Article Text

Download PDFPDF

Infectious complications of rheumatoid arthritis and psoriatic arthritis during targeted and biological therapies: a viewpoint in 2020
  1. Olivier Lortholary1,2,
  2. Mario Fernandez-Ruiz3,4,
  3. John W Baddley5,
  4. Oriol Manuel6,
  5. Xavier Mariette7,8,
  6. Kevin L Winthrop9
  1. 1 Paris University, Necker Pasteur Center for Infectious Diseases and Tropical Medicine, IHU Imagine, Necker Enfants malades University Hospital, APHP, Paris, France
  2. 2 Institut Pasteur, National Reference Center for Invasive Mycoses and Antifungals, Molecular Mycology Unit, CNRS UMR 2000, Paris, France
  3. 3 Unit of Infectious Diseases, Hospital Universitario "12 de Octubre", Instituto de Investigación Sanitaria Hospital "12 de Octubre" (imas12), School of Medicine, Universidad Complutense, Madrid, Spain
  4. 4 Spanish Network for Research in Infectious Diseases (REIPI RD16/0016), Instituto de Salud Carlos III, Madrid, Spain
  5. 5 University of Maryland School of Medicine, Division of Infectious Diseases, Baltimore, Maryland, USA
  6. 6 Infectious Diseases Service and Transplantation Center, University Hospital and University of Lausanne, Lausanne, Switzerland
  7. 7 Rheumatology, Assistance Publique-Hôpitaux de Paris (AP-HP), Hôpitaux universitaires Paris-Sud – Hôpital Bicêtre, Le Kremlin Bicêtre, France
  8. 8 Université Paris-Sud, Center for Immunology of Viral Infections and Auto-immune Diseases (IMVA), Institut pour la Santé et la Recherche Médicale (INSERM) UMR 1184, Université Paris-Saclay, Le Kremlin Bicêtre, France
  9. 9 Oregon Health Sciences University, Portland, Oregon, USA
  1. Correspondence to Professor Olivier Lortholary, Paris University, 75015 Paris, France; olivier.lortholary{at}


Biological therapies have improved the outcomes of several major inflammatory, autoimmune and also neoplastic disorders. Those directed towards cytokines or other soluble mediators, cell-surface molecules or receptors or various components of intracellular signalling pathways may be associated with the occurrence of infections whose diversity depends on the particular immune target. In this context and following a keynote lecture given by one of us at the European League Against Rheumatism meeting on June 2018, a multidisciplinary group of experts deeply involved in the use of targeted and biological therapies in rheumatoid and psoriatic arthritis decided to summarise their recent vision of the immunological basis and epidemiology of infections occurring during targeted and biological therapies, and provide useful indications for their management and prevention.

  • rheumatoid arthritis
  • psoriatic arthritis
  • infections
  • vaccination
  • DMARDs (biologic)

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Biological therapies have improved the outcomes of several major inflammatory, autoimmune and also neoplastic disorders. They include biological response modifiers such as immunoglobulins, interleukins and colony stimulating factors, various modalities of gene and targeted therapies. The latter are directed towards cytokines or other soluble mediators, cell-surface molecules or receptors or various components of intracellular signalling pathways and are classically divided into therapeutic monoclonal antibodies and small-molecule enzyme inhibitors. In addition, monoclonal antibodies which started to appear 30 years ago have become more sophisticated overtime with more pronounced immunological impact and improved pharmacology. Due to their marked efficacy on host immunological pathways, their use may be associated with the occurrence of infections whose diversity and severity depends on the particular immune target. Such diversity can be anticipated by the knowledge issued from primary immune deficiencies (eg, interleukin (IL)-17 targeted therapy and risk of mucocutaneous candidiasis reminiscent of what is observed in STAT1 gain-of-function mutations) or from experimental models of infection using genetically modified animals. It should however be noted that the frequency and severity of infectious complications, leading to increased hospitalisations and even death, either opportunistic or not, community or healthcare related depend not only on a particular biological therapy but also on host’s physiological characteristics, medical history, other immunosuppressive therapies given and their duration and epidemiology. Of note, infections occur mostly within the first year of biological therapy use, and major concerns are severe bacterial infections, mycobacterial and fungal diseases, herpes zoster and hepatitis B virus (HBV) reactivation and finally travel-associated infections.

In addition, the occurrence of some infections may not be evidenced even through large pivotal studies, thereby emphasising the marked contribution of registries and population-based data sources for better assessing the infectious risk and its preventive management.

In this context and following a keynote lecture given by one of us (OL) at the European League Against Rheumatism (EULAR) meeting in 15 June 2018, a multidisciplinary group of experts deeply involved in the use of targeted and biological therapies in rheumatoid and psoriatic arthritis (RA and PsA) decided to summarise, through a narrative review, their recent vision of the immunological basis and epidemiology of infections occurring during targeted and biological therapies, and provide useful indications for their management and prevention.

Functional classification of biologics and current management of rheumatoid arthritis and psoriatic arthritis

Since the first Food and Drug Administration approval of infliximab for the treatment of RA over 20 years ago, the biologic armamentarium for treating RA and PsA has exploded and continues to grow to this day. While agents blocking tumour necrosis factor (TNF)-α revolutionised treatment options, a number of novel therapeutics have been approved in the last decade, including synthetic small molecules such as Janus kinase (JAK) inhibitors. Agents that modulate interleukin (IL)-6, IL-12/23 common p40 subunit, and IL-17 via either direct cytokine inhibition or inhibition of the cytokine receptor have all proven useful in either RA or PsA or both. Historically, treatment paradigms in both diseases have advocated starting with non-biologic disease-modifying agents with or without short-term low-dose prednisone. Traditionally, the first biologic used in patients who did not achieve disease control with non-biologics has been an anti-TNF-α agent. While the majority of patients starting anti-TNF-α therapy achieve a reduction in disease activity, a substantial proportion either do not maintain or achieve low disease activity such that the initial anti-TNF therapy is discontinued and another biologic is tried in its place. While historically a second anti-TNF-α therapy would be employed, recent guidelines and practice suggest that any of the other biologics and/or JAK inhibitors could be used for patients non-responding or intolerant of their initial anti-TNF-α agent.1–3 In fact, the most recent guidelines make no distinction between which targeted therapy can be used first line for those failing non-biologic disease-modifying antirheumatic drugs (DMARDs).1 For RA, agents beyond TNF-α blockers include the CD80/86-binding T-cell modulator abatacept or anti-IL-6 agents (tocilizumab or sarilumab), the peripheral B-cell depletion agent rituximab, or the JAK inhibitors tofacinitib, baricitinib and upadicitinib. PsA offers similar choices in addition to ixekizumab and secukinumab, and which inhibit IL-17 signalling, and ustekinumab that blocks the shared p40 subunit of IL12/23. For those with milder forms of disease, the small-molecule phosphodiesterase-4 inhibitor apremilast is also an option.4 These therapies and their recommended use are summarised in table 1.

Table 1

Approved and recommended biologic and small molecular therapies for rheumatoid arthritis (RA) and/or psoriatic arthritis (PsA)

Current view of infectious risk of biologics in RA or PsA

Before describing the specific increased risks of infection with each of these targeted therapies, several important general considerations should be considered: RA or PsA per se is associated with an increased infectious risk, the population included in randomised clinical trials (RCTs) is highly selected for registration purposes, numerically relatively low, with a limited exposure to targeted therapy. In spite of that, even it is never statistically significant even after a first meta-analysis performed by the Cochrane library, the rate of infections or of serious infections is, in most of the trials, numerically superior in the active arm than in the placebo arm, which is in line with the theory.5 The increased risk of serious infections with biologics in RA was confirmed in a more recent meta-analysis with standard dose (OR 1.31, 95% CI 1.09–1.58) and high dose (OR 1.90, 95% CI 1.50 to 2.39), although not with low dose (OR 0.93, 95% CI 0.65 to 1.33).6

For all the licensed targeted therapies, this slight increased risk is acceptable regarding the risk of an alternative treatment, usually higher doses of steroids or classical immunosuppressive drugs like methotrexate or hydroxychloroquine. Steroids, besides increasing the risk of diabetes, cardiovascular events, osteoporosis, also increase the risk of infections even given at a very low daily dose of 5 mg/day.7 Thus, the infectious risk of any new targeted drug has to be analysed regarding what would be the infectious risk of pursuing or increasing steroids rather than giving this new targeted drug.

Conversely to RCT, most of the registries, which include a bigger size, heterogeneous, but real-life population, found that the increased risk of infections with targeted therapies compared with classical treatments is limited to the first 6–12 months after the start of the drug. Curiously, if the analysis is made later, the infectious risk is usually lower with biologics than with the classical treatments. This apparent paradox is due to what is called ‘the healthy drug survivor effect’.8 At the beginning of a treatment with a biologic, several parameters are shared for increasing the risk of infections: the inflammatory disease is more active, requires more cotreatment including steroids and, at the top of that, there is the specific infectious risk of the new biologic that is introduced. Overtime, activity of the disease decreases, steroids are tapered or stopped and, mainly, the patients having developed serious adverse events with the biologics will have discontinued it. Because of this bias, it makes little sense to analyse the evolution of rate of serious infections over time in the long-term extensions of RCTs, what is frequently done by the pharmaceutical companies.

The long-term follow-up is not very useful for assessing the risk of classical infections but remains important for assessing a specific risk of rare infection, like, now classified for rheumatologists, opportunistic infections,9 linked to the mechanism of action of the biologic, except in case of reactivation of a latent infection that occurs early, like reactivation of latent tuberculosis infection (LTBI) with TNF-α inhibitors.10

The surveillance and prevention of the infectious risks of targeted therapies must be adapted to each mechanism of action and to specific populations. For example, even if tuberculosis (TB) was not seen in phase III studies with TNF-α inhibitors (TNFIs) because of the low incidence of this condition in countries where the studies were conducted, it was expected that inhibiting TNF-α could lead to an increased risk of developing TB. After the risk was demonstrated, the screening of LTBI was proposed for any new biologic drugs. But has it a sense with B-cell targeted therapies or anti-IL-6 therapies? Not sure regarding the mechanism of action of these drugs. Likewise, clinicians have to be aware of the risk of some specific infectious events linked to the mechanism of action of the targeted drugs: mucosal candidiasis with anti-IL-17 therapies regarding the specific role of IL-17 in this anti-Candida defence,11 bacterial infections with encapsulated microorganisms with the repetition of treatment with B-cell targeted therapies leading to hypogammaglobulinaemia (HGG), herpes zoster (HZ) with the use of JAK inhibitors especially in people originating from certain subregions of Asia like Japan or Korea.12–14 Thus, we will move towards a more personalised prevention and surveillance of the infectious risk of targeted therapies depending on the mechanism and on the population.

In this context of personalised medicine, one important question is whether it should be considered to restart the same biologic agent in patients with recent serious infections instead of swapping to another DMARD class with a different mechanism of action. Here again, there is no general rule. It will depend on the type of infection, on the mechanism of action of the drug, on the other available drugs for the considered disease and, of course, on the willingness of the patients to restart a drug possibly having giving him/her a side effect. Just two examples: (1) a patient having done reactivation of latent TB after adalimumab may restart adalimumab after 3 months of anti-TB treatment. It will be the first choice if he suffers from spondyloarthritis with recurrent uveitis. If he/she suffers from RA and may come back in endemic areas, the choice will rather be to switch to etanercept or to a drug with another mechanism of action. (2) In a patient with RA treated by rituximab and developing Streptococcus pneumoniae infection favoured by a rituximab-induced hypo-IgG at 4 g/L, the choice will be to replace rituximab (RTX) by a drug with another mechanism of action.

Current view on major infections induced by TNF-α inhibitors

TNF-α inhibitors (TNFIs) represent important treatment advances for patients with RA and PsA. TNF-α is involved in a number of adaptive immune responses, including formation of granulomas, development of phagosomes, activation and differentiation of macrophages, and immune response against viral pathogens.15 TNFIs inhibit these biological activities by binding with high affinity to the cytokine itself or by blocking the binding of TNF-α with its receptors.15 In general, patients receiving TNFI are at increased risk for mycobacterial and fungal infections, particularly TB and histoplasmosis. Theoretically, TNFI would also increase susceptibility to other intracellular pathogens and viruses.

Clinical experience with infection risk and TNFI has been derived from meta-analyses of RCTs, open-label extension studies, registries and cohort studies16–31 (table 2). Patients with LTBI receiving TNFI therapy for RA or PsA have approximately a fourfold increase in the risk of developing active TB as compared with control patients (table 2).16–18 However, there may be different risks between monoclonal Ab and the soluble receptor probably linked to a differential effect on granuloma due to a different stability of binding to membrane TNF. Two recent meta-analyses have confirmed an increased risk for TB among patients with inflammatory bowel disease (IBD) receiving TNFI, with this association being particularly evident in the presence of Crohn’s disease.32 33 Many countries have implemented LTBI screening and treatment guidelines in these populations, and in general, the risk of TB is quite low with appropriate screening; however, incomplete LTBI screening and treatment procedures have been observed in RCTs and registries and lack of adherence to guidelines has been associated with increased TB incidence.16 22 34

Table 2

Summary of estimated risks for infection associated with TNFI reported in selected meta-analyses and registries

Clinical trials and observational studies with TNFI have evaluated the risk of serious infections, but have often failed to capture specifics of infectious syndromes or causative agents. Typically, pneumonia and soft tissue infections are the most common serious infections observed among patients receiving TNFI, similar to the pre-biologic era.35 When comparing patients on TNFI with those receiving conventional DMARDs, studies focused on the first year of therapy show increased risk of infection, with adjusted rate ratios ranging from 1.5 to 5.0; however, those with longer follow-up periods yielded contrasting results.7 31–38 A meta-analysis of 106 RCTs of targeted therapies (mostly TNFI) in patients with RA demonstrated a twofold increase in the risk of serious infections, particularly when high doses were used, compared with traditional DMARDs.16 A meta-analysis evaluating patients with PsA reported a crude OR for infection of 1.18 (95% CI 1.05 to 1.33) in patients exposed to TNFI (vs not exposed).39 In contrast, a meta-analysis including patients with ankylosing spondylitis failed to show an increased risk of infection with TNFI compared with untreated controls.38 Among patients with IBD, a recent meta-analysis concluded that the risk of serious non-opportunistic infections was not increased.33

Smaller case series have described the risk of specific infections with TNFI.40–42 For example, one study described a fourfold increased risk of severe listeriosis with TNFI in comparison with the general population.41 In addition, incidence of legionellosis was 37-fold higher in patients receiving TNFI.43 Other granulomatous bacterial infections with TNFI, including nocardiosis and non-tuberculous mycobacteria, have been reported.40 44

TNFI and HZ risk has been evaluated, but with conflicting results. A US cohort study among patients with RA, IBD or other inflammatory diseases did not demonstrate an increased risk of HZ among TNFI users.45 In contrast, data from European registries suggest an average twofold increase in the risk of HZ.46–48 Differing practices in the use of corticosteroid therapy between European and US practitioners have been suggested as an explanation of the conflicting results.49

TNFI-mediated effects may influence the odds of clearance or reactivation of chronic hepatitis B virus (HBV) infection. Cases of HBV reactivation while on therapy have been reported for different indications,50–52 prompting specific recommendations for monitoring and treatment.53–56 The risk for worsening of hepatitis C virus (HCV) infection is quite low with use of TNFI.56–58

The impact of TNFI on risk of fungal infection has been described with reports of endemic fungal infections, cryptococcosis, aspergillosis, and Pneumocystis jirovecii pneumonia (PCP). However, the concomitant use of other immunosuppressive therapies, especially corticosteroids, renders risk interpretation problematic.59–64 A meta-analysis of RCTs in patients with RA found that the use of biologic DMARDs did not significantly increase the risk of invasive fungal infection (OR 2.58, 95% CI 0.68 to 11.91) or PCP (OR 1.77, 95% CI 0.42 to 7.47).65

Box 1

Recommendations for prevention and management of infections induced by TNFI

  • TNFI should be discontinued, at least temporarily, upon the occurrence of serious infection. Therapy should not be restarted until infection has been treated and clinical response is achieved.

  • Screening for LTBI should be performed before starting TNFI and providers should assess risk factors and exclude active TB. Anti-TB therapy should be offered to patients diagnosed with LTBI in order to reduce the risk of progression to active TB.

  • TNFI-mediated effects may influence the risk of clearance or reactivation of HBV infection. Screening for chronic HBV infection, based on the detection of both hepatitis B surface antigen (HBsAg) and hepatitis B core antibody (anti-HBc), should be performed before starting TNFI.

  • Antibacterial or antifungal prophylaxis is not routinely recommended for TNFI patients. However, age-appropriate vaccinations for Streptococcus pneumoniae, influenza and other pathogens should be administered. Some live-virus vaccines (ie, varicella-zoster virus or measles–mumps–rubella) may be contraindicated in patients receiving TNFI.

  • LTBI, latent tuberculosis infection; TNFI, tumour necrosis factor-α inhibitor.

Current view on infections induced by other biologics than TNFI

Two IL-6-targeted agents (tocilizumab (RoActemra, Roche) and sarilumab (Kevzara, Sanofi)) are currently approved for the treatment of RA in patients with inadequate response or intolerance to conventional synthetic DMARDs. In addition, some other agents (clazakizumab) are in different stages of clinical development. In view of the proinflammatory and pleiotropic functions displayed by IL-6, the functional blockade of this cytokine—whose signalling pathway is shared by other mediators that also use the protein gp130 as a common signal transducer66—would impact on the host capacity for generating acute phase responses and protective immunity against pathogens, including Th17 differentiation and long-lived plasma cell generation.67 68 Clinical experience derived from pivotal RCTs and observational registries among patients with RA (which is by far greater for tocilizumab than for the more recently approved sarilumab) supports the notion that IL-6-targeting agents increase the risk of infection to a degree similar to TNFI and other biological DMARDs.69–72 A large meta-analysis published in 2011 (based on 9414 person-years of exposure to tocilizumab) reported a rate of serious infection of 4.9 per 100 patient-years among participants treated with 8 mg/kg (the currently approved dose for RA) as compared with 3.5 per 100 patient-years among those receiving placebo. Pneumonia and cellulitis were the most commonly reported syndromes.70 Since IL-6 plays a key role in the acute-phase response by the liver, it has been proposed that the diagnosis of infections occurring in patients on tocilizumab therapy may be masked due to the sustained suppression of C reactive protein (CRP) synthesis.73 Prior therapy with TNFI was associated with an increased rate of infection.74 Risk factors for developing infectious complications on tocilizumab therapy are similar to those identified for other biological DMARDs (ie, older age, chronic lung disease and long-term corticosteroids). On the other hand, neutropenia of all grades has been found to occur more frequently in patients treated with tocilizumab or sarilumab than placebo.74 75 The clinical relevance of this finding, which would be explained by the margination of neutrophils rather than peripheral sequestration or incomplete granulopoiesis,76 is unclear since the incidence of serious infection in RCTs did not seem to correlate with low neutrophil counts.75 76 As expected in view of the older age and higher disease burden usually exhibited by real-world patients as compared with RCT participants, the incidence of infection reported from open-label multicentre registries and population-based studies is sensibly higher, reaching 9.0–14.9 episodes per 100 patient-years.77 78 Episodes of opportunistic infection among tocilizumab-treated RA patients include TB, PCP, cryptococcosis and HZ.72 74 The risk of active TB in a focused meta-analysis including RCTs and long-term extension studies seems to be lower for tocilizumab than for anti-TNF-α agents,78 although a postmarketing surveillance carried out in Japan found a similar rate between both therapies.79 Regarding HZ, the estimated incidence rate in a large population-based study on US Medicare data (2.15 episodes per 100 patient-years) was similar to that observed with other biological agents80 (table 3).

Table 3

Estimated risks and incidence rates for infection associated with non-TNFI biological DMARDs approved in RA and reported in selected meta-analyses

Different IL-17-targeted agents have been approved for patients with plaque psoriasis or PsA, including anti-IL-17A (secukinumab (Cosentyx, Novartis) and ixekizumab (Taltz, Eli Lilly)) and anti-IL-17 receptor A (IL-17RA) monoclonal antibodies (brodalumab (Siliq or Kyntheum, Valeant Pharmaceuticals)). The biological functions of the IL-17 family of cytokines (comprised of IL-17B to IL-17F in addition to the main effector member IL-17A) include neutrophil granulopoiesis and chemotaxis, expression of antimicrobial peptides (β-defensin-2) and macrophage stimulation. More importantly, IL-17A and IL-17F constitute the signature cytokines of the Th17 T-cell response.81 82 Akin to other biological agents, the clinical effects of the therapeutic blockade of the IL-17A/IL-17RA pathway may be anticipated in view of the phenotype associated with certain primary immunodeficiencies.83 Chronic mucocutaneous candidiasis results from non-sense mutations in the IL17A gene and other genes involved in the IL-17 signalling pathway, or from the development of anti-IL-17 autoantibodies. These patients suffer from persistent and recurrent Candida infections of the mucosal, nail and skin surfaces.84 Accordingly, patients with PsA treated with IL-17-targeted agents are exposed to a dose-dependent increased risk of mucocutaneous candidiasis, which usually are mild to moderate in severity.85–88 A systematic review reported rates of Candida infection of 4.0%, 1.7% and 3.3% for patients treated with brodalumab, secukinumab and ixekizumab, respectively, mostly in the form of oral or vulvovaginal candidiasis.11 Some cases of leucopenia and neutropenia have been observed in pivotal RCTs, although all of them were of grade 1 severity and did not correlate with the occurrence of infection. Of note, the risk for serious opportunistic infections, such as tuberculosis, appears not to be increased in patients treated with IL-17-targeted agents.69

Three JAK inhibitors are approved for the treatment, in combination with methotrexate, of RA in patients who have not responded or are intolerant to at least one DMARD. Tofacitinib (Xeljanz, Pfizer) inhibits JAK1, JAK2 and JAK3 (and, to a lesser extent, TyK2), whereas baricitinib (Olumiant, Eli Lilly) has a more selective action on JAK1 and JAK2 (91) and the recently approved upadacitinib (Renvoq, AbbVie) has largely selective action at JAK1. The JAK family plays a crucial role in immune cell signalling and differentiation via the signal transducer and activator of transcription (STAT) pathway. Indeed, loss-of-function mutations in the JAK3 gene induce a clinical phenotype of severe combined immunodeficiency.89 In accordance, JAK inhibitors exert a deleterious impact on adaptive immunity by decreasing CD4+ T- cell expansion and Th1 and Th17 differentiation, among other effects.90 91 A pooled analysis from phase II–III RCTs and long-term extension studies on tofacitinib reported an overall rate of serious infection of 3.1 per 100 patient-years, with age, diabetes, corticosteroid therapy, low lymphocyte counts and tofacitinib dose acting as independent risk factors. The most common sites were pneumonia and skin and soft tissues, and most patients required temporary discontinuation of therapy.92 The incidence rate for HZ (3.9 per 100 patient-years) was higher than those estimated for other infectious events, with approximately 8% of patients experiencing disseminated or multidermatomal forms. Geographical variations for HZ have been reported, with higher incidence rates in Asian individuals as compared with white or black patients. Patient age ≥65 years also increases the risk of HZ, whereas no clear impact has been shown for the tofacitinib dose used (5 mg twice daily vs 10 mg two times per day).92 In addition, the incidence of tofacitinib-associated HZ seems to be increased for patients with RA that receive concurrent corticosteroid therapy.93 After excluding uncomplicated HZ, TB was the more common opportunistic infection found among tofacitinib-treated RA patients (0.21 cases per 100 patient-years), although the risk greatly varied according to the regional background incidence rate.94 Other reported opportunistic infections included oesophageal candidiasis, cytomegalovirus disease and PCP.95 Similar infection risks have been reported for both baricitinib and upadacitinib within clinical trial data.96–98 Is it too early to suggest that the risk of HZ seems lower with filgotinib, the more recent JAK inhibitor still pending approval, than with the three other agents and to speculate on the reason for this preliminary finding (that would point to a particularly selective action on JAK1.99 100

One cytotoxic T-lymphocyte-associated antigen-4-immunoglobuln G (CTLA-4-IgG) fusion protein (abatacept (Orencia, Bristol-Myers Squibb)) has been approved for the treatment of RA, juvenile idiopathic arthritis and active PsA. Abatacept blocks the interaction between T cells and antigen presenting cells by binding to both CD80 (B7-1) and CD86 (B7-2), thus preventing CD28 interaction. Available data suggest that abatacept, which has a distinct mechanism of action upstream of other agents, does not induce a relevant increase in the risk of infection. An observational cohort study that included patients with RA that were initiated on a non-TNFI agent from 2010 to 2017 reported adjusted relative risks for abatacept of 0.94 (95% CI 0.81 to 1.08) and 1.00 (95% CI 0.88 to 1.14) as compared with rituximab and tocilizumab, respectively.101 A retrospective observational study based on insurance claims databases concluded that the risk of infection-related hospitalisations and associated costs among TNFI-experienced patients with RA was actually lower with abatacept than TNFI and other agents (such as tocilizumab, rituximab or tofacitinib).102 The lack of an apparent impact on the incidence of infection is consistent for opportunistic pathogens, including HZ, dermatomycosis, candidiasis and endemic mycoses.103 Although the occurrence of post-transplant lymphoproliferative disorder involving the central nervous system seems to be increased among solid organ transplant recipients treated with belatacept (another structurally similar CTLA-4-IgG fusion protein),104 the incidence of malignancy among patients with RA does not significantly differ between abatacept and other biological DMARDs.103

Rituximab (MabThera, Roche, MabThera, Genentech and biosimilars) is a genetically engineered chimeric monoclonal antibody that targets CD20 on B-cells. It is approved, in combination with methotrexate, for the treatment of adult patients with RA with moderately to severely active RA who have inadequate response to TNFI. Since neither B-cells precursors nor plasma cells express CD20, the use of rituximab does not immediately impair immunoglobulin production, although HGG may subsequently occur with cumulative treatment courses, with an incidence rate estimated at 2.7 events per 100 patient-years after a mean interval of 64 months.105 As detailed below, the presence of baseline HGG at the initiation of therapy increases the subsequent risk of infection.106 107 Although the development of rituximab-induced HGG leads to a well-established susceptibility to bacterial pathogens, available evidence suggests that rituximab does not impact on the overall risk of infection in the RA population, as suggested by a meta-analysis that included data from more than 9000 patients.108 A recent meta-analysis that pooled data from 7 studies and 3480 patients did not show significant differences either between rituximab associated with methotrexate and methotrexate alone in terms of the occurrence of infection.109 In addition to infections caused by capsulated bacteria, opportunistic infections in the form of PCP110 and chronic enteroviral meningoencephalitis111 have been reported. In this line, the incidence of PCP has been reported to be significantly higher with rituximab than with TNFI (adjusted HR (HR) 3.2, 95% CI 1.4 to 7.5).112

Finally, the oral phosphodiesterase-4 inhibitor apremilast (Otezla, Celgene Corporation) is approved in Europe, Canada and the USA for the treatment of active PsA. The occurrence of serious infections in pivotal RCTs on apremilast was rare, with comparable rates across different exposure groups.113 114 Such a favourable safety profile has been subsequently confirmed in observational post-marketing studies, with incidence rates for HZ (0.9 cases per 100 patient-years) and HCV reactivation actually lower than those observed for TNFI and other DMARDs.115

Box 2

Recommendations for prevention and management of infections induced by other biologics than TNFI

  • IL-6-targeting agents increase the overall risk of infection to a degree similar to TNFI and other biological DMARDs. The diagnosis of infectious episodes may be hindered due to the suppression of CRP synthesis while on therapy.

  • Although the risk of progression from LTBI to active TB seems to be lower than for TNFI, screening for LTBI should be performed before starting IL-6-targeting agents. Anti-TB therapy should be offered to patients diagnosed with LTBI.

  • IL-17-targeting agents are associated with a dose-dependent increased risk of mucocutaneous candidiasis, usually in the form of oral or vulvovaginal candidiasis and mild to moderate in severity.

  • JAK inhibitors increase the risk of HZ (including disseminated or multidermatomal forms), with geographical variations in incidence rates. The risk is higher for patients ≥65 years and those receiving concomitant corticosteroid therapy. Antiviral prophylaxis with (val)aciclovir may be individually considered in patients with multiple risk factors.

  • The development of rituximab-induced HGG may lead to an increased susceptibility to capsulated bacteria. Age-appropriate vaccinations for S. pneumoniae, Haemophilus influenzae type B and influenza should be administered.

  • Rituximab may influence the risk of reactivation of HBV infection. Screening for chronic HBV infection should be performed before starting therapy. Antiviral prophylaxis while on therapy and for at least 12–18 months after the last dose should be offered to HBsAg-positive patients, with preference given for agents with high genetic barrier to HBV resistance (eg, entecavir). Monitoring for HBV reactivation should be performed for at least 12 months after the end of prophylaxis. Prophylaxis (usually with lamivudine) should be offered to HBsAg-negative/anti-HBc-positive patients to prevent reactivation of resolved HBV infection.

  • DMARD, disease-modifying antirheumatic drug; HBV, hepatitis B virus; HZ, herpeszoster; LTBI, latent tuberculosis infection; TNFI, tumour necrosis factor-α inhibitor.

Prediction of infectious risk in patients with RA or PsA

The prediction of infectious complications in patients with RA or PsA receiving targeted and biological therapies remains a challenge. First, underlying disease may influence the risk of infection even before the start of the immunosuppressive regimen. Second, there are no enough validated biomarkers that can be routinely used to predict the risk of infection, and most studies are only based on the assessment of clinical risk factors for infection. Finally, given the relatively low rate of infectious complications in this population, intervention strategies for preventing infections are difficult to implement.

There has been an interest in evaluating the usefulness of monitoring of lymphocyte subpopulation counts and immunoglobulin levels for predicting the risk of infection. The type of marker used depends on the mechanism of the immunosuppressive drug: patients receiving tofacitinib are monitored by T-cell counts, while rituximab needs monitoring of immunoglobulin levels, but not B lymphocytes, since they are completely or very deeply suppressed in almost all patients. In patients with RA receiving tofacitinib, a total lymphocyte count <500 cells/mm3 was associated with an increased risk of serious infections, although only a minority of patients in this cohort had profound lymphopenia.116 Another study evaluating tofacitinib showed that a CD8+ T-cell count ≤211 cells/mm3 at baseline predicted the development of clinically significant infections.117 The presence of HGG (IgG levels of less than 6–7 g/L) at the time of initiation of rituximab was associated with a significantly higher risk for severe infection (usually pneumonia and skin and soft tissue infection) in two cohorts of patients with autoimmune diseases with an HR of 2.36 in one study106 and an OR of 4.9 in the second one.118 Low B-cell counts were not an identified risk factor for infection in these studies. Other significant clinical predictors of severe infection included age, the presence of an autoimmune disease other than RA and diabetes.106 It is not known whether these patients may benefit from antibacterial or antiviral prophylaxis (eg, with co-trimoxazole or acyclovir). Because these tests are easy to perform and inexpensive, they can be used in a case by case basis, but it needs further validation before their use in the routine clinical practice.

A clinical score named RABBIT Risk Score for Infections has been developed to predict the risk of infection in patients with rheumatic diseases.119 This score includes age, Health Assessment Questionnaire (HAQ) score, history of severe infection in the last 12 months, chronic lung disease, chronic kidney disease, number of previous treatments with DMARD, current dose of steroids and current drug used (TNFI, abatacept, rituximab, tocilizumab or non-biologic DMARD). The score was developed and validated in two different cohorts including several thousands of patients. An online, easy-to-use, calculator of the score is found in the website (

Some specific tests for measuring the risk of infection have also been assessed in patients with rheumatic diseases. A test measuring ATP after non-specific stimulation (ImmunKnow) was evaluated in a cohort of 134 patients with RA mostly on infliximab. The mean ATP level was significantly lower in patients with infection compared with those without infection.120 Also, the use of indirect surrogate markers for immunosuppression, such as EBV DNAemia, may help identify patients at higher risk for infection in these populations.121 The multibiomarker disease activity (MBDA) test is a biomarker assay composed by 12 proteins including CRP, serum amyloid A, interleukin 6, matrix metalloproteinase-1 and -3, among others. A higher score of the MBDA has robustly been associated with a significant higher rate of serious infection events.122 Although interesting, the performance of these tests for individualising preventive strategies has not been validated and thus the benefit of their routine use in the clinical practice needs yet to be proven.

Prevention of biologics-induced infections

There are few interventional trials assessing preventive strategies against infection in patients with RA or PsA, so most data come from observational cohorts and case series. Besides the use of antiviral prophylaxis for patients with hepatitis B during rituximab therapy123 and the need for therapy for LTBI in patients receiving TNFI,9 most recommendations are of low evidence, and therefore risk and benefits of these measures need to be taken into consideration. Ideally, all patients with RA or PsA diseases should benefit from an infectious disease evaluation before the start of targeted and biological therapies. A prolonged careful revision of the exposure history (eg, travel history or exposure to tuberculosis) and evaluation of immunisation history should be obtained. The European League Against Rheumatism (EULAR) has recently updated their recommendations for vaccination of adults with rheumatic diseases124 and the reader is referred to this document to have a comprehensive guidance on the use of vaccines in these populations.

A vaccination catch-up schedule should be performed depending on vaccination records and seroprotection. This includes influenza vaccine, dTP booster and pneumococcal vaccine hepatitis B/A should be proposed to patients with risk factors.124 125 There is no clear evidence on the best anti-pneumococcal vaccination strategy in patients receiving biologics. Most guidelines recommend a single dose of 13-valent pneumococcal conjugate vaccine (PCV-13) followed by a 23-valent polysaccharide vaccine (PPV-23) 8 weeks later.124 While PPV-23 increases the coverage of vaccine serotypes, it does not boost the immune response of previous PCV-13 vaccination126 and may be associated with vaccine hyporesponse,127 thus potentially jeopardising future anti-pneumococcal vaccine responses. On note, in case of previous vaccination with PPV-23, PCV-13 needs to be administrated at least 1 year later. There are no data on the efficacy of booster doses of PCV-13.128 Non-live vaccines can be administered before or during immunosuppressive therapy, but a better response is expected if vaccination is performed before the introduction of immunosuppressive drugs. Some studies have shown an appropriate immunogenicity of non-live vaccines in rheumatic patients, particularly for influenza.128 Live vaccines are contraindicated in patients under targeted and biological therapies, and should be administered as soon as possible in the initial assessment. Live vaccines include varicella, zoster live vaccine and MMR. Vaccination with yellow fever vaccine should be proposed before starting the immunosuppressive therapy if it is expected that the patient travels to endemic areas. A new inactivated zoster vaccine (Shingrix) has been approved for its use in those over age 50 and it is not contraindicated in immunocompromised patients, although there are very little data on its safety or efficacy in patients with rheumatologic or other inflammatory diseases.129 Table 4 shows the recommended schedule for vaccination in patients before and during immunosuppressive therapy. Table 5 shows the period after discontinuation of immunosuppressive drugs for vaccinating with live vaccines. Patients with a positive TST or IGRA test should be investigated for active tuberculosis. In case of a negative work-up, a treatment for latent tuberculosis should be initiated, particularly if an anti-TNF-α drug is started,31 current regimens include 9-month isoniazid, 4-month rifampin130 and 3-month isoniazid/rifampin. Rifampin-based regimens are better tolerated and more likely to be completed given their shorter course, and they should be used preferentially if possible. There is no need to complete the whole antimycobacterial preventive regimen before the introduction of the immunosuppressive drugs, and they can be started at soon after starting preventive therapy.31 130

Table 4

Vaccination strategies for patients with RA or PsA receiving targeted and biological therapies

Table 5

Estimated waiting period for vaccination with live vaccines after discontinuation of immunosuppressive drug

There are no interventional trials to assess the efficacy of co-trimoxazole prophylaxis in rheumatologic patients receiving immunosuppressive drugs. Given the overall low rate of opportunistic infection in this population, universal prophylaxis is not currently recommended.31 130 A simple rule of using co-trimoxazole prophylaxis in patients with high-dose steroids (>30 mg for >1 month) or those with combination immunosuppressive therapy, particularly in patients with low CD4+ T-cell counts, may target those patients at higher risk for opportunistic infections. A study involving more than 700 patients with RA showed that age >65 years, coexisting pulmonary disease and use of glucocorticoids were associated with PCP.131 Giving co-trimoxazole prophylaxis to these patients with at least two of those risk factors decreased significantly the incidence of PCP in this population. Antiviral prophylaxis with acyclovir/valacyclovir for prevention of HSV and VZV infections is not routinely recommended either, but may be introduced in patients with recurrent infections.130 131 Prevention of hepatitis B reactivation in patients receiving immunosuppressive therapy is a key measure and specific guidelines in different populations have been published. The higher risk for reactivation and subsequent complications are in any HBsAg-positive patient or in anti-HBc-positive patients receiving rituximab, and these patients should receive antiviral prophylaxis with entecavir or tenofovir (mainly for Ag HBs) or lamivudine (mainly for anti-HBc). Because of the risk of late reactivation, antiviral therapy should be continued for at least 12–18 months after the discontinuation of rituximab. Anti-HBc-positive patients without evidence of active disease (ie, HBsAg negative and HBV DNA negative) receiving other immunosuppressive therapy should be monitored with HBV DNAemia every 1–3 months. In patients dependent on rituximab with low IgG and serious infections, supplementation with intravenous immunoglobulin may be considered.


The recently expanded field of biotherapies for the accurate management of RA and PsA has now rendered more complex the contribution of the Infectious Diseases specialist to the management of sepsis in the rheumatic setting. In addition, we acknowledge the relative value of pivotal trials to detect rare infectious events contrasting with the more interesting contribution of postmarketing surveillance and open-label extension studies. Nevertheless, the role of comorbidities and steroids in combination with prior or ongoing other targeted and biological therapies led to complexities in the assessment of specific attributable risk of infection. Finally, we claim for the need for multicentre registries and multidisciplinary approaches, for new vaccines trials in RA and PsA and for better defining when and how biologics can be restarted after severe infections.


View Abstract


  • Handling editor Josef S Smolen

  • Contributors OL designed the review, wrote introduction and conclusion, and wrote the final version. All coauthors wrote one paragraph each.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

  • Patient consent for publication Not required.

  • Provenance and peer review Not commissioned; externally peer reviewed.