Portopulmonary hypertension (POPH), the presence of pulmonary hypertension in association with portal hypertension, is a known complication of chronic liver disease.[1-9] Prospective studies and case-control studies have documented that POPH occurs in approximately 5% to 6% of patients with advanced liver disease. In patients with portal hypertension, the association with pulmonary hypertension is seen in 2% to 6%.[10, 11] The incidence of POPH in patients referred for liver transplantation (LT) is 4% to 6%. Although Doppler echocardiography has proven to be a useful noninvasive screening tool for the detection of POPH, right heart catheterization (RHC) remains the gold standard for diagnosis.[3, 13, 14] The criteria for POPH include a mean pulmonary artery pressure (mPAP) > 25 mm Hg, a pulmonary vascular resistance (PVR) > 240 dyne·second·cm−5, and a pulmonary capillary wedge pressure (PCWP) < 15 mm Hg (measured during RHC).[1-9] In the setting of LT, the presence of POPH is associated with poor outcomes.[1, 4, 15] An elevated risk of morbidity and mortality due to right heart failure has been reported in patients with moderate (mPAP ≥ 35 mm Hg and < 45 mm Hg) to severe POPH (mPAP ≥ 45 mm Hg).[16-19] Mortality rates of 50% and 100% have been reported for patients with moderate and severe POPH, respectively. Krowka et al. demonstrated that patients with POPH have significantly poorer outcomes (with respect to survival rates and freedom from all-cause hospitalizations) than patients with idiopathic pulmonary hypertension, although patients with POPH have better hemodynamics at the time of diagnosis. mPAP values > 35 mm Hg are considered a contraindication for LT. This is the reason that patients with moderate to severe POPH are treated with vasodilator therapy before LT (ie, to lower mPAP to a value < 35 mm Hg because this is not associated with an additional intraoperative risk). Without medical intervention, a mean survival of approximately 15 months can be expected for patients with POPH.[4-6, 11, 18, 21, 22] In order to improve survival, patients with POPH should be treated with medication or undergo LT. A noninvasive estimation of the systolic pulmonary artery pressure (sPAP), documented during echocardiography, allows us to screen LT candidates as part of their pretransplant evaluation.[7, 13, 14, 21] Although it is generally agreed that LT candidates should be screened with echocardiography, different sPAP cutoff values at which patients should be referred for RHC are used at different centers. At our institution, an sPAP cutoff value of 30 mm Hg is used to detect POPH at an early stage in accordance with the results of the prospective study by Colle et al in 2003. With a cutoff value of 30 mm Hg, positive and negative predictive values of 59% and 100%, respectively, were reported for 165 LT candidates undergoing successive echocardiography and RHC. In other words, repeated sPAP measurements ≤ 30 mm Hg rule out POPH. By using this low cutoff value, which detects all degrees of POPH, we aim to reduce to a minimum the risk of missing patients with POPH during the preoperative evaluation. However, a drawback of this low cutoff value is the high number of false positives. In this study, the accuracy of Doppler echocardiography was assessed with different cutoff values, which ranged from 30 to 50 mm Hg.
Portopulmonary hypertension (POPH), a complication of chronic liver disease, may be a contraindication to liver transplantation (LT) because of the elevated risk of peritransplant and posttransplant morbidity and mortality. Because POPH is frequently asymptomatic, screening with echocardiography is recommended. The only reliable technique, however, for diagnosing POPH is right heart catheterization (RHC). The aims of this study were to evaluate the current estimated systolic pulmonary artery pressure (sPAP) cutoff value of 30 mm Hg and to determine a better cutoff value. One hundred fifty-two patients underwent pretransplant echocardiography between January 2005 and December 2010. These echocardiographic results were compared with pulmonary artery pressures measured during the pretransplant workup or at the beginning of the transplantation procedure (both by catheterization). With a cutoff value of 30 mm Hg, 74 of the 152 patients met the criteria for POPH on echocardiography, although the diagnosis was confirmed in only 7 patients during catheterization; this resulted in a specificity of 54%. It would have been more accurate to use a cutoff value of 38 mm Hg, which had a maximal specificity of 82% and, at the same time, guaranteed a sensitivity and negative predictive value of 100%. With the incorporation of the presence or absence of right ventricular dilatation, the specificity even increased to 93% for this new cutoff value. In conclusion, the prevalence of POPH was 4.6% among LT candidates in this study. We can recommend that LT candidates with an sPAP > 38 mm Hg should be referred for RHC. With the cutoff value increased from 30 to 38 mm Hg, the number of patients undergoing invasive RHC during their evaluation could be safely reduced. Liver Transpl 19:602–610, 2013. © 2013 AASLD.
mean pulmonary artery pressure
mean right atrial pressure
pulmonary capillary wedge pressure
pulmonary vascular resistance
right heart catheterization
right ventricular end-diastolic diameter
systolic pulmonary artery pressure
systemic vascular resistance
PATIENTS AND METHODS
The medical records of 152 patients who underwent an adequate pretransplant evaluation at Ghent University Hospital from January 2005 to December 2010 were retrospectively analyzed. These 152 patients were among 371 patients evaluated for LT. Seventy-six patients were removed from the transplant waiting list or died during this time interval. Twenty-six of the remaining 295 LT candidates were minors and were excluded from the study. Several re-LT procedures, revisions, and the absence of an adequate preoperative evaluation were also considered exclusion criteria. Patients with POPH were included in the study, even though they had not undergone LT up to that point. The remaining 152 study patients included 100 males and 52 females. The mean age was 58 ± 11 years (range = 22-73 years). The indications for transplantation are summarized in Table 1. Echocardiography was performed routinely (every 3 months) as part of the pretransplant workup for all candidates. For the analysis of transthoracic echocardiography as a screening test for the detection of POPH, pulmonary pressures, estimated during echocardiography, should ideally be compared with hemodynamic data measured during RHC. At our hospital, a cutoff value of 30 mm Hg was used, so patients with an sPAP ≤ 30 mm Hg did not undergo RHC, and this led to a number of missing values. These cases were corrected through the analysis of hemodynamic data measured during LT (more specifically, after the induction of general anesthesia and before the abdominal incision). The ethical committee of the Faculty of Medicine and Health Sciences, Ghent University, approved the protocols.
|Hepatitis C virus||20||13|
|Primary sclerosing cholangitis||10||7|
|Hepatitis B virus||9||6|
|Primary biliary cirrhosis||5||3|
Formal RHC was performed with a Swan-Ganz catheter inserted into the right or left jugular vein. The systolic and diastolic pressures, mPAP, PCWP, and mean right atrial pressure (mRAP) were measured. The cardiac output (CO) was determined with the thermodilution method. The systemic vascular resistance (SVR) was calculated as follows:
PVR was calculated as follows:
The transpulmonary gradient (TPG) was calculated as follows:
In every patient who underwent transplantation, additional pulmonary artery pressure and wedge pressure measurements were performed because at our center all transplant recipients receive a pulmonary artery catheter after the induction of anesthesia. End expiratory measurements during controlled minute ventilation were enabled by monitor-incorporated software. Pulmonary artery pressure measurements taken after the postinduction stabilization period were possibly lower than awake measurements, but this affected only patients with an sPAP ≤ 30 mm Hg before induction because these were the only patients who did not undergo RHC before transplantation. POPH was defined as an mPAP > 25 mm Hg, a PVR > 240 dyne·second·cm−5, and a PCWP < 15 mm Hg.
All patients were evaluated with 2-dimensional Doppler transthoracic echocardiography on a GE Vivid 7 machine (General Electric Co., GE Healthcare, Wauwatosa, WI). Tricuspid regurgitation was used to calculate the right ventricular sPAP as an estimation of sPAP with the following formula:
The latter value depends on the degree of vascular filling of the inferior vena cava.
An assessment of the right ventricular morphology requires the integration of multiple views. Although different methods of quantitative echocardiographic right ventricle assessment have been described, in clinical practice, the study of the right ventricle morphology is mainly qualitative. In this study, the right ventricular end-diastolic diameter (RVEDD) was measured at the parasternal long-axis view. Measurements greater than 3.3 cm were considered right ventricular dilatation.
Receiver operating characteristic curve analysis was used to determine the ideal cutoff value with maximal sensitivity and specificity. The proportions of true positives, true negatives, false positives, and false negatives were determined, and the sensitivity, specificity, positive predictive value, negative predictive value, positive likelihood ratio, and accuracy of Doppler echocardiography were calculated with different cutoff values. The statistical analysis was performed with SPSS 19.0 (SPSS, Inc., Chicago, IL). After a Shapiro-Wilk test of normality, the following statistical tests were used: a 1-way analysis of variance, the Kruskal-Wallis test, the chi-square test, and the Pearson correlation test. P values < 0.05 were considered to be statistically significant.
Analysis of the Actual sPAP Cutoff Value of 30 mm Hg
Patients were categorized into 3 groups according to the combined results of echocardiography (with an sPAP cutoff value of 30 mm Hg) and RHC. Group 1 included 7 patients who met the criteria for POPH (sPAP > 30 mm Hg) on echocardiography and whose diagnosis was confirmed by RHC (true POPH). All patients had evidence of right ventricle dilatation on echocardiography. Two patients had abnormal tricuspid annular plane systolic excursion values as a sign of right ventricle dysfunction. Interventricular septal flattening (a D shape) was documented in 5 patients. Two patients underwent transplantation successfully with decreased pulmonary pressures after combined oral vasodilator therapy (phosphodiesterase type 5 inhibitors and endothelin 1 receptor antagonists). The therapy was continued in the posttransplant phase to date in one patient, whereas in the other transplant patient, even oral vasodilator therapy could be stopped. These patients were doing very well 6 and 5 years after LT, respectively. A third patient died during LT because of perioperative complications that induced uncontrolled pulmonary hypertension. The other 4 patients were still under treatment with oral vasodilator therapy. One was awaiting LT. The other 3 patients were not listed because the severity of their POPH and their general condition were considered contraindications to LT. The individual characteristics of these 7 patients are listed in Table 2. Group 2 consisted of 67 patients with pulmonary hypertension on echocardiography for whom the diagnosis of POPH was ruled out by RHC (false POPH). Group 3 included the remaining 78 patients who did not meet the criteria according to echocardiography or RHC (no POPH). The numbers of true positives, false positives, true negatives, and false negatives were 7, 67, 78, and 0, respectively. For an overview of the sensitivity, specificity, positive predictive value, negative predictive value, positive likelihood ratio, and accurary value, see Table 3.
|Patient Number||Age at Transplant (Years)||Sex||sPAP (mm Hg)||mPAP (mm Hg)||PCWP (mm Hg)||PVR (dyne·second·cm−5)||TPG (mm Hg)||CO (L/minute)||Cardiac Index (L/Minute/m2)||Transplant||Time on Waiting List (Days)||Outcome at ≤90 Days||Outcome at >90 Days|
|7 Patients With POPH on Echocardiography Confirmed by RHC|
|25 Patients With POPH on Echocardiography (With a Cutoff Value of 38 mm Hg) Not Confirmed by RHC|
|30 mm Hg||35 mm Hg||38 mm Hg||40 mm Hg||45 mm Hg||50 mm Hg|
|Positive predictive value (%)c||10||14||22||21||33||46|
|Negative predictive value (%)d||100||100||100||99||99||99|
|Positive likelihood ratioe||2.2||3.3||5.6||5.4||10.8||17.2|
Determination and Analysis of a New Cutoff Value
A new cutoff value with maximal sensitivity and specificity that guaranteed a 100% negative predictive value was determined with receiver operating characteristic curve analysis. The area under the curve represents the accuracy of the Doppler screening test with a specific cutoff value, and a value of 1 corresponds to the ideal cutoff value. An sPAP cutoff value of 38 mm Hg had an area under the curve of 0.974 and was associated with a sensitivity of 100%, which corresponded to a specificity of 82% (see Fig. 1).
With the new cutoff value of 38 mm Hg, the patients were again categorized into 3 groups (true positives or true POPH, false positives or false POPH, and true negatives or no POPH). Seven patients met the criteria for POPH on echocardiography and during RHC (see Table 2 for individual characteristics). In 25 patients for whom POPH was suspected on echocardiography, the diagnosis of POPH was ruled out by RHC. All 25 patients underwent LT. None of them developed pulmonary hypertension later. The individual characteristics of these patients are listed in Table 2 as well. All patients in groups 1 and 2 had an adequate tricuspid regurgitant jet on echocardiography. In 120 patients who did not meet the criteria for POPH on echocardiography, the absence of POPH was confirmed by RHC. At the end of the study, all underwent transplantation. Ten patients had an adequate tricuspid regurgitant jet on echocardiography and a calculated sPAP ≤ 38 mm Hg. In the remaining 110 patients, a tricuspid regurgitant jet was absent, and sPAP could not be calculated. Of great importance is the absence of false negatives. The associated sensitivity, specificity, positive and negative predictive value, positive likelihood ratio, and accuracy value are summarized in Table 3.
Analysis of Higher Cutoff Values
For higher cutoff values, the numbers of true and false positives and true and false negatives were also calculated. A summary of the sensitivities, specificities, positive and negative predictive values, positive likelihood ratios, and accuracy values can be found in Table 3. Notable is the loss of a 100% negative predictive value, which was, on the contrary, secured for cutoff values up to 38 mm Hg.
Analysis of the Impact of Including Right Ventricular Morphological Features on the Accuracy of Doppler Echocardiography
In order to reduce the number of false positives and to enhance the specificity of Doppler echocardiography, we added for any sPAP cutoff value a variable of right ventricular morphology. More specifically, we analyzed the impact of taking the presence or absence of right ventricular dilatation into account. Right ventricular dilatation was defined as an RVEDD > 3.3 cm. Patients with an RVEDD ≤ 3.3 cm were considered to have no POPH. Patients with an RVEDD > 3.3 cm and an estimated sPAP greater than the tested sPAP cutoff value were considered to be suspicious for POPH. According to the catheterization measurements, they were categorized as having true or false POPH. A summary of the sensitivities, specificities, positive and negative predictive values, positive likelihood ratios, and accuracy values can be found in Table 4.
|30 mm Hg||35 mm Hg||38 mm Hg||40 mm Hg||45 mm Hg||50 mm Hg|
|Positive predictive value (%)c||18||25||41||40||50||54|
|Negative predictive value (%)d||100||100||100||99||99||99|
|Positive likelihood ratioe||4.5||6.9||14.5||13.8||20.7||24.9|
Comparison of the Hemodynamic Parameters
A comparison of the systemic hemodynamic investigations in the 3 groups (true positives, false positives, and true negatives) is shown in Table 5. mPAP, PVR, and TPG differed significantly among the 3 groups and reflected the criteria on which the differentiation between true and false POPH is based. PCWP was lower in group 1, although this difference did not reach a significant level.
|Variable||Group 1: True POPH (n = 7)||Group 2: False POPH (n = 25)||Group 3: No POPH (n = 120)||P Value|
|Mean arterial pressure (mm Hg)||99 ± 11||89 ± 16||89 ± 13||NS|
|mPAP (mm Hg)||47 ± 16||21 ± 7||18 ± 7||<0.001|
|PCWP (mm Hg)||11 ± 4||15 ± 7||13 ± 6||NS|
|PVR (dyne·second·cm−5)||532 ± 242||82 ± 54||84 ± 38||<0.001|
|TPG (mm Hg)||36 ± 16||6 ± 3||5 ± 2||<0.001|
Correlation Between mPAP and PCWP
In groups 2 and 3, mPAP and PCWP were significantly correlated (r2 = 0.93, P < 0.001). In group 1 (confirmed POPH), no correlation between mPAP and PCWP existed (r2 = 0.08, P = 0.86). This difference in correlations between group 1 and groups 2 and 3 represents one of the main criteria for differentiating between true and false POPH.
POPH is known to be associated with an elevated risk of intraoperative and postoperative morbidity and mortality due to right heart failure. In this study, POPH was seen in 4.6% of the LT candidates. Because POPH is frequently asymptomatic, systematic screening is recommended.[4, 13, 14, 21] Previous studies have described the successful use of transthoracic Doppler echocardiography as a noninvasive technique for the detection of POPH as part of the pretransplant evaluation.[3, 7, 21, 25] However, there is considerable variation at different centers in the sPAP cutoff values at which patients should be referred for RHC. At our hospital, a cutoff value of 30 mm Hg for the estimated sPAP, which is measured during echocardiography, is used for referring patients for RHC to confirm or to rule out the diagnosis of POPH; this value was previously determined in a prospective study published in 2003 by Colle et al. They reported positive and negative predictive values of 59% and 100%, respectively, for 165 LT candidates undergoing successive echocardiography and RHC. Our institution chose to follow this approach in an attempt to reduce the risk of missing patients with POPH during the preoperative evaluation to a minimum. A drawback to date is the high number of false positives, which is represented by a low positive predictive value. Here, a positive predictive value of only 10% was calculated for the currently used cutoff value of 30 mm Hg. A false-positive test result can be interpreted as follows: a significant proportion of the false positives had an mPAP > 25 mm Hg measured during RHC, but in contrast to true POPH, this elevated value of mPAP was associated with a PCWP ≥ 15 mm Hg. In groups 2 and 3 (false POPH and no POPH, respectively), a highly significant correlation between mPAP and PCWP was documented. In group 1 (true POPH), there was no correlation. This difference in correlations between group 1 and groups 2 and 3 represents one of the main criteria for differentiating between true and false POPH (or no POPH). Showing the importance of RHC, Krowka et al. demonstrated that only 66% of patients with an sPAP > 50 mm Hg on echocardiography also have an elevated PVR. In our study, only 58% (7/12) of the patients with an sPAP > 50 mm Hg on echocardiography also had an elevated PVR. RHC is considered the only reliable tool for differentiating between patients with increased PVR (true POPH) and patients with normal PVR (false POPH), for whom an elevated mPAP is a result of hyperdynamic circulation, which is frequently seen in patients with cirrhosis. A mildly increased pulmonary pressure is found in 20% to 50% of patients with cirrhosis because of increased CO with normal PVR.[9, 27] This differentiation between true and false POPH is of great importance because there is no elevated risk of intraoperative and postoperative morbidity or mortality in patients with an increased mPAP due to a hyperdynamic circulation; this is in contrast to patients with true POPH. Krowka and Golbin and Krowka also proposed TPG as a hemodynamic parameter for diagnosing true POPH. An elevated TPG (>12 mm Hg) correlates with an elevated PVR and reflects the severity of the obstruction of pulmonary blood flow. Therefore, including TPG as one of the diagnostic criteria is another way of distinguishing an elevated mPAP in the setting of POPH from an elevated mPAP caused by hyperdynamic flow and volume overload. Consequently, we can argue that RHC is the only method for identifying true POPH. In contrast to its low specificity, the cutoff value of 30 mm Hg is also associated with maximal sensitivity and a negative predictive value of 100%. With this cutoff value, no patients with POPH would be missed during the preoperative workup. In summary, echocardiography with a cutoff value of 30 mm Hg can be considered a highly sensitive and safe screening method for the detection of POPH, although it has important limitations in terms of its specificity and positive predictive value. In this context, a new cutoff value, associated with a higher specificity yet guaranteeing a 100% negative predictive value, was determined.
With the aid of a receiver operating characteristic curve, a better cutoff value of 38 mm Hg was determined, and it was associated with a sensitivity and specificity of 100% and 82%, respectively. With the actual cutoff value increased from 30 to 38 mm Hg, the number of false positives could be safely reduced. With a cutoff value of 30 mm Hg, 74 patients underwent RHC, whereas if a cutoff value of 38 mm Hg had been used, only 32 patients would have been referred for RHC. In contrast to the use of 30 mm Hg, at which 44% (67/152) were wrongly referred, only 16% (25/152) would have been wrongly referred for RHC with a cutoff value of 38 mm Hg. Here we can conclude that with an increase in the cutoff value from 30 to 38 mm Hg, the number of patients referred for RHC could be safely reduced.
Some hospitals use higher cutoff values up to 50 mm Hg, which were also previously studied.[31, 32] According to the statistical results of this study, cutoff values > 38 mm Hg may be associated with a higher risk of missing patients with POPH during the preoperative phase (no 100% negative predictive values). Using a cutoff value of 50 mm Hg, Cotton et al. determined positive and negative predictive values of 37.5% and 91.9%, respectively. Kim et al. reported positive and negative predictive values of 74% and 97%, respectively, with a cutoff value of 50 mm Hg. In other words, an sPAP > 50 mm Hg predicts moderate to severe POPH in 3 of 4 LT candidates. In our opinion, the use of cutoff values > 38 mm Hg should not be recommended because of the elevated risk of morbidity and mortality due to right heart failure. On the other hand, the specificity, positive predictive value, and accuracy improve as the cutoff value increases. This advantage, however, does not offset the increased risk of missing patients with POPH. If patients with POPH were missed during the pretransplant evaluation, elevated pulmonary pressures would, however, be detected at the time of LT. This situation should be avoided anyway because the intraoperative detection of POPH implies the discontinuation of surgery and the loss of the donor liver.
We can conclude that a cutoff value of 30 mm Hg means too many false positives. A safe alternative is to increase the cutoff value to 38 mm Hg, which guarantees a 100% negative predictive value. Using higher cutoff values (eg, 50 mm Hg) means the loss of the 100% negative predictive value, although the specificity would be higher (95%). Therefore, an sPAP < 38 mm Hg can be used to rule out POPH, whereas higher cutoff values are more specific for the diagnosis of POPH. Whether patients with an sPAP between 38 and 50 mm Hg need to be referred for RHC varies from institution to institution. We advocate a safe approach with a minimal risk of missing patients with POPH during the preoperative evaluation.
Of course, additional echocardiographic findings and morphological and functional right ventricle parameters (a dilated right ventricle, evidence of right ventricle dysfunction, or septal flattening) that can be seen as indirect signs of significant pulmonary hypertension need to be incorporated into the pretransplant assessment. In order to further reduce the number of false positives, we also tested the impact of adding the presence or absence of right ventricular dilatation on the accuracy of Doppler echocardiography for detecting POPH. Right ventricular dilatation was defined as an RVEDD > 3.3 cm. With the incorporation of this extra variable into the screening test, the number of false positives dramatically dropped, and this resulted in increased specificity. For example, the specificity increased from 82% to 93% with the new cutoff value of 38 mm Hg.
Another serious problem is the de novo development of POPH: patients who do not have POPH at the time of their evaluation but are found to have acquired elevated pulmonary pressures at the time of transplantation. Previous studies have shown that POPH is progressive and can develop within time periods as short as 2 to 3 months. Normal findings during echocardiography do not exclude the possibility of POPH developing in the future. This is the reason that the onset of new symptoms (eg, dyspnea) while being on the waiting list for LT requires control echocardiography. In this study, no patients (0%) with de novo POPH were identified by the 3-month anticipatory screening.
A limitation of this study may be the fact that measurements of the right heart pressure were undertaken in dissimilar settings: as part of the preoperative evaluation by elective catheterization, at the beginning of transplantation (after general anesthesia but before abdominal incision), or both. Here we compensated by screening patients with echocardiography every 3 months and by ensuring that the time interval between the last moment of screening and the moment of LT was at most 3 months.
In summary, POPH occurs in approximately 4.6% of LT candidates. Transthoracic Doppler echocardiography is considered a highly sensitive screening test for the detection of POPH during the preoperative evaluation. The current sPAP cutoff value of 30 mm Hg, at which patients are referred for RHC to confirm or rule out the diagnosis of POPH, leads to a high number of false positives, and this results in a low specificity and a low positive predictive value. In this context, we investigated opportunities to increase the actual cutoff to a value associated with a lower number of false positives and a higher specificity yet guaranteeing a 100% negative predictive value. With an increase in the current cutoff value of 30 mm Hg, the number of patients referred for RHC could be safely reduced. On the basis of the results of this study, we recommend that patients with an sPAP > 38 mm Hg (as measured during echocardiographic screening) be referred for RHC. The incorporation of right ventricular morphology variables (eg, the presence of right ventricular dilatation) is another way of further increasing the specificity of Doppler echocardiography screening for POPH.