Trouble with tumor necrosis factor α inhibitors, not just tuberculosis
Article first published online: 1 JUN 2010
Copyright © 2010 by the American College of Rheumatology
Arthritis Care & Research
Volume 62, Issue 6, pages 770–774, June 2010
How to Cite
Racunica, T., Cassidy, D., Cicuttini, F. and Hall, A. (2010), Trouble with tumor necrosis factor α inhibitors, not just tuberculosis. Arthritis Care Res, 62: 770–774. doi: 10.1002/acr.20245
- Issue published online: 1 JUN 2010
- Article first published online: 1 JUN 2010
- Manuscript Accepted: 29 OCT 2009
- Manuscript Received: 22 MAY 2009
It is established that tumor necrosis factor α (TNFα) inhibitors increase the risk of activating latent tuberculosis (TB) (1). When TB complicates TNFα inhibitor therapy, the recommended break in TNFα inhibitor therapy then raises issues in the management of the underlying rheumatic condition (1). There is a further issue with TNFα inhibitors worth exploring. When TB does complicate TNFα inhibitor therapy, the use of TNFα inhibitors and/or TB infection or simply the presence of TB antigens may alter the immunology of the underlying disease or predispose to new autoimmune disease. We report here on a case of a patient who developed extrapulmonary TB during treatment with infliximab for psoriatic arthritis (PsA) with sacroiliitis. During appropriate TB treatment, following cessation of infliximab and dose reduction of prednisolone, the patient developed anterior uveitis and bilateral retinal vasculitis. Although several etiologies for the eye condition are plausible, the authors propose it is related to a localized immune response in the eye in the setting of controlled TB infection: so-called “Eales' disease.”
A 29-year-old Asian woman with PsA involving mainly spinal and axial joints and a 10-year history of psoriasis had commenced infliximab 5 mg/kg. Due to the axial involvement, the usual loading regimen for ankylosing spondylitis of infusions at weeks 0, 2, and 6, then every 6 weeks, was employed (Figure 1). At week 12, the patient attended her fourth infusion when she reported feeling very well with resolution of her joint symptoms and psoriasis, yet her inflammatory markers were significantly elevated (C-reactive protein level 33 mg/liter and erythrocyte sedimentation rate 81 mm/hour). One week later, she experienced persisting lower abdominal pain and mild dysuria. She denied fevers and her general physical examination was unremarkable. Investigations, including a midstream urine and urinary chlamydia and gonorrhea polymerase chain reaction (PCR), were normal or negative. Over the subsequent week her symptoms worsened and pelvic ultrasound revealed moderate ascites.
Past history was unremarkable except for PsA, for which she had previously taken oral medications, including celecoxib, methotrexate (20 mg weekly), and cyclosporine (150 mg daily). The patient had immigrated to Australia from Hong Kong 6 years earlier, employed as an accountant. Six weeks prior to presentation she had visited Hong Kong for 4 weeks. She had no known exposure to TB. Prior to commencement of infliximab, TB screening had been performed using recommendations at that time. A chest radiograph was reported as normal, tuberculin skin test (5 IU of purified protein derivative) was 0 mm, and an interferon-γ release assay, QuantiFERON-TB Gold test, was indeterminate due to a poor mitogen response.
On examination the patient was febrile (38.9°C) with a sinus tachycardia (120 beats/minute), and her blood pressure was 115/70 mm Hg. Her abdomen was generally tender with shifting dullness consistent with ascites. There were small bilateral pleural effusions. Undilated fundal examination was normal.
An ultrasound-guided ascitic aspirate was negative for organisms and TB PCR. There were mesothelial and inflammatory cells noted. A chest radiograph at this time showed multiple small nodules. The patient had no cough and even with inducement, no sputum could be obtained. A bronchoalveolar lavage was obtained revealing normal cytology, negative gram stain, no acid-fast bacilli (AFB), and negative TB PCR. Computed tomography of the chest, abdomen, and pelvis showed moderate ascites and apical scarring in the left lung. The patient continued experiencing high fevers yet remained hemodynamically stable.
On admission to the hospital, methotrexate and infliximab were ceased. Despite the lack of microbiologic confirmation, peritoneal TB was suspected; thus, after 12 days of investigations she was commenced on a 4-drug TB treatment, as well as pyridoxine 25 mg orally daily. This included: isoniazid 300 mg, rifampin 600 mg, ethambutol 15 mg/kg (800 mg), and pyrazinamide 25 mg/kg (1,250 mg), all taken orally daily. Four days later, a targeted peritoneal biopsy was performed that revealed necrotizing granulomatous inflammation and AFBs. A repeat interferon-γ release assay was positive.
Four days after commencing anti-TB treatment, the patient developed a maculopapular rash over her extremities and trunk thought to be a drug reaction to isoniazid or rifampin. She then commenced prednisolone 25 mg orally daily. Isoniazid and rifampin doses were reduced (to 200 mg and 300 mg, respectively). After 7 days, her rash and fevers settled. One month later, the isoniazid and rifampin doses were increased (to 300 mg and 450 mg, respectively). The patient tolerated this and was compliant with her medications. However, because over time the patient developed worsening of joint pains (shoulders, knees, ankles, and sacroiliac joints), methotrexate 20 mg/week orally was recommenced and intraarticular corticosteroid injections were administered. This resulted in good control of joint symptoms except for the sacroiliitis, which remained particularly disabling. There was also poor control of her cutaneous psoriasis.
Two months later, fully sensitive Mycobacterium tuberculosis was confirmed in sputum and peritoneal samples so pyrazinamide and ethambutol were ceased. The prednisolone dosage was reduced to 12.5 mg orally daily with the intent to wean and cease.
Approximately 3 months after commencing TB treatment and 4 months after ceasing infliximab, the patient presented with a painful, red right eye. There were no changes to visual acuity, but a cellular exudate (3+ cells and flare) in the anterior chamber of the right eye and perivascular retinal changes (retinal vasculitis) bilaterally were noted (Figure 2). The patient was diagnosed with right anterior uveitis and bilateral retinal vasculitis. An anterior chamber aspirate was performed; however, the limited fluid available only allowed for TB PCR testing, which was negative. Her antinuclear antibodies (ANAs) were weakly positive at 1:160 and double-stranded DNA (dsDNA) and perinuclear and classic antineutrophil cytoplasmic antibodies were negative.
Immune-related phenomena without active infection at the eye after cessation of infliximab and reduction in prednisolone in the setting of controlled TB were considered the most likely responsible factors contributing to the ocular condition. The patient was started on prednisolone eye drops every 2 hours with oral prednisolone increased to 25 mg daily. Her anti-TB treatment remained unchanged. Ocular signs resolved over the following 4 weeks. Prednisolone was then reduced over 6 months. Both topical and oral prednisolone were ceased without recurrence of ocular symptoms. After 6 months of anti-TB treatment, the patient recommenced infliximab with planned anti-TB treatment for 12 months. To date, 14 months after commencing anti-TB treatment, the patient has ceased her anti-TB medications, and continues receiving infliximab and methotrexate. She remains free of TB and ocular symptoms and her PsA is once again well controlled.
This report describes a noteworthy case of latent TB activation during infliximab therapy for PsA with the additional complication of uniocular anterior uveitis and bilateral retinal vasculitis. This complication arose in the setting of current effective anti-TB treatment, corticosteroid withdrawal, and previous infliximab use. A noninfectious immune-related etiology is thought to be the most likely mechanism underlying the ocular disease. In this case report, we would like to discuss the diagnostic and therapeutic dilemma of the ocular disease and highlight the difficulty in diagnosing this condition when TB complicates TNFα inhibitor therapy.
The ocular disease presented a diagnostic and therapeutic dilemma (Table 1). TB uveitis is a common manifestation of ocular TB infection, although any part of the eye can be involved (2). Ocular TB was thought to be unlikely because there were no other signs of active TB in this patient who had been treated with appropriate anti-TB therapy for over 2 months, and the ocular TB PCR was negative. Although other tissue samples that subsequently grew TB on culture were also TB PCR negative in this patient, ocular symptoms resolved with immunosuppression and no change in anti-TB treatment. Another opportunistic infection was considered; however, the clinical appearance was atypical of cytomegalovirus (CMV), herpesvirus, or fungal infections.
|Other opportunistic infection|
|Related to seronegative arthropathy|
|Tumor necrosis factor α inhibitor induced|
|Or systemic lupus erythematosus|
|Antituberculosis medication adverse effect|
|Immune reconstitution inflammatory syndrome|
Anterior uveitis is a well-known complication of the spondylarthritides, and indeed, the ocular symptoms may have been an expected complication of PsA. The presence of bilateral retinal vasculitis was not in keeping with the acute anterior uveitis typically associated with a seronegative arthritis. Posterior uveitis, with or without macular edema, is described in patients with PsA, but in the largest published series outlining uveitis in PsA, no retinal vasculitis was described (3).
Given the timing of this presentation, a noninfectious adverse effect solely due to infliximab would seem unlikely. There are reports of TNFα inhibitor–related uveitis but not retinitis, and the uveitis resolves on cessation of the TNFα inhibitor (4).
An adverse effect from the anti-TB medications is also possible. There are reports of uveitis with rifabutin, a rifamycin similar to rifampin (5). However, uveitis has not been associated with rifampin except in active TB infection. Ethambutol causes an optic neuritis, a different ocular presentation. No reports of isoniazid, pyrazinamide, or pyridoxine causing retinitis and anterior uveitis were found in the literature. The onset was over 3 months after commencing these medications and 6 weeks after ceasing the ethambutol and pyrazinamide. Also, ocular symptoms were controlled despite the continued treatment with rifampin, isoniazid, and pyridoxine. Therefore, an adverse effect to anti-TB medications was thought to be less likely.
An immune basis for the ocular symptoms was suspected based on the lack of obvious infection or other diagnosis as discussed above. The onset of ocular symptoms followed the reduction of prednisolone. Ocular symptoms resolved with corticosteroid treatment without any sequelae, even with subsequent cessation of corticosteroids. It seems likely that the TB infection and/or infliximab altered the immune response so that when the prednisolone dose was reduced, immune-based ocular symptoms became apparent. From a literature search, two possible immune syndromes, Eales' disease and immune reconstitution inflammatory syndrome (IRIS), localized to the eye could account for the patient's ocular symptoms. In fact, the two different immune syndromes may be the same in this patient.
TB infection without use of TNFα inhibitors has been associated with a retinal vasculitis termed Eales' disease (6). The similarities between this patient and patients with Eales' disease are striking in appearance as well as the association with TB. Eales' disease is an ill-defined syndrome of retinal perivasculitis (6). Tuberculous etiology has been frequently suggested, but it is unclear whether this condition represents active ocular TB or a hypersensitivity reaction manifested on the retina. The appearance of this patient's retina was typical of that described with Eales' disease: perivascular cuffing, a “periphlebitis,” predominantly in the peripheral retina with some exudates and areas of venous occlusion (6). In later stages of the disease, neovascularization can occur with risk of vitreous hemorrhage and blindness. Treatment depends on severity ranging from no pharmacologic therapy to systemic glucocorticoids, and laser photocoagulation for neovascularization or vitrectomy for hemorrhage (6).
Another possible etiology for this patient's eye disease is IRIS. This syndrome is controversial and reports relate to patients with acquired immunodeficiency syndrome receiving highly active antiretroviral therapy (HAART), where the CD4 T cells were significantly depressed and recovered with therapy (7). It is described as paradoxical worsening of preexisting symptoms from infection following the initiation of HAART for the human immunodeficiency virus (HIV). The disorder is thought to be inflammatory rather than infectious, although preexisting infections may have been treated or remain subclinical and later unmasked by the host's regained capacity to mount an immune response (7). Associated infections include Mycobacterium, Cryptococcus, CMV, Pneumocystis jiroveci, and the JC virus. However, there are reports of similar paradoxical inflammatory syndromes in individuals not infected with HIV following treatment for TB or leprosy. The IRIS-like syndromes occurred in individuals not infected with HIV after corticosteroid withdrawal, postchemotherapy neutrophil recovery, post–stem cell transplant, or in withdrawal of immunosuppression in transplant patients with Cryptococcus infection (8–11).
Paradoxical immune reactions occurring in patients treated with infliximab who developed TB have also been reported (12, 13). TB onset varied from 1 month to 2 years with the onset of the IRIS-like syndrome occurring up to 16 months after commencing anti-TB treatment. All patients ceased infliximab at the time of TB diagnosis. Sites included lymph node inflammation after miliary TB, and pulmonary and perianal TB in a patient with Crohn's disease. Management included no treatment, nonsteroidal antiinflammatory medications, or prednisolone, with all of the patients' symptoms eventually resolving. The common factor in the individuals not infected with HIV presenting with a paradoxical inflammatory condition was the association of TB or leprosy. Our patient had an ocular presentation that was very similar to Eales' disease. It is worth considering that Eales' disease may be a subset of IRIS affecting the eye and associated with TB infection, or at least the presence of TB antigens.
It is also well known that TNFα inhibitors can induce ANAs, dsDNA antibodies, antiphospholipid antibodies, and in rare occasions, a lupus-like syndrome (14). The timing of this presentation (4 months after ceasing the infliximab) and the presence of uveitis and retinal vasculitis rather than cotton wool spots would not be consistent with a lupus retinitis. However, what is not clear is why TNFα inhibitors can induce this immune-based phenomena and the mechanism. Suggested theories regarding the mechanism include 1) the promotion of humoral autoimmunity by inhibiting cytotoxic T cell induction that would normally suppress autoreactive B cells, 2) inducing antibody or complement-dependent cellular cytotoxicity of cells expressing TNFα, and 3) enhancing apoptotic processes, thereby promoting the release of autoantigens and the consequent development of autoantibodies to nuclear and cytoplasmic compounds (14).
Furthermore, the eye is a privileged site in terms of immunity (15). This is due to a number of factors that are not completely understood. The eye has a relative physical barrier, the blood–retinal barrier, which limits passage of molecules based on size. Additionally, there is the complex interplay of cytokines and inhibitory molecules that dampen or prevent immune and inflammatory responses within the eye. The third mechanism is thought to involve a deviant response of T cell immunity in the eye with induction of CD4+ and CD8+ regulatory T cells that seem to assist in the immune-dampening process (15). So then why do some patients experience autoimmune uveitis? Perhaps both despite and due to its immune privilege, the eye is subject to autoimmune attack. In the traumatized eye, whether due to injury or infection, antigens may be released to draining lymph nodes that result in a systemic cross-sensitivity to ocular antigens. Alternatively, a peripheral microbial stimulus may create cross-reactivity to ocular antigens. Why this occurs only in some individuals may relate to the genetic predisposition to mount a Th1 cell–dominant adaptive response (15). In this case, M tuberculosis is likely to have been the infectious insult causing cross-reactivity to ocular antigens, and the process was not inhibited either due to a genetic predisposition or other as yet unknown factors. Additionally, it is not clear what role infliximab may have had on the cytokine processes involved in immune privilege.
This case highlights the difficulty in discerning the causative factor(s) of uveitis and retinitis when TB complicates TNFα inhibitor therapy. The unanswered question remains as to whether infliximab contributed to the condition independent to TB reactivation. Regardless, TB is associated with TNFα inhibitor therapy and when both do occur, interpretation of ocular complications is difficult. Although the differential diagnosis is extensive, inappropriate activation of the immune system whether due to the TNFα inhibitor or to the TB infection (active or treated) plus corticosteroids or interactions of any or all of these factors was thought to be the most likely contributing factors in this case. The possibility of aberrant immune reactions should be considered in patients ceasing or taking a break from TNFα inhibitors due to TB complicating therapy.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Racunica had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Acquisition of data. Racunica, Cassidy, Cicuttini, Hall.
Analysis and interpretation of data. Cicuttini, Hall.