Wilson's disease in children: 37-Year experience and revised King's score for liver transplantation



Wilson's disease (WD) is a rare liver-based disorder of copper metabolism. Prognostic criteria described by our group in 1986 to predict death without transplantation have not been universally validated. The clinical features of 88 children were reviewed, retrospectively in 74 and prospectively in 14. Data from the retrospectively recruited patients that died or survived on long-term chelation were used to evaluate the validity of our old scoring system and to devise a new prognostic index, then assessed in the 14 prospectively recruited patients. Using the old scoring system, 5 children scoring ≥ 7, the cutoff value for death without transplantation, survived, whereas 4 scoring ≤ 7 died (sensitivity 87% and specificity 90%). A new index based on serum bilirubin, international normalized ratio, aspartate aminotransferase (AST), and white cell count (WCC) at presentation identified a cutoff score of 11 for death and proved to be 93% sensitive and 98% specific, with a positive predictive value of 88%. When the new index was evaluated prospectively in 14 patients, it predicted the need for transplantation in only the 4 who required it, although 1 child with a score of 11 survived on medical treatment. In conclusion, the new Wilson Index is more sensitive and specific in predicting mortality without transplantation than the old scoring system, but needs to be validated in a larger number of patients. (Liver Transpl 2005;11:441–448.)

Wilson's disease (WD) is an autosomal recessive disorder of copper metabolism with a prevalence of 1 in 30,000 in the general population.1 The gene defect has been identified with mutations in the ATP7B gene located on the long arm of chromosome 13. The clinical manifestations are secondary to accumulation of copper in various organs, the most common presentations being hepatic or neuropsychiatric. In children, the disease usually presents after 3 years of age with either incidental discovery of abnormal liver function tests or as chronic liver disease and rarely as acute hepatic failure. Copper chelating agents are effective except in patients with very advanced disease and decompensated cirrhosis or with fulminant liver failure where liver transplantation appears to be the only effective mode of treatment. While it is not difficult to decide for transplantation in the presence of encephalopathy, a complication generally lethal in WD, the criteria for transplantation in patients presenting with acute liver failure (ALF) without encephalopathy are more controversial. In 1986 Nazer et al.,2 from our unit, in a combined pediatric and adult population of WD devised a scoring system to predict the outcome of patients presenting with liver decompensation. Over the years, this scoring system has been used by other centers with variable results.3 In the present study, we have retrospectively reviewed all children with WD treated in the Paediatric Liver Service at King's College Hospital between 1967 and 2000 to critically analyze the validity of the Nazer2 scoring system. The revised score was then tested prospectively on children presenting with WD between 2001 and 2003.


WD, Wilson's Disease; ALF, acute liver failure; INR, international normalized ratio; AST, aspartate aminotransferase; WCC, white cell count.

Patients and Methods

The medical records of 74 children with WD from 62 families diagnosed and treated in our hospital from January 1967 to December 2000 were retrospectively reviewed. A total of 57 patients were referred because of liver disease, whereas 17 were asymptomatic siblings diagnosed during family screening. The diagnosis of WD in the probands was based on the presence of liver disease and at least 2 of the following criteria:

  • 1Positive family history
  • 2Low serum ceruloplasmin (<0.2 g/L)
  • 3Elevated liver copper (>250 mg/g dry weight)
  • 4Presence of Kayser-Fleischer rings
  • 5Elevated baseline 24-hour urinary copper excretion (>1 μmol/24 hours)
  • 6Elevated 24-hour urinary copper excretion following administration of 2,500-mg doses of penicillamine (>25 μmol/24 hours)4
  • 7Coombs' negative hemolytic anemia

Genetic mutational analysis for WD was only available in 6 patients and was not used for diagnosis or for analysis. Other causes of liver disease such as autoimmune liver disease, chronic viral hepatitis, α1-anti-trypsin deficiency, and other metabolic conditions were excluded by appropriate investigations. ALF was defined as the presence of coagulopathy (international normalized ratio [INR] > 2) with acute presentation of signs and symptoms of liver disease, with or without encephalopathy.

The 57 symptomatic children were divided in 3 groups:

  • Group 1: Children on long-term chelation (n = 32)

  • Group 2: Patients who died soon after diagnosis (n = 15)

  • Group 3: Liver transplant recipients (n = 10)

Statistical Analysis

Statistical analysis was performed with NCSS 2000 (Number Cruncher Statistical Software Systems, Kaysville, UT) and the CIA program (Confidence Interval Analysis 2000, Southampton, UK). Comparisons between outcomes in the symptomatic groups were made using 1-way analysis of variance and the Kruskal-Wallis multiple comparisons test. Categorical data were analyzed using chi-square or Fisher's exact test.

In order to develop a prognostic index, the data from children in Groups 1 and 2 were used (n = 47). To identify significant differences between those who were alive and those who died, univariate analysis was performed on continuous data with 1-way analysis of variance, and contingency tables were used for discrete variables. The continuous variables were entered into receiver-operating curve analysis and the area under the receiver-operating curve was calculated to evaluate the prediction accuracy of the previously determined best predictors for outcome. The closer the area under the curve is to 1.0, the better the variable. The aim was to achieve an area under the curve greater than 0.90 with as small as possible standard error, but all those higher than 0.70 were chosen for further examination. To calculate cutoff values, the highest sensitivity and specificity were selected (using the largest likelihood ratio would bias towards the positive side of the equation). To find out which variables were associated between a categorical dependent variable (outcome) and the set of independent (explanatory) variables, logistic regression was used. The parameters evaluated as predictors were data obtained on admission: demographic information; standard laboratory tests including the serum levels of copper, zinc, ceruloplasmin, immunoglobulins, and autoantibodies; and categorical data such as type of presentation, presenting features, and outcome. A 2-proportions analysis was conducted to determine the sensitivity and specificity of the new index using prospectively collected data from children presenting with WD between January 2001 and September 2003.


Children With Symptomatic Liver Disease

Presenting Clinical Features

The median age at diagnosis of the 57 patients with liver disease was 11.9 (range, 5.9–17.9) years, and 35 were male. A total of 27 (47.3%) presented with ALF, 17 of whom had encephalopathy. Jaundice at presentation was seen in 41, abdominal pain in 24, ascites in 19, and hepatosplenomegaly was present in 18. Other gastrointestinal and neurological symptoms are presented in Table 1. Further presenting features were: peripheral edema in 13, dark urine in 10, fever in 10, epistaxis in 6, pruritus in 6, and gynecomastia and joint pain in 4 patients each.

Table 1. Retrospective Study—Gastrointestinal and Neurological Features at Presentation in 57 Symptomatic Children With WD
Abdominal distension16Lethargy22
Vomiting13Behavioral changes5
Pale stool8Headaches4
  Developmental delay1

Laboratory Findings

Demographic, clinical, and diagnostic laboratory data at presentation are shown in Table 2. Of the 57 symptomatic children, 17 (30%) had Coombs' negative hemolytic anemia, 44 (77.2%) low ceruloplasmin, 54 (95%) elevated urinary copper excretion, 20 (69%) of 29 positive penicillamine challenge (i.e., urinary copper excretion >25 μmol/L after administration of 500 mg of penicillamine every 12 hours during the 24-hour urine collection),4 and 14 (82.4%) of 17 had elevated liver copper content. In the 10 children who underwent transplantation, the explanted liver histology showed massive necrosis in 2 and cirrhosis with widespread necrosis in 8.

Table 2. Retrospective Study—Clinical and Diagnostic Data of the 57 Symptomatic Patients With Wilson Disease at Presentation (Median [Range])
 Group 1 (n = 32)Group 2 (n = 15)Group 3 (n = 10)
  1. * P = 0.03 vs. Group 2.

Age (years)11.8 [5.9–16.2]11.9 [8.2–17.9]13.2 [8.5–16.3]
Acute liver failure (number)41310
KF rings (number)2086
Ceruloplasmin (g/L)0.09 [0–0.60]0.16 [0.06–0.82]0.09 [0.05–0.20]
Urine copper (μmol/24 hours)33.7 [7.1–122]38.4 [8.3–51]25 [9.3–192]
Liver copper (μg/g dry wt)730 [105.8–2,358]263.6 [28–825]932 [5–1,100]
Penicillamine challenge (μmol/24 hours)35.9 [1.2–381.6]32.4 [30.8–60.1]Not done
Coomb's negative haemolytic anemia593

Other laboratory indices at presentation are presented in Table 3. There was no significant difference in serum immunoglobulin G (IgG) and IgA levels between the groups, but children who died had a significantly higher serum IgM concentration (median, 4.95; range, 0.8–7.05 g/L) than those who underwent transplantation and those on long-term chelation (median, 1.65 [range, 1.03–1.8] and 2.05 [0.07–6.6] g/L, respectively; P < 0.05 for all).

Table 3. Retrospective Study—Laboratory Indices at Presentation (Median [Range]) of the 57 Symptomatic Children With Wilson Disease
 Group 1 (n = 32)Group 2 (n = 15)Group 3 (n = 10)
  • *

    P < 0.05 vs. Groups 2 & 1.

  • P < 0.05 vs. Group 2.

  • P < 0.05 vs. Groups 3 & 1.

INR1.6 [0.9–4.3]*3.4 [1.5–8]3.45 [1.63–15]
AST (IU/L)120 [9–634]233 [96–1,560]353 [77–694]
Total Bilirubin (μmol/L)23.5 [3–807]*480 [29–928]453 [112–1,053]
GGT (IU/L)102.5 [3–384]82 [21–241]78 [18–133]
ALP (IU/L)331 [30–931]141 [10–440]201 [5–924]
Albumin (g/L)29 [19–47]23 [15–33]27.5 [18–34]
Hb (g/L)11.1 [7.6–13.4]9.4 [5.6–14]10.6 [8.3–12.9]
WCC (109/L)5.7 [2.2–15.3]*10.2 [6.9–18]9.6 [3.85–30.3]
Platelets (109/L)119.5 [12–316]96 [65–300]106 [34–271]
IgG17.45 [2.7–41.1]18.25 [10.7–25]12.4 [11–14.7]
IgA3.6 [0.07–8.15]5.2 [2.59–7.19]2.7 [0.89–3.72]
IgM2.05 [0.07–6.6]4.95 [0.8–7.05]1.65 [1.03–1.8]
ALP/SBR11.0 [0.37–115.7]*0.29 [0.12–4.9]0.36 [0.09–8.25]

Autoantibodies were positive in 2 children who died (anti-smooth muscle antibody, 1 of 10 and 1 of 40) and in 9 children alive on long-term chelation (1 = gastric parietal cell (1 of 10), 1 = antinuclear antibody (antinuclear antibody, 1 of 10), 5 = anti-smooth muscle antibody (median, 1 of 10 [range, 1 of 10 to 1 of 40]) and 2 = antinuclear antibody (1 of 10 and 1 of 40) and anti-smooth muscle antibody (both 1 of 10) positive.


In Group 1 (children on long-term chelation, follow-up median 11.78 [range, 1.45–34.2] years), 12 received penicillamine (10 mg/kg twice daily to a maximum of 500 mg) and 20 penicillamine and zinc sulfate (250 mg once daily), administered 6 hours apart. Minor side effects occurred in 2 children (rash and drowsiness), which abated when the dose was altered. A total of 7 children were converted to trientine (10 mg/kg twice daily) due to bone marrow suppression in 3 and hematuria in 2, while for 1 girl who was attempting to become pregnant, trientine was thought to be less teratogenic than penicillamine. In children who were intolerant to penicillamine, the drug was reduced to 5 mg/kg twice daily; if symptoms persisted the drug was replaced with trientine. A total of 1 child developed a rash and was treated with zinc sulfate monotherapy for 5 years, but a repeat liver biopsy revealed increased liver copper and ongoing liver damage, i.e., inflammatory and fibrotic activity, hence trientine was added.

In Group 2 (children who died soon after presentation without transplantation), 4 died before chelation therapy could be started, 9 received penicillamine only, and 2 received penicillamine and zinc sulfate. Duration of chelation was a median of 18 (range, 1–50) days. A total of 1 child developed a rash that resolved without discontinuing treatment, while 1 child developed hemolytic anemia after starting penicillamine and was converted to trientine. Of the 10 children in Group 3 (patients who underwent transplantation), only 5 received chelation before transplantation (penicillamine in 3 and penicillamine and zinc sulfate in 2) for a median of 7 (range, 1–20) days.


In Group 1, tests of liver function normalized in 20 of 32 children at variable times after the onset of chelation therapy: INR after a median 1.8 (range, 0–12.2) years; aspartate aminotransferase (AST) after a median 0.97 (range, 0–9) years; and bilirubin at a median of 0 (range, 0–2.3) years. A total of 2 children who had initially responded to chelation died of sepsis and encephalopathy 70 and 114 days after diagnosis. Both were nonadherent to chelation therapy. None of the children on long-term chelation in the present study, 20 of whom were on penicillamine and zinc, has required liver transplantation over a median follow-up period of 11.8 (range, 1.45–34.2) years. Other associated complications were asymptomatic portal hypertension in 1, bile duct calculus in 1, bleeding duodenal ulcer and cholecystectomy for gallstone in 1, growth delay in 1, ulcerative colitis in 1, and recurrent epistaxis in 1. A total of 2 girls have become pregnant and have delivered healthy babies.

In Group 2, median interval to death from presentation was of 10 (range, 0–50) days after admission. All 15 presented with ALF, 13 with encephalopathy and 2 without. A total of 13 children presented before liver transplantation became routinely available in the United Kingdom, and the remaining 2 died waiting for an organ to become available. A total of 8 children died of renal failure and persisting hypotension, 4 of gastrointestinal bleeding, 2 of respiratory failure secondary to pulmonary hemorrhage, and 1 of bacterial peritonitis.

All the children in Group 3 presented with ALF, 6 with encephalopathy. Transplantation occurred at a median of 3.5 (range, 0–20) days following admission. A total of 2 children died 54 and 58 days after transplantation, of multiorgan failure and fungal septicemia, respectively. A total of 2 children underwent retransplantation 7 months and 8 years after the first transplant for chronic rejection. A total of 1 child developed a biliary stricture and hepatitis B infection 1 month after transplantation but has normal graft function.

Asymptomatic Siblings

The 17 asymptomatic siblings (median age at diagnosis 10.4 [range, 2.6–16.9] years; 10 male) were from 13 families. AST was elevated in 11 (65%) and gamma-glutamyl transferase in 6 (35%), whereas alkaline phosphatase, albumin, and INR were normal in all. Ceruloplasmin was low in 15 (88.2%), 24-hour urine copper excretion >1 μmol/24 hours in 12 (70.6%), positive response to penicillamine challenge in 5 of 9 (55.6%), and elevated liver copper in 7 of 8 (87.5%). A total of 10 children were treated with penicillamine and 7 with penicillamine and zinc. A total of 3 children experienced rash, headaches, and stammering during the first year of treatment, which resolved following a reduction of the dose, whereas 2 were converted to trientine because of nausea and musculoskeletal pain. Liver function normalized in 15 at a median interval of 283 days (range, 35 days to 6.7 years) after diagnosis. At last follow-up 14 years after diagnosis, 1 child still has elevated AST (133 IU/L). During follow-up, 6 children developed clinical symptoms: speech problems in 2, ulcerative colitis in 1, and neuromuscular dysfunction in 1 child who was noncompliant. A total of 2 children required therapy for substance abuse. A total of 1 patient has had a successful pregnancy.

Development of a New Prognostic Score

After 1986, indication for transplantation for WD in our unit was determined according to the original scoring system,2 modified to replace prothrombin time with INR (Table 4).5

Table 4. Modified Nazer's Wilson Disease Outcome Score*
 Bilirubin (μmol/L)AST (IU/L)INR
  • *

    PT converted to INR. ≥7 = increased risk of mortality without liver transplant.5


A score of 7 and above suggested an high risk of mortality if liver transplantation was not performed. When the score was tested in those children from the current series who were seen before transplantation became available or who were listed for transplantation but did not undergo transplantation because of lack of a donor organ, 5 children with a score ≥7 survived and 4 with a score <7 died. The sensitivity and specificity of the Nazer scoring system was 87% and 90%, with a positive predictive value of 72%. In view of this, the data from the symptomatic children who survived with chelation treatment (Group 1) and those who died without liver transplantation (Group 2) were reviewed and analyzed (as described in Patients and Methods) to find a more sensitive predictor of mortality. Following calculation with receiver-operating characteristic curves (Fig. 1) and multivariate analysis, a combination of AST, albumin, bilirubin, INR, and white cell count (WCC) was found to be a better predictor of mortality than the original score (Table 5). A score ≥11 using the new Wilson Index (Table 6) had a sensitivity and specificity of 93% and 97%, and positive predictive value and negative predictive values of 92% and 97%, respectively. All children who received a liver transplant had a score ≥11.

Figure 1.

Receiver-operating characteristic (ROC) curve for the analytes in the Wilson Index.

Table 5. Receiver-Operator Characteristic Curve Predictors of Mortality
AnalyteAUC95% CIDied/Alive
  1. Abbreviations: AUC, area under the curve; CI, confidence interval.

Table 6. New Wilson Index for Predicting Mortality
ScoreBilirubin (μmol/L)INRAST (IU/L)WCC (109/L)Albumin (g/L)

Prospective Evaluation of the Wilson Index

The new Wilson Index was applied prospectively to 14 further symptomatic children with WD (median age 12.9 [range, 6.6–15.8 years]; 10 male) diagnosed in our unit between January 2001 and September 2003. Main features at presentation were hepatosplenomegaly in 9, jaundice in 6, fever in 6, abdominal pain in 5, and ascites in 4. Diagnostic features and laboratory indices are presented in Table 7. A total of 4 children presented acutely and all received liver transplantation at a median of 18 (range, 6–27) days after admission; all are alive and well at a median of 23.2 (range, 14.7–36.5) months after surgery. The remaining 10 children received penicillamine (3), penicillamine and zinc sulfate (5), or trientine (2). A total of 2 children were converted to trientine during follow-up because of neutropenia in 1 and proteinuria in 1. All children on chelation therapy responded to treatment, with INR normalizing at a median of 387 (range, 375–409) days and AST at a median of 183 (47–531) days.

Table 7. Diagnostic Features, Laboratory Values, and Wilson Index of the 14 Children in the Prospective Evaluation Group (Median [Range])
 All patients (n = 14)Chelation therapy (n = 10)Transplantation (n = 4)
  • *

    P < 0.05 compared to the chelation group.

KF rings present (number)532
Ceruloplasmin (g/L)0.11 [0.05–0.27]0.1 [0.05–0.27]0.14 [0.12–0.15]
Urine copper (μmol/2 hours)10.5 [1.6–109]4.5 [1.6–97]23.5 [19.2–109]
Liver copper (μg/g dry wt)761 [262–1,318]368 [262–445]1,184 [1,077–1,318]*
Penicillamine challenge (μmol/24 hours)37.7 [2.3–187.2]17.5 [2.3–63.2]93.9 [41–187.2]*
Total bilirubin (μmol/L)18.5 [5–692]15 [5–407]360 [120–692]*
INR1.43 [1–5.8]1.1 [1–2.1]3.24 [2.04–5.8]*
AST (IU/L)89 [32–777]77 [32–777]243.5 [94–312]
WCC (109/L)7.64 [3.68–17.92]6.4 [3.68–9.4]11.89 [7.89–17.92]*
Albumin (g/L)31 [23–47]38 [25–47]25.5 [23–28]*
Wilson Index score5 [1–16]2 [1–11]13.5 [11–16]*

All the children who underwent transplantation had a Wilson Index of 11 or above. However 1 child with a score of 11 responded to chelation therapy alone (Fig. 2). None of the children who responded to medical treatment had a score >6. When applied to the prospective data, the new Wilson Index had a sensitivity and specificity of 75% and 91%, respectively.

Figure 2.

Receiver-operating characteristic (ROC) curve on and distribution of scores calculated by the Wilson Index when applied to the prospective evaluation group.


The clinical presentation of WD in this retrospective study reflects the referral pattern of a specialized tertiary center. Thus, onset with ALF was more common than in previous reports, affecting about 33% of the patients compared to a reported figure of less than 20% of WD patients presenting with liver disease.6 In contrast to previous reports, we have not observed a prevalence of females among patients with ALF.7 Since in an animal model of WD it has been suggested that sex hormones may be the cause for the higher prevalence of ALF presentation in females,8 it is possible that our inability to confirm this observation in a relatively large number of pediatric patients is due to the fact that most of them were prepuberal.

As previously reported, we confirm that the diagnosis of WD is not always easy, particularly in patients who present at a young age and with ALF. Ceruloplasmin has been shown to have a sensitivity and specificity of 82.4% and 94.4%, serum copper of 50% and 88.4%, and liver copper of 75% and 61.5%, respectively.4 In the present series, 20% of patients had a normal ceruloplasmin level, a figure similar to previous reports.9 Ceruloplasmin is an acute phase reactant and may increase in ALF as a reflection of hepatic inflammation,6 though in our series ceruloplasmin was normal in 4 of 27 children presenting with ALF or fulminant hepatic failure. The best diagnostic sign for WD remains the presence of Kayser-Fleischer rings. Although they can be found also in patients with prolonged cholestasis due to various causes, WD does not usually present with prolonged cholestasis. The combination of Kayser-Fleischer rings and low ceruloplasmin is probably the best guide to a rapid diagnosis,6, 10 but unfortunately Kayser-Fleischer rings are uncommon in pediatric series, being absent in 50% of the children in the present series.

Due to the difficulty in reaching the diagnosis of WD, a number of other laboratory parameters have been developed. A ratio between alkaline phosphatase and serum bilirubin of <2.0 was shown to differentiate WD from other causes of ALF in adults,11 but it is not always helpful in children, probably because of the effect of bone-derived alkaline phosphatase.12 The urinary copper excretion after penicillamine challenge was also developed as an aid to confirm the diagnosis and was shown to be 88.2% sensitive and 98.2% specific.4 Although helpful, the test is cumbersome and difficult to perform in severely ill children; in the present series, we were unable to do it in any of the children who died or required urgent transplant. In addition, the penicillamine challenge gave normal results in 9 of 29 tested children who presented with less severe disease and responded to chelation. This observation may reflect the difficulty in obtaining an accurate 24-hour urine collection in young children.

We encountered some difficulty also in the diagnosis of WD among siblings of the index patients. Urinary copper after penicillamine challenge and liver copper were only elevated in 55% and 87% of the patients, respectively, and ceruloplasmin was low in 88%. The diagnostic difficulties for this group of patients are well highlighted by other reports and emphasize the importance of an extended workup for these patients.13, 14

This retrospective review confirms the efficacy of chelation therapy in WD. Of the 49 symptomatic children who were treated with long-term chelation, 47 are alive up to 37 years after commencing treatment. Although it has been suggested that early administration of penicillamine in patients with WD presenting with ALF may lessen the need for transplantation,15 this has not been our experience; 16 of the children who died or required transplantation were treated with penicillamine, with or without zinc. We have observed side effects attributable to penicillamine in 18 children (38% of patients treated with this drug). These subsided in 6 children following a reduction of the dose, while 12 required conversion to trientine.

After the first reports of the efficacy of zinc in inhibiting the intestinal absorption of copper and in increasing metallothionein that binds copper,16 we have used a combination of zinc sulfate and penicillamine in 36 patients with WD, administrating the 2 drugs 6 hours apart, to avoid chelation of zinc by penicillamine. Because of the severity of the liver disease in our symptomatic patients and the lack of data on the use of zinc monotherapy in this group of patients, none was treated with zinc alone except a child who was switched to zinc alone for a few years because of penicillamine side effects, and eventually required addition of trientine because of high liver copper content and ongoing disease activity in a follow-up liver biopsy. Whether the asymptomatic siblings would have responded to zinc alone as well as to penicillamine, or penicillamine and zinc, remains unanswered. Zinc monotherapy has been found to be an effective, nontoxic alternative to penicillamine in asymptomatic siblings by other authors.16 The response to the penicillamine/zinc chelation treatment in our series appears to be excellent, even in the presence of severe liver damage, having been successful in 4 children who presented with ALF. Furthermore, unlike other studies,17–19 none of the children on long-term chelation in the present study, most of whom were taking penicillamine and zinc, has required liver transplantation over a median follow up period of 11.8 (1.45–34.2) years.

Liver transplantation has an important role in the management of WD presenting with decompensated hepatopathy. The mortality of those patients who present with fulminant hepatic failure, i.e., ALF and encephalopathy, is 100%,11 and they should be immediately listed for urgent transplantation. Although it has been suggested that liver transplantation should also be considered in those patients on chelation therapy who do not achieve normal liver function after 3 months of treatment,17 our data show that, at least in pediatrics, transplantation soon after diagnosis has very little, if any, role to play in WD not presenting with ALF. Thus, none of our patients on long-term chelation therapy has required a liver transplant during follow-up, even though some have taken a long time to normalize liver function and others continue to have abnormal transaminases up to 14 years after diagnosis. Particularly difficult is the decision of when to list a patient with ALF without encephalopathy for transplant, because chelation therapy can rescue some of them, as shown in the present series, avoiding the complications of lifelong antirejection treatment. In order to predict the outcome of WD patients presenting with ALF without encephalopathy, Nazer et al.2 developed a prognostic score to be used at presentation. The score was developed using a mixed population of adults and children and was based upon simple laboratory parameters. When the score was used in other centers, however, its utility was questioned.1 Indeed, when we applied the Nazer score2 to the retrospective series in the present study, 5 children who were predicted to die survived and 4 who were predicted to live died. Using more sophisticated statistical methods in a relatively large series of exclusively pediatric patients, we have been able to reevaluate and improve the original score. The new Wilson predictive Index was 93% sensitive, 98% specific, and had a positive predictive value of 93%. More importantly, all children with a score >11 died without transplantation, whereas all those with a score <11 survived, showing that the Wilson Index is helpful in identifying children in whom liver transplantation is indicated.

Of the 4 parameters (bilirubin, INR, AST, and WCC), used to devise the index, the first 3 are conventional measures of liver function and are easy to relate to the severity of liver disease. Elevated WCC has been used in various scoring systems that predict outcome of patients admitted to intensive care, such as the Acute Physiology and Chronic Health Evaluation score.20 WCC was also found to be useful in predicting outcome of ALF in children.21 Elevated WCC may be a marker of occult infection or a surrogate marker of an as yet unidentified factor that predicts severity of liver failure irrespective of the etiology of liver disease. Evaluation of the new Wilson Index in a small prospective cohort has confirmed its validity. The new scoring system, with its important clinical implications, should be validated in a multicenter study involving a large number of patients.