Autoimmune hepatitis (AIH) is a progressive inflammatory disease of the liver with an unknown pathogenesis. It predominantly affects prepuberal female children and is characterized by the presence of autoantibodies and hypergammaglobulinemia (1). Progression to cirrhosis or liver failure may occur. Combination of prednisone at a dosage of 2 mg/kg/day, not exceeding 60 mg, and azathioprine at a dosage of 1.5 to 2 mg/kg/day is the treatment of choice, and its favorable effects have been well established in large series of patients (2–6). However, this conventional treatment sometimes fails to induce remission (3,5,7), and it is not clear which clinical or laboratory parameters can predict the response. Furthermore, serious complications related to high prednisone doses are often reported. These adverse effects are associated with elevated morbidity and poor compliance (5). Therefore, other drugs with different modes of action may be required to control this disorder. We have already reported on the use of cyclosporine (CyA) as an initial treatment in a pilot study where biochemical remission was obtained in most cases during the first 6 months of therapy and in all of them by the end of the first year. Early results indicated that side effects were few, transient and well tolerated (8). Here, we report the outcome of children in this series and analyze potential predictors of the time to remission.
PATIENTS AND METHODS
The study was conducted prospectively, under a protocol described elsewhere (8). In the period from January 1994 to March 2001, 84 patients from 5 pediatric centers (4 in Argentina and 1 in Canada) were included consecutively as cases having definite AIH, according to the scoring system of the “International Autoimmune Hepatitis Group” (1,9). None of the patients had received immunosuppressive medication previously.
Patients presenting with fulminant or subfulminant liver failure or with other causes of chronic hepatitis, such as hepatitis B or C, or Wilson disease were excluded, as were those with any history of exposure to potentially hepatotoxic agents. The CyA (Neoral, Novartis) was started at a dosage of 4 mg/kg/day, divided into 3 equal daily doses for the first 6 months. The whole blood concentration was 250 ± 50 ng/mL for 3 months. Thereafter, in patients with aminotransferase levels lower than 3 times their normal values, CyA was reduced. Prednisone was initiated after 6 months of CyA at a dosage of 0.3 mg/kg/day or 0.5 mg/kg/day in children weighing less or more than 20 kg, respectively. Azathioprine was started at the same time as prednisone at a dosage of 1 to 2 mg/kg/day. The CyA was gradually decreased for a period of 15 days. Upon discontinuation of CyA at the seventh month, prednisone was then reduced slowly, and azathioprine was kept at the same dosage.
Clinical and Laboratory Evaluation
Complete clinical examination and laboratory evaluation were undertaken including liver function tests, blood urea nitrogen and creatinine level, urinalysis, blood count, serum protein electrophoresis and indirect immunofluorescence for the detection of autoantibodies: smooth muscle antibodies, antinuclear antibodies and liver-kidney microsome type 1 (LKM1) antibodies. Reactivity of LKM1 autoantibodies against CYP2D6 was tested using Western blotting. The diagnosis of portal hypertension (PHT) was considered by the presence of splenomegaly and ultrasonographic findings of the widening of the free border of the lesser omentum to a diameter of more than 1.7 times that of the aorta and confirmed by endoscopy. No histological picture compatible with bile duct obstruction was observed in the 59 liver biopsies obtained before starting CyA treatment. Fibrosis on biopsy was staged according to previously published histological criteria (1,10). Remission was defined as the absence of clinical symptoms in the presence of normal aminotransferase levels.
Subjects were controlled weekly during the first month, every 2 weeks during the second month and once a month thereafter. This study was accepted by the Ethics Review Board of all participating institutions.
The multivariate Cox proportional hazards approach was adopted as an approximation to model the risk of a delay in remission as a function of time, in absolute terms and under various covariate adjustments (11). A total of 68 potential predictor variables were submitted for statistical analysis. The results were considered significant if P < 0.05.
Unadjusted time to remission was estimated by the Kaplan-Meier method (12). Statistical analyses were performed with SAS software (SAS Institute Inc, Cary, NC) (13).
Table 1 enumerates the basic clinical and laboratory features at presentation of the 84 children being considered in the present follow-up. Median follow-up time was 29 months. This represents a follow-up period of 58.3% longer than the 12-month period reported initially (8).
Among the 80 children who were positive for smooth muscle antibodies, 43 (53.7%) were also positive for antinuclear antibodies. Only 4 patients (4.7%) were LKM1-positive. All patients had elevated aminotransferase levels upon entry into the study, and total serum bilirubin level was increased in 60 cases. Albumin was below 30 g/L in 24, and prothrombin time was less than 50% in 12 children. Serum immunoglobulin G exceeded the amount of 20 g/L in 73 patients. Extensive fibrosis (≥4) was observed in 53.5% of liver biopsies.
Thirteen patients (15%) were withdrawn from the study (median time, 7 months), 12 because of lost to follow-up and 1 was transplanted 5 months after the beginning of treatment because of persistent liver failure (Fig. 1). This case corresponds to a 13-year-old girl who presented liver failure at the time of diagnosis and responded to CyA; alanine aminotransferase (ALT) levels were 10 times higher than the normal values at the start of treatment and twice as much as the normal values after 3 months on CyA. However, no improvement in prothrombin time was ever attained, even after 2 months of treatment with prednisone and azathioprine. This case is regarded as a treatment failure. Among the other patients, 8 had already attained remission (median time, 5.5 months), and 5 did not exhibit remission while in the trial. Of these 5 patients, 4 were lost to follow-up (median time, 3 months).
The remaining 71 patients (85%) were maintained on treatment (median follow-up time, 9 months). Among the 79 patients (94.1%) who went into remission, aminotransferase levels were normalized in 57 (72.2%) within 6 months, under CyA administration.
The estimated median remission time for all 79 patients in the study was 4.7 months (95% confidence interval, 3.8–5.4 months). However, the remission rate was not constant. As seen in Figure 2, after an initial delay phase of approximately 40 days, remissions accumulated steadily for the next 7 months. Around that time, with 66 of the patients (79%) already in remission, the process tapered off, with remissions being increasingly more difficult for the remaining patients.
No statistically significant differences were found when the final outcome was compared between patients with a high (≥4) or low (≤3) stage of fibrosis on liver biopsy at onset. Sixty-eight variables describing demographic and biomedical data were assessed regarding their involvement in the remission process. The study variables displayed an intricate network of associations, among themselves and with varying phases of remission. As a result, some of these variables competed with and excluded each other as predictors of the probability of modifying the time to remission. The joint presence at baseline of total serum bilirubin level of greater than 1.2 mg/dL and PHT was found to be the simplest and most powerful predictive combination of a delay in remission. As Table 2 shows, total serum bilirubin level of greater than 1.2 mg/dL and PHT have strong independent predictive capability, even after adjusting for each other.
Tolerance of CyA was satisfactory. Transient hypertrichosis was seen in 55% (46/84) of the children, and a moderate degree of gingival hypertrophy was seen in 39% (33/84) of the patients. Transitory elevation of creatinine levels (8 cases in one of the visits) and hypertension (3 cases) were detected once in 1 control of each patient during the first months of follow-up. Both complications were mild. One patient developed transient glucose intolerance during the period of daily prednisone treatment. Tolerance of azathioprine was uneventful, and it was not necessary to modify the standard dosage of 1.5 mg/kg/day. No patient was excluded from the protocol because of side effects. No patient relapsed during the follow-up.
The large series of children with AIH treated by CyA presented in this paper confirm the value of cyclosporine as an initial therapy (8). Under this protocol of treatment, aminotransferase levels were normalized in all of the 71 patients who remained in the study. Remission was observed in 69% of them within the first 6 months of CyA therapy. The risk factors that we considered are quantitatively intertwined, and several multivariate models are capable of summarizing the remission experience of the children in our study. Our statistical results seem to favor the combination of joint presence at baseline of total serum bilirubin level of greater than 1.2 mg/dL and PHT as the most powerful and parsimonious anticipation of a delay in remission. By extension, patients with more severe liver disease upon diagnosis, as evidenced by the presence of jaundice, PHT, total serum bilirubin level of greater than 1.2 mg/dL, prothrombin time of less than 50%, albumin of less than 33 g/L and gammaglobulin of greater than 42 g/L may require significantly more prolonged treatment to achieve complete normalization of serum ALT levels. In our protocol, median follow-up time of the 71 patients (85%) remaining in the study was 2 years and 5 months. After CyA withdrawal, no relapse occurred under continuing prednisone and azathioprine. These results indicate that CyA leads to complete and persistent remission of the liver inflammatory process in a variable period. Some large pediatric series of patients treated with prednisone or prednisone and azathioprine showed percentages of remission similar to those in this paper (5,6). However, induction of remission by CyA in series was not responsible for severe side effects, similar to those described with high doses of prednisone, such as deadly infections and psychosis (5,6). In our patients, type 1 AIH was more frequent than type 2 (80/4); nevertheless, no differences in response to treatment were seen in either type. Recently, Debray et al. (14) also demonstrated the effectiveness of CyA in a larger number of patients with type 2 AIH in whom ALT activities normalized within 6 months, confirming that either type of AIH responds equally well to immunosuppressive treatment (2,3,6,14). In addition, this retrospective study confirmed the importance of considering CyA as initial therapy in very sick patients.
One of the major aims of our investigation was to protect children and adolescents from the untoward effects of steroids using CyA as primary immunosuppressant. Thus, CyA was administered 3 times a day for only 6 months to avoid its long-term side effects (15). This way of administering CyA may explain the good outcome in the course of the disease and adverse events. In adult patients, CyA also seemed to be safe and effective. In an openlabel trial (16), 19 patients with AIH were treated with CyA (10 patients had been treated with corticosteroids either unsuccessfully or with intolerable side effects, and 9 were treatment-naive). Fifteen patients (79%) completed a 26–week course of therapy and successfully tolerated the medication with minimal side effects. In the other 4 cases, CyA had to be discontinued because of side effects (1 patient did not respond even after adding prednisolone). Altogether, these results point to CyA as a promising treatment to control the liver inflammatory process in AIH, and the ensuing remission is well maintained by low doses of prednisone and azathioprine. CyA could be considered as an alternative to steroid-azathioprine initial therapy for AIH at any age, with mild and well-tolerated adverse effects. Patients with hepatic failure, signs of PHT and high total serum bilirubin level at onset could need an initial treatment with several immunosuppressive agents because rapid control of the inflammatory process is mandatory. It remains a future challenge to conduct a controlled study, classifying patients according to a quantifiable measure of the severity of their liver disease, so as to ascertain the best therapy, dose and period of treatment for each individual case.
The authors thank all of the AIH children who participated in the study and their families.
This investigation was supported in part by Novartis Pharmaceuticals of Argentina.
1. Johnson PJ, McFarlane IG. Meeting Report: International Autoimmune Hepatitis
Group. Hepatology 1993; 18:998–1005.
2. Maggiore G, Bernard O, Hadchouel M, et al
of autoimmune chronic active hepatitis in childhood. J Pediatr 1984; 104:839–844.
3. Maggiore G, Bernard O, Homberg J, et al
. Liver disease associated with anti–liver-kidney microsome antibody in children. J Pediatr 1986; 108:399–404.
4. Maggiore G, Porta G, Bernard O, et al
. Autoimmune hepatitis
with initial presentation as acute hepatic failure in young children. J Pediatr 1990; 116:280–282.
5. Maggiore G, Veber F, Bernard O, et al
. Autoimmune hepatitis
associated with anti-actin antibodies in children and adolescents. J Pediatr Gastroenterol Nutr 1993; 17:376–381.
6. Gregorio G, Portmann B, Reid F, et al
. Autoimmune hepatitis
: 20 years of experience. Hepatology 1997; 25:541–547.
7. Czaja AJ. Diagnosis, prognosis
of classical autoimmune chronic active hepatitis. In: Krawitt EL, Wiesner RH, editors. Autoimmune Liver Diseases. New York, NY: Raven Press; 1991. pp. 143–166.
8. Álvarez F, Ciocca M, Cañero-Velasco C, et al
. Short-term cyclosporine
induces a remission of autoimmune hepatitis
in children. J Hepatol 1999; 30:222–227.
9. Álvarez F, Berg P, Bianchi F, et al
. International Autoimmune Hepatitis
Group Report: review of criteria for diagnosis of autoimmune hepatitis
. J Hepatol 1999; 31:929–938.
10. Desmet V, Gerber M, Hoofnagle J, et al
. Classification of chronic hepatitis: diagnosis, grading and staging. Hepatology 1994; 19:1513–1520.
11. Cantor A. Extending SAS survival analysis techniques for medical research
. Cary, NC: SAS Institute Inc, 1997:208.
12. Lee ET. Statistical Methods for Survival Data Analysis
. New York: John Wiley & Sons Inc, 1992:250–262.
13. SAS/STAT User's Guide, Version 8.2. Cary, NC: SAS Institutes Inc.
14. Debray D, Maggiore G, Girardet J, et al
. Efficacy of cyclosporine
A in children with type 2 autoimmune hepatitis
. J Pediatr 1999; 135:111–114.
15. Paradis K, El Edreesi M, St-Vil D, et al
kinetics in pediatric liver transplant recipients: impact of dose fractionation on true glomerular filtration rate. Transplant Proc 1994; 26:2777–2778.
16. Malekzadeh R, Nasseri-Moghaddam S, Kaviani M, et al
. Cyclosporin A is a promising alternative to corticosteroids in autoimmune hepatitis
. Dig Dis Sci 2001; 46:1321–1327.