Patients being treated with biological therapies are at increased risk for serious infections, including opportunistic infections. Although more is known about opportunistic infection risk with older biologics, such as antitumor necrosis factor drugs, there is less knowledge of opportunistic infection risk with newer biological therapies. The incidence of certain opportunistic infections (tuberculosis, herpes zoster, pneumocystosis) has been rigorously evaluated in large observational studies. However, data are more limited for other infections (histoplasmosis, nontuberculous mycobacteria). Infectious morbidity and mortality may be preventable with screening and prophylaxis in select populations.
Key points
- •
The risk of opportunistic infections is increased with use of biological therapies.
- •
The risk of tuberculosis (TB) is increased with anti–tumor necrosis factor (TNF) therapy, and monoclonal antibodies (infliximab, adalimumab) have a higher risk of TB reactivation than etanercept.
- •
Anti-TNF therapy and tofacitinib are associated with an increased risk of zoster in patients with immune-mediated inflammatory diseases.
Introduction
Treatment of immune-mediated inflammatory diseases (IMIDs) with biological therapies has resulted in substantial improvement in patient symptoms and has slowed the natural progression of these often-debilitating conditions. Although these therapies have improved the quality of life for many patients, a consequence of biological therapies has been an increased risk of opportunistic infection (OI). In many patients, depending on the underlying disease, there already may be an increased baseline risk of infection independent of disease-modifying therapy. A recent meta-analysis of 70 trials including more than 32,000 patients identified an overall increased risk of OIs at 1.7 excess infections per 1000 patients treated with biologics. After the US Food and Drug Administration (FDA) approved infliximab for the treatment of Crohn disease (CD) in 1998, much of the early knowledge on the risk for OIs in the postmarketing period has come from spontaneous reporting and relied heavily on point-of-care diagnoses. Since that time, newer biological agents with variable mechanisms of action have been approved for a variety of conditions.
Previous reports have shown an increased risk of OIs, serious infections, and hospitalization among users of biologics. The French RATIO (Research Axed on Tolerance of Biotherapies) study evaluated nontuberculosis OI risk in patients on anti–tumor necrosis factor (TNF) therapy for any indication and found that infliximab (odds ratio [OR] 17.6) and adalimumab (OR 10.0) carried an increased risk for OIs compared with etanercept. The US Safety Assessment of Biologic Therapy (SABER) study found a higher rate of nonviral OIs among a large cohort of new users of TNF inhibitors (n = 33,324) versus those initiating therapy with nonbiological disease-modifying antirheumatic drugs (DMARDs). In the study’s rheumatoid arthritis (RA) cohort, new infliximab users experienced a higher rate of nonviral OIs compared with both nonbiological DMARD users (adjusted hazard ratio [aHR] 2.6) and etanercept users (aHR 2.9).
In contrast, a smaller (n = 570) prospective Japanese study evaluating OI incidence in patients with inflammatory bowel disease (IBD) found no increased risk of OIs among patients on infliximab over a 12-month period but did show a risk with other immunosuppressants and increasing age. A Japanese anti-TNF agent switch study found increased incidence of OIs in the first year of treatment with TNF inhibitors, similar to findings in a retrospective Spanish study showing an increased risk of OIs in the first year of therapy with infliximab. Additional TNF antagonists, such as certolizumab pegol and golimumab, as well as targeted drugs with differing mechanisms of action, such as belimumab, rituximab, tocilizumab, ustekinumab, abatacept, anakinra, and tofacitinib, have also been studied; but data are limited to controlled trials.
A major challenge in studying and defining OI risk lies in providing a workable OI case definition. OIs are often difficult to define, and reaching a consensus definition across studies within the area of biological therapy has proven challenging. Although OIs have been more consistently defined within certain diseases, such as human immunodeficiency virus (HIV) infection, this is not the case for biologics. A recent review sought to define OIs in the setting of biologics and provide case definitions for specific candidate pathogens. Although the investigators did reach consensus, they noted that prior attempts to define OIs with the use of biologics have been inconsistent, resulting in wide-ranging OI risk estimates. Herein, the authors review the risk of OIs in biological therapy, with a focus on several major OIs and the most rigorously studied biologics.
Introduction
Treatment of immune-mediated inflammatory diseases (IMIDs) with biological therapies has resulted in substantial improvement in patient symptoms and has slowed the natural progression of these often-debilitating conditions. Although these therapies have improved the quality of life for many patients, a consequence of biological therapies has been an increased risk of opportunistic infection (OI). In many patients, depending on the underlying disease, there already may be an increased baseline risk of infection independent of disease-modifying therapy. A recent meta-analysis of 70 trials including more than 32,000 patients identified an overall increased risk of OIs at 1.7 excess infections per 1000 patients treated with biologics. After the US Food and Drug Administration (FDA) approved infliximab for the treatment of Crohn disease (CD) in 1998, much of the early knowledge on the risk for OIs in the postmarketing period has come from spontaneous reporting and relied heavily on point-of-care diagnoses. Since that time, newer biological agents with variable mechanisms of action have been approved for a variety of conditions.
Previous reports have shown an increased risk of OIs, serious infections, and hospitalization among users of biologics. The French RATIO (Research Axed on Tolerance of Biotherapies) study evaluated nontuberculosis OI risk in patients on anti–tumor necrosis factor (TNF) therapy for any indication and found that infliximab (odds ratio [OR] 17.6) and adalimumab (OR 10.0) carried an increased risk for OIs compared with etanercept. The US Safety Assessment of Biologic Therapy (SABER) study found a higher rate of nonviral OIs among a large cohort of new users of TNF inhibitors (n = 33,324) versus those initiating therapy with nonbiological disease-modifying antirheumatic drugs (DMARDs). In the study’s rheumatoid arthritis (RA) cohort, new infliximab users experienced a higher rate of nonviral OIs compared with both nonbiological DMARD users (adjusted hazard ratio [aHR] 2.6) and etanercept users (aHR 2.9).
In contrast, a smaller (n = 570) prospective Japanese study evaluating OI incidence in patients with inflammatory bowel disease (IBD) found no increased risk of OIs among patients on infliximab over a 12-month period but did show a risk with other immunosuppressants and increasing age. A Japanese anti-TNF agent switch study found increased incidence of OIs in the first year of treatment with TNF inhibitors, similar to findings in a retrospective Spanish study showing an increased risk of OIs in the first year of therapy with infliximab. Additional TNF antagonists, such as certolizumab pegol and golimumab, as well as targeted drugs with differing mechanisms of action, such as belimumab, rituximab, tocilizumab, ustekinumab, abatacept, anakinra, and tofacitinib, have also been studied; but data are limited to controlled trials.
A major challenge in studying and defining OI risk lies in providing a workable OI case definition. OIs are often difficult to define, and reaching a consensus definition across studies within the area of biological therapy has proven challenging. Although OIs have been more consistently defined within certain diseases, such as human immunodeficiency virus (HIV) infection, this is not the case for biologics. A recent review sought to define OIs in the setting of biologics and provide case definitions for specific candidate pathogens. Although the investigators did reach consensus, they noted that prior attempts to define OIs with the use of biologics have been inconsistent, resulting in wide-ranging OI risk estimates. Herein, the authors review the risk of OIs in biological therapy, with a focus on several major OIs and the most rigorously studied biologics.
Tuberculosis
Incidence and Drug-Specific Risk
Disease due to Mycobacterium tuberculosis remains a major cause of morbidity and mortality, with an estimated 9.6 million cases of incident tuberculosis (TB) worldwide in 2014 according to the most recent World Health Organization’s Global Tuberculosis Report. TB has been increasingly reported in patients receiving treatment with biological therapies for a variety of indications since the early 2000s, emphasizing its importance as an opportunistic pathogen. In 2001, Keane and colleagues reported 70 cases of TB in patients on infliximab received through the FDA Adverse Event Reporting System as of May 2001. Cases developed a median of 12 weeks after initiation of therapy, underscoring the need to screen for latent TB infection (LTBI) and disease, especially in areas of high endemicity.
Early studies from North America and Europe found an increased incidence of TB associated with initiation of anti-TNF-α therapy ( Table 1 ). Surveillance studies from Spain, based on data from the BIOBADASER (Spanish Society of Rheumatology Database on Biologic Products) database established in 2000 for the long-term follow-up of patients with rheumatic diseases on biological therapy, estimated the incidence of TB associated with infliximab to be 1893 cases per 100,000 patients compared with 21 cases per 100,000 inhabitants. Follow-up studies in 2005 and 2007 reported an overall decrease in incidence after the introduction of official screening recommendations by Spanish health authorities. Importantly, the probability of developing active TB was 7 times higher when guidelines were not followed.
| Reference, Year | Study Type | Location and Period | Person-Years | TB Rate (per 100,000 Person-Years) |
|---|---|---|---|---|
| Tubach et al, 2009 RATIO | Prospective cohort | France, 2004–2006 | 57,711 | 116.7 |
| Dixon et al, 2010 BSRBR | Prospective cohort | Great Britain, 2001–2008 | 28,447 (on drug) a | 95 |
| Kim et al, 2011 | Retrospective cohort | South Korea, 2002–2009 | 1784 | 561 |
| Lee et al, 2013 | Retrospective cohort | South Korea, 2002–2011 | 1717 | 519 |
| Winthrop et al, 2013 | Retrospective cohort | United States, 2000–2008 | 20,330 | 49 |
| Abreu et al, 2013 | Retrospective cohort | Portugal, 2001–2012 | NR | 1337 (INF), 792 (ADA), 405 (ETN) |
| Yoo et al, 2014 | Retrospective cohort | South Korea, 2005–2011 | 231 | 1300 |
| Baddley et al, 2014 SABER | Retrospective cohort | United States, 1998–2007 | 22,275 | 36 |
| Arkema et al, 2015 | Prospective cohort | Sweden, 2002–2011 | 48,228 | 39.4 |
a The investigators used 2 models to report TB cases. On drug refers to TB diagnosis being made while patients are actively receiving TNF inhibitor therapy.
In the 2009 French RATIO study, 69 validated TB cases were collected based on spontaneous reporting, with the denominator the estimated number of person-years (py) of receipt of anti-TNF therapy in France from 2004 to 2006. The study included patients being treated for RA, spondyloarthropathies, IBD, psoriasis, and Behçet disease. With regard to specific agents used, 36 patients received infliximab, 28 received adalimumab, and 5 were treated with etanercept. Standardized incidence ratios were higher for the monoclonal antibody (mAb) biological class compared with etanercept. A subsequent case-control analysis identified the use of either infliximab or adalimumab versus etanercept as an independent risk factor for TB.
In 2010, Dixon and colleagues addressed the risk of TB between 3 anti-TNF agents (infliximab, adalimumab, etanercept) in patients with RA. Data were from the British Society for Rheumatology Biologics Register (BSRBR), a prospective observational study. A total of 13,739 patients were included, 3232 in the traditional DMARD cohort and 10,712 in the anti-TNF cohort. Forty cases of active TB were reported in the anti-TNF cohort, whereas no cases were reported in the DMARD cohort. Infliximab and adalimumab were associated with a higher rate of TB (136 and 144 events per 100,000 py, respectively) compared with etanercept (39 events per 100,000 py).
A 2011 Korean study reviewing TB in patients with ankylosing spondylitis (AS) found an increased incidence rate of TB in TNF inhibitor (TNFi)-naïve patients versus the general population. Among the patients with AS, no increase in risk was observed between those who were TNFi-naïve and those exposed to biologics. However, a lower risk of TB was observed among users of etanercept compared with the mAb class. In 2013, Lee and colleagues performed a retrospective, longitudinal cohort study from 2002 to 2011 at a large Korean hospital. Among 509 patients treated with TNF antagonists, 9 (1.8%) developed active TB (incidence rate of 519 per 100,000 py). This finding represented a 6.4 times higher incidence of TB than the general population. This study also reported a lower but significant incidence of nontuberculous mycobacterial (NTM) lung disease (230.7 per 100,000 py).
A US study by Winthrop and colleagues identified more than 8000 anti-TNF users in the United States, most of which had RA, through automated pharmacy records from 2000 to 2008. Estimated anti-TNF–associated TB incidence rates were 49 per 100,000 py compared with 2.8 per 100,000 py in the general population. A retrospective matched case-control study from Portugal reported 25 cases (out of n = 765) of active TB diagnosed in patients on anti-TNF treatment. As noted elsewhere in the literature, patients with TB were more likely to be receiving infliximab (64% of cases) as compared with adalimumab (24%) and etanercept (3 cases), demonstrating an incidence rate of 1337,792 and 405 per 100,000 py, respectively.
A recent retrospective Korean study evaluated the incidence of TB in patients on anti-TNF therapy within a 6-month period. Of 175 cases reviewed in patients with a variety of IMIDs, TB was diagnosed in 3 patients. Although the incidence of TB was estimated to be 18-fold higher than that of the general population, the investigators noted that this might be an overestimation due to underreporting of TB in the general population.
More recently, a large prospective population-based national cohort study from Sweden reported the risk of TB in biological-naïve (n = 37,982) and biological-exposed (n = 10,800) patients with RA from 2002 to 2011. Each patient with RA was matched with up to 5 general population comparators (n = 175,972). When compared with the general population, the biological-naïve patients with RA had a 4-fold increased risk of incident TB. A 2.5-fold higher risk of TB was observed in the biological-exposed patients compared with those who were biological-naïve.
Prevention
Reactivation of LTBI seems to be the predominant reason for the development of disease in low-prevalence countries, with some studies showing an increased risk at the early initiation of therapy. Abreu and colleagues noted that 17 of 25 cases of active TB identified in their Portuguese cohort had been previously diagnosed with LTBI; a large proportion had some combination of negative screening tests, interferon gamma release assay (IGRA), tuberculin skin testing, or chest radiograph, highlighting that the diagnosis of LTBI is not always straightforward. It is important to note that the likelihood of a false-negative screening test may increase in the setting of active treatment with immunosuppressive therapy, before starting biologics. Early reports about the risk of TB and biological therapy have led to the development of guidelines regarding screening and prophylaxis of TB. In some low-prevalence TB countries, screening guidelines have resulted in a reduction in TB incidence.
Nontuberculous mycobacterial disease
Although TB has been established as an important consequence of immunosuppression due to biological therapies, fewer studies have reported on the increased risk of NTM disease in patients on biologics. In the United States, a country with low background rates of TB, the prevalence of NTM disease seems to be increasing. Within the SABER cohort study, approximately 11% of the 80 cases of nonviral OIs reported were due to NTM disease. However, the risk between traditional and biological DMARDs was similar. Although a risk differential may exist between TB and NTM disease, the investigators point out the potential for confounding by indication bias, whereby those at risk for NTM infections (eg, severe lung disease) may not be prescribed biological DMARDs.
Data extracted from the large North American health maintenance organization Kaiser Permanente Northern California (KPNC) estimated a higher risk of NTM disease (74 cases per 100,000 py) in those on TNFi compared with the general population and unexposed patients with RA. It has been suggested that NTM disease may be a more important consideration than TB before starting anti-TNF therapy, especially in countries with low TB prevalence. In this KPNC study, a higher proportion of patients with NTM disease died compared with those developing active TB. This finding could be partly due to increased adherence to TB screening guidelines in recent years, a predisposition to pulmonary NTM among patients with RA, or the difficulty in treating pulmonary NTM disease.
Additional investigation is needed in this area, as NTM disease prevalence increases. No screening guidelines exist for patients at risk for infections due to these organisms, and identifying patients at risk is challenging. Providers must remain vigilant for NTM disease development, especially among aging patients with RA with chronic lung disease. No chemoprophylaxis is currently indicated for prevention.
Herpes zoster
Incidence and Drug-Specific Risk
Herpes zoster (HZ), or shingles, due to the reactivation of latent varicella zoster virus (VZV), remains a burdensome disease with potential for long-term disability due to postherpetic neuralgia (PNH). In the United States, there are approximately 1 million new cases diagnosed each year, with those aged 50 years and older being most affected. National incidence of disease has been estimated to be about 3.2 cases per 1000 py overall, with increasing risk in the aging population. Reactivation of latent VZV has been linked to a diminished varicella-specific cell-mediated immunity in the elderly and immunosuppressed. Multiple studies on the incidence of HZ in those on biological therapy have been reported, with a wide range of findings ( Table 2 ).
| Reference, Year | Study Type | Location and Period | Person-Years | HZ Rate |
|---|---|---|---|---|
| McDonald et al, 2009 | Retrospective cohort | United States, 1998–2005 | ∼71,000 | 10.6/1000 py a |
| Strangfeld et al, 2009 RABBIT | Prospective cohort | Germany, 2001–2006 | 2588 (ETN), 3524 (INF/ADA) | 8.9 (ETN), 11.1 (INF/ADA)/1000 py |
| Garcia-Doval et al, 2010 | Prospective cohort | Spain, 2000–2008 | 15,389 b | 32/100,000 py b |
| Winthrop et al, 2013 | Retrospective cohort | United States, 1998–2007 | 28,392 c | 10.9/1000 py c |
| Galloway et al, 2013 BSRBR | Prospective cohort | United Kingdom, 2001–2009 | 17,048 d | 1.6/100 py d |
| Winthrop et al, 2014 | Retrospective cohort | United States, 2014 | 290 (low dose), 290 (high dose) e | 2.1, 5.2/100 py e |
| Yun et al, 2015 f | Retrospective cohort | United States, 2006–2011 | 7614 (ABA), 3611 (RTX), 3135 (INF), 2638 (ADA), 2229 (ETN), 839 (TOC), 774 (CTL), 683 (GOL) (per 100 py) | 1.87 (ABA), 2.27 (RTX), 1.82 (INF), 1.74 (ADA), 2.15 (ETN), 2.15 (TOC), 2.45 (CTL), 1.61 (GOL) (per 100 py) |
Related posts:
Infection and Malignancy in Rheumatic Diseases
Infection and Malignancy in Rheumatic Diseases
Infection and Malignancy in Rheumatic Diseases
Immune-Related Adverse Effects of Cancer Immunotherapy— Implications for Rheumatology
Malignancy and Janus Kinase Inhibition
The Risk of Progressive Multifocal Leukoencephalopathy in the Biologic Era
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
