Allopurinol Salvage Therapy in Pediatric Overlap Autoimmune Hepatitis‐Primary Sclerosing Cholangitis With 6‐MMP Toxicity

Abstract

Thiopurines have been shown to be efficacious for the treatment of autoimmune disorders including inflammatory bowel disease (IBD) and arthritis as well as for the treatment of malignancies. They have also conventionally been used for maintenance of remission in autoimmune hepatitis (AIH) after induction with steroids (1). The safety profiles of 6-mercaptopurine (6-MP) and its prodrug azathioprine (AZA) have long been established in adults and children. Studies found the incidence of hepatotoxicity from thiopurines to be 2.8% to 10% in patients with IBD (2,3). AZA and 6-MP are metabolized via 3 divergent pathways (Fig. 1). The active metabolite, 6-thioguanine (6-TG), is produced by hypoxanthine phosphoribosyltransferase. Xanthine oxidase leads to the production of the inactive metabolite, 6-thiouracil. Thiopurine methyltransferase (TPMT) also produces an inactive metabolite, 6-methylmercaptopurine (6-MMP). Some patients have preferential shunting of AZA/6-MP toward excess of 6-MMP over 6-TG, and this may lead to hepatotoxicity. These metabolites can be measured in serum. A 6-TG level above 235 pmol/8 × 108 erythrocytes is considered to be in the therapeutic range, whereas levels above 450 pmol/8 × 108 erythrocytes have been associated with bone marrow toxicity. Hepatotoxicity correlates with elevated 6-MMP, more than 5700 pmol/8 × 108 erythrocytes (4). These values were confirmed in a meta-analysis by Osterman et al (5) in IBD and shown to be useful in children with AIH by Rumbo et al (6).FIGURE 1: AZA metabolism. AZA and 6-MP are metabolized via 3 divergent pathways. The active metabolite 6-TG is produced by HPRT with some intermediate steps. HPRT activity first leads to the production of 6-TIMP, which can then either go toward 6-TG by IMPDH and GMPS or be methylated by TPMT into 6-MMPR. XO leads to the production of the inactive metabolite, 6TU. In the third pathway, TPMT also produces an inactive metabolite, 6-MMP. Allopurinol acts by inhibiting XO as indicated by the solid gray diagonal line. It may also have other mechanisms that have yet to be uncovered (8). AZA = azathioprine; GMPS = guanosine monophosphate synthetase; HPRT = hypoxanthine phosphoribosyltransferase; IMPDH = inosine monophosphate dehydrogenase; 6-MMP = 6-methylmercaptopurine; 6-MMPR = 6-methylmercaptopurine ribonucleotides; 6-MP = 6-mercaptopurine; 6-TG = 6-thioguanine; 6-TIMP = 6-thiosine 5'-monophosphate; TPMT = thiopurine methyltransferase; 6-TU = 6-thiouracil; XO = xanthine oxidase (4,5,10,13,17).Allopurinol, a xanthine oxidase inhibitor, has been used in IBD to attain therapeutic 6-TG levels. Combination use of allopurinol with a thiopurine in adult and pediatric patients with IBD was shown to have clinical benefit with improved disease control and decreased steroid use without hepatotoxicity (7–10). AZA is a first-line therapy in patients with AIH because it has few adverse effects, is easy to monitor, and has high compliance with once-daily dosing. It is clinically important to be able to use AZA in a higher proportion of patients by combining it with allopurinol. To date, there are no published case reports/series of pediatric patients with AIH being treated with allopurinol-thiopurine combination therapy. We discuss the successful use of such therapy in 3 children with AIH. METHODS The database of children who received care at the Mount Sinai Pediatric Liver Program in the last 20 years was searched for patients with AIH. Clinical information of children with serology-and biopsy-proven AIH who were treated with combination therapy was reviewed. Three were identified and their charts reviewed. RESULTS AZA dose, autoimmune markers, 6-TG/6-MMP, transaminases, white blood cell count, and length of follow-up are summarized in Table 1.TABLE 1: Summary of cases: sequence of events for each case goes from left to rightCase 1 An 8-year-old girl was diagnosed with AIH type I (immunoglobulin G [IgG] 1596 mg/dL, ASMA 1:640) at 3 years of age by biopsy and with large duct primary sclerosing cholangitis (PSC) at 4 years of age by biopsy and magnetic resonance cholangiopancreatography (MRCP), essentially overlap syndrome. She was treated with steroids and AZA at 1.5 mg/kg. As the steroid dose was tapered during 3 months, her transaminases became elevated: alanine transaminase (ALT) 108/aspartate aminotransferase (AST) 166 IU/L. γ-Glutamyl transferase (GGT) was elevated to 924 IU/L. She was found to have a nontherapeutic 6-TG (121 pmol/8 × 108 erythrocytes) and an elevated 6-MMP (6179 pmol/8 × 108 erythrocytes). Based on our successful use of rapamycin in controlling AIH in a posttransplant setting, she transitioned to rapamycin, and ursodeoxycholic acid was added secondary to changes in PSC (11). She failed to respond after 7 months of treatment, despite attaining therapeutic rapamycin levels (4.5–8.8 μg/L). Her ALT remained elevated at 78 IU/L, AST 120 IU/L, and GGT 192 IU/L. Allopurinol was then given 25 mg/day, and AZA was restarted at a reduced dose of 0.5 mg/kg. Her liver enzymes and GGT normalized after 3 months: ALT 55 IU/L, AST 46 IU/L, and GGT 79 IU/L. She continues to do well with stable liver enzymes and no adverse effects after 34 months follow-up. Case 2 An 11-year-old boy was diagnosed at 8 years of age with ulcerative colitis treated with mesalamine and at 10 years of age with overlap AIH type I (IgG 1290 mg/dL, anti-smooth muscle antibody [ASMA] 1:45) and PSC diagnosed by biopsy and MRCP. He had biochemical remission with steroids and ursodeoxycholic acid (ALT 68/AST 31 IU/L) after 4 months of therapy. Soon after starting 6-MP and increasing the dose to 2 mg/kg, however, the transaminases became elevated: ALT 285/AST 117 IU/L. 6-MMP levels were in toxic range (5605 pmol/8 × 108 erythrocytes) and 6-TG subtherapeutic (123 pmol/8 × 108 erythrocytes). Therefore, allopurinol was given at 50 mg/day and the AZA dose was lowered to 0.7 mg/kg. His liver enzymes normalized after 4 months: ALT 33/AST 30 IU/L. GGT was never elevated. He has had no adverse effects and remains well after 8-month follow-up. Case 3 A 6-year-old girl was diagnosed with small duct PSC at 3 years of age by biopsy and MRCP. She was initially treated with ursodeoxycholic acid with improvement of her GGT from 533 to 93 IU/L and her transaminases: ALT from 262 to 98 IU/L and AST from 200 to 84 IU/L (12). After a period of noncompliance, her transaminases peaked: ALT 912/AST 680 IU/L. GGT increased to 491 IU/L. She was given the same therapy, but her transaminases did not normalize after 6 months. Overlap AIH type I (IgG 1400 mg/dL, ASMA 1:320) was confirmed by biopsy and serology. Remission was partially attained with steroid treatment: ALT 104/AST 76 IU/L. AZA was added at 1 mg/kg, and the transaminases remained elevated despite an escalation of AZA to 4 mg/kg in 6 months. It was then discovered that she had high 6-MMP (7341 pmol/8 × 108 erythrocytes) with subtherapeutic 6-TG (158 pmol/8 × 108 erythrocytes). She was treated with a reduced dose of AZA (1.1 mg/kg) combined with allopurinol, 50 mg/day, leading to remission after 4 months of therapy (ALT 42/AST 52/GGT 35 IU/L). She continues in remission with no adverse effects after 8-month follow-up. DISCUSSION Allopurinol has a history of being used to alter thiopurine metabolite levels in patients who have undergone renal transplantation and later in patients with IBD. Allopurinol inhibits xanthine oxidase. It may also affect TPMT activity; however, neither in vitro nor in vivo studies have shown this to be true (7,8). Other unproven theories state that allopurinol may be inhibiting an unidentified cofactor involved in thiopurine metabolism (8). We successfully treated 3 pediatric patients with AIH receiving allopurinol-thiopurine combination therapy. This has not been reported in the literature. These patients were treated similarly as were the patients with IBD in the literature (7–10,12), including monitoring and avoidance of potential adverse effects of combination therapy. Allopurinol adverse effects include allergic reactions including a hypersensitivity-related hepatitis, rashes, gastrointestinal upset, and bone marrow suppression (13). Bone marrow suppression is noted to occur in 0.6% of patients (13). This was especially important in using allopurinol in combination with AZA. In our series, we did not encounter any adverse effects and monitored weekly complete blood cell counts initially. Dose reduction of AZA is the main measure used to avoid neutropenia. We reduced the dosage to 30% of the maximal dose. This is consistent with the studies in adult and pediatric patients with IBD in which the dose reduction was 25% (14) and 30% to 40% (10), respectively. Thiopurine use in patients with AIH may be more complicated given that these patients have underlying liver disease. Elevations in liver transaminases may have been due to nonresponse to treatment. Thus, use of metabolite levels helped in differentiating patients who are noncompliant, underdosed (low 6-TG and 6-MMP), refractory (therapeutic 6-TG and nontoxic 6-MMP), or have preferential shunting to 6-MMP (low 6-TG and toxic 6-MMP). All of our patients had subtherapeutic 6-TG and elevated 6-MMP with an elevated ALT. Patients who are found to be refractory or intolerant to thiopurines may be treated with other immunosuppressants such as mycophenolate mofetil, tacrolimus, or cyclosporine (1,11,15). After starting combination therapy, all of the patients had biochemical remission, therapeutic 6-TG, and low 6-MMP. Our youngest patient, 3 years old at diagnosis, reached a higher dose of AZA (4 mg/kg) before starting combination therapy. A similarly high dosage was discovered in the youngest patient in a pediatric IBD case series by Rahhal (10). An earlier study by Grossman et al (16) showed that children younger than 6 years old with IBD required higher doses of AZA to achieve a therapeutic effect. Age-related differences in metabolism such as absorption, bioavailability, and higher TPMT activity may contribute to this effect. Thus, it is especially important to monitor thiopurine metabolites during therapy in younger children. It was intriguing that all of the patients in this series had overlap AIH and PSC. They also had AIH type I, which is characteristically associated with overlap syndrome (12). Treating the autoimmune component of the disease with steroid induction and immunosuppressants such as AZA for maintenance of remission in combination with ursodeoxycholic acid is the standard of care (1). The fact that all of our patients needing allopurinol-thiopurine combination therapy had overlap AIH/PSC poses the question of some unknown alteration in metabolism of thiopurines that may be more prevalent in this group of patients. During 20 years at Mount Sinai Medical Center we have treated 65 pediatric patients with AIH and 12 with overlap AIH/PSC. No patients with isolated AIH have failed AZA therapy secondary to hepatoxicity. As our understanding of PSC, overlap syndrome, and thiopurine metabolism increases, it will be interesting to see whether a common link exists. In summary, we recommend that children with overlap AIH and PSC receiving thiopurine therapy, who present with elevated liver enzymes, have measurements of 6-TG and 6-MMP. These children may have preferential thiopurine metabolism leading to 6-MMP hepatotoxicity and subtherapeutic 6-TG. In this case series, all 3 patients had overlap AIH and PSC. Combination therapy with a xanthine oxidase inhibitor and a lower dose of thiopurine is well tolerated and resulted in maintenance of remission in this subset of patients. It is likely that isolated patients with AIH who present in a similar manner would respond to combination therapy as well.