Opportunistic Infections Associated with TNF-Α Treatment

REVIEW

Opportunistic infections associated with TNF-α treatment

Robert Orenstein & The therapeutic agents known as TNF-α inhibitors have been widely adopted as effective Eric L Matteson† and standard therapy for many rheumatic diseases. Since their introduction into clinical †Author for correspondence practice, there has been concern that these agents that blunt host immunity to intracellular Mayo Clinic College of pathogens would lead to the development of opportunistic infections. Early reports of Medicine, Division of Rheumatology, Department of extrapulmonary tuberculosis, listeriosis, Pneumocystis jiroveci pneumonia and invasive Internal Medicine, fungal diseases seemed to confirm this association. Prospective and retrospective studies, 200 First Street SW registries, adverse reporting databases and experience from clinical practices indicate at Rochester, MN 55905, USA least a twofold risk of serious bacterial infections with TNFs versus standard DMARDs but data Tel.: +1 507 284 8450; are limited on opportunistic infections (OIs). This article will review the available data on OIs Fax: +1 507 284 0564; [email protected] describing these risks and studies that have been done to reduce that risk.

Opportunistic infections (OIs) are infections innate immune system [2]. It is critical in the for- caused by organisms that ordinarily do not lead to mation and maintenance of granulomas and disease unless the host is immunodeficient, when production of IFN-γ. The complete inhibition they may cause significant morbidity and mortal- of TNF may allow for the dissolution of granu- ity [1]. The predisposition to OIs often relates to lomas and inability to maintain latency. On the an inherited, acquired or medication-induced other hand, less potent inhibition may affect defect in immune function. Often, multiple granuloma formation but not maintenance, thus defects are present due to the underlying illness of allowing for acute infection but not reactivation. the host and its treatment. The host defects TNF is required for defense against intracellu- induced by various immunosuppressive agents lar pathogens, is involved in the Th1-mediated differ and their effects may be multiplicative. One immune response and affects apoptosis. In mice, of the challenges in defining opportunistic TNF induces apoptosis of macrophages, which infections in this group of patients is their under- may provide sanctuary sites to intracellular lying risk for granulomatous and intracellular organisms. infections, even in the absence of therapy. In these mice, the absence of TNF-α may Many OIs are due to intracellular pathogens, lead to dissolution of granulomas that previously whereby the principal host defect lies in the initi- contained mycobacteria [3,4]. Infliximab may ation of the cellular immune response. These induce apoptosis in monocytes potentially inhib- include tuberculosis (TB), atypical mycobacte- iting the memory T-cell immune response for rial infection, salmonellosis, listeriosis, invasive maintenance of granulomas [5,6]. Studies in mice and endemic fungal disease, legionellosis, para- in whom the ability to produce TNF-α was sitic diseases (Strongyloides, Leishmania, Toxo- genetically blocked demonstrated enhanced sus- plasma) and opportunistic viruses such as ceptibility to intracellular pathogens and dissem- cytomegalovirus (CMV) varicella zoster virus ination of previously contained granulomatous (VZV), herpes simplex virus (HSV) and infections [7]. Epstein–Barr virus (EBV). TNF is important in chemokine regulation; TNF-α inhibitor therapy has become widely increasing adhesion molecule expression; mediat- used in the management of autoimmune disor- ing macrophage apoptosis and limiting excessive ders, including rheumatoid arthritis, Crohn’s dis- type 1 immune activation during intracellular ease, ankylosing spondylitis and psoriatic infection [8,9]. It has a direct impact on the innate arthritis. The currently available anti-TNF immune system by affecting specific Toll-like Keywords: adalimumab, agents infliximab, etanercept and adalimumab receptors (TLRs) and decreasing the presentation etanercept, infliximab, opportunistic infections, vary in structure, pharmacology, effect duration, of antigens to T-cells [10]. Although we tend to TNF-α inhibitors specific targets, effectiveness and side effects. generalize about the role of TNF-α, the specific TNF plays several critical roles in the host actions upon inflammatory and immune func- part of immune response: affecting cytokine regulation, tion may vary depending upon the component neutrophil recruitment, T-cell activation and the which is blocked. Blockade of the soluble portion

of TNF activity may inhibit the inflammatory Despite this apparent low risk, there are response without adversely impacting the innate numerous case reports and series of serious and immune response to infection [11,12]. Conversely, opportunistic infections in recipients of TNF blockade of both transmembrane and soluble blocking agents. In an attempt to clarify the con- TNF may blunt both. fusion between serious infection rates reported The risk of infections associated with TNF from clinical trials and that of clinical practice, blockade may also be related to genomic and Salliot and colleagues looked at a large-clinic pharmacodynamic factors. Specific poly- practice of patients with rheumatic diseases morphisms in the TNF-α gene may predispose treated before and after the use of TNF inhibi- to more serious or frequent infections in tors, using the patients themselves as controls. In certain individuals, which may also be dose this study of 709 patients, they found the inci- dependent [13]. dence of serious infections in the TNF inhibitor- Because of TNF-α’s central role in host treated group was increased threefold. They cal- defenses, several clinically important questions culated a number needed to harm (NNH) of impact those receiving TNF-blocking agents: only 14 for the first year of treatment [22]. The • Are there more infections in patients receiving serious infections were evenly distributed these agents? amongst the three agents. Serious skin and soft- tissue infections and pulmonary infections pre- • Which specific infections are most frequently dominated. Bacterial infections accounted for seen? 74.5%, viral for 10.6%, mycobacterial for 4.2%, • Are these infections serious or opportunistic? parasitic for 2.1% and fungal for 2.1% of serious infections. In a multivariate model, the risk fac- Serious infections & TNF inhibitors tors most often associated with serious infections The precise degree to which TNF inhibitors are were previous joint surgery and high previous associated with increased infection risks is diffi- cumulative steroid dose. cult to ascertain due to a host of study design In a meta-analysis of randomized controlled limitations. From experience in clinical practice, trials of infliximab and adalimumab in RA, the it is apparent that patients receiving these agents overall risk of serious infections in TNF inhibi- develop serious and disseminated infections tor-treated patients was twice that of those often early in their use. treated with standard therapies, even excluding While this review focuses on OIs, any infec- the granulomatous infections, although the rela- tion that requires hospitalization, intravenous tive risk of OIs could not be assessed [23]. In a anti-infectives or causes significant morbidity or stratified analysis according to dose, higher doses death is defined as a serious infection. The fre- of adalimumab and infliximab appeared to be quency and types of infection may vary with the associated with an increased risk. These authors host’s underlying disease. Patients with RA have calculated the NNH was 59 within a treatment an increased baseline risk of serious infection period of 3–12 months [23]. 1.8-times that of non-RA patients [14]. In a large Case reports and passive surveillance provide primary care cohort of patients with polyarthri- another window to the infectious morbidity asso- tis, the overall infection incidence was greater ciated with these agents. In a review of the US than 2.5-times that of the general population. FDA’s MedWatch Adverse Event Reporting Sys- Smoking, corticosteroid use and a positive rheu- tem for reports of granulomatous infections asso- matoid factor were independent predictors of ciated with infliximab or etanercept use, from infection-related hospitalization. Those with all 1998 to 2002, a total of 639 granulomatous three risk factors were seven-times more likely to infections were reported among 197,000 patients be hospitalized [15]. who had received infliximab and 113,000 etaner- In clinical trials of the three TNF inhibitors, cept respectively [24,25]. Of these patients, 40% the observed serious infection event (SIE) rate is received concomitant immunosuppression with two to six serious infections/100 patient-years either methotrexate or corticosteroids. The over- [16–21]. The incidence of serious infections was all rate of granulomatous infection was 4.1 per 100 patient-years (1%) with etanercept 129/100,000 treated patients for infliximab and [16], four per 100 patient-years with adalimumab 60/100,000 for etanercept. The most frequently and both similar to placebo [17]. The overall seri- reported granulomatous infections were tubercu- ous infection risk from most clinical trials is losis (TB) followed by histoplasmosis, candidia- approximately five events/100 patient-years. sis, listeriosis, nontuberculous mycobacteria and

568 Future Rheumatol. (2007) 2(6) futurefuture sciencescience groupgroup Opportunistic infections associated with TNF-α treatment – REVIEW

aspergillosis. The median time to onset of infec- Bacterial opportunists tion was 40 days for the infliximab group and One of the most commonly reported intra- 236 days for etanercept [24,25]. Of infliximab- cellular bacterial infections in these patients is associated infections, 72% occurred within the listeriosis, occurring in three per 10,000 patient- first 90 days of treatment, compared with 28% years [33]. Listeria monocytogenes is a facultative for etanercept [24,25]. The same was true for TB, intracellular food-borne pathogen that affects with the median time to onset of 190 days in the hosts with impaired cellular immunity. TNF-α infliximab arm and 511 days in the etanercept protects mice against listerial infection and TNF group; 44% of the infliximab-associated TB cases blockade may lead to overwhelming listerial occurred within 90 days of treatment versus 10% infection. Several reports of disseminated listeri- for etanercept [24,26]. Granulomatous infections osis and deaths have been reported in persons were three-times more frequent in the infliximab receiving TNF blockade [33,34]. These cases group versus etanercept. occurred between 4 and 290 days after receipt of Several large patient registries were established the first infliximab infusion with most occurring in the late 1990s in Sweden, Germany, Spain, following the sixth infusion [35]. Because this France and Great Britain to monitor the use, subset at risk cannot be easily identified and the effectiveness and safety of these agents [26–29]. overall event rate is low, antimicrobial prophy- The British Society of Rheumatology (BSR) Bio- laxis for listeria cannot be routinely recom- logics Registry is the largest reported prospective mended. Patients receiving TNF antagonists observational cohort study of TNF-inhibitor use should be counseled to avoid soft cheeses, in RA patients. Rates of serious infections were unpasteurized milk products and ‘ready-to-eat’ compared, including site-specific and bacterial meat products to decrease their exposure to intracellular infections in 7664 patients who Listeria. Disseminated salmonellosis is another received a TNF inhibitor for RA with a compar- food-borne opportunistic infection occasionally ison group of 1354 on only DMARDS. A total reported in patients on anti-TNF therapy [36,37]. of 525 serious infections occurred in the TNF- treated group compared with 56 in the DMARD Mycobacterial infections group with a median follow-up of approximately Both atypical and tuberculous mycobacterial 1 year. The incidence rate ratio (IRR) was infections have been described with the TNF 1.28 [30]. When controlled for disease severity, inhibitors. In early experiments, it was found baseline steroid use and smoking there was no that TNF-α deficient mice had poorly formed difference in infection risk. However, when granulomas and extensive necrosis when infected adjusted for just the first 90 days of treatment with mycobacterium tuberculosis [3]. Thus, it with the TNF inhibitor, the adjusted incidence was anticipated that disseminated mycobacterial rate rose to 4.6 [31]. There were 19 intracellular infections might occur in humans receiving these (opportunistic) bacterial infections – ten TB, agents. As of December 2004, 82 cases with two Legionella, three Listeria, one M. fortuitum etanercept (27.1/100,000 patients), 633 and three Salmonella – all of which occurred (33/100,000 patients) with infliximab and 15 only in the TNF inhibitor-treated group [30,31]. (27.1/100,000 patients) with adalimumab had In the prospective cohort of RA patients from been reported worldwide [16–18,38]. Germany (the Rheumatoid Arthritis Observation Tuberculosis incidence rates from MedWatch of Biologic Therapy [RABBIT] cohort), infec- data are: 54 cases/100,000 patients treated with tions occurred in 204/1529 (13%) of patients infliximab, and 28/100,000 patients treated with overall; 15% in the etanercept, 21% in the inflixi- etanercept [24,25] compared with the baseline rate mab, and 6% among the controls who received of TB in the USA of 6.2/100,000 [3,39]. Several of DMARDS alone [29]. The relative risk of serious the early clinical trials enrolled patients in coun- infections was 2.7–2.8-times higher in the TNF tries with high endemic TB rates. The incidence inhibitor-treated group compared with those rate of TB in trials of infliximab was 0.4%. TB treated with DMARDS alone. Another study risk in patients receiving adalimumab appears to using administrative claims data found the risk of be dose-dependent; the calculated incidence rate hospitalization for infection was twofold higher in of TB in patients receiving adalimumab prior to the TNF-treated group and fourfold higher in the PPD screening was 1.3/100,000 patient-years and initial 6 months of treatment as compared with 0.19/100,000 patient-years after screening [38]. the methotrexate (MTX) alone group [32]. Most of The median time to TB onset was 167 days and these infections were pneumonias and cellulitis. 65% cases were extrapulmonary [21,38,40]. futurefuture sciencescience groupgroup www.futuremedicine.com 569 REVIEW – Orenstein & Matteson

In a review of the US National Databank for the IFN-γ ELISPOT assay, (QuantiFERON® Rheumatic Diseases of 10,782 patients with RA TB-Gold Cellestis Ltd.), has not yet been vali- during 1998–1999 and 6460 infliximab-treated dated in this population. However, it appears to patients during 2000–2002, the TB incidence have improved sensitivity in immunosuppressed rate prior to TNF-inhibitor use was hosts (including HIV) and might prove useful in 6.2/100,000 patient-years versus 52.5/100,000 those who have received Bacillus of Calmette and patient-years with infliximab use [41]. The US Guérin (BCG) immunization [47]. baseline TB incidence rate for years 1999 and If tuberculosis develops while receiving a TNF 2000 was 6.4 and 5.8 per 100,000 patient-years, inhibitor, it is recommended that the anti-TNF respectively [41]. These numbers differ in Europe agent be discontinued and a standard course of where endemic TB rates are higher. In an analy- antituberculous therapy be administered [42,48]. sis of data from 2000 to 2004 in Spain, the base- Anti-TNF therapy should not be resumed until line TB rate was 21/100,000 patients and the the completion of therapy if possible. If the rate among 4102 patients treated with etaner- underlying disease is progressive, TNF inhibi- cept, adalimumab and infliximab was 20-fold tion should be withheld for at least 2 months higher than the background rate [26]. Following after antituberculous therapy has been started the implementation of recommended TB and there has been an adequate response. screening, rates fell 78% from 522/ to Patients who then resume TNF inhibitors must 117/100,000 [26]. be closely monitored for disease recurrence. Phy- In most patients treated with TNF inhibitors, sicians should also be aware of the potential for a TB presents as a reactivation disease. In cases paradoxical worsening of TB after initial associated with infliximab, the median age was improvement to anti-TB therapy following dis- 57 years, and the median time to onset was continuation of TNF inhibitors [49]. This may be 12 weeks versus 46 weeks for etanercept, and similar to the immune reconstitution observed 30 weeks for patients on adalimumab [24,25,38,39]. in antiretroviral treated HIV patients [50]. Only a small number of patients reported a past history of TB exposure. Two-thirds of the Histoplasmosis patients had extrapulmonary disease (EPTB) Macrophage apoptosis plays an important role in and 25% had disseminated disease (DTB) [38,42]. control of endemic fungal infections. Resolution This contrasts with the usual 18% EPTB and of histoplasma infection is dependent upon an 2% DTB in non-HIV-associated TB. The diag- effective cellular immune response. In vitro inf- nosis of TB in these patients may be difficult due liximab inhibits T-cell proliferation when alveo- to atypical and extrapulmonary presentations lar macrophages are the antigen-presenting cells and in some circumstances poor granuloma and results in a significant reduction in IFN-γ formation [38]. production [51,52]. In murine models, TNF-α As latent TB infection may present a signifi- blockade after infection results in fatal dissemi- cant risk to patients on TNF inhibitors, US nated histoplasmosis [53]. A total of 39 cases of guidelines recommend screening for TB risk with histoplasmosis associated with infliximab and a focused history, purified protein derivative three with etanercept had been reported to the (PPD) skin test and a chest radiograph. A PPD MedWatch as of September 2002 and additional skin test of more than 5 mm induration should cases have been noted since [24,25,54,55]. be considered as evidence of tuberculous infec- Histoplasmosis presents from 1 week to tion. In these patients, active TB should be 6 months after initiating anti-TNF treatment. excluded and, if so, receive preventive therapy for Patients may present with an acute and fulmi- latent TB infection (LTBI) for 9 months. Two- nant course characterized by fever, malaise, step TB skin testing is not recommended. cough, dyspnea and interstitial pneumonitis [54]. Because many patients qualifying for TNF-inhib- The infection may mimic the underlying inflam- itor therapy have been on immunosuppression, matory disease being treated (Crohn’s, RA) [51]. which increases the risk of a falsely negative PPD, Patients treated with infliximab may be at higher a negative test needs to be interpreted with cau- risk for developing histoplasmosis than those tion [43]. Screening of at-risk patients appears to treated with etanercept or adalimumab [24]. have reduced the incidence rate of TB [26,44,45]. In Although most cases of histoplasmosis RA patients treated with steroids who had a posi- during anti-TNF therapy have been diagnosed tive PPD skin test, the incidence of TB was via bronchoalveolar lavage or lung biopsy, the reduced by 70% by this strategy [46]. The role of urine histoplasma antigen may be a potentially

570 Future Rheumatol. (2007) 2(6) futurefuture sciencescience groupgroup Opportunistic infections associated with TNF-α treatment – REVIEW

useful tool in those suspected of disseminated reduces the morbidity from coccidioidomycosis disease and for follow-up [51]. Its role in moni- [60]. It is recommended that patients receiving toring at risk patients in highly endemic areas is TNF inhibitors from endemic areas for coccidi- not established. oidomycosis have a screening chest radiograph Patients who develop histoplasmosis should and coccidioides serologies. Unlike histoplas- stop their TNF inhibitors until they have com- mosis, targeted antifungal prophylaxis with flu- pleted their histoplasma treatment. Initial ther- conazole has been effective in transplant apy should include a liposomal formulation of recipients with a history of coccidioidomycosis amphotericin B for 1–2 weeks followed by itra- infection or positive coccidioidomycosis serol- conazole at 200–400 mg daily for 1 year [56]. ogy in endemic areas [60]. More data are needed Itraconazole levels should be monitored to to determine whether fluconazole should be ensure efficacy. If the patient had a positive uri- considered in patients receiving TNF inhibitors nary histoplasma antigen, this should be fol- who are at increased risk for this infection. lowed to ensure adequate therapy and monitor Those who acquire infection should discon- for relapse. Since histoplasmosis is a rare event, tinue their anti-TNF therapy and be treated even in chronically immunosuppressed patients with azole therapy. Whether anti-TNF therapy living in endemic areas, and because screening can be safely re-instituted with or without azole by serology has not proven predictive of who secondary prophylaxis is unknown and not will acquire the disease, the role of screening advised in cases of CNS coccidioidomycosis and prophylaxis remains unclear [57,58]. It is where the risk for relapse is high [61]. prudent to advise patients receiving TNF-α inhibitors to avoid activities that increase their Pneumocystis jiroveci pneumonia risk of exposure to histoplasma, such as explor- Though rarely reported in patients only on TNF ing caves or cleaning chicken coops and to wear inhibitors, there are potential mechanisms by a mask when working in soil. Further data are which this may occur [62,63]. Cases of Pneumo- needed to optimize preventive strategies. cystis pneumonia (PCP) have occurred in patients on infliximab and five on etanercept [64–66]. The Coccidioidomycosis FDA-Adverse Event Reporting System (AERS) Like histoplasmosis, this granulomatous infec- safety database contained 84 cases of PCP in tion has been reported in endemic areas in patients receiving infliximab between January patients treated with TNF inhibitors. The esti- 1998 and December 2003. Most of these mated annual risk of coccidioidomycosis in Tuc- patients were on other concomitant immuno- son, AZ, USA is 3%. The majority of these suppressives, including prednisone, MTX, aza- infections are asymptomatic. In patients with thioprine, 6-mercaptopurine and cyclosporin A, impaired cell-mediated immunity, 7% per year all of which have previously been associated with are symptomatic [59]. The risk of developing PCP. The mean time to onset of PCP was symptomatic coccidioidomycosis on TNF 21 days; 27% died [66]. The time to onset was inhibitors may be increased by fivefold com- within 1 month of initiation of infliximab and pared with patients suffering from inflammatory 2 months of etanercept therapy, suggesting these arthritis [59]. The cumulative incidence of coc- patients were already moderately immunosup- cidioidomycosis in RA, reactive arthritis and pressed. At present, PCP prophylaxis is not war- patients with psoriatic arthritis is 1%. ranted unless other risk factors for this infection Coccidioidomycosis in TNF inhibitor-treated are present. patients mimics TB with a median onset at week 7 [59]. Of the 13 patients with coccidioidomyco- Other reported infections sis reported by Bergstrom, 12 received infliximab A variety of other opportunists, including Toxo- and one etanercept. Of 12 patients, 11 were tak- plasma, Nocardia, Cryptococcus, atypical myco- ing MTX. All 13 developed pneumonia, four bacteria, Leishmania, Bartonella, Legionella, had disseminated disease, 11 of 13 had positive Aspergillus, Sporothrix and Microsporidiosis have coccidioidal serology, five were hospitalized and been reported in patients receiving TNF two died [59]. inhibitors [67]. Screening patients receiving TNF inhibitors In the few reported cases of aspergillosis, the in endemic areas for coccidioidomycosis is con- disease was localized and invasive within the troversial. In transplant patients, screening in lungs; its onset was rapid, within 5 days and cul- hyperendemic areas and use of azole prophylaxis ture positive within 1 week. Cases have occurred futurefuture sciencescience groupgroup www.futuremedicine.com 571 REVIEW – Orenstein & Matteson

in patients on infliximab therapy alone [68,69]. therapy is unknown. For those requiring lifelong Several cases of unusual and severe complications immunosuppression, lifelong antivirals may such as Brachiola myositis [70] and disseminated be needed. sporotrichosis [71] illustrate the need for appro- Hepatitis C infection may also coexist with priate counseling regarding exposure risks, and a the patient’s underlying disease, such as RA or meticulous search for infections prior to intensi- psoriasis. Hepatitis C screening is prudent, fication of immunosuppression in patients with although several reports indicate no worsening, inflammatory diseases. and potential benefit of etanercept in some patients with hepatitis C infection [75,80]. Viral diseases A variety of opportunistic and nonopportunistic Conclusion viral diseases have been reported in patients Patients receiving TNF inhibitors may be at risk receiving these drugs including CMV [72], for opportunistic infections. Some of these infec- hepatitis B and C [73–75], EBV [76], recurrent tions may be prevented by personal protective herpes simplex, varicella zoster [77,78] and severe measures, some via immunization and others by molluscum contagiosum [79]. It is not clear that antimicrobial prophylaxis (Table 1 & Box 1). A pre- these occur more frequently in patients receiv- treatment assessment of infectious risks and ing TNF inhibitors than with other forms of directed risk reduction may be reasonable. The immunosuppression. development of monitoring systems to deter- CMV disease is rarely encountered [72]. The mine the risk factors and absolute risk for OIs diagnosis of CMV is usually by tissue biopsy, will help to better target preventive measures pp65 antigenemia or quantitative PCR. It has including antimicrobial prophylaxis. Current been suggested that patients on TNF inhibitors recommendations based principally upon expert who demonstrate elevated quantitative measures opinion [1,81,82] suggest a pretreatment chest of CMV receive treatment with antiviral ther- radiograph and PPD skin test only. It is recom- apy; however, there is a lack of data to support mended that all age-appropriate immunizations this recommendation [72]. be given prior to initiation of anti-TNF therapy. The risk for zoster is increased with pred- Live virus vaccinations should be avoided [83]. As nisone doses above 10 mg daily, the use of cyclo- patients with rheumatic disease treated with phosphamide and the combination of both a TNF inhibitors are at increased risk for respira- DMARD and a TNF inhibitor. There appears to tory infections and because of increased morbid- be no higher risk in patients on TNF inhibitors ity due to pneumococcal disease, the annual alone [78]. Severe disease is treated with intrave- inactivated influenza vaccine, and pneumococcal nous acyclovir whereas mild or localized disease vaccine are recommended [84,85]. may be treated with an oral agent such as vala- Patients need to be aware that their physician cyclovir, famciclovir or acyclovir. Since there are may recommend interruption of therapy for no data on use of zoster vaccine (Zostavax®) in serious infections. How long to hold them immunosuppressed hosts, this new vaccine remains unclear, especially for OIs. Most author- should not be used. ities suggest holding these agents until the OI Patients who are candidates for TNF-inhibi- has been cleared. tor therapy should be screened for hepatitis B and immunized prior to immunosuppression. Future perspective For those with chronic hepatitis, hepatology As this review illustrates, there remain many consultation is needed. The European Associa- unanswered questions regarding the infectious tion for the Study of the Liver (EASL) recom- risks associated with these therapies. There will mends that antiviral therapy begins 2–4 weeks be a proliferation of newer agents affecting other prior to the start of a TNF inhibitor in patients immunomodulatory pathways and therapeutic with inactive hepatitis B [75]. The duration of indications will continue to expand. As in

Table 1. Precautions when using TNF-blocker therapy. Contraindications Use precaution Active Hep B or C, or treated with Hepatitis B/C, treated Liver cirrhosis HIV, treated Active serious infection/sepsis active HSV and HIV Avoid live vaccines

572 Future Rheumatol. (2007) 2(6) futurefuture sciencescience groupgroup Opportunistic infections associated with TNF-α treatment – REVIEW

Box 1. Opportunistic infections: preventive measures. patients with RA and Crohn’s as well as psoriasis and various forms of spondyloparthopathies Indicated: routine vaccinations with inactivated vaccine or toxoid may have inadequate long-term responses to • Diphtheria, tetanus and pertussis individual TNF inhibitors, we will begin to see • Hepatitis A and B the use of sequential or combination therapy •Influenza with other biological agents. Whether the infec- • Haemophilus influenzae type B tious risks are the same with sequential or combination therapies will need to be determined. • Inactivated polio We will begin to explore the risk across the Contraindicated diverse populations receiving these agents. Pre- • Live-attenuated vaccines such as measles, mumps, rubella, varicella, treatment risk assessment will be standardized rotavius, varicella (Varivax), zoster (Zostavax), inhaled influenza (FluMist), and preventive measures will be provided for yellow fever, oral typhoid, oral polio those at highest risk. Lastly, our experience thus far is limited with respect to long-term adverse transplantation, future research will help us bet- effects. We will continue to monitor for the ter define those at greatest risk of infectious development of immunomodulatory virus morbidity and predict the timing of such events. induced malignancies. Though at first this will involve traditional risk- factor analysis, this area will be ripe for genomic Financial & competing interests disclosure exploration. We will likely be able to search for Dr Matteson has received grant support from Wyeth, Amgen specific polymorphisms that might place and Centocor and has served in an advisory capacity to patients at risk for infectious complications. Amgen and Centocor. The authors have no other relevant Future studies will also better define the phar- affiliations or financial involvement with any organization macologic and pharmacogenomic parameters or entity with a financial interest in or financial conflict which might reduce infectious risks. We will with the subject matter or materials discussed in the manu- better define the impact of specific agents on script apart from those disclosed. infection risk, including the role of TNF-recep- No writing assistance was utilized in the production of tor specificity. Since a significant proportion of this manuscript.

Executive summary Risk of opportunistic infection with anti-TNF treatment

•The TNF-α inhibitors have begun to revolutionize the treatment of inflammatory diseases. Like their predecessors, these new agents carry an increased risk of serious and granulomatous infections, the timing and frequency of which may vary between the specific agents. • Patients receiving TNF inhibitors will be at variable risk for opportunistic infection, depending upon their underlying disease and prior immunosuppression. • Granulomatous and intracellular infections may occur with TNF inhibition, and the risks for opportunistic infections are cumulative with multiple immunosuppressives. • Infections may present atypically and be disseminated at time of diagnosis. The first 3–6 months appear to be the period of greatest risk. Measures for infection risk reduction • Patients in whom anti-TNF therapy is being initiated should undergo pretreatment screening for exposure risks, including TB. • Appropriate immunizations including pneumonia vaccination, seasonal influenza vaccination and hepatitis B vaccination in patients at risk for this infection should be performed prior to initiation of therapy. • Patients with infection should undergo an aggressive diagnostic evaluation that considers the possibility of opportunistic infection. Future perspective • Many practical questions remain unanswered such as the most appropriate screening, optimal timing of vaccines, need for discontinuation of TNF inhibitors, and role of opportunistic infection monitoring and prophylaxis.

futurefuture sciencescience groupgroup www.futuremedicine.com 573 REVIEW – Orenstein & Matteson

Bibliography 12. Zalevsky J, Secher J, Ezhevsky SA et al.: meta-analysis of rare harmful effects in Papers of special note have been highlighted as Dominant negative inhibitors of soluble randomized controlled trials. JAMA 295, either of interest (•) or of considerable interest (••) TNF attenuate experimental arthritis 2275–2285 (2006). to readers. without suppressing innate immunity to •• Systematic review of the risks associated 1. Furst DE, Breedveld FC, Kalden JE et al.: infection. J. Immunol. 179, 1872–1883, with infliximab and adalimumab. The key Updated consensus statement on biological (2007). finding is the twofold increased risk of agents, specifically tumor necrosis factor α •• Describes why selective TNF inhibition serious infections and a number needed to blocking agents and interleukin-1 receptor may decrease infectious morbidity. harm of 59 within the first 3–12 months antagonist, for the treatment of rheumatic 13. Westhovens R, Yocum D, Han J et al.: The of treatment. diseases, 2005. Ann. Rheum. Dis. 64(Suppl. safety of infliximab, combined with 24. Wallis RS, Broder M, Wong J, IV), IV2–IV14 (2005). background treatments, among patients Beenhower D: Granulomatous infections 2. Dinarello C: Anticytokine therapeutics and with rheumatoid arthritis and various due to tumor necrosis factor blockade: infections. Vaccine 21, S24–S34 (2003). comorbidities. A large randomized, placebo- correction. Clin. Infect. Dis. 39, 1254–1255 3. Gardam MA, Keystone EC, Menzies R controlled trial (The START trial). Arthritis (2004). et al.: Anti-tumour necrosis factor agents Rheum. 54, 1075–1086 (2006). 25. Wallis RS, Broder M, Wong J, and tuberculosis risk: mechanism of action 14. Doran MF, Crowson CS, Pond GR, Beenhower D: Granulomatous infections and clinical management. Lancet Infect. Dis. O’Fallon WM, Gabriel SE: Frequency of associated with tumor necrosis factor 3, 148–155 (2003). infection in patients with rheumatoid antagonists. Clin. Infect. Dis. 38, 4. Lin PL, Pawar S, Zinovik A et al.: TNF and arthritis compared with controls. Arthritis 1261–1265 (2004). chemokine expression in granulomas on Rheum. 46, 2287–2293 (2002). • Original descriptive review of the US FDA non-human primates infected with 15. Franklin J, Lunt M, Bunn D, Symmons D, adverse event database describing M. tuberculosis. Presented at: Annual Silman A: Risk and predictors of infection granulomatous infections in patients Meeting of the Infectious Diseases Society of leading to hospitalization in a large primary- receiving TNF inhibitors. America. Boston, MA, USA (Abstract) care-derived cohort of patients with 26. Carmona L, Gomez-Reino JJ, (2004). inflammatory polyarthritis. Ann. Rheum. Rodriguez-Valverde, BIOBADASER 5. Wallis RS, Ehlers S: Tumor necrosis factor Dis. 66, 308–312 (2007). Group: Effectiveness of recommendations to and granuloma biology: explaining the 16. Enbrel®, Package insert. Immunex prevent reactivation of latent tuberculosis differential infection risk if etanercept and Corporation, CA, USA. infection in patients treated with tumor infliximab. Semin. Arthritis Rheum. 17. Humira®, Package insert. Abbott necrosis factor α antagonists. Arthritis 34(Suppl. 1), 34–38 (2005). Laboratories, Il, USA. Rheum. 52, 1766–1772 (2005). 6. Dinarello CA: Differences between anti- 18. Remicade®, Package insert. Centocor, Inc., 27. Dixon WG, Hyrich KL, Watson K, tumor necrosis factor α monoclonal PA, USA. Symmons DPM, Silman AJ and the BSR antibodies and soluble TNF-receptor 19. Sandborn WJ, Loftus EV: Balancing the Biologics Register: Serious infection rates in antagonists in host defense impairment. risks and benefits of infliximab in the patients receiving biologic therapy in the J. Rheum. Suppl. 74, 40–47 (2005). treatment of inflammatory bowel disease. United Kingdom: results from the BSR 7. Deepe GS, Gibbons RS: T cells require Gut 53, 780–782 (2004). Biologics Register. Presented at: EULAR. tumor necrosis factor-α to provide 20. Colombel JF, Loftus EV Jr, Tremaine WJ Vienna, Austria, June 8–11 protective immunity in mice infected with et al.: The safety profile of infliximab in (Abstract OP0094) (2005) . Histoplasma capsulatum. J. Infect. Dis. 193, patients with Crohn’s disease: the Mayo 28. Tubach F, Salmon-Ceron D, Ravaud P, 322–330 (2006). Clinic experience in 500 patients. Mariette X, RATIO study group: The 8. Keane J: TNF-blocking agents and Gastroenterology 126, 19–31 (2004). RATIO observatory: French registry of tuberculosis: new drugs illuminate an old 21. Klareskog L, van der Heijde D, de Jager JP, opportunistic infections, severe bacterial topic. Rheumatology 44, 714–720 (2005). Gough A, Kalden J, Malaise M et al., the infections, and lymphomas complicating 9. Zganiacz A, Santosuosso M, Wang J et al.: TEMPO study investigators: Therapeutic anti-TNF α therapy. Joint Bone Spine 72, TNF-α is a critical negative regulator of type effect of the combination of etanercept and 456–460 (2005). 1 immune activation during intracellular methotrexate compared with each treatment 29. Listing J, Stranfeld A, Kary S et al.: bacterial infection. J. Clin. Invest. 113, alone in patients with rheumatoid arthritis: Infections in patients with rheumatoid 401–413 (2004). double blind randomized controlled trial. arthritis treated with biologic agents. 10. Netea MG, Radstake T, Joosten LA et al.: Lancet 363, 675–681 (2004). Arthritis Rheum. 52, 3403–3412 (2005). Salmonella septicemia in rheumatoid 22. Salliot C, Gossec L, Ruyssen-Witrand A 30. Dixon WG, Watson K, Lunt M et al.: Rates arthritis patients receiving anti-tumor et al.: Infections during tumor necrosis of serious infection, including site-specific necrosis factor therapy. Association with factor α blocker therapy for rheumatic and bacterial intracellular infection in decreased interferon γ production and Toll- diseases in daily practice: a systematic rheumatoid arthritis patients receiving anti- like receptor 4 expression. Arthritis Rheum. retrospective study of 709 patients. tumor necrosis factor therapy. Arthritis 48, 1853–1857 (2003). Rheumatology 46, 227–234 (2007). Rheum. 54, 2368–2376 (2006). 11. Alexopoulou L, Kranidioti K, Xantharea S 23. Bongartz T, Sutton AJ, Sweeting MJ, 31. Dixon WG, Symmons DPM, Lunt KD et al.: Transmembrane bound TNF protects Buchan I, Matteson EL, Montori V: Anti- et al.: Serious infection following anti-tumor mutant mice against intracellular bacterial TNF antibody therapy in rheumatoid necrosis factor α therapy in patients with infection and autoimmune diseases. Eur. arthritis and the risk of serious infections rheumatoid arthritis. Arthritis Rheum. 56, J. Immunol. 36, 2768–2780 (2007). and malignancies. Systematic review and 2896–2904 (2007).

574 Future Rheumatol. (2007) 2(6) futurefuture sciencescience groupgroup Opportunistic infections associated with TNF-α treatment – REVIEW

•• Largest prospective cohort study showing a recommendations for assessing risk and for complicating immunotherapy with tumor 4.6-fold relative risk of infection within the managing Mycobacterium tuberculosis necrosis factor α antagonist’s infliximab and first 3 months of TNF-inhibitor use. infection and disease in patients due to start etanercept. Arthritis Rheum. 46, 2565–2570 32. Curtis JR, Patkar N, Xie A et al.: Risk of anti-TNF-α treatment. Thorax 60, 800–805 (2002). serious bacterial infections among (2005). 55. Nakelchik M, Mangino JE: Reactivation of rheumatoid arthritis patients exposed to 44. Gomez-Reino JJ, Carmona L, Valverde VR, histoplasmosis after treatment with tumor necrosis factor α antagonists. Arthritis Mola EM, Montero MD: Treatment of infliximab. Am. J. Med. 112, 78–79 (2002). Rheum. 56, 1125–1133 (2007). rheumatoid arthritis with tumor necrosis 56. Wheat JL, Freifeld AG, Kleiman MB et al.: 33. Gluck T, Linde HJ, Scholmerich J, factor inhibitors may predispose to Clinical practice guidelines for the Muller-Ladner U, Fiehn C, Bohland P: significant increase in tuberculosis risk: a management of patients with Anti-tumor necrosis factor therapy and multicenter active-surveillance report. histoplasmosis. 2007 update from the Listeria monocytogenes infection: report of Arthritis Rheum. 48, 2122–2127 (2003). Infectious Diseases Society of America. Clin. two cases. Arthritis Rheum. 46, 2255–2257 45. Perez JL, Kupper H, Spencer-Green GT: Infect. Dis. 45, 807–825 (2007). (2002). Impact of screening for latent TB prior to 57. Vail GM, Young RS, Wheat LJ, Filo RS, 34. Slifman NR, Gershon SL, Lee KH, initiating anti-TNF therapy in North Cornetta K, Goldman M: Incidence of Edwards ET, Braun MM: Listeria America and Europe. Presented at: The histoplasmosis following allogeneic bone monocytogenes infection as a complication of Annual Congress of the European League marrow transplant or solid organ transplant treatment with tumor necrosis factor α Against Rheumatism. Vienna, Austria, in a hyperendemic area. Transpl. Infect. Dis. neutralizing agents. Arthritis Rheum. 48, 8–11 June 8–11 (Abstract OP0093) (2005). 4, 148–151 (2002). 319–324 (2003). 46. Hernandez-Cruz B, 58. McKinsey DS, Wheat LJ, Cloud GA et al.: 35. Ellerin T, Rubin RH, Weinblatt ME: Ponce-de-leon-Rosales S, Itraconazole prophylaxis for fungal Infections and anti-tumor necrosis factor α Sifuentes-Osornio J, infections in patients with advanced human therapy. Arthritis Rheum. 11, 3013–3022 Ponce-de-leon- Garduno A, immunodeficiency virus infection: (2003). Diaz-Jouanen E: Tuberculous prophylaxis in randomized, placebo-controlled double 36. Cornillie F, Shealy D, D’Haens G et al.: patients with steroid treatment and systemic blind study. Clin. Infect. Dis. 28, Infliximab induces potent anti- rheumatic disease. A case control study. 1049–1056 (1999). inflammatory and local immunomodulatory Clin. Exp. Rheumatol. 17, 81–87 (1999). 59. Bergstrom L, Yocum DE, Ampel NM et al.: activity but no systemic immune 47. Lalvani A, Richeldi L, Kunst H: Interferon γ Increased risk of coccidioidomycosis in suppression in patients with Crohn’s disease. assays for tuberculosis. Lancet Infect. Dis. 5, patients treated with tumor necrosis factor α Aliment. Pharmacol. Ther. 15, 463–473 322–323 (2005). antagonists. Arthritis Rheum. 50, (2001). 48. American Thoracic Society, Centers for 1959–1966 (2004). 37. Fu A, Bertouch JV, McNeil HP: Diseases Control, Infectious Diseases 60. Blair JE, Douglas DD, Mulligan DC: Early Disseminated Salmonella typhimurium Society of America: Treatment of results of targeted prophylaxis for infection secondary to infliximab treatment. tuberculosis. Am. J. Respir. Crit. Care Med. coccidioidomycosis in patients undergoing Arthritis Rheum. 50, 3049 (2004). 167, 603–662 (2003). orthotopic liver transplantation within an 38. Claudepierre P, Deprez X, Goupille P et al., 49. Vidal CG, Fernandez SR, Lacasa JM et al.: endemic area. Transpl. Infect. Dis. 5, 3–8 Club Rheumatismes et Inflammation, Paradoxical response to antituberculous (2003). French Society of Rheumatology: Anti-TNF therapy in infliximab treated patients with 61. Dewsnup DH, Galgiani JN, Graybill JR α therapy and safety monitoring: clinical disseminated tuberculosis. Clin. Infect. Dis. et al.: Is it ever safe to stop azole therapy for tool guide. Joint Bone Spine 72(Suppl. 1), 40, 756–759 (2005). Coccidioides immitis meningitis? Ann. Intern. S1–S58 (2005). 50. Narita M, Ashkin D, Hollender E, Med. 124, 305–310 (1996). 39. Fleischman R, Yocum D: Does safety make Pitchenik A: Paradoxical worsening of 62. Thomas CF, Limper AH: Pneumocystis a difference in selecting the right TNF tuberculosis following antiretroviral therapy pneumonia. N. Engl. J. Med. 350, antagonist? Arthritis Res .Ther. 6(Suppl. 2), in patients with AIDS. Am. J. Respir. Crit. 2487–9248 (2004). S12–S18 (2004). Care Med. 158, 157–161 (1998). 63. Ding K, Shibui A, Yang Y, Takamoto M, 40. Winthrop KL, Siegel JN, Jereb J, Taylor Z, 51. Wood KL, Hage CA, Knox KS et al.: Matsuguchi T, Sugame K: Impaired Iademarco MF: Tuberculosis associated with Histoplasmosis after treatment with anti- recognition by Toll-like receptor 4 is therapy against tumor necrosis factor α. TNF therapy. Am. J. Crit. Care Med. 167, responsible for exacerbated murine Arthritis Rheum. 52, 2968–2974 (2005). 1279–1283 (2003). Pneumocystis pneumonia. Microbes Infect. 7, 41. Wolfe F, Michaud K, Anderson J, 52. Hage CA, Wheat LJ, Twigg HL, Knox KS: 195–203 (2005). Urbansky K: Tuberculosis infection in Infliximab does not affect dendritic cells’ 64. Tai TL, O’Rourke KP, McWeeney M, patients with rheumatoid arthritis and the mediated lymphoproliferative response to Burke CM, Sheehan K, Barry M: effect of infliximab therapy. Arthritis Rheum. Histoplasma capsulatum. Clin. Infect. Dis. 41, Pneumocystis carinii pneumonia following a 50, 372–379 (2004). 1685–1687 (2006). second infusion of infliximab. Rheumatology 42. Centers for Disease Control and Prevention: 53. Smith JG, Williams DM, Graybill JR: 41, 951–952 (2002). Tuberculosis associated with blocking agents Tumor necrosis factor-α plays a role in host 65. Velayos FS, Sandborn WJ: Pneumocystis against tumor necrosis factor-α – California, defense against Histoplasma capsulatum. carinii pneumonia during maintenance anti- 2002–2003. MMWR 53, 683–686 (2004). J. Infect. Dis. 162, 1349–1353 (1990). tumor necrosis factor α therapy with 43. Ormerod LP, British Thoracic Society 54. Lee JH, Slifman NR, Gershon SK et al.: infliximab for Crohn’s disease. Inflamm. Standards of Care Committee: BTS Life-threatening histoplasmosis Bowel Dis. 10, 657–660 (2004).

futurefuture sciencescience groupgroup www.futuremedicine.com 575 REVIEW – Orenstein & Matteson

66. Kaur N, Mahl TC: Pneumocystis jiroveci rheumatoid arthritis treated with infliximab 82. Ledingham J, Wilkinson C, Deighton C: (carinii) pneumonia after infliximab and low dose methotrexate. Ann. Rheum. British Thoracic Society recommendations therapy: a review of 84 cases. Dig. Dis. Sci. Dis. 62, 686–687 (2003). for assessing risk and managing tuberculosis 52, 1481–1484 (2007). 75. Calabrese LH, Zein N, Vassilopoulos D: in patients due to start anti-TNF-α 67. Crum NF, Lederman ER, Wallace MR: Safety of antitumor necrosis factor therapy in treatments. Rheumatology 44, 1205–1206 Infections associated with tumor necrosis patients with chronic viral infections: hepatitis (2005). factor α antagonists. Medicine 84, 291–302 C, hepatitis B, and HIV infection. Ann. 83. Orenstein R: Travel in patients receiving (2005). Rheum. Dis. 63(Suppl. II), II18–II24 (2004). TNF-α inhibitors. Travel Med. Infect. Dis. 3, 68. Warris A, Bjorneklett, Gaustad P: Invasive 76. Balandraud N, Roudier J, Roudier C: What 105–109 (2005). aspergillosis associated with infliximab are the links between Epstein–Barr virus, 84. Elkyam O, Caspi D, Reitblatt T, therapy. N. Engl. J. Med. 14, 1099 (2001). lymphoma, and tumor necrosis antagonism Charboneau D, Rubins JB: The effect of 69. DeRosa FG, Shaz D, Campagna AC, in rheumatoid arthritis? Semin. Arthritis tumor necrosis factor blockade on the Dellaripa PF, Khettry U, Craven DE: Rheum. 34(Suppl. 1), 31–33 (2005). response to pneumococcal vaccination in Invasive pulmonary aspergillosis soon after 77. Swindell K, Goldfarb J, Danziger-Isakov L, patients with rheumatoid arthritis and therapy with infliximab, a tumor necrosis Sabella C: Herpes zoster after infliximab for ankylosing spondylitis. Seminars Arthritis factor α neutralizing antibody: a possible Crohn’s disease in adolescent patients. Rheum. 33, 283–288 (2004). healthcare associated case? Infect. Control Presented at: 42nd Annual Meeting of the 85. Fomin I, Caspi D, Levy V, et al.: Hosp. Epidemiol. 24, 7 (2003). Infectious Diseases Society of America. Boston, Vaccination against influenza in 70. Coyle CM, Weis LM, Rhodes LV et al.: MA, USA, September 30 – October 3 rheumatoid arthritis patients: the effect of Fatal myositis due to the microsporidian (Abstract 999) (2004). disease modifying drugs, including TNF α Brachiola algerae, a mosquito pathogen. 78. Wolfe F, Michaud K: Rates and risk factors blockers. Arthritis Rheum. 65, 191–194 N. Engl. J. Med. 351, 42–47 (2004). for herpes zoster in patients with (2006). 71. Gottlieb GS, Lesser CF, Holmes KK, rheumatoid arthritis. Presented at: EULAR. Wald A: Disseminated sporotrichosis Vienna, Austria, June 8–11 Affiliations associated with treatment with (Abstract OP0149) (2004). • Robert Orenstein, DO immunosuppressants and tumor necrosis 79. Cursiefen C, Grunke M, Dechant C, Mayo Clinic College of Medicine, Division of factor α antagonists. Clin. Infect. Dis. 37, Antoni C, Junemann A, Holbach LM: Infectious Diseases, Department of Internal 838–840 (2003). Multiple bilateral eyelid molluscum Medicine, 200 First Street SW, Rochester, 72. Haerter G, Manfras BJ, de Jong-Hesse et al.: contagiosum lesions associated with TNF α- MN 55905, USA Cytomegalovirus retinitis in a patient treated antibody and methotrexate therapy. Am. Tel.: +1 507 255 8459; with anti-tumor necrosis factor α antibody J. Ophthalmol. 134, 270–271 (2002). Fax: +1 507 255 7767; therapy for rheumatoid arthritis Clin. Infect. 80. Wallis RS, Kyambadde P, Johnson JL et al.: [email protected] Dis. 39, E88–E94 (2004). A study of the safety, immunology, virology • Eric L Matteson, MD 73. Esteve M, Saro C, Gonzales-Huix F, and microbiology of adjunctive etanercept Mayo Clinic College of Medicine, Division of Forne M, Viver JM: Chronic hepatitis B in HIV-1 associated tuberculosis. AIDS 18, Rheumatology, Department of Internal reactivation following infliximab therapy in 257–264 (2004). Medicine, 200 First Street SW, Rochester, Crohn’s disease patients: need for primary 81. Mariette X, Salmon D: French guidelines for MN 55905, USA prophylaxis. Gut 1363–1365 (2004). diagnosis and treating latent and active Tel.: +1 507 284 8450; 74. Ostuni P, Botsios C, Punzi L et al.: tuberculosis in patients with RA treated with Fax: +1 507 284 0564; Hepatitis B reactivation in a chronic TNF blockers. Ann. Rheum. Dis. 62, 791 [email protected] hepatitis B surface antigen carrier with (2003).

576 Future Rheumatol. (2007) 2(6) futurefuture sciencescience groupgroup