ARTICLE |
a Divisions of Gastroenterology, Hepatology, and Nutrition
b Respiratory Medicine
c Cardiology
d Nuclear Medicine, Hospital for Sick Children, Toronto, Ontario, Canada
| ABSTRACT |
|---|
|
|
|---|
METHODS. Pulse oximetry was undertaken in children with liver disease, and those with oxygen saturation
97%, those with cirrhosis, and those with clinically severe portal hypertension from other causes underwent contrast-enhanced echocardiography for detection of intrapulmonary vascular dilations. Patients with intrapulmonary vascular dilation underwent arterial blood gas analysis and technetium-99m–labeled macroaggregated albumin scan.
RESULTS. Oxygen saturation was measured in 301 children and was
97% in 8. These 8 and an additional 18 patients with cirrhosis or portal hypertension underwent contrast-enhanced echocardiography. Seven (27%) patients had intrapulmonary vascular dilation detected by contrast-enhanced echocardiography; 2 of these patients had abnormal arterial blood gas analysis and thus met diagnostic criteria for hepatopulmonary syndrome (representing 8% of patients with cirrhosis or severe portal hypertension). Both patients with hepatopulmonary syndrome had abnormal pulse oximetry. Technetium-99m–labeled macroaggregated albumin scans for 6 patients showed a median 6.5% (range: 4%–12%) tracer uptake outside the lungs.
CONCLUSIONS. Hepatopulmonary syndrome occurs in an important minority of children with cirrhosis or severe portal hypertension. Additional studies should be undertaken to determine the importance of intrapulmonary vascular dilation without hepatopulmonary syndrome and the impact of hepatopulmonary syndrome on the outcomes of affected children.
Key Words: hepatopulmonary syndrome portal hypertension liver disease contrast-enhanced echocardiography
Abbreviations: HPS—hepatopulmonary syndrome IPVD—intrapulmonary vascular dilatations PHT—portal hypertension ABG—arterial blood gas CEE—contrast-enhanced echocardiography 99mTc-MAA—technetium-99m–labeled macroaggregated albumin A-a gradient—alveolar-arterial oxygen gradient AIH—autoimmune hepatitis
The term "hepatopulmonary syndrome" (HPS) was coined in 1990 to describe the association of liver disease, hypoxemia, and intrapulmonary vascular dilations (IPVD).1 The prevalence of HPS in adults with cirrhosis is reported to range from 4% to 29%.2–7 In children, 2 retrospective studies described the prevalence of HPS to be 9% to 20% for children with biliary atresia and 0.5% for children with portal vein thrombosis.8,9 Other case reports confirmed the occurrence of HPS in children with chronic liver disease, portal hypertension (PHT), and congenital anomalies of the portal vein.10–29 Information on the impact of HPS on clinical outcome is scarce; some children have shown a deterioration in arterial oxygenation during medium-term follow-up, and 25% to 46% have died.8,9,12 The prevalence of HPS derived from systematic and prospective testing of a cohort of children with liver disease has not been previously reported to our knowledge.
The diagnosis of HPS in a child with liver disease is established by demonstration of hypoxemia or elevated alveolar-arterial oxygen gradient on arterial blood gas (ABG) analysis and the presence of intrapulmonary shunting using contrast-enhanced echocardiography (CEE) or technetium-99m–labeled macroaggregated albumin (99mTc-MAA) perfusion scan.5,30 Routine use of ABG analysis to screen children for HPS is problematic because of the invasive nature of this test; however, pulse oximetry is well accepted by children, and an oxygen saturation of
97% measured by pulse oximetry identifies hypoxemia proved by arterial blood gas analysis (PaO2 <70 mmHg) in adults with cirrhosis, with sensitivity of up to 100% and specificity of
65%.31–33 Although pulse oximetry is not a perfect replacement for ABG analysis, its use to screen for HPS in children would potentially identify all children who have cirrhosis and meet the hypoxemia diagnostic criterion for HPS and would significantly reduce the number of children who do not have HPS and might otherwise be subjected to more invasive tests. Because little is known about the prevalence and impact of HPS in children, we undertook this prospective study to identify the prevalence of HPS in children with liver disease.
| METHODS |
|---|
|
|
|---|
Procedures
All children were screened using pulse oximetry, which was performed according to a standardized procedure with the child breathing room air in a sitting or standing position. The pulse oximetry threshold for the identification of hypoxemia was
97%.31
Children with pulse oximetry
97% and all children with cirrhosis or noncirrhotic severe PHT underwent CEE to identify evidence of IPVD. CEE was performed by 1 operator who was blinded to the clinical and oximetry data. After confirmation of normal cardiac anatomy, 10 mL of agitated normal saline solution was rapidly injected into the patient's venous cannula and echocardiographic images were digitally recorded for analysis. The times at which contrast appeared in the right ventricle and in the left ventricle and the number of cardiac cycles between these times were recorded. An abnormal CEE scan was defined by the appearance of contrast in the left ventricle >3 cardiac cycles after its appearance in the right ventricle (appearance after 20 cardiac cycles was considered normal).5
Children with an abnormal CEE scan then had a 99mTc-MAA perfusion scan. With the child in the upright or standing position, 0.05 mCi/kg 99mTc-MAA was injected into a peripheral venous cannula. After 20 minutes, quantitative whole-body imaging was performed with a dual-head
camera. Quantitative evaluation of relative uptake was determined using regions of interest drawn over the brain, kidney, soft tissues, and lungs. The shunt index (expressed as a ratio of the uptake in the lungs to the uptake in the total body minus kidney activity) was measured. An abnormal 99mTc-MAA scan was defined as lung uptake <93% of the total body uptake.30
ABG analysis was performed in children with an abnormal CEE. With the child sitting and breathing room air, an arterial blood sample was obtained to measure the PaO2 and to calculate the alveolar-arterial oxygen gradient (A-a gradient). An abnormal arterial blood oxygen was defined as a PaO2 <70 mmHg or an A-a gradient of >15 mmHg.34
A chest radiograph was obtained for patients with pulse oximetry
97% and in patients with abnormal CEE. Respiratory consultation was undertaken when there were respiratory symptoms or signs or when the child had an abnormal CEE or an abnormal chest radiograph.
Children were considered to have IPVD when the CEE was reported as abnormal. In cases in which the CEE was inconclusive, the 99mTc-MAA scan was used to clarify the presence of IPVD. HPS was considered present in a child when there was evidence of IPVD and a PaO2 <70 mmHg or an A-a gradient of >15 mmHg. The algorithm for testing used in this study is shown in Fig 1.
|
| RESULTS |
|---|
|
|
|---|
|
97% in 8 children (6 at 97%, 1 at 94%, and 1 at 85%): 5 with cirrhosis, 2 with noncirrhotic severe PHT, and 1 with apparently mild chronic hepatitis B liver disease. None of these 8 children had respiratory symptoms. Six of them had a normal chest radiograph; minor changes of prominent pulmonary vasculature were noted in 2 children.
According to the study protocol, CEE was required in 28 children (9%, 13 boys); the test was refused by 2 of these patients, 1 with cirrhotic biliary atresia and 1 with cirrhotic AIH. Indications for CEE were the presence of cirrhosis in 21 children (5 also with low oxygen saturation), significant PHT in 4 children (3 with nodular regenerative hyperplasia after chemotherapy for leukemia and 1 with congenital hepatic fibrosis), and low oxygen saturation in 1 child with apparently mild liver disease as a result of chronic hepatitis B. Among the 21 children with cirrhosis, 13 had a classification of Child-Pugh class B cirrhosis (none had class C), 7 had AIH and/or primary sclerosing cholangitis, 6 had biliary atresia, 2 had had liver transplantation (1 de novo AIH, 1 chronic rejection), 2 had Alagille syndrome, 2 had cryptogenic cirrhosis, 1 had Wilson disease, and 1 had
-1 antitrypsin deficiency.
CEE was abnormal in 7 (27%) of 26 patients, 6 with Child-Pugh class B and 1 with class A cirrhosis (Table 2). ABG analysis in these 7 patients with abnormal CEE showed abnormalities in 2 patients; 1 child had a PaO2 of 80 mmHg and an A-a gradient of 25 mmHg, and the second child was hypoxemic with PaO2 of 49 and an A-a gradient of 72 mmHg. Both of these patients had shown abnormal pulse oximetry (97% and 85%, respectively). Thus, 2 patients had HPS, representing 8% of the patients with cirrhosis or other severe liver disease with significant PHT. The other 6 children with abnormal pulse oximetry had normal CEE and otherwise normal respiratory investigations; their oxygen saturation values were normal at subsequent follow-up assessments; therefore, no diagnosis was identified for the initial low oxygen saturation levels.
|
Of the 305 patients assessed, 39 had already undergone liver transplantation; 28 seemed well, 7 had chronic rejection, 1 had de novo AIH, 1 had PHT as a result of parenchymal disease and arteriovenous fistula, and 2 had recurrence of their original disease (parenteral nutrition-associated cholestasis and primary sclerosing cholangitis). One patient with severe chronic rejection and PHT (Child-Pugh class B) had a low oxygen saturation value of 85% and a positive CEE result; ABG analysis was abnormal, and HPS was therefore diagnosed. This patient was relisted for liver transplantation. One other posttransplantation patient with de novo autoimmune hepatitis, severe liver disease, and PHT (Child-Pugh class B) had normal pulse oximetry but a positive CEE result; because ABG analysis was normal, a diagnosis of isolated IPVD was reached. The remaining posttransplantation patients did not meet the criteria for undergoing CEE, apart from 1 with severe chronic rejection and PHT (Child-Pugh class B), who refused to undergo CEE or other testing.
| DISCUSSION |
|---|
|
|
|---|
The prevalence of HPS that we have measured in children with cirrhosis or severe PHT lies within the range of 4% to 29% reported by previous studies of adults with cirrhosis. Using a PaO2 of
80 mmHg as a marker for HPS (as suggested by Schenk et al2) would have correctly identified both of our affected patients; however, the prevalence of HPS measured in this study may underestimate the true prevalence because of the lack of ABG analysis in all of the children recruited. Although pulse oximetry has acceptable accuracy in the diagnosis of hypoxemia, it will not detect patients whose early oxygenation abnormality consists of a high A-a gradient but not a low arterial PaO2. According to a proposed grading scale for severity of HPS, pulse oximetry would fail to identify all patients with mild HPS (positive CEE result, A-a gradient
15 mmHg, and PaO2
80 mmHg) and some with moderate HPS (positive CEE result, A-a gradient
15 mmHg, and PaO2
60 and <80 mmHg).36 For these reasons, adult guidelines recommend screening at-risk groups using ABG analysis initially36; however, we expected that the majority of our pediatric patients would not readily accept the arterial puncture required for this approach.
In the absence of a gold standard test, a combination of clinical criteria are used to make a diagnosis of HPS. Although it is clear from previous studies that the use of different diagnostic thresholds for these clinical criteria has a significant effect on the measured prevalence of HPS, consensus has now largely been reached for appropriate diagnostic criteria36; however, there is concern that the diagnostic precision may be suboptimal and that HPS may be only 1 of a number of possible causes of the simultaneous presence of liver disease, intrapulmonary shunting, and hypoxemia.37 Among adults with cirrhosis, hypoxemia is seen in approximately one third of patients,38 and coexisting pulmonary diseases that may contribute to hypoxemia and that may be associated with intrapulmonary shunting are common, such as chronic obstructive airway disease. Chronic lung diseases are far less common among children with liver disease; however, the implications for the diagnosis of HPS must be carefully considered when children are identified with chronic lung disease, such as the chronic lung disease that persists after preterm birth and in children with cystic fibrosis.
Among our patients in this study, clinical respiratory assessment and chest radiograph did not suggest any alternative respiratory diagnosis in children with hypoxemia, isolated IPVD, or HPS. With the exception of 1 child with HPS, the abnormal pulse oximetry measurements in the children in our study suggested only minimally abnormal oxygenation, and many of these children had returned to normal pulse oximetry values on follow-up assessments. Four of 5 patients with IPVD and both patients with HPS had bilateral interstitial markings on chest radiograph, which have previously been described in association with IPVD.38 No pleural effusions or other radiographic abnormalities were noted.
We found a proportion of children with IPVD without ABG abnormalities, similar to previous adult studies.5 The significance of this has yet to be determined in adults and will be addressed in these and other children by future follow-up studies. Small case series show that the abnormal CEE findings tend to persist and are therefore unlikely to represent false-positive scans.39,40 The effect on morbidity and mortality is unknown.
We showed that the results of 99mTc-MAA scans correlated poorly with CEE abnormality and ABG analysis; however, previous studies showed that 99mTc-MAA scans correlate well with PaO2 and A-a gradient and can be used as a method for quantitative assessment of HPS severity and to predict outcome after liver transplantation.41,42 The suggestion that CEE has greater sensitivity for the identification of IPVD compared with 99mTc-MAA scan43 must therefore be further studied in children, although the absence of radiation exposure has led many centers to adopt CEE as their preferred diagnostic test for IPVD.
Screening for HPS and assessment of its severity have been recommended among adults who have cirrhosis and may be candidates for liver transplantation.36 This recommendation arises from the influence that HPS has on the likelihood of survival both before and after liver transplantation. In a study of 111 adults with cirrhosis, the median survival was 11 months for those with HPS compared with 41 months for those without HPS, independent of other predictors, such as Child-Pugh class and age.6 HPS improves in the majority of patients after liver transplantation, although its negative impact measured in adults who undergo transplantation includes an increase in both postoperative complications and mortality.44
Similar recommendations for screening children with HPS must await additional studies of the natural history of HPS in children and a clearer understanding of the degree to which its presence should influence listing priority. The limited information on the impact of HPS on clinical outcome in children does not include comparison with matched control subjects without HPS.8,9,12 Several pediatric cases that demonstrated reversal of HPS after liver transplantation have been reported, although these reports raise the suggestion of more frequent postoperative complications.8,9,12,19,21 Successful management of HPS in children has been reported with both transjugular intrahepatic portosystemic shunt and inhaled nitric oxide18,24 and after surgical correction of congenitally abnormal portal or hepatic venous drainage.10,11
| CONCLUSIONS |
|---|
|
|
|---|
| ACKNOWLEDGMENTS |
|---|
| FOOTNOTES |
|---|
Address correspondence to Simon C. Ling, MB ChB, Division of Gastroenterology, Hepatology and Nutrition, Hospital for Sick Children, 555 University Ave, Toronto, Ontario M5G 1X8, Canada. E-mail: simon.ling{at}sickkids.ca
The authors have indicated they have no financial relationships relevant to this article to disclose.
| REFERENCES |
|---|
|
|
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||