Volume 16, Number 4

Review Article

Congenital Porto-Systemic Shunt as the Major Cause of Galactosemia

Nobuo Sakura, MD, PhD; Nobuyuki Mizoguchi, MD; Hiroaki Ono, MD, PhD; Yutaka Nishimura, MD; Kumiko Naito, MD

Address reprint requests to the Department of Pediatrics, Hiroshima University School of Medicine, Kasumi 1-2-3, Minami-ku, Hiroshima, 734-8551, Japan (Dr Sakura).

Please give Japanese version time to download. Click on Japanese characters.

Abstract

The clinical presentation and outcome of 15 neonates with porto-systemic (PS) shunt detected by mass screening (Paigen method) for galactosemia are reviewed. Routine screening for galactosemia initially recognized fourteen patients, but one patient was discovered by the presentation of multiple skin hemangiomas and subsequent re-screening for galactosemia. The majority of patients were excluded by enzyme assay for hereditary galactosemias and diagnosed by ultrasonography and /or angiogram as having PS shunts in the neonatal period. Galactose is effectively extracted from portal blood by liver and PS shunts result in hypergalactosemia. The neonates with PS shunts were separated into 3 types. Two patients had congenital absence of portal vein (CAPV) with PS shunt by ductus venosus Arantii. Four had extrahepatic porto-left renal venous (PRV) shunt. Nine neonates presented with intrahepatic porto-hepatic venous (PV) shunts associated with or without multiple hepatic hemangiomas. At diagnosis, galactosemia was present in all patients, but the degree of hypergalactosemia was varied among 3 types (over 16mg/dl in CAPV, 20~6 mg/dl in PV shunts, 10~8 mg/dl in PRV shunt). The bile acids (TBA) are also absorbed by liver and plasma level of total bile acids (TBA) was increased in all patient (297~117 µ M in CAPV, 138~51 µ M in PV shunts, 50~ 46 µ M in PRV shunt). Hepatic hemangiomas regressed in all, but associated PV shunts persisted in 3 cases after 1 to 12 years of follow up. CAPV maintained in nature and PRV shunt was not closed, and hyperammonemia and hypermanganemia presented as late complications for the patients with persistent PS shunts. These results indicate that PS shunt should be suspected in all neonates with increased levels of galactose and TBA, when hereditary galactosemias were excluded. The prognosis varied among 3 types of PS shunts. Careful assessment must be warranted for complications.

Introduction

It is well known that inherited deficiencies of galactose-metabolizing enzymes resulted in hereditary galactosemias, but increasing evidence has accumulated that porto-systemic (PS) shunts caused hypergalactosemia.^1,2 In our screening programs, the most common cause of persistent hypergalactosemia is PS shunts rather than hereditary galactosemias. To date we have identified 15 cases with PS shunts detected by neonatal mass-screening for galactosemia.^2-5 In this report, we have reviewed the clinical manifestations and outcome of these cases, and insisted that mass-screening for galactosemia by Paigen method is useful for the early diagnosis of PS shunts during the neonatal period.

Patients and Methods

Fifteen patients (6 male and 9 female) were diagnosed as having PS shunts from about 350,000 neonates screened between 1987 and 1999. Ultrasonography or angiography was used to establish the definitive diagnosis.

Blood galactose was estimated by Paigen method and hereditary galactosemias were excluded by assays of galactose metabolizing enzyme activities in red blood cells. Galactokinase activity was assayed by our HPLC method,⁶ galactose-1-phosphate uridyltransferase according to Hansen and Mayes,⁷ and UDP-galactose epimerase by the method of Gitzelmann and Steinmann.⁸

Plasma total bile acids (TBA) and blood ammonia were estimated using commercial kit method. The level of blood manganese was determined by atomic absorption spectrophotometry.

Galactose

The majority of the patients, except one, showed hypergalactosemia at the age of 5 days by the Paigen method for neonatal mass screening. However, an exceptional case was detected by re-screening for galactosemia (galactose; 20 mg/dl) at 25 days of age because of multiple cutaneous hemangiomas, which indicates the presence of hepatic hemangiomas with PV shunt as our acknowledgement.⁴

At the time of diagnosis, 14 patients had increased levels of blood galactose. Their degree was varied with portal systemic flow. In two patients with CAPV, galactose was markedly increased ( over 16 mg/dl).⁴ Nine patients with PV shunt had various degree of hypergalactosemia (over 20~8 mg/dl) based on the degree of PV shunt.^2,5 One neonate showed normal level of galactose (6 mg/dl), while her PV shunts were regressing at the time of her first visit to us. Four patients with PRV shunt had mild increases in their blood galactose concentration (10~8 mg/dl)³ (Table 1). Blood galactose was rapidly normalized with restricted diet, but still elevated after challenge test in the patients with persistent PS shunts.

Total Bile Acids

All of the patients had increased levels of TBA. The concentration of TBA was the highest in the patients with CAPV (over 100 M; normal infant under 30 µ M),⁴ but lowest in those with PRV shunt (around 50 µ M).³ Neonates with PV shunts had varying TBA levels based on their shunt flow (Table 1), and TBA levels decreased to the normal range after PV shunts regressed.⁵

Galactose Metabolizing Enzyme

Galactokinase, galactose-1-phosphate uridyltransferase and UDP-galactose epimerase activities were normal in all the patients.^2-5

Ultrasonography/Angiography

Portal vein imaging studies were performed in all 15 patients. Three distinct imaging patterns were noted: (1) intrahepatic PV shunts with or without hepatic hemangiomas (PV shunts); (2) absence of intrahepatic portal vein (CAPV); and (3) PS shunt between extrahepatic portal vein and left renal vein (PRV shunt).

Nine patients with PV shunts and four patients with PRV shunt were examined using ultrasonography. In 9 former patients, three patients were presented with PV and AV shunts, and six with only PV shunts. Eight of 9 patients with PV shunts had multiple hepatic hemangiomas (Fig 1), and 6 of them had cutaneous hemangiomas. The number of cutaneous hemangiomas ranged from over 100 to only 1. One patient had only intrahepatic PV shunts without hemangiomas.⁵

It was difficult to diagnose PRV shunt by ultrasonography in 2 previous patients with "hypergalactosemia with unknown cause". PRV shunt was not recognized in these patients until the age of 2 years and 6 months (Fig 2). However, two other patients were correctly diagnosed as having PRV shunt by ultrasonography during the neonatal period.³ In all of the patients with PRV shunt, intrahepatic portal vein was normally maintained while extrahepatic portal vein shunted with the left renal vein. Color Doppler ultrasonography confirmed that the venous blood flowed from the portal vein into the left renal vein.

In two patients with CAPV, ultrasonography showed absence of intrahepatic portal vein and a patent ductus venosus Arantii (Fig 3), which were confirmed by angiography.

In a total of 6 cases of both CAPV and PRV shunt, four patients presented with congenital heart anomalies. None of the PV shunts had congenital heart anomalies (Table 2), but they developed heart failure in one case due to AV shunts.²

Ammonia and Manganese

Two patients with CAPV developed hyperammonemia, which was treated by oral lactulose. One patient with PV shunt, which persisted after hepatic hemangiomas regressed, also had hyperammonemia. Hypermanganemia was detected in four patients (Table 3). Basal ganglia lesion were detected by magnetic resonance imaging in two patients, one patient with PV shunt (12 years of age) and the other with CAPV(10 years of age), possibly because of hypermanganemia.

Prognosis

CAPV and PRV shunt have remained and several complications developed (hyperammonemic encephalopathy, hypermanganemic damage of basal ganglion, nodular hepatic hyperplasia).

In 7 patients with hepatic hemangiomas and PV shunts, hepatic hemangiomas regressed as well as did PV shunts until one year of age.⁵ However, in three of them, PV shunts persisted after hemangiomas disappeared and hyperammonemia developed as a complication (Table 3) .

Of three patients with multiple hemangiomas and PV & AV shunts, one patient developed heart failure due to AV shunt, which must be treated by embolization.² In another patient, AV and PV shunts naturally regressed as did hemangiomas.⁵ The other patient is under observation.

Discussion

CAPV and PRV shunts are extremely rare malformation and are difficult to diagnose. These anomalies are often not identified in the neonatal period, and consequently the age at diagnosis varies.^9,10 Infantile hepatic hemangiomas with PV or AV shunts can be diagnosed when they manifest as congestive heart failure or cutaneous hemangiomas. However, our neonates with hepatic portal vein anomalies typically manifested galactosemia, increased TBA, and normal galactose metabolizing enzymes. This clinical triad can make the diagnosis of portal vein anomalies even during the neonatal period.

Congenital PS shunts can be categorized into two types. One is extrahepatic PS shunt with or without CAPV. The shunt vessels either exist or develop from embryonic venous remnants, ductus venosus or spleno-renal venous anastomosis. The other is intrahepatic PS shunt with or without hepatic hemangiomas. These congenital PS shunts can be diagnosed late in life as porto-systemic encephalopathy develops.^10,11 However, the diagnosis of PS shunts is difficult during the neonatal period and often made during preoperative staging of a liver mass, or at autopsy.⁹ Therefore, it is significant to detect PS shunts by neonatal mass screening. Moreover, early diagnosis of PS shunts can prevent early or late complications.

Mass screening for galactosemia by Paigen method can identify various hepatic lesions (PS shunts, hepatitis or bile tract atresia) as well as hereditary galactosemia. Gitzelmann et al. first described hypergalactosemia caused by portal vein anomaly (CAPV),¹ and we reported another type of portal vein anomaly (multiple hepatic hemangiomas and PV & AV shunts) could be detected by hypergalactosemia.² Galactose is effectively extracted from portal blood by liver cells (first pass effect), and portal vein anomalies induce hypergalactosemia because galactose cannot access the liver through the portal vein. In these patients, galactose metabolizing enzyme activities were normal. Therefore, the PS shunts were believed to cause galactosemia. We propose to name this clinical condition "shunt galactosemia".

Galactose concentration was markedly increased in patients with CAPV and mildly increased in patients with PRV shunt. No intrahepatic portal blood flow was detected in patients with CAPV, while portal blood flow was maintained but PS shunt flow existed from the extrahepatic portal vein to the left renal vein in patients with PRV shunt. The degree of hypergalactosemia was dependent on the flow through the PS shunt. Furthermore, the blood galactose concentration varied in neonates with hemangiomas in proportion to the PV shunt flow.

The presence of bile acids is limited within the enterohepatic circulation. The bile acids are also absorbed by liver via first pass through the portal blood circulation.⁴ All of our patients had increased levels of TBA. The presence of both hypergalactosemia and increased levels of TBA are highly suggestive of PS shunts.

TBA levels must be examined in all of hypergalactosemic neonates detected by Paigen method. However, some cases transiently presented with both manifestations, because ductus venosus was closed in the infantile period, not in the neonatal period.^12,13 In fact, our screening program has identified several cases with transient increases of galactose and TBA, both of which were normalized within the age of one month or two.¹² Galactose was effectively normalized with restricted diet, but the patients with persistent PS shunts maintained increased levels of TBA even during the diet therapy. The assessment of TBA level is appropriate to follow up the patients with PS shunts.

The presence of cutaneous hemangiomas was suggestive for hemangiomas with PV shunts. However, not all the cases manifested as multiple cutaneous hemangiomas.^5,9 The number of cutaneous hemangiomas was varied among the cases (over 100 to 0). In the absence of cutaneous hemangiomas, hypergalactosemia is useful for the diagnosis.²

Heart failure or congenital heart anomaly is also suggestive of PS shunts, but not all of the neonates with PS shunts had congestive heart failure. Hypergalactosemia was prevalent in neonatal period even if the neonate did not have an overt heart failure. Congenital heart anomalies can be associated with CAPV or congenital PRV shunt,¹⁰ and the concomitant existence of hypergalactosemia and congenital heart anomalies strongly suggest congenital PRV shunt or CAPV.

The prognosis is different among 3 types of PS shunts. The patients with CAPV developed several complications, such as hyperammonemia , hypermanganemia, and hepatic nodular hyperplasia.^10,15 Ammonia and manganese are also effectively extracted from portal blood by liver.¹⁴ PS shunts result in increases of ammonia and manganese. Drugs prevent Hyperammonemia, but a restricted diet is the only effective therapy for hypermanganemia. In patients with PRV shunt, fortunately galactose is mildly increased because shunt flow is limited. However, it should be reserved for children with PRV shunt enough to evaluate hypermanganemia in persistent association of small PS shunt.

Hepatic hemangiomas, like cutaneous tumors, spontaneously regressed late in infancy, but PV shunts have resided even after hemangiomas are involuted.⁵ In the case of persistent PV shunts, hyperammonemia and hypermanganemia must be carefully assessed. AV shunts are associated with hepatic hemangiomas in some cases, and must be clarified by ultrasonography. Not all the infants with AV shunts associated with hemangiomas had congestive heart failure, but hepatic artery embolization could control life-threatening congestive heart failure in one case of AV shunts.

During the same period of this study, we identified 8 patients with hereditary galactosemias (one heterozygous for galactokinase deficiency and 7 with UDP-galactose epimerase deficiency) and 2 patients with liver disorders ( hepatitis and bile tract atresia ). Contrary to our expectation, the incidence of hereditary galactosemia was lower than that of hypergalactosemia associated with PS shunts. In Hiroshima prefecture, the major cause of galactosemia is PS shunts. The number of cases has accumulated in other prefectures and a united view has obtained that increased levels of galactose and TBA can provide a diagnostic clue to the presence of a PS shunt.

1. Gitzelmann R, Arbenz UV, Willi UV. Hypergalactosemia and portosystemic encephalopathy due to persistence of ductus venosus Arantii. Eur J Pediatr. 1992;151:564-568. [MEDLINE]
2. Matsumoto T, Okano R, Sakura N, et al. Hypergalactosemia in a patient with portal-hepatic venous and hepatic arterio-venous shunts detected by neonatal screening. Eur J Pediatr. 1993;152:990-992. [MEDLINE]
3. Mizoguchi N, Sakura N, Ono H, Naito K, Hamakawa M. Congenital porto-left renal venous shunt as a cause of galactosemia. Int Pediatr. 2001;16. [MEDLINE]
4. Sakura N, Mizoguchi N, Eguchi T, et al. Elevated plasma bile acids in hypergalactosemic neonates: a diagnostic clue to portosystemic shunts. Eur J Pediatr. 1997;156:716-718. [MEDLINE]
5. Ono H, Mawatari H, Mizoguchi N, Eguchi T, Sakura N. Clinical features and outcome of eight patients with intrahepatic porto-venous shunts detected in neonatal screening for galactosemia. Acta Paediatr. 1998;87:631-634. [MEDLINE]
6. Mizoguchi N, Eguchi T, Sakura N, Ueda K. Erythrocyte galactokinase assay with high performance liquid chromatography. Clin Chim Acta. 1993;216:145-151. [MEDLINE]
7. Hansen RG, Mayes JS. Galactose-1-phosphate uridyltransferase. Methods Enzymol. 1966;9:713-718.
8. Gitzelmann R, Steinmann B. Uridine diphosphate galactose 4-epimerase deficiency. Helv Paediatr Acta. 1973;28:497-510.
9. Boon LM, Burrows PE, Paltiel HJ, et al. Hepatic vascular anomalies in infancy: A twenty-seven-year experience. J Pediatr. 1996;129:346-354. [MEDLINE]
10. Guariso G, Fiorio S, Altavilla G, et al. Congenital absence of the portal vein associated with focal nodular hyperplasia of the liver and cystic dysplasia of the kidney. Eur J Pediatr. 1998;157:287-290. [FULL TEXT]
11. Raskin NH, Price JB, Fishman RA. Portal-systemic encephalopathy due to congenital intrahepatic shunts. N Engl J Med. 1964;270:225-229.
12. Ono H, Mawatari H, Mizoguchi N, Eguchi T, Sakura N, Hamakawa M. Transient galactosemia detected by neonatal mass screening. Pediatr Inter. 1999;41:281-284. [MEDLINE]
13. Fugelseth D, Lindemann R, Liestol K, Kiserud T, Langslet A. Ultrasonographic study of ductus venosus in healthy neonates. Arch Dis Childh. 1997;77:F131- F134. [MEDLINE] [FULL-TEXT]
14. Nolte W, Wiltfang J, Schindler CG, et al. Bright basal ganglia in T1-weighted magnetic resonance images are frequent in patients with portal vein thrombosis without liver cirrhosis and not suggestive of hepatic encephalopathy. Hepatol. 1998;29:443-449. [MEDLINE]
15. Uchino T, Matsuda I, Endo F. The long-term prognosis of congenital portosystemic venous shunt. J Pediatr. 1999;136:254-256. [MEDLINE]