each patient underwent transthoracic Doppler echocardiography screening and pretransplantation pulmonary consultation. Based on preliminary screening data and liver transplantation outcome,1, 10 transplantation candidates with an increased right ventricular systolic pressure >50 mm Hg underwent subsequent right heart catheterization to determine specific pulmonary hemodynamic patterns and document the existence of POPH. Prospective data collection and retrospective data analyses were reviewed and approved by the Mayo Institutional Review Board.

POPH Diagnostic Criteria.

POPH is defined by hemodynamic criteria determined via catheterization. Recently, investigators have followed the diagnostic criteria proposed by the European Respiratory Society–European Association for the Study of the Liver (ERS-EASL) Task Force on Hepatic-Pulmonary Vascular Disorders of POPH4: (1) MPAP ≥25 mm Hg; (2) PVR ≥240 dynes/s/cm−5; and (3) PAOP <15 mm Hg.

Transthoracic Doppler Echocardiography.

A complete transthoracic echocardiographic examination was performed for routine pretransplantation evaluation, including two-dimensional, continuous-wave color Doppler recordings. Imaging on various models of commercially available equipment included tricuspid regurgitant peak velocity determinations via the apical, parasternal, and subcostal views. Interpretations were documented by staff cardiologists in the Mayo Clinic Echocardiography Laboratory. The right ventricular systolic pressure (RVSP) was determined using the modified Bernoulli equation:

equation image

where TR (m/s) = tricuspid regurgitant peak velocity.

Right atrial pressure estimate was determined via echocardiographic assessment of the inferior vena cava size and degree of collapse with respiration15, 16 (Table 1).

Table 1. Right Atrial Pressure Estimate
Right Atrial PressureInferior Vena Cava Diameter/Collapse
 4 mm HgDiameter < 15 mm
10 mm Hg15 < Diameter < 20 mm; > 50% collapse
14 mm Hg15 < Diameter < 20 mm; < 50% collapse
20 mm HgDiameter > 20 mm; no collapse

Right Heart Catheterization.

Catheterization was conducted using a triple-lumen, balloon-tipped thermodilution catheter in all patients with RVSP >50 mm Hg. Percutaneous vascular assess was accomplished via the internal jugular vein when the international normalized ratio was <1.5 IU. Cardiac outputs were determined via thermodilution using 10 mL of 5% dextrose solution (average of 3 assessments). Standard pressure measurements included determinations of mean right atrial pressure, pulmonary artery systolic pressure, MPAP, and PAOP with the pressure transducer positioned at the midaxillary line. PAOP was measured at the end of expiration and was an estimate of central volume reflecting the left ventricular end-diastolic volume. PVR was calculated using the standard formula:

equation image

The difference between MPAP and PAOP is referred to as the transpulmonary gradient (TPG).

Liver Disease Severity.

Severity of liver disease was determined using Child-Turcotte-Pugh (CTP) classification and model for end stage liver disease (MELD) score. A CTP score (range 5-15) is based upon three serum parameters (total bilirubin, prothrombin time, and albumin) and two qualitative, clinical parameters (ascites and encephalopathy). CTP class A (score 5-6) is less severe than CTP class C (score 10-15).5 The MELD score is quantitative, being derived from a formula with the variables serum total bilirubin, international normalized ratio, and serum creatinine.17 The CTP classification and MELD score determinations were obtained based on data available at the time of the right heart catheterization in a retrospective manner.

Statistical Analysis.

All data were presented as the mean ± SD. Pearson-product linear correlations were calculated. Two-tailed t tests and ANOVA were used for mean comparisons. Bland-Altman analysis was used to describe RVSP (echocardiography) versus pulmonary artery systolic pressure (catheterization) variation.10


Doppler Echocardiography

At the time of data analysis, 1,235 consecutive liver transplantation candidates had undergone echocardiography per the algorithm (Fig. 1). RVSP could be determined in 958 (77.6%) patients, and 104 (10.9%) patients had RVSP >50 mm Hg with right heart catheterization conducted a mean of 5 days following Doppler echocardiography. Three individuals had RVSP >50 mm Hg (55, 75, and 95 mm Hg), but each died prior to right heart catheterization.

Figure 1.

Overall results in patients who had screening Doppler echocardiography. Only patients with estimated RVSP >50 mm Hg underwent right heart catheterization. Catheterization results are summarized by various hemodynamic categories. RVSP, right ventricular systolic pressure; PVR, pulmonary vascular resistance; MPAP, mean pulmonary artery pressure; TPG, transpulmonary pressure gradient; PAOP, pulmonary artery occlusion pressure.

Echocardiography–Right Heart Catheterization Comparison

For all patients with RVSP >50 mm Hg, the correlation between RVSP and catheterization results was moderate at .60 (Fig. 2A); for those with POPH the correlation was worse (.43). Bland-Altman analysis indicated that RVSP tended to overestimate catheterization results at higher RVSP pressures (Fig. 2B).

Figure 2.

Relationship between echocardiography-derived estimated RVSP and right heart catheterization results, specifically the measured pulmonary artery systolic pressure. (A) Linear Pearson-product correlation between PASP and RVSP. (B) Bland-Altman plot displaying the variation between the mean of the two estimates of pulmonary artery systolic pressure (RVSP + PASP)/2 and the difference between PASP and RVSP. The discordance between RVSP and PASP appears to worsen with increasing RVSP. PASP, pulmonary artery systolic pressure; RVSP, right ventricular systolic pressure.

Right Heart Catheterization: Subgroups

Patient characteristics and catheterization results (n = 101) categorized by PVR are summarized in Table 2 and Fig. 1. Within each PVR category, further subgroup classification by pressure (MPAP), flow (CO), central volume status (PAOP), and TPG are shown in Tables 3 and 4. The resistance–pressure and resistance–flow relationships are shown in Fig. 3A-D.

Table 2. Patient Characteristics Categorized by PVR (dynes/s/cm−5) Determined via Right Heart Catheterization
 AllPVR <240PVR >240P Value
  • Abbreviatons: NS, not significant; HepC, hepatitis C; HCC, hepatocellular carcinoma; AC, alcoholic cirrhosis; PBC, primary biliary cirrhosis; PSC, primary sclerosing cholangitis; CC, cryptogenic cirrhosis; NASH, nonalcoholic steatohepatitis; MELD, model for end stage liver disease; INR, international normalized ratio; CTP, Child-Turcotte-Pugh; TIPS, transjugular intrahepatic portosystemic shunt.

  • *

    Included ZZ antitrypsin deficiency and congential hepatic fibrosis. PVR subgroups were not associated with specific diseases (P = .29).

  • Either spontaneous (n = 1 in the PVR >240 group) or surgical (n = 8).

Total10135 (35%)66 (66%) 
Male 4413 (30%)31 (70%)NS
Female 5722 (39%)35 (61%) 
Age, yr53 ± 1055 ± 951 ± 1 
Liver disease    
 HepC  83 5 
 HepC/HCC  32 1 
HepC/AC 14311 
 AC 22517 
PBC 115 6 
 PSC  64 2 
 CC 16511 
 NASH 102 8 
 Other* 116 5 
MELD score14.0 ± 6.215.4 ± 7.313.3 ± 5.5.13
 Total bilirubin4.6 ± 8.06.4 ± 10.73.7 ± 6.1NS
 Serum creatinine1.2 ± .71.2 ± .81.2 ± .6NS
 INR1.3 ± .31.4 ± .41.3 ± .3NS
CTP class    
 A 3811 (29%)27 (71%)NS
 B 3611 (31%)25 (69%) 
 C 2713 (48%)14 (52%) 
Portacaval shunts  91 8 
Splenectomy  61 5 
TIPS  44 0 
Beta-blocker use 22814 
MPAP >35 mm Hg 7410 (13%)64 (85%) 
Table 3. Pulmonary Hemodynamic Patterns Associated With Advanced Liver Disease (Mean Right Heart Catheterization Data; n = 101)
  • NOTE. All patients had RVSP >50 mm Hg via echocardiography (normal <35 mm Hg).

  • Abbreviations: MPAP, mean pulmonary artery pressure; CO, cardiac output; PVR, pulmonary vascular resistance; PAOP, pulmonary artery occlusion pressure; TPG, transpulmonary pressure gradient; POPH, portopulmonary hypertension; RVSP, right ventricular systolic pressure.

  • *

    Hemodynamic parameter differences within high-flow and POPH subgroups (P < .001).

  • Volume differences within high-flow and POPH subgroups (P < .001).

High flow (↓PVR) (n = 35)31 + 98.6 + 2.6*142 + 5816 + 6*16 + 7
 Normal volume (n = 20)28 + 88.2 + 2.3154 + 6012 + 217 + 7
 Increased volume (n = 15)34 + 109.1 + 3.0125 + 5221 + 414 ± 7
POPH (↑PVR) (n = 66)49 + 11*6.1 + 2.0533 + 247*12 + 637 + 11*
 Normal volume (n = 50)48 + 115.9 + 2.0571 + 25710 + 338 + 11
 Increased volume (n = 16)53 + 96.8 + 2.0407 + 17121 + 534 ± 10
Table 4. Cardiac and Renal Findings in Patients With Increased PAOP (n = 31)
Patient no.PAOPLVEF (%)Diastolic DysfunctionAbnormal Liver FunctionAtrial Volume*Renal ClearanceCRTPGPVRMPAP
  • Abbreviations: PAOP, pulmonary artery occlusion pressure; LVEF, left ventricular ejection fraction; CR, serum creatinine; TPG, transpulmonary pressure gradient; PVR, pulmonary vascular resistance; MPAP, mean pulmonary artery pressure; ND, not done.

  • *

    Abnormal = abnormally increased.

  • Abnormal = abnormal via creatinine clearance or iothalamate clearance.

Abnormal PVR (n = 16)          
Normal PVR (n = 15)          
Figure 3.

Resistance–pressure–flow relationships via right heart catheterization. Each patient had RVSP >50 mm Hg during screening. (A) PVR versus CO for all patients (n = 101). As the resistance to pulmonary arterial flow increases, a reduction in blood flow as measured by CO is noted. (B) PVR versus MPAP for all patients (n = 101). Increasing resistance to pulmonary arterial flow is associated with increasing MPAP. Above a PVR >800, a few patients appear to show a relative reduction in MPAP, suggesting failure of the right ventricle as PVR increases. (C) MPAP versus CO in only those patients with increased PVR (POPH) (n = 66). A nonlinear relationship (CO worsens with increasing MPAP) suggests worsening right ventricular function. (D) TPG (MPAP–PAOP) versus CO for all patients (n = 101). All patients above the PVR = 240 dynes/s/cm−5 line had increased PVR; all patients below the PVR = 240 line had normal PVR. The patients between PVR = 120 dynes/s/cm−5 and PVR = 240 dynes/s/cm−5 merit close follow-up. PVR, pulmonary vascular resistance; MPAP, mean pulmonary artery pressure.


Increased PVR (≥240 dynes/s/cm−5) was documented in 66/101 (65%) patients and was greater than those with normal PVR (533 ± 247 vs. 142 ± 58; P <. 001) (Table 2). MPAP was significantly higher in those with increased PVR versus those with normal PVR (49 ± 11 vs. 31 ± 9 mm Hg; P < .011). Mean CO and PAOP were lower in patients with increased PVR (P < .001); mean TPG was significantly higher in those with increased PVR (P < .001). An inverse relationship between PVR and CO was noted, with a decreasing CO associated with increasing PVR (Fig. 3A).


Increased MPAP (≥25 mm Hg) was documented in 90/101 (90%) patients. Of the 66 patients with POPH, 64 (97%) patients had moderate to severe POPH (MPAP ≥35 mm Hg with abnormal PVR). In 13% of patients, MPAP ≥35 mm Hg was associated with normal PVR. In patients with MPAP <25 mm Hg (11/101 [11%]), PVR ranged from 60 to 205 dynes/s/cm−5 and the CO range was 4.5 to 10.2 L/min. As a group, MPAP tended to increase with increasing PVR (Fig. 3B).


There was a poor correlation between CO and most pulmonary hemodynamic parameters (Table 5). CO tended to increase, peak, and then decline with increasing MPAP (Fig. 3C).

Table 5. MELD Score Correlations With Right Heart Catheterization Pulmonary Hemodynamics in POPH (n = 66)
  1. Abbreviations: MPAP, mean pulmonary artery pressure; CO, cardiac output; PAOP, pulmonary artery occlusion pressure; TPG, transpulmonary pressure gradient; PVR, pulmonary vascular resistance; RVSP, right ventricular systolic pressure; PASP, pulmonary artery systolic pressure; MELD, model for end stage liver disease.



The subgroup of 31 patients (31%) with increased PAOP (>15 mm Hg; median 19 mm Hg with range 16-34 mm Hg) was analyzed in terms of left ventricular function (ejection fraction) and left atrial volume (by echocardiography), as well as renal status (by creatinine or iothalamate clearance). Abnormal PVR was documented in 16/31 (50%) patients; the maximum PVR was 772 dynes/s/cm−5, and the maximum MPAP was 76 mm Hg (Table 4). The mean ejection fraction was 65 ± 6 %; (range 50%-77%). Diastolic dysfunction was documented in 5/31 (15%) patients. Increased left atrial volume (cm3/m2 > 30) was noted in 24/31 (68%) patients; abnormal creatinine or iothalamate clearance was noted in 9/31 (29%) patients.


An abnormal TPG was noted in every patient with abnormal PVR (n = 66). An abnormal PAOP was documented in 16/66 (24%) of these patients (Fig. 1). Mean PVR of 407 ± 171 dynes/s/cm−5 was noted in patients with increased PAOP (Table 2).

Right Heart Catheterization: Relationship With Severity of Liver Disease

There was no significant relationship between the liver disease diagnoses or severity (as measured by CTP A, B, C classification or the MELD score and MPAP or PVR (Table 2). No correlation was noted between MELD scores and any measure of pulmonary hemodynamics (Table 5).

Pulmonary Hemodynamics at Time of Liver Transplantation

In the cohort with RVSP >50 mm Hg, liver transplantation was subsequently conducted in 27/101 (27%) patients; 6 patients required pre- and posttransplantation prostacyclin therapy, with no transplantation hospitalization mortality. During the time of this algorithm, 906 patients have undergone liver transplantation. In prostacyclin-treated patients with RVSP <50 mm Hg or indeterminate RVSP, no patient was subsequently noted to have MPAP ≥50 mm Hg at the time of liver transplantation—an absolute contraindication to transplantation at our institution that would have necessitated cancellation of surgery.


Screening echocardiography before liver transplantation has been espoused by other investigators,11, 13 is supported by an ERS-EASL Hepatic-Pulmonary Vascular Disorder Task Force,4 and is a practice guideline from the American Association for the Study of Liver Disease.12 Our prospective algorithm, the largest series reported to date, merits comment from four perspectives: POPH echocardiography screening shortcomings, the necessity for catheterization to accurately characterize pulmonary pressure–flow–volume–resistance relationships, the potential clinical importance of pulmonary hemodynamic subgroups, and MELD score correlation with pulmonary hemodynamics.

POPH Echocardiography Screening Shortcomings.

The purpose of echocardiography screening before liver transplantation is to identify patients with clinically significant POPH before the surgery, thus preventing the surprise identification of moderate to severe POPH at the time of transplantation. Although our prospective screening identified patients with a range of increased RVSP for all-cause pulmonary hypertension, echocardiography alone was not adequate in characterizing the specificity for or severity of POPH. However, the algorithm did identify all patients with moderate to severe POPH (MPAP >35 mm Hg) before transplantation.

The overall correlation between RVSP and pulmonary systolic pressure measure during catheterization was reasonable, but weaker in the subgroup with POPH. In that regard, our data are similar to findings from Cotton et al.11 It would appear that in both studies, as well as in our previous analysis, the discordance between RVSP and catheterization results worsens as the screening echo-derived RVSP increases.10

RVSP was not obtained in approximately 22% of patients screened due to the inability to accurately detect a measurable tricuspid systolic regurgitant flow. Such limitation is not unusual, as noted in previous studies in patients with liver disease11, 13 wherein RVSP could not be determined in 18% to 68% of patients studied. In those patients, the decision to proceed to catheterization should be based on the qualitative findings of the echocardiogram and clinical picture. An enlarged or dilated right ventricle would be concerning for significant pulmonary hypertension and the need for further investigation. The contribution of right atrial pressure estimates to the calculation (and variability) of RVSP determinations via the Bernoulli formula may not be trivial, ranging from 4 to 20 mm Hg at our institution. Such variation in right atrial pressure estimates can significantly affect reported RVSP values, as well as the threshold of RVSP used to determine the need for right heart catheterization.16, 17

The screening RVSP cutoff (>50 mm Hg) algorithm described herein did offer a practical approach in deciding which transplantation candidates should proceed to invasive hemodynamic measurements. In our experience, using an RVSP cutoff of >30 mm Hg or >40 mm Hg would have resulted in catheterization being conducted in 50% and 20% of screened patients, respectively (unpublished Mayo Clinic data). In the current algorithm, 10.9% of screened patients underwent invasive measurements.

Necessity for Right Heart Catheterization.

Although screening echocardiography can identify patients with all-cause pulmonary hypertension, our data suggest that catheterization is necessary to characterize specific hemodynamic patterns and pressure–flow–resistance relationships. Importantly, these patterns may require different therapeutic approaches based on volume status. Approximately 33% of patients with RVSP >50 mm Hg had normal pulmonary arterial resistance to blood flow (calculated PVR) associated with a high-flow state (increased CO) and/or increased central volume (elevated PAOP) causing increased MPAP. Despite abnormal echocardiography findings, these patients did not fulfill current diagnostic criteria for POPH. These data are similar to the experience reported by Colle et al.,13 in which 7/17 (41%) patients with abnormal screening echocardiography did not satisfy POPH criteria based on catheterization data. Arguably, such patients would not require aggressive pulmonary antihypertensive therapy with such agents as prostacyclin, endothelin antagonists, or phosphodiesterase inhibitors. The risk in proceeding to liver transplantation in such patients would not appear to be increased from a pulmonary hemodynamic perspective as supported by previous reports.1, 11, 13

Catheterization allowed calculation of a transpulmonary gradient (MPAP–PAOP). Importance of the transpulmonary gradient relates to the scenario in which increased pulmonary venous pressure (measured via PAOP) may exist in addition to and further worsen an increased MPAP. In other words, pulmonary venous hypertension possibly adds to the existing pulmonary artery hypertension problem. In normal individuals, simply increasing pulmonary venous pressures would not necessarily increase pulmonary vascular resistance due to recruitment of other pulmonary vessels. Such recruitment may be limited or nonexistent in POPH.

Most patients with increased PVR and MPAP associated with abnormal PAOP (n = 16) (Table 4) had volume overload (suggested by left atrial volume index >30 cm3/m2) with or without systolic dysfunction (ejection fraction <60%) or diastolic dysfunction (defined echocardiographically by an abnormal mitral inflow pattern). Chronic renal insufficiency (defined by creatinine clearance or markedly elevated serum creatinine) was noted in two patients (one required dialysis). Each patient in this subgroup had elevated TPG suggesting the combined venous and arterial abnormality. Accepted diagnostic criteria for POPH requires that the PAOP should be <15 mm Hg.4 In our opinion, increased PAOP and TPG should not preclude the diagnosis of POPH, simply due to increased PAOP. These patients should be considered as a separate subgroup of POPH. Modification to the ERS-EASL diagnostic criteria for POPH (replace PAOP with TPG parameter) merits further study. Diuresis with pulmonary vasodilator therapy would still be a careful management option in this setting.

Most patients with normal PVR, increased MPAP and abnormal PAOP (n = 15) (Table 4) had volume overload associated with diastolic dysfunction or renal insufficiency. The mean MPAP and TPG in this subgroup were significantly less than that noted in patients with increased PVR. In short, careful cardiorenal study is needed in suspected POPH patients with increased PAOP. Salt restriction, diuretic therapy with or without attention to diastolic function and optimal fluid management would seem appropriate therapies.

Relative Importance of Pulmonary Hemodynamics and Subgroups.

Elevated MPAP has been the most important parameter in discussing the multicenter diagnostic implications of POPH.5, 8 However, various reasons for elevated MPAP exist in the setting of liver disease (hyperdynamic circulatory state, increased central volume). Our data clearly demonstrate that elevated MPAP alone is not specific for the diagnosis of POPH, supporting the current diagnostic criteria that must include PVR.4 The key concept concerning POPH, in our opinion, is the increased resistance to pulmonary arterial flow as measured by the calculated entity PVR. The prospective importance of stratification based on hemodynamic subgroups based on resistance, pressure, flow, and volume has yet to be determined. However, our preliminary data suggest that MPAP in the setting of increased PVR and PVR alone may have long-term prognostic implication irrespective of liver transplantation candidacy.18 To our knowledge, no prospective studies have been published that address the relative importance of pulmonary hemodynamic parameters (MPAP, PVR, CO, and TPG) as each relates to the prognosis of POPH in nontransplanted patients. In the largest retrospective study accomplished in the pre-prostacyclin and transplantation treatment era, survival within a POPH cohort (n = 26) was not related to any of these parameters as reported by Hadengue et al.7

Regarding these parameters and contraindication to transplantation, limited data have suggested that those candidates with MPAP <35 mm Hg do well intraoperatively and have excellent long-term survival.1, 18 Those with MPAP ≥35 mm Hg and increased PVR have increased risk for intraoperative death during transplantation surgery and postoperative mortality.2, 5, 8 Cardiac output was not identified as a prognostic factor.5, 8 No prognostic conclusions can be stated from previous studies3, 5, 7, 8 concerning TPG.

This algorithm correctly identified all patients before transplantation that might require significant pharmacological intervention, namely those identified with moderate to severe POPH (MPAP ≥35 mm Hg and increased PVR). Importantly, no patient was found to have clinically significant POPH at the time of operation (MPAP >50 mm Hg) that necessitated cancellation of surgery.

MELD Score Correlations With Pulmonary Hemodynamics.

Poor correlation was noted between MELD scores calculated at the time of catheterization and all measured parameters of pulmonary hemodynamics. These data support the premise that severity of hepatic dysfunction for transplantation priority has no relationship to the severity of POPH. Whether POPH patients with moderate to severe POPH should have adjusted MELD scores in an attempt to improve survival is an issue that requires further study. Clearly, these data suggest that MELD scores would not offer additional insight into posttransplantation as it relates to varying degrees of POPH.

Several limitations of this study should be noted. The current analysis was limited to accurate diagnostic classification and does not allow conclusions regarding outcome. With additional patient follow-up and careful consideration of therapeutic interventions including liver transplantation, characterizations of our long-term outcomes will be forthcoming in a separate analysis. Despite our hemodynamic characterization, this study did not address the specific etiologies of POPH, which remain speculative. These include the embolic effects of surgical or spontaneous portocaval shunts, splenectomy, increased circulating levels of endothelin-1, or possible genetic polymorphisms. In addition, the effect of beta-blockade in this cohort is unknown. Controlled withdrawal of beta-blockers has been shown to increase CO and improve 6-minute walk, symptoms, and PVR (with no change in MPAP).19 As previously reported, the hemodynamic patterns in POPH (especially the increased CO) do appear different from those noted in idiopathic pulmonary artery hypertension.20 RVSP could not be determined in 22.4% of patients; therefore, we did not determine specific hemodynamic patterns in such patients. The choice of RVSP >50 mm Hg for screening was based on a small series at our institution after much education of our cardiology colleagues as to the importance of accurate, noninvasive right heart hemodynamic assessment. The expense (approximately $6,000 per catheterization), potential morbidity (due to thrombocytopenia and coagulopathy) and backup with pulmonary hemodynamic measurements at the time of transplantation also shaped our criteria selection. The variable times between the conduct of echocardiography and catheterization due to patient travel convenience/expense were factors that we could not control.

In conclusion, although screening echocardiography may suggest POPH, discordance between echocardiography and right heart catheterization results was significant in this cohort. Catheterization was necessary to accurately diagnose and determine the severity of POPH, as well as identify other causes for pulmonary hypertension (high flow and/or increased central volume) in the setting of liver disease. MELD scores correlated poorly with all measured parameters of pulmonary hemodynamics in patients with POPH. Future studies involving long-term outcome in POPH patients relating parameters of pulmonary vascular resistance, flow, and pressure to evolving therapies that include liver transplantation will be of interest.