Disease-modifying antirheumatic drugs (DMARDs), including methotrexate and biologic agents, reduce joint inflammation and improve functional outcomes in rheumatoid arthritis, but are associated with significant side effects.1 Biologic agents, in particular Tumor Necrosis Factor (TNF) α inhibitors, increase the risk of severe infections.1 In 2008, the US Food and Drug Administration (FDA) issued a warning about the increased risk of systemic fungal infections in patients treated with TNFα inhibitors.2 The FDA has not issued a similar warning for abatacept, which has a mechanism of immunosuppression that is distinct from TNFα inhibitors. While abatacept has previously been associated with some fungal infections, only one case of abatacept-associated, non-disseminated histoplasmosis has been reported in the literature to date.3 Here, we report a case of abatacept-associated disseminated histoplasmosis. The patient provided verbal consent for the publication of her case.

CASE

A 62-year-old woman presented to the hospital in autumn 2016 with a 10-day history of abdominal pain, nausea, and fevers. Her medical history was remarkable for asthma and rheumatoid arthritis that had been treated with etanercept for several years. Six months before admission, she was transitioned from etanercept to adalimumab for improved symptom control. Three months before admission, she presented to the clinic with arthralgias and joint stiffness: She was switched to methotrexate and abatacept, and adalimumab was discontinued. She also received a 5-day course of prednisone for an asthma exacerbation in the month prior to admission. She lived in the Dominican Republic before immigrating to the USA in the 1980s. Five weeks after starting abatacept and 7 weeks prior to admission, she visited an underground lagoon in Mexico.

At the time of presentation, she was febrile to 103 °F and appeared markedly uncomfortable. Her abdomen was distended and diffusely tender without rebound tenderness or guarding. Her musculoskeletal exam was unremarkable. Her complete blood count, basic metabolic panel, and total bilirubin were normal. Liver enzymes were ALT 146 U/L, AST 298 U/L, and alkaline phosphatase 338 U/L. An abdominal CT scan was normal.

Throughout her hospitalization, her abdominal pain and fevers continued. Blood cultures were repeatedly negative. AST and ALT remained stably elevated while alkaline phosphatase peaked at 776 U/L, with direct hyperbilirubinemia to 1.6 mg/dL. Viral hepatitis serologies and serum PCR testing for Epstein Barr Virus (EBV), Cytomegalovirus (CMV), Herpes Simplex Virus (HSV), Varicella Zoster Virus (VZV), Babesia, and Ehrlichia/Anaplasma were unrevealing. Antibody tests for Human Immunodeficiency Virus (HIV) and Lyme disease were negative. An interferon gamma release assay for tuberculosis was negative. Serum 1,3-beta-d-glucan, a nonspecific test for fungal infections, was elevated to 126 pg/mL, and galactomannan, a test for Aspergillus, was normal. The elevated beta-d-glucan increased our suspicion for a fungal process.

MRI of the abdomen was obtained to rule out an infectious process not visualized on CT, and to evaluate the liver given her rising alkaline phosphatase and bilirubin. The study revealed splenomegaly with innumerable T2 hypointense lesions with delayed contrast enhancement, and hepatomegaly with marked periportal edema. Repeat 1,3-beta-d-glucan, obtained 4 days after the initial result, was elevated further at 237 pg/mL. A liver biopsy demonstrated multiple non-necrotizing granulomas with multinucleated giant cells. Budding yeast forms in the center of the granulomas, consistent with Histoplasma capsulatum, were identified by silver stain (Fig. 1). Urine and serum Histoplasma antigens, which were pending at the time of liver biopsy, returned positive at 16 ng/mL (with ≥ 0.5 ng/mL being positive) and elevated beyond the measurement of the assay, respectively. CNS infection was excluded by lumbar puncture.

Figure 1
figure 1

Multiple non-necrotizing granulomas with multinucleated giant cells are identified in the patient’s liver (a, ×20 and b, ×40, hematoxylin and eosin stain). Methenamine silver staining highlights budding yeast forms in the center of the granulomas, which are consistent with Histoplasma capsulatum (c, ×40, and d, ×100, arrows).

A diagnosis of disseminated histoplasmosis was made on the ninth day of her admission, and she was started on liposomal amphotericin B. Methotrexate and abatacept were discontinued. She was transitioned to itraconazole after 8 days of amphotericin treatment, because of worsening renal function, and discharged after a 19-day hospitalization. Five months after discharge, her symptoms completely resolved, yet urine and serum Histoplasma antigens remained positive. She continued on itraconazole and remained off biologic agents. Periodic rheumatoid arthritis flares were treated with oral corticosteroids, except for one severe flare 10 months after diagnosis, which resulted in restarting methotrexate. Her rheumatologists have not excluded reinitiating abatacept in the future.

DISCUSSION

Histoplasmosis is a fungal infection caused by inhalation of Histoplasma capsulatum microconidia, which are endemic to the Ohio and Mississippi River valleys in the USA, Central America, and parts of southern Europe, Asia, and Africa.4 Bird and bat guano supports the growth of Histoplasma; disruption of contaminated soil and exposure in caves in endemic areas can lead to infection.4 Infections in immunocompetent individuals are often asymptomatic. Disseminated histoplasmosis presents in patients who are unable to mount an effective cell-mediated immune response to Histoplasma, including those with advanced HIV, hematologic malignancies, solid organ transplant recipients, and patients on immunomodulatory biologic agents, such as TNFα inhibitors.

Abatacept is a commonly used immunosuppressive biologic agent with a mechanism of action distinct from TNFα inhibitors, and with an unknown risk of disseminated histoplasmosis. Abatacept inhibits T cell activation by competing for the costimulatory signal on antigen-presenting cells (APCs). Complete T cell activation requires an interaction between costimulatory molecules CD80 or CD86 on APCs and CD28 on T cells. Cytotoxic T lymphocyte antigen 4 (CTLA4) is an inhibitory signal on the surface of T cells with a stronger affinity for CD80 and CD86 than CD28. CTLA4 can inhibit complete T cell activation by competing for costimulatory signals. Abatacept is a recombinant fusion protein including the extracellular domain of CTLA4.5 By inhibiting full T cell activation, abatacept prevents the production of cytokines important to the progression of rheumatoid arthritis but also to killing intracellular organisms, such as Histoplasma.

The patient in this case report was diagnosed with disseminated histoplasmosis that was associated with the use of abatacept. To our knowledge, treatment with abatacept has not been previously reported as a risk factor for disseminated histoplasmosis. Long-term prednisone use poses an infectious risk; however, our patient only received a 5-day course of prednisone prior to symptom onset. Although disseminated histoplasmosis in patients on methotrexate alone has been reported,6 a systematic review including 106 trials found that standard-dose biological agents carried a higher risk of serious infection compared to methotrexate.7 It is possible that the patient’s prior use of TNFα inhibitors may have contributed to her susceptibility to disseminated disease. The half-life of adalimumab is 10–20 days,8 and her presumptive exposure to Histoplasma in Mexico occurred 5 weeks, approximately two half-lives, after discontinuing adalimumab. Although residual immune suppression from adalimumab may have contributed, treatment with full-dose abatacept is a more likely explanation for the development of disseminated histoplasmosis in this case.

The patient’s travel to an underground lagoon in Mexico 1 month prior to symptom onset is highly suggestive of primary infection. Given that the patient previously lived in the Dominican Republic where Histoplasma is endemic, reactivation cannot be excluded as the cause of the patient’s presentation. However, reactivation would have been more probable during the period when the patient was treated with TNFα inhibitors, which unlike abatacept have a well-characterized association with disseminated histoplasmosis.2 More generally, reactivation is not favored as the mechanism of histoplasmosis infections in patients on TNFα inhibitors.9

Impaired cellular immunity by abatacept can lead to infectious complications other than histoplasmosis, albeit at a low rate. A 2008 integrated safety analysis of patients on abatacept demonstrated that the incidence of serious infections was slightly higher in patients on abatacept as compared to placebo (3.47 vs. 2.41 events/100 person-years).10 Opportunistic infections occurring in these patients included tuberculosis (0.06 events/100 person-years), aspergillosis (0.02), blastomycosis (0.01) and systemic candidal infection (0.01).10 By comparison, TNFα inhibitors are associated with a risk of tuberculosis of 0.15 and 0.09 events/100 person-years for infliximab and adalimumab, respectively.11 Although the rates of these systemic infections on abatacept are low, the increased risk and severity of the illnesses suggest clinicians should be aware of the possibility of these conditions in patients on abatacept.

Subsequent studies have questioned the risk of serious infections in patients taking abatacept. A 2009 meta-analysis including five placebo-controlled trials in patients taking abatacept observed common infections including pulmonary infections, bacteremias, and gastrointestinal infections. However, there was no significant increase in the rate of serious infections, defined as life-threatening infections requiring intravenous antibiotics or hospitalization, with abatacept treatment compared to placebo.12 A single case of pulmonary aspergillosis and one case of unconfirmed tuberculosis were reported.12 Although the risk of tuberculosis appears low in patients taking abatacept, screening for latent tuberculosis is recommended when initiating this medication.13

The evidence indicating whether it is safe to resume biologic agents during or after treatment for histoplasmosis is scant. In a case series of seven patients treated for histoplasmosis, no patients relapsed after TNFα inhibitors were resumed (follow-up period 1–8 years).9 However, three of these patients were maintained on itraconazole during the follow-up period. The Infectious Diseases Society of America (IDSA) recommends itraconazole treatment until clinical and laboratory findings normalize.14 Since 10–15% of patients with disseminated histoplasmosis relapse, lifelong itraconazole therapy may be necessary for patients who require continued immunosuppressive therapy.14 Per IDSA, patients diagnosed with active histoplasmosis up to 2 years prior to initiating immunosuppression may require itraconazole prophylaxis, but the duration of prophylaxis needed is not clear.14

The safety of restarting biologic agents after a serious infection is contingent not only on the patient’s immune status, but also the nature of the initial infection. For instance, it has been shown that it is safe to resume TNFα inhibitor treatment following completion of an antituberculous regimen.15 Immunosuppressed individuals infected with histoplasmosis can relapse without reexposure.4 For this reason, providers should proceed with caution when considering the continuation of biologic agents in patients treated for histoplasmosis.

For rheumatoid arthritis patients with a prior serious infection, the American College of Rheumatology conditionally recommends abatacept over TNFα inhibitors.16 This recommendation is based on a single comparative study of biologic therapies used after a serious infection in patients treated with TNFα inhibitors.17 The study found that patients taking abatacept after their initial infection were less likely to be re-hospitalized for a subsequent infection compared to patients taking infliximab.17 Although this finding indicates abatacept may be a safer alternative to other biologic therapies for rheumatoid arthritis, our case demonstrates that providers should remain alert for possible serious fungal infections on this medication.

When deciding to restart an immunosuppressive agent in the setting of a serious infection, the potential side effects of the immunosuppressive agents must be balanced against those of the antibiotics. In addition to increased risk for serious infection, patients on abatacept had higher rates of autoimmune adverse events, most commonly psoriasis and vasculitis, when compared to patients given placebo.18 Patients on abatacept had similar rates of malignancy as patients given placebo, but compared to the US general population, patients taking abatacept had higher rates of lymphoma.18 Methotrexate is frequently associated with hepatic and gastrointestinal toxicity, and long-term prednisone can lead to Cushing’s syndrome, osteoporosis, and an increased risk of cardiovascular disease.19,20,21 Itraconazole, which our patient will take long term, is generally well-tolerated and most commonly causes gastrointestinal side effects.22 Accordingly, determining which agent to choose when restarting immunosuppression requires weighing the severity of rheumatoid arthritis symptoms, concern for a relapse of histoplasmosis, and the side effect profiles of various available DMARDs and antibiotics.

Patients taking immunomodulating medications may be at risk for serious fungal infections, including histoplasmosis, and this disease should be considered in patients on TNFα inhibitors or abatacept who present with fever of unknown origin. Gathering a comprehensive travel history to evaluate for proximate travel to endemic areas is particularly important, as recent exposure is the likely cause of histoplasmosis for patients on these medications. The safety of resuming biologic agents following successful treatment of the infection should be evaluated on a case-by-case basis.