Safety of Anti-Tumor Necrosis Factor Therapies in Arthritis Patients

Purpose. Inflammatory and rheumatic arthritis remain leading causes of disability worldwide. The arthritis therapeutic area commands the largest market for the prescription of biological and non-steroidal anti-inflammatory drugs (NSAID). Yet biotechnology and pharmaceutical companies conducting research and providing therapeutics in this area frequently face challenges in patient safety. The purpose of our study was to assess safety of anti-tumor necrosis factor therapies in arthritis patients. Methods: The present study systematically reviews adverse events of biologicals alone or in the presence of NSAIDs and other immunosuppressant therapeutics such as disease-modifying antirheumatic drugs (DMARD). We assessed the rheumatology literature that included clinical trials with anti-tumor necrosis factor (TNF) biologicals and case reports published between 2010 and 2014. Results: Currently approved anti-TNF biologicals in arthritis include the monoclonal antibodies infliximab, adalimumab, certolizumab pegol and golimumab, and the fusion protein etanercept. The most frequently-reported adverse event was infection. We grouped the adverse reactions as immune-mediated, hypersensitivity syndrome reactions including cutaneous and hepatic manifestation, neurological, hematological, and malignancy. Discussion: Most adverse events are due to the failure of host immunological control, which involves susceptibility to the drug itself, or de novo infection or reactivation of a latent bacterial or viral infection, often with a different expression of disease. Drug-induced liver injury associated with anti-TNF biologicals must be kept in mind when evaluating patients with increased liver enzymes. Conclusion: Risk assessment in individuals undergoing treatment with biologicals represents a step towards achieving a personalized medicine approach to identify those patients that will safely benefit from this therapeutic approach. Patients and physicians must be alert of anti-TNF agents as potential causes of druginduced liver injury and monitor the therapies. Personalizing therapeutic pharmacovigilance promises to optimize benefits while minimizing side effects. This article is open to POST-PUBLICATION REVIEW. Registered readers (see “For Readers”) may comment by clicking on ABSTRACT on the issue’s contents page.


INTRODUCTION
The intracellular destruction of pathogens by phagocytes provides the first line of defense against bacterial and viral infections.The apoptosis of phagocytic macrophages induces pro-inflammatory cytokines, leukotriene and prostaglandins that localize the infection at the site of entry.The attraction of leukocytes to tissues is essential for inflammation.The process is controlled by chemokines, which have an important role in the pathophysiology of inflammatory diseases.Tumor necrosis factor (TNF)-α is a pivotal cytokine that acts upon both the innate and the adaptive immune systems (1)(2)(3).At the same time, cytokines and chemokines represent potential targets for therapy.Cytokine technology uses human receptor elements linked to functional proteins to create potent soluble inhibitors of cytokine function.
There is a multitude of potential drug targets in the human autoimmune diseases field, which creates a great opportunity to reverse or prevent inflammation-induced tissue damage.Exposure to anti-TNF therapeutics impairs the production of T helper (Th) 1 cytokines.In patients with rheumatoid arthritis (RA), therapeutic intervention with anti-TNF biologicals restores the low proliferative responses of peripheral blood mononuclear to mitogens, thus suggesting reversibility of this process.Treatment with anti-TNF biologicals in RA led to accumulation of Th1 CD4 + T cells in peripheral blood (4).Occasionally, treatment with anti-TNF biologicals in RA increased peripheral T cell reactivity to several microbial antigens with a significant increase in the production of interferon (5).In inflammatory rheumatic diseases, the proinflammatory effects of cytokines lead to inflammation, while the anti-inflammatory effects of biologicals at the level of the cartilage and osteocytes re-establish the balance between the proand anti-inflammatory messages.At the same time, there is a significant need for rapid target validation and post-market pharmacovigilance studies.
Clinically, TNF-α inhibitors have shown efficacy in inflammatory and autoimmune disorders.However, by increasing the reactivity of peripheral T cells to specific antigens, TNF-α also stimulates the antimicrobial defense mechanism (6).TNF-α is a key cytokine in the defense system against infectious diseases, and the efficacy of TNF-α inhibitors is paralleled by susceptibility to a variety of infections (7)(8)(9).Most reports to date have described increased susceptibility to intracellular pathogens in patients with underlying chronic infections.TNF-α-mediated pathways regulate the molecular interactions between cellular and viral factors within cells.Failure of host immunological control involves reactivation of latent infections.Nonetheless, TNF-α inhibitors have displayed a reasonable safety profile in the setting of some chronic viral infections, and in certain circumstances have demonstrated adjunctive activity in the treatment of these infections.Given the high prevalence of chronic viral infections in patients who are candidates for anti-TNF therapy and the potential for these agents in the treatment of chronic illness, additional studies are needed to assess the risks and benefits of such therapy in the individuals exposed to biologicals.While clinical trials of biologicals in RA note only slight infections, post-marketing analysis has clearly demonstrated an increased susceptibility to infections, especially those caused by Mycobacterium tuberculosis (8).

MATERIALS AND METHODS
The present study systematically reviews adverse events of biologicals alone or in the presence of NSAIDs and other immunosuppressant therapeutics such as DMARDs.We carried out a comprehensive search on Medline.We used Ovid and PubMed mesh database.Additionally, we assessed the rheumatology literature that included clinical trials with anti-TNF biologicals and case reports published between 2010 and 2014.
Incidence rates of AEs are compiled from data reported in clinical trials of individuals with RA, AS or JIA using the terms "infliximab", "adalimumab", "etanercept", "certolizumab" or "golimumab," and "clinical trial".Further searches were performed for each section separately in order to get a more detailed picture of the AEs under investigation, using terms such as "infection," "hypersensitivity", "anaphylaxis", "cutaneous", "allergy", "lupus", "autoimmune," "rash," "neurological", "neuropathy", "demyelinating," "progressive multifocal leukoencephalopathy," "cytopenia", "anemia", "leukopenia", "neutronpenia", "thrombocytopenia", "granulocytopenia," "pancytopenia," "lung", "respiratory", "interstitial," "pulmonary," "liver failure," "hepatotoxicity," "hepatitis," "cardiac failure," "malignancy," "cancer" or "neoplasm," along with the name of each drug.Inclusion criteria were studies performed in human patients with rheumatic conditions treated with anti-TNF agents published between 2010present.Nevertheless, there are limitations in the interpretation, since data collection by various centers may introduce great variability in the outcomes.Moreover, there is patient heterogeneity depending on the inclusion/exclusion criteria of the study, as well as, the variants taken into analysis or the aim of the study.An additional source of variability is the statistical methods used especially in clinical trials.Thus, the data interpretation differs from one publication to the other.

Adverse Events
The rate of AEs in patients with inflammatory arthritis treated with anti-TNF biologicals in placebo-controlled clinical trials are shown in Table 1.Different treatment regimens included anti-TNF biologicals or the DMARD like methotrexate (MTX), in combination with placebo or with one another.The main AEs of IFX were infections, infusion reactions and abnormalities in hepatic enzymes levels, while the presence of anti-drug antibodies (ADA) and autoantibodies like antinuclear antibodies (ANA) and anti-doublestranded DNA (anti-dsDNA) antibodies was associated with a loss of efficacy and a higher risk of AEs.Incidence of AEs were similar between IFX and placebo in patients co-treated with MTX, with a trend towards higher incidences of infections and infusion reactions between IFX monotherapy and placebo (10)(11)(12)(13)(14)(15).For ETN, the main AEs were infections, abnormalities in hepatic enzymes levels, neutropenia and infusion reactions.The rates of AEs were similar between patients treated with ETN or with MTX.Infusion reactions occurred more frequently during ETN injections compared to placebo injections (16)(17)(18)(19).The main AEs of ADM were infections and infusion reactions.The incidence of infusion reactions was higher for ADM than placebo (20).Rates of serious AEs were low in CZP patients.A trend towards higher odds of developing serious AEs was noted for CZP compared to placebo, primarily due to higher odds of serious infections, regardless of MTX cotreatment (21-23).A dose-dependent trend towards higher rates of AEs was observed in patients treated with GLM, owing particularly to higher rates of infusion reactions and serious infections (24-26).These were consistent with the warnings included in the product information of the drugs (27-31).

Infections
Anti-TNF biologicals are recognized as risk factors for serious infections (27-31).Anti-TNF biologicals (especially IFX and ADM) are disproportionately associated with infections in the French research axed on tolerance of biotherapies registry and the Portuguese spontaneous reporting database (32,33).In RA cohorts specifically, IFX use was a risk factor for opportunistic infections compared to ETN and MTX (32,34).Concomitant immunosuppressant treatment, especially steroids, represents an additional risk factor for opportunistic infections among patient receiving anti-TNF treatment (32,34).
While similar patterns were reported in these studies, the individual odds ratios of developing infections varied, as they included patients of different ethnic backgrounds from different areas of the world, patients with RA exclusively or patients with any condition treated by anti-TNF biologicals, as well as different treatment regimens depending on dose, frequency or physician preference.Adherence to several guidelines can help lead to a safer anti-TNF treatment course (35).These include screening for active or latent tuberculosis (TB), as well as vaccination against hepatitis B virus (HBV), varicella zoster virus, annual influenza, and human papilloma virus in young females.Opportunistic bacterial and fungal infections have been documented, some with serious AEs (35).

Tuberculosis
Reactivation of latent TB or de novo TB infection is one of the most important infectious AEs in patients treated with anti-TNF biologicals.In patients, treatment with anti-TNF biologicals (IFX, ADM or ETN) increased the risk of developing TB compared to the general population (36)(37)(38).Rates of TB were also higher in RA patients not exposed to anti-TNF treatment compared to the general population (38).Among patients treated with anti-TNF biologicals, IFX and ADM carry risks similar to one another, while monoclonal antibodies are associated with higher odds of developing TB than ETN (39,40).
As over one third of the world population may carry latent TB infection which can lead to active disease, especially under conditions of immunosuppression, it is important to detect and treat this infection prior to commencing anti-TNF treatment (41,42).Guidelines for TB screening in future anti-TNF patients include detailed clinical history, physical examination, chest radiograph, and a tuberculin skin test (43).Exposure to anti-TNF biologicals increased the risk of developing active TB in a Taiwanese RA sample with a low incidence of TB screening (44).In contrast, screening excluded TB with a high degree of certainty (sensitivity 0.83, specificity 0.74, positive predictive value 0.29 and negative predictive value 0.97) in a small sample of patients with inflammatory rheumatic diseases treated with anti-TNF biologicals in an area with a high endemic rate (45).Among patients born in regions with low endemic TB rates, being a healthcare worker might be a risk factor of contacting the disease (46).De novo contamination was also observed in patients receiving anti-TNF treatment when coming into contact with infected individuals or traveling to regions with high endemic rates (47).

Non-tuberculosis Bacteria and Fungi
Non-TB bacterial infections include non-TB mycobacteriosis, listeriosis, legionellosis, staphylococcemia, salmonellosis, nocardiosis and pneumocystosis, with presentations including cutaneous manifestations, pulmonary presentation (predominantly pneumonia) and neurological manifestations (predominantly meningitis).Anti-TNF biologicals were risk factors for non-TB mycobacterial diseases compared to both RA patients not treated with these agents and to the general population.IFX and ADM were risk factors compared to ETN.RA was a risk factor among nonusers of anti-TNF biologicals compared to the general population, while older age and RA were risk factors for non-TB mycobacterial disease among anti-TNF users (38).Treatment with anti-TNF biologicals is a risk factor for cutaneous infections.Skin infections generally only have cutaneous manifestations, but may also present systemic features such as fever (48).
The risks of developing Legionella pneumophila or Listeria monocytogenes infections were also higher among anti-TNF treated patients compared to the general population in European samples (36,49,50), while IFX or ADM use was a risk factor for legionellosis compared to ETN (50).
Fungal infections include pneumocystosis, histoplasmosis, aspergillosis, cryptococcosis, candidiasis, actinomycosis, blastomycosis and coccidioidosis, with predominantly pulmonary presentation.In a comprehensive review of the MEDLINE and PubMed databases, invasive fungal infections associated with IFX occurred after a median 55 days and a median 3 infusions after initiating treatment, while occurring after a median 144 days in ETN patients.Invasive fungal infections in anti-TNF patients were lethal in 32.2% of the cases in which outcome was available (51).
Pneumocystis jirovecii colonization was detected in 25.6% of RA, AS and psoriatic arthritis patients treated with anti-TNF biologicals, of which half occurred with IFX.Risk factors include corticosteroids use, MTX use and IFX duration of >3 years (52).The onset of Pneumocystis pneumonia in the Food and Drug Administration Adverse Event Reporting System database occurred after a mean 21 ± 18 days since starting IFX, and after a mean 2.1 ± 1.3 IFX infusions.Pneumocystis pneumonia resulted in death in 27.4% of these cases (53).IFX was also a risk factor for coccidioidomycosis in a sample of inflammatory arthritis patients.The cumulative incidence of coccidioidomycosis in this sample was 1.1% (2.8% among IFX patients and 0.5% among patients receiving other medications) (54).

Viruses
Viral infections include HBV, cytomegalivirus, varicella zoster virus, herpes simplex virus, Epstein-Barr virus, and John Cunningham virus.In these cases, the affected organs are those in which the virus proliferates predominantly.Viruses are the predominant infections in pediatric and adolescent JIA patients, suggesting likely childhood diseases such as infections by members of the herpes virus family (55).
The main concern regarding viral infections in anti-TNF patients is infections and reactivation of HBV.As such, anti-TNF biologicals should be avoided in chronic HBV infection, classified as positive by the presence of HBV surface antigen, and in resolved HBV infection, classified as positive by the presence of anti-HBV core and/or anti-HBV surface antibodies (56).Treatment with anti-TNF was a risk factor for HBV seroconversion in a sample of RA patients (57).ADM was associated with a low HBV risk in an RA population, which could mean either that ADM suppresses HBV reactivation, or that ADM is not used out of concern in individuals with a high risk of HBV infection (58).While a diagnosis of RA was a risk factor for slightly elevated aspartate aminotransferase (AST) and alanine aminotransferase (ALT) in an RA sample with preexisting HBV infection, these did not exceed the upper limit of normal (59).Viral reactivation was not observed in small samples of RA patients with resolved HBV infection treated with anti-TNF agents (60,61).
Monoclonal antibodies were risk factors for varicella zoster virus infections compared to ETN (62).IFX or ETN were not risk factors for Epstein-Barr viremia in peripheral blood mononuclear cells in RA patients.Instead, a positive correlation was observed between Epstein-Barr viral load and disease activity in RA patients (63).

Parasites
The main parasitic infection is leishmaniasis (Leishmania sp.), which generally includes visceral or cutaneous presentation.Also infection with Strongyloides stercoralis with pulmonary and gastrointestinal symptoms are described (32).

Infusion Reactions
Anti-TNF biologicals are disproportionately associated with general disorders and administration site conditions in the Portuguese spontaneous reporting database (33).Infusion reactions accounted for 257 of 920 (27.9%) cutaneous AEs reported in a large sample of patients with chronic inflammatory rheumatic conditions treated with IFX, ADM or ETN, including erythema, urticaria, eczema or rash, which can be accompanied by pain and swelling.Such reactions occur with all anti-TNF biologicals, primarily in the first month of treatment (48,64).
Monoclonal antibodies are structurally immunonegenic as they contain sections that can be recognized by the immune system as non-self epitopes, leading to the production of specific ADAs (65).The presence of IgE ADAs was associated with acute infusion reactions in IFX patients (66).Anaphylaxis is an example of acute infusion reaction.The most commonly accepted mechanism of anaphylaxis reaction involves IgEmediated type I acute hypersensitivity.This hypersensitivity reaction is characterized by the release of mediators from mast cells, basophils and recruited inflammatory cells.Symptoms of anaphylaxis can occur between a few minutes to a few hours after exposure to the drug, and can include any combination of cutaneous, cardiovascular, respiratory distress, laryngeal edema and severe bronchospasm, gastrointestinal and neurologic/muscular symptoms, and severe hypotension (65,67,68).
IgE ADAs were also significantly associated with positive skin tests results in patients with previous immediate hypersensitivity reaction to IFX.In turn, skin tests positivity was also associated with early and severe reactions.These results support the hypothesis that infusion reactions to IFX are IgE-mediated, while skin tests with IFX preparations provide a predictive tool for these AEs (69,70).This was demonstrated in a study in which positive skin tests were reported in all cases of urticaria and in 5 of 8 cases of anaphylaxis.The remaining 3 of 8 cases of anaphylaxis occurred at the first dose, suggesting non-IgE anaphylaxis (69).These are also known as anaphylactoid reactions.Anaphylactoid reactions have symptoms similar to anaphylaxis reactions, but occur in an IgE-independent manner (65,67).
An overall rate of acute infusion reactions of 5.8% was observed in a retrospective analysis of 135 RA patients treated with IFX over a 9 years period (71).In this study, most infusion reactions were mild or moderate in intensity, and most occurred early during IFX treatment.The areas most commonly affected by acute infusion reactions were the neck and the head, followed by the skin.The most common of these were pruritis, headache, facial flushing and chest tightness (71).The most common delayed infusion reactions, occurring 1-14 days post-infusion, were moderate in severity, and included chest/respiratory symptoms, general manifestations and skin manifestations.These occurred after a mean 7.2 days.Most delayed infusion reactions also occurred during the early phases of treatment (71).
Risk factors for acute hypersensitivity reactions in patients treated with anti-TNF monoclonal antibodies include patient-specific factors such as disease, atopic phenotype and concomitant use of immunosuppressants, as well as drug-specific factors such as dose, duration, number of infusions and route of administration.The highest rate of IFX infusion reactions was found in RA patients, suggesting that infusion reactions may be, at least in part, mediated by the same mechanisms as the disease itself.While immunosuppressants may help limit the development of ADAs and thus of infusion reactions, the role of prophylaxis with acetaminophen and antihistamines is less clear.The dose and frequency of IFX administration are closely related to ADAs formation, and thus with infusion reactions, as ADAs develop primarily in patients receiving lower doses or intermittent treatment (65,67).
Desensitization was performed in a series of patients with clinically-observed hypersensitivity reactions to anti-TNF biologicals and positive skin tests.A standard 12-step protocol was used for desensitization, involving the administration of increasing doses at 15 min intervals.Manifestations occurring during desensitization were less frequent and less severe in nature.In cases in which hypersensitivity occurred during desensitization, patient-specific modifications were made to the protocol, usually involving the administration of antihistamines, and the protocol was repeated (68).ADM desensitization was performed using gradually increasing doses ranging from 5 to 15 mg at a time.Dose escalations occurred every 30 min, for a cumulative dose of 40 mg.The 40 mg were administered quicker using larger doses during subsequent infusions, and ADM was accompanied by antihistamines and antileukotrienes during the first 4 treatment sessions (79).

Immune System Disorders
Immune system disorders encompass AEs in which the immune system overreacts to non-threatening foreign substances or against the body's own tissues.Anti-TNF biologicals are disproportionately associated with immune system disorders (33).Anti-TNF biologicals are associated with autoimmune disorders, particularly cutaneous disorders, drug-induced psoriasis, inflammatory bowel disease and autoimmune hepatitis (64,84,85).Immune system disorders described in recent case reports are detailed in Table 3.

Autoimmune Adverse Events
Lupus-like syndrome is a common autoimmune AE.The incidence of lupus-like syndrome was 1.75% in a small sample of spondyloarthritis patients treated with IFX, ADM or ETN (100).Specifically, the incidence of IFX-induced systemic lupus erythematosus was 1.3% in a recent systematic review (101).A high incidence of autoantibodies including ANAs, anti-dsDNA antibodies, anticardiolipin antibodies, anti-histone antibodies, anti-nucleosome antibodies and antineutrophil cytoplasmic antibodies positivity was noted in these patients (100, 101).Several recent case reports describe drug-induced lupus in patients treated with anti-TNF biologicals (102)(103)(104)(105)(106)(107)(108)(109)(110)(111)(112).While manifestations were varied, a common feature was the presence of autoantibodies.Levels of autoantibodies decreased after discontinuation of anti-TNF biologicals in most patients, and even returned to normal in some (102).Autoantibodies were negative in a patient diagnosed with discoid lupus erythematosus (112).
The first step in the treatment of drug-induced systemic lupus erythematosus is the discontinuation of the offending drug, while treatment with corticosteroids and immunosuppressive agents may be required to achieve full symptoms resolution (113).The pathogenesis of systemic lupus erythematosus is believed to involve a shift from Th1 cytokines towards Th2 cytokines, brought about by TNF-α inhibition with anti-TNF biologicals.This cytokines shift is then thought to bring about the production of autoantibodies and the development of systemic lupus erythematosus.Alternatively, TNF-α inhibition prevents apoptosis through decreased CD44 production, preventing the clearance of nuclear debris and apoptotic neutrophils, and thus promoting the production of antibodies against DNA and nuclear components (113).
Another common autoimmune AE of anti-TNF biologicals is drug-induced psoriasis.This was recently described in a review of the literature and of the French pharmacovigilance database.The majority of cases involved pustular lesions.The most commonly affected sites were the palms and/or soles.All of IFX, ADM and ETN were involved, with IFX accounting for more cases than the other two drugs.Recurrence of psoriasis can occur at the same location as the initial reaction or at a different site.It can further occur with a different anti-TNF biological (114).
Ten cases of psoriasiform eruptions are described in patients with inflammatory arthritis (RA and spondylo-arthropathy) treated with ETN or ADM.The condition presented pustular, plaque or guttate morphology.Plantar plaques or pustules are described in 4 of 10 (40.0%) cases.Nail involvement was not observed.Topical treatment was provided in 4 of 10 (40.0%) cases.The initial anti-TNF biological was interrupted in 8 of 10 (80.0%) cases.Among these, switching to a different anti-TNF biological led to either no improvement or only partial improvement.Further stopping anti-TNF biologicals led to complete symptoms resolution, or at least partial improvement of psoriasiform eruptions.Partial improvement was obtained with topical ointments without modifying treatment in the remaining 2 of 10 (20.0%) cases (115).Psoriasis lesions are also described in several recent case reports (87, 107,[116][117][118][119][120][121]. Autoimmune AEs with cutaneous manifestations also included intermediate bullous and cicatricial pemphigoid (122) and granuloma annulare (123).Other autoimmune AEs in patients treated with anti-TNF biologicals included de novo or exacerbation of autoimmune hepatitis (124)(125)(126)(127)(128) and antiphospholipid syndrome (129, 130).IFX is believed to lead to the unmasking of autoimmune hepatitis by binding to transmembrane TNF on the cell surface, thus inducing apoptosis and leading to the release of nucleosomes (131).

Neurological Adverse Events
The majority of neurological AEs reported in relation to IFX, ADM, CZP or ETN in the Food and Drug Administration Adverse Event Reporting System occurred in RA patients, with 393 of 772 (50.9%) events over a 10 years period.ETN and IFX were associated with the majority of these events (132).In RA patients specifically, the most common neurological AEs were central nervous system (CNS)/spinal demyelination, optic neuritis, peripheral neuropathy and facial palsy, with low incidences of transverse myelitis, leukoencephalopathy, other demyelinating disease, cerebrovascular disease, encephalopathy and CNS infections (132).Neurological adverse events are described in Table 4.
Among 33 reports of demyelinating disorders in French patients with rheumatic diseases treated with anti-TNF biologicals over a 3 years period, IFX was used in 15 (45.4%) patients, ETN in 12 (36.4%)and ADM in 6 (18.2%).Two thirds of these cases involved the CNS (encephalic involvement, transverse myelitis, optic neuritis) and one third the peripheral nervous system (chronic inflammatory demyelinating poly-radiculoneuropathy, Guillain-Barré syndrome).Cerebrospinal fluid analysis revealed raised protein levels in 4 patients, immunoglobulin oligoclonal bands in 11 and/or pleiocytosis in 4, while it was normal in 6 patients with CNS involvement.Cerebrospinal fluid analysis showed raised protein levels in 9 of 10 (90.0%) of patients with peripheral nervous system involvement in which analysis was performed (133).White matter lesions were common findings on magnetic resonance imaging in patients with CNS involvement.Anti-TNF biologicals were discontinued in all patients with CNS involvement and in 10 (90.9%) patients with peripheral nervous system involvement.Treatment with glucocorticoids was initiated in 15 (68.2%) and complete resolution was noted in 12 (54.5%),with partial resolution in 8 (36.4%) and no changes in the remaining 2 patients (9.1%) with CNS involvement.Intravenous immunoglobulins were administered in 8 patients, with 2 (25.0%) complete and 6 (75.0%) partial recoveries in patients with peripheral nervous system involvement.The remaining 2 patients did not receive any treatment, with neurological symptoms remaining either stable or improving.The anti-TNF treatment was continued in one patient, with neurological symptoms remaining stable.Two (25.0%) of the patients treated with intravenous immunoglobulins relapsed, one upon introduction of ADM and one without any subsequent anti-TNF treatment (133).Demyelination events occurred in 3 patients treated with GLM 100 mg for rheumatological indications in a large clinical trial.The incidence of demyelination events in GLM was 0.12 per 100 patient-years (134).
Progressive multifocal leuko-encephalopathy (PML) is a subacute CNS infection associated with the destruction of oligo-dendrocytes by John Cunningham virus.John Cunningham virus reactivation occurs under conditions of immunosuppression.Symptoms include confusion, motor impairment, impaired coordination, speech disorders and visual disturbances, while seizures are uncommon.Demyelinated areas are visible in the parietal and occipital regions.Diagnostic confirmation is usually obtained by identification of John Cunningham virus in cerebrospinal fluid or by brain biopsy (151, 152).The incidence of PML is low, and it is comparable between the general population and RA patients (152).A total of 34 PML cases were confirmed among autoimmune rheumatic diseases patients in the Food and Drug Administration Adverse Event Reporting System database.Among these, 15 patients were exposed to biological agents and 19 patients were exposed to other anti-inflammatory agents such as azathioprine, cyclosporin A and prednisone equivalents.One case is reported in an RA patient treated with IFX (153).IFX and ADM have been associated with PML in a recent review of the Canada Vigilance and World Health Organization adverse event databases (154).Two recent cases of PML are described (155, 156).A case of encephalitis associated with Epstein-Barr virus infection is also reported (129).

Hematological Adverse Events
Hematological AEs are important safety considerations of anti-TNF biologicals.Cytopenia describes a condition marked by a reduction in the number of blood cells, and it can include anemia, leukopenia, neutropenia, thrombocytopenia, granulocytopenia or pancytopenia.Transient neutropenia (neutrophil count <1.50×10 9 /L) is the predominant non-malignant hematological complication in patients treated with anti-TNF biologicals.Risk factors include a history of neutropenia on other medications or a low baseline neutrophil count.However, neutropenia can develop without co-medication as well.ETN was the biological most often associated with this hematological complication, and consequently most cases were reported in RA patients (157).Hematological adverse events in recent case reports are detailed in Table 5.
Neutropenia is described in 5 RA patients treated with ETN (158).In two patients, neutrophil counts dropped from baseline until they reached a low but stable level during treatment.ETN was continued and the patients are being monitored.A third patient experienced a significant drop in the neutrophil count while on ETN.A rebound in the neutrophil count to a low but stable level was observed upon ETN discontinuation and a switch to ADM.Mild neutropenia is described in a patient treated with ETN plus MTX.ETN was well tolerated after gradual MTX discontinuation.Neutropenia occurred in a fifth ETN patient, and any attempt to re-introduce ETN failed due to recurrent neutropenia (158).Neutropenia is also detailed in several case reports in which neutrophil counts improved following discontinuation of anti-TNF biologicals (159-161).Two patients had a history of neutropenia while receiving DMARDs, therefore the influence of co-medication such as MTX cannot be discounted (159).Another patient had no history of abnormal hematological parameters during MTX treatment (161).
Leukopenia is described in two patients treated with ETN and DMARDs.No improvement in the leukocytes count was achieved when the DMARDs were discontinued, but leukocyte levels normalized when ETN was discontinued (162, 163).
In a recent review, thrombocytopenia (platelet count <150×10 9 /L) was observed predominantly in anti-TNF patients without comedication, thus suggesting that this hematological complication is likely a consequence of these biologicals.IFX and ETN were associated with thrombocytopenia (157).Thrombocytopenia is described in two patients.Platelet counts rebounded soon after discontinuation of anti-TNF biologicals (164).
Lymphocytosis followed by neutropenia is described in a patient treated with ADM plus MTX.Discontinuation of MTX led to resolution of lymphocytosis, while neutropenia also resolved soon after ADM discontinuation.Therefore, ADM was only responsible for neutropenia in this patient (165).
Anti-TNF biologicals such as IFX, ADM and GLM improve anemia in RA and AS patients.On the other hand, ETN has no significant effects on hemoglobin levels (166, 167).
Pancytopenia is described in two patients treated with ETN plus MTX (168, 169).One of the patients developed severe multilobar pneumonia, although a pathogenic organism could not be definitively identified, likely due to prophylactic antibiotic treatment.The patient died of multi-organ failure resulting from septic shock of respiratory (169).
Pancytopenia associated with hemophagocytic syndrome complicated a case of drug-induced lupus (170).

Adverse Events Affecting Other Organs
Adverse events affecting other organs in recent case reports are detailed in Table 6.

Lung Disease
Anti-TNF biologicals are further associated with respiratory, thoracic and mediastinal disorders (33).Interstitial lung disease (ILD) describes a large group of pulmonary conditions associated with inflammation and fibrosis.
Non-infective respiratory, thoracic and mediastinal disorders are reported in 8 of 201 (4.0%) patients with longstanding RA treated with ADM (171).The incidence of ILD was 0.6% in a large postmarketing surveillance sample of RA patients treated with ETN (172).Treatment with anti-TNF biologicals (IFX, ADM and ETN) is thus a risk for the development of ILD in RA patients.RA itself and MTX co-treatment represent additional risk factors for ILD, while age ≥65 years, a history of ILD, and concomitant immunosuppressants are risk factors for fatal ILD in RA patients (173-175).Cases of ILD in CZP-treated patients have also been recently described, while serious noninfectious pulmonary AEs are reported in clinical trials in RA patients receiving GLM in the presence of MTX only (176).Lung biopsies are often needed to confirm ILD (175).The most common classifications of ILD were interstitial pneumonia, nonspecific interstitial pneumonia, organizing pneumonia, diffuse alveolar damage and lymphoid interstitial pneumonia (175).
ILD is described in 3 patients (177)(178)(179).Progressive resolution of symptoms occurred in a patient treated with prednisone (179), while little recovery was observed in two patients upon treatment with methylprednisolone (177,178).Differences could be explained by more extensive affected areas in the latter two patients, who were also older (177,178).In another study, exposure to ETN led to exacerbation of the condition in a patient with a history of ILD (180).
Upon CZP exposure, a patient with a previous episode of pneumonitis while on MTX treatment developed fatal fibrosing alveolitis with respiratory failure (181).Diffuse alveolar hemorrhage is described in two patients (182,183).Interstitial pneumonia (184)(185)(186)(187) and acute pneumonitis (188,189) were diagnosed elsewhere.Interstitial pneumonia was fatal in an elderly patient with a history of ILD treated with ETN (187).Lupus occurred in the presence of organizing pneumonia in another patient (190).

Hepatitis
Hepatic AEs in patients treated with anti-TNF biologicals are usually classified as infectious AEs, as they result from reactivation of viral hepatitis B. Alternatively, anti-TNF biologicals may also uncover autoimmune hepatitis, while instance of symptomatic, severe acute hepatitis are rare (191).Ghabril et al. (192) identified a total of 34 cases of drug-induced liver injury (DILI) associated with the use of TNF-α antagonists between 2003 and 2011 (26 IFX, 4 ETN and 4 ADM).
The incidence of elevations in ALT and AST over the upper limit of normal occurred in 5.9% of RA patients in a large sample treated with IFX, ADM or ETN (193).IFX, ADM and to a certain degree ETN were risk factors for ALT and AST elevations >2 times over the upper limit of normal compared with DMARDs (193,194).Asymptomatic AST and ALT elevations >10 times over the upper limit of normal were observed in a patient treated with ETN plus MTX.Liver function test results normalized after ETN interruption, but recurred with ADM.IFX was well tolerated despite persistently positive ANA (195).
Acute hepatitis with positive ANA and anti-dsDNA antibodies is reported in two patients.DILI was diagnosed in both patients as features of autoimmune hepatitis or sclerosing cholangitis were absent in biopsy, despite persistently high autoantibodies (196,197).

Sarcoidosis
Sarcoidosis is a granulomatous disorder that can affect multiple organs.Neurosarcoidosis is described in an RA patient treated with ETN (198).The mechanism of neurosarcoidosis is believed to involve the expansion of inflammatory meningitis into the brain parenchyma or into the spinal cord (199).Manifestations of neurosarcoidosis depend on the affected neuroaxis.For example, infiltration of granulomas into leptomeningeal and intraparenchymal structures can lead to cranial nerve palsies, basal meningitis or endocrine dysfunction, the consequences of which may be peripheral neuropathies and sensorimotor polyneuropathy (200).Cases of thoracic and pulmonary sarcoidosis are described elsewhere (201,202).Additional cases of granulomatous hepatitis and granulomatous interstitial nephritis are also presented (203,204).

Cardiovascular Effects of anti-TNF Biologicals
Based on data from Medicare and drug benefit programs (1994)(1995)(1996)(1997)(1998)(1999)(2000)(2001)(2002)(2003)(2004), the use of anti-TNF biologicals (ETN and IFX) was associated with an increased risk of heart failure compared to MTX in elderly (≥65 years) RA patients (205).Both ETN and IFX were risk factors for heart failure in a sample of 2121 younger (<50 years of age) RA patients.The risk with anti-TNF biologicals was comparable to that of DMARDs, which can be attributed in part to a relatively low overall incidence in this younger population (206).Based on this evidence, older age is a risk factor for heart failure in RA patients treated with anti-TNF biologicals.Furthermore, elderly anti-TNF patients were found to have a 4.2-fold higher risk of death from heart failure compared to MTX patients (205).Elsewhere, RA disease activity was the main risk factor for the 3 years incidence of heart failure, and overall anti-TNF biologicals generally shows more beneficial than detrimental effects with respect to the risk of heart failure, owing primarily to the reduction in the inflammatory activity of RA (207,208).The incidences of tachyarrhythmias and bradyarrhythmias were not different between IFX and placebo in a sample of 75 spondyloarthritis or RA patients (209).Furthermore, anti-TNF biologicals have positive effects on cardiovascular health by improving metabolic parameters, at least in the short term (210).
Cardiovascular AEs are described in a few RA patients treated with IFX and ETN (211)(212)(213)(214). Supraventricular tachycardia occurred within 3 hours of the 8 th IFX infusion in the first of these patients (211).A significant decrease in cardiac output (7.04 ± 2.3 to 6.12 ± 2.1 L/min) and in stroke volume (91 ± 29.0 to 83 ± 28.8 mL/beat), with non-significant increases in systolic blood pressure, diastolic blood pressure and total peripheral vascular resistance, also occurred as an infusion reaction in an IFX patient (212).Dilated cardiomyopathy was induced by IFX after 6 months of treatment in the third patient (213).Severe heart failure, reversible upon ETN discontinuation, developed in an AS patient (214).

Relationship between Anti-TNF Biologicals and Cancer
The relationship between anti-TNF biologicals and cancer is controversial.The following section provides a brief analysis of the current knowledge.Anti-TNF biologicals were disproportionately associated with benign, malignant or unspecified neoplasms (33).However, inflammation is a known risk factor for cancer, such that the increased risk of cancer observed in RA cohorts treated with anti-TNF agents could be a result of the underlying disease (215).
Overall, the risk of cancer is similar in RA patients treated with anti-TNF biologicals and the general population.Using data from randomized controlled clinical trials, IFX and ADM were associated with an increased risk of malignancies compared to placebo, particularly a non-significant trend towards a higher incidence of lymphomas (216,217).While the risk of lymphoma, particularly Hodgkin's lymphoma, appears increased in RA patients treated with anti-TNF biologicals, patients with RA carry a 2-3-fold higher risk of lymphomas compared to the general population.Risk factors for lymphoma include RA, predominantly in individuals with positive rheumatoid factor (218,219).When adding biologicals to the equation, no increased risk of cancers is usually found between patients exposed to anti-TNF agents or placebo.It is interesting to note that while anti-TNF monoclonal antibodies may be risk factors, no significant differences exist between anti-TNF agents and placebo when IFX, ADM and ETN are considered together.Looking at individual cancers separately and individuals biologicals separately would thus likely offer a more appropriate means to compare anti-TNF and placebo, yet no significant differences are apparent due to the relatively low incidence of these AEs (215).
Using data from 33 placebo-controlled trials, treatment with anti-TNF agents was not associated with an increased risk of cancer compared to placebo in RA patients treated for up to 2 years.A trend towards an increased risk of non-melanoma skin cancers was however observed (220).As such, treatment with anti-TNF biologicals is a risk factor for cutaneous malignancies (64).Skin neoplasms include melanoma, and non-melanoma cancers such as basal cell carcinoma (48).IFX and ETN are associated with non-melanoma skin cancer and with melanoma based on data from the US National Data Bank for Rheumatic Diseases (221).Anti-TNF biologicals were associated with non-melanoma skin cancer in a recent meta-analysis, but not with other types of cancer (222).Overall, these reports suggest that RA patients may be predisposed to higher rates of malignancies than the general population, especially lymphomas, while long-term exposure to biologicals appears safe.An alternative explanation could be that anti-TNF biologicals may cause cancer on their own but may also decrease the risk of cancer associated with the chronic inflammatory environment.Based on current knowledge, no definitive conclusions could be drawn with regards to the risk of cancer in anti-TNF patients, and studies with longer follow-up times may help elucidate the relationship, if any.

DISCUSSION
The present review discussed AEs in which anti-TNF biologicals were incriminated.However, additional factors such as co-medication should be considered when analyzing the causes of AEs.The majority of RA patients that are treated with anti-TNF biologicals are also taking NSAIDs as painkillers or as additional therapies.
Several NSAIDs play a prominent role in the history of idiosyncratic hepatotoxicity.For example, diclofenac was shown to produce hepatic injury since the 1980s.In 15-20% of patients taking the drug, aminotransferase levels are markedly increased (223)(224)(225)(226)(227)(228)(229)(230).Clinically, acute disease resembles acute viral hepatitis and chronic injury is similar to chronic hepatitis (227,231).At the histological level, the main lesion has been hepatic necrosis.The necrosis is non-zonal although it tends to be more marked in zone 3. Female gender is a risk factor for hepatotoxicity susceptibility.In addition, patients with osteoarthritis have a significantly higher incidence of hepatic injury than RA patients (230).Most cases have presented with the syndrome of acute hepatitis characterized by jaundice and to a varying degree by fatigue, anorexia, nausea and vomiting.Fever, rash and eosinophilia are uncommon, but were recorded in one report (232).As with other drug-induced hepatocellular injury, massive necrosis with fulminant hepatic failure and death were noted (230,233).Chronic active hepatitis was also described with diclofenac.The mechanism of this reaction was delayed-onset hypersensitivity (234).Rostom et al. (235) reviewed NSAIDs hepatoxicity in randomized controlled trials in arthritis patients.Laine et al. (236) further studied liver injury associated with diclofenac in a large sample of arthritis patients in a long-term prospective clinical trial, and concluded that this is not an uncommon AE.Moreover, Pilotto et al. (237) reported NSAIDs-related gastrointestinal bleeding.Diclofenac was associated with hospitalization in a case-crossover study, including peptic ulcer, bleeding and perforation (238).
The major side effects of NSAIDs which are the gastrointestinal complications, renal disturbances and cardiovascular events are presented in the seminal review by Harirfoorosh et al., (239).Arthritis patients using pain killers and NSAIDs are prone to cardiovascular disease (240) and renal damage (241).However, the population treated for RA presents a heterogeneity in patient population in terms of specific disease involvement, and ethnic diversity of the patients.Moreover, age group is an important parameter.The majority of patients are elderly individuals that, in addition to RA, may present another underlying disease (diabetes, hypertension, obesity) that may direct them to ADRs.Also, the type and small number of ADRs, when compared to the general population taking the same medication, make any conclusion regarding the risk of NSAIDs incrimination in interaction with biologicals almost impossible to judge.In many of these studies, the inflammation and inflammatory activators that exist continuously in RA is ignored.The target of both biologicals and NSAIDs is inflammation.This phenomenon is important since inflammation is by itself a risk factor of cardiovascular events.Biologicals and NSAIDs thereby, both should reduce the cause for increase cardiovascular ADRs.However, no study can identify the impact of separate component in an adverse event.
Gastrointestinal complications are frequent during the NSAIDs therapy for RA (239).Several factors including a history of gastric mucosa damage, age over 60 years, previous exposure and ADR to NSAIDs, and/or drug interactions with corticosteroids or anticoagulants increase the risk of developing gastrointestinal side effects.These factors may also influence the interaction between biologicals and NSAID.
Ibuprofen is a widely-used NSAID with antipyretic and analgesic properties that can produce an unpredictable hypersensitivity syndrome reaction likely caused by a combination of metabolic and immunologic factors.Immune-mediated components, such as T cell cytokines and chemokines, can exacerbate cellular responses and create complex pathways that lead to a variety of clinical manifestations including severe cutaneous reaction and of drug-induced liver injury (242)(243)(244).
Treatment of RA and related conditions with low doses of MTX may lead to steatosis.MTXinduced fibrosis can appear during treatment.Rarely, cirrhosis was reported (245,246).A case of MTX anaphylaxis allowed ETN to be continued in a patient co-treated with these two agents (247).
One of the most worrisome conditions is DILI.The clinical syndrome of acute liver damage relates, at least in part, to the apparent mechanism of injury.Hepatic injury induced by large single overdose of intrinsically toxic drugs, (e.g., acetaminophen) develops within 24 to 72 hours of intake and usually, is accompanied by renal failure.Regular intake of some toxic drugs (e.g., methotrexate) leads to slowly evolving chronic disease.Liver damage due to hypersensitivity-type of idiosyncrasy to NSAIDs, it is not dosedependent.Usually appears after one to five weeks of taking the drug unless there has been previous exposure and is preceded or accompanied by systemic features that are hallmarks of hypersensitivity (243).Hepatic injury attributable to metabolic idiosyncrasy may appear after weeks to months of taking the drug in the normally prescribed dose and usually presents without the systemic features.Organs other than the liver may be involved in the syndrome of DILI as the result of selective injury or as part of hypersensitivity reaction (244).Regardless of underlying disease, hepatotoxicity of NSAIDs was described in many cases (248).
However, RA per se may be responsible for susceptibility to certain NSAIDs.Patients with JIA and RA appear to be more vulnerable than others, and there is a positive correlation between activity of the underlying disease and susceptibility to the hepatic injury.Hepatic damage induced by aspirin occurs in 50% or more of patients with high but normal blood levels of the drug (> 15 mg/dl).This level can be achieved by the dose given in active RA.The injury is intrinsic.Bilirubin levels are normal or slightly elevated and jaundice is noted in less than 5% of cases.Values of aminotransferases are elevated, 5-40-fold.Biopsy has shown focal necrosis with mild inflammatory response in the portal areas and sinusoids and ultrastructural changes (249)(250)(251)(252)(253). The abnormality disappears promptly on stopping the drug.Overdoses of aspirin lead to microvesicular steatosis (249,254).
In a study reported by the Acute Liver Failure Study Group, prolonged administration of high doses of NSAIDs for RA resulted in irreversible acute liver failure that led to liver transplant or death (255).
Other cytotoxic drugs have been employed in the treatment of RA, although not as widely used (251).The effects of DMARDs on the liver, especially those of MTX, are important to consider as possible triggers for hepatotoxicity.Glycogen inclusions in the nuclei are described as prominent in DILI.These features are common findings in the liver of diabetics but can be also seen in nondiabetic individuals.Particular attention has been paid to the association with prolonged MTXexposure.Thus, the steatosis produced by MTX leads primarily to hepatomegaly as a clinical manifestation (256).
The diagnosis of drug-induced injury can often be difficult.The relationship between drugingestion and toxicity is not always clear.Patients may be taking multiple medications making identification of the offending agent difficult, and they may have concomitant diseases, which can produce similar clinical and laboratory features.Characteristics suggesting drug toxicity include good health prior to ingesting the drug, clinical illness or biochemical abnormalities developing after beginning the drug, and after the drug is withdrawn.If an immunologic reaction is suspected, the illness will generally recur upon reintroduction of the offending substance.However, rechallenge is not advised.
Dose and frequency related adverse events have not been reported in the clinical trials with anti-TNF medication.Infections generally tend to occur during the first 1-2 years of treatment with anti-TNF (32).Therefore after a prolonged immune suppression the initial infection is reactivated.However, the majority of non-viral opportunistic infections occurred in the first 6 months of treatment with anti-TNF (33).Regarding cutaneous AEs primarily appear in first few months of treatment (64).Specific infusion-related severe anaphylactic reactions related to the presence of anti-IFX ADAs; scheduled dosing is thus preferred over episodic dosing in order to avoid ADAs (66).Autoimmune AEs are also associated with the presence of autoantibodies (101).
Neurological AEs occurred between a few months to a few years of anti-TNF treatment (same wide range of treatment duration for all of IFX, ADM and ETN) (133).Haematologic effects such as thrombocytopenia, neutropenia, thrombosis occurred between a few weeks to several months of anti-TNF treatment (157).
We conclude that the number and severity of the ADRs have no relationship with dose or length of anti-TTNF treatment.The only exception is immune-mediated AEs, which tend to occur more frequently in the presence of ADAs or autoantibodies.The presence of ADA occur more frequently when treatment is not administered continuously.
As a result of the frequent co-treatment with anti-TNF biologicals and DMARDs or NSAIDs, the effects of all therapeutics need to be considered, and thus the superimposed effects of drug-drug interactions cannot be discounted for both immunemediated mechanisms and those resulting from abnormal drug metabolism and clearance.Pharmacovigilance practice is vital for the successful development, marketing and defence of pharmaceutical products.
The scope of the search for safe anti-TNF therapeutics should be broadened to the molecular level and cover various receptors involved in the side effects and also to explore interaction with other therapeutics that might be given concomitantly.In addition, to predict various side effects of anti-TNFs, identification of readily measured biomarkers is of therapeutic interest.We recommend laboratory monitoring of TNF levels in blood and the ADA to different anti-TNF therapeutics.
To avoid possible NSAID hypersensitivity we recommend the use of lymphocyte toxicity assay to the specific NSAID.

CONCLUSION
Adverse drug reactions remain a persistent concern in the prescription of many commonly-used therapeutic agents in clinical practice.Anti-TNF biologicals are efficient and generally safe in the treatment of patients with RA.Increased experience resulting from their extensive use has verified their overall safety profile with favorable risk-benefit ratio.However, caution is strongly advised in the administration of biological therapies.The use of anti-TNF therapies in combination with NSAIDs by individuals with chronic liver disease can provoke development of hepatotoxicity.Physicians should monitor patients for infections.Moreover, anti-TNF agents per se and in combination with other medication administered are potential causes of drug-induced, cardiovascular, gastrointestinal and liver injury.Monitoring remains the cornerstone of early diagnosis and effective management.From the point of view of pharmaceutical companies and health authorities, post-marketing surveillance continues to be paramount with every newly introduced agent.From the perspective of physicians and their patients, pharmacovigilance and personalized medicine provide the necessary tools for optimizing therapy while minimizing side effects.

Table 1 .
Incidence of adverse events in clinical trials

Table 3 .
Immune-mediated and autoimmune adverse events

Table 6 .
Adverse events in other organs