Comparison of relapse-free survival following an initial relapse for those treated with steroids alone vs. steroids and an immunomodulatory agent (Adapted from Hart et al.  with permission from the BMJ Publishing Group)
In this series, almost one-quarter (n = 9/41) of all patients needed to discontinue the immunomodulatory agent due to side effects (including bacteremia, myelosuppression, and nausea/vomiting). This rate of discontinuation is comparable to rates in other disease populations, such as inflammatory bowel disease . Depending on the indication for drug discontinuation, substitution with a different thiopurine (i.e., 6-mercaptopurine) or mycophenolate mofetil can be considered (this was successful in 6/9 patients in the previous series).
The decision to use azathioprine as the primary steroid-sparing agent was based on previous experience in management of autoimmune hepatitis, which does not clearly share a common pathogenic mechanism with AIP. Recently, Schwaiger et al. published a study using a murine model showing that an alternative immunomodulatory agent may be more efficacious than azathioprine . They developed a model in MRL/Mp female mice that when exposed to polyinosinic:polycytidylic acid induces an inflammatory condition that mimics the histologic and phenotypic features of human type 1 AIP. Treatment with cyclosporine or rapamycin reduced histologic damage to the pancreas to a greater degree than azathioprine achieved. Thus, these data suggest that a different immunomodulatory agent may be more clinically efficacious than azathioprine, which has been the most studied in AIP. However, whether or not the findings from their animal model are relevant in humans remains to be demonstrated.
Immunomodulatory agents, such as azathioprine, may have a modest benefit in maintenance of remission in patients with AIP. However, almost half of the treated patients develop disease relapse during treatment, are unable to be weaned from steroids, or require discontinuation due to side effects. Therefore, additional treatment options are needed. It is possible that alternative immunomodulatory agents (other than thiopurines) may be more effective than azathioprine; however, this remains to be demonstrated in humans.
Rituximab (RTX) is a monoclonal antibody to the CD20 antigen on premature and mature B cells. It is used to treat a variety of conditions including B-cell lymphoma, rheumatoid arthritis, and ANCA-associated vasculitis; however, its use in pemphigus vulgaris and orbital pseudolymphoma are most pertinent to the treatment of AIP. Pemphigus vulgaris is a bullous skin disease in which pathogenic IgG4 antibodies (antikeratinocyte cell-surface antibodies) mirror disease activity that responds to RTX . Orbital pseudolymphoma is a condition that was traditionally treated with radiation treatment. However, it has also been demonstrated that treatment with RTX successfully induces remission in most patients . Subsequent studies demonstrated that many patients with orbital pseudolymphoma frequently have elevated serum IgG4 levels, and now, orbital pseudolymphoma is recognized as a manifestation of IgG4-RD .
Treatment of AIP with RTX was first reported in a patient with relapsing IgG4-related sclerosing cholangitis . He had refractory disease requiring multiple courses of high-dose steroids and developed a flare during maintenance treatment with 6-mercaptopurine. Core pancreatic biopsies revealed a modest population of CD20-positive B cells in the inflamed tissue, suggesting RTX would address at least a significant component of the inflammatory infiltrate. Thus, RTX was initiated, and subsequent cholangiography was markedly improved with almost complete resolution of liver test abnormalities. Other disease activity including pancreatic and orbital enlargement also resolved with RTX treatment. Since this initial patient, we have continued to use RTX in difficult-to-treat patients who either are resistant to an immunomodulatory agent or are unable to tolerate induction treatment with high-dose steroids.
Currently, the two largest series describing the use of RTX in AIP/IgG4-RD are from the United States, including one from Massachusetts General Hospital (MGH) (Boston, MA) and the other from Mayo Clinic (Rochester, MN). The series published from MGH details the treatment of 10 patients with IgG4-RD, with predominantly extra-pancreaticobiliary involvement (one patient had pancreas involvement and two had biliary involvement) . The dosing protocol consisted of two infusions of 1,000 mg intravenously on days 0 and 15. Nine of the 10 patients demonstrated significant clinical improvement within one month of therapy. All 6 patients on steroid treatment at the time of RTX initiation could be successfully tapered off of steroids. However, within 6 months of follow-up, four (40 %) patients required additional RTX treatment for disease recurrence.
In the Mayo Clinic series, a total of 12 patients were treated with a series of RTX infusions spanning a period of 2 years . One patient had isolated biliary involvement, while the other 11 patients had pancreatic involvement with or without biliary disease. The treatment protocol was similar to that used to treat B-cell lymphoma, in which patients received 4 weekly infusions (375 mg/m2), followed by a maintenance infusion every 3 months for 2 years (12 total doses). Using this regimen, 10/12 patients had a complete, clinical remission (symptomatic, biochemical, and radiographic) within 3 months of initiating treatment. Four of these patients were unable to tolerate high-dose steroids and were treated with RTX alone, demonstrating that it is an alternative induction agent to steroids. One patient had a partial response but had persistent liver test abnormalities due to another diagnosis. No patients developed a disease relapse during treatment.
Although RTX has only been used in small numbers of individuals with AIP, it has been extensively used in large numbers of patients with rheumatoid arthritis and B-cell lymphoma, so the side effect profile is well characterized. The most common side effect is a cytokine-mediated infusion reaction. This occurs in up to 10 % of patients during the first infusion and are typically mild. Otherwise, an immunologically mediated hypersensitivity reaction can produce cutaneous-, respiratory-, or joint-related symptoms. Myelosuppression (particularly lymphopenia) is an important consideration, and the median onset of abnormalities is 14 days following an infusion. However, delayed-onset cytopenias can also develop, so routine monitoring of complete blood counts is advisable during and after treatment. Finally, infectious complications are an important consideration, with infections occurring in up to 30 % of patients. Most infections are not serious, but one notable exception is progressive multifocal leukoencephalopathy (PML). PML has been described after RTX treatment, and although it is exceedingly rare, it is almost universally fatal and therefore should be discussed with patients prior to starting treatment .