Abstract
Tumor necrosis factor (TNF) inhibitors were the first biologic drugs prescribed for the treatment of spondyloarthritis (SpA) and rheumatoid arthritis (RA). Although they provide significant improvement of signs and symptoms, TNF inhibitors need to be used frequently for a long period of time. The analysis of the follow-up of the largest national biologics registries has shown that the most important adverse effect of TNF inhibitors is infection, which is significantly higher than the non-biologic treatment group; reactivation of latent tuberculosis is three to four times more frequent in patients using monoclonal antibodies than soluble receptors. The only cancer site more frequent to be associated with TNF inhibitors in RA and SpA is the non-melanoma skin cancer. Paradoxical reactions do occur during anti-TNF treatment mainly in SpA, such as new manifestations or flares of acute uveitis, new onset of psoriasis, such as palmoplantar pustulosis, or new onset or flares of inflammatory bowel disease, which occurs especially during etanercept treatment.
Introduction
Spondyloarthritis (SpA) is a group of rheumatic diseases dominated by spinal symptoms, classified as axial SpA or by peripheral arthritis, classified as peripheral SpA . Axial SpA is subdivided into two types, ankylosing spondylitis (AS), which requires radiographic changes of the sacroiliac joints (according to the Modified New York Criteria) , and the non-radiographic axial SpA, which is mainly based on a combination of clinical symptoms, the presence of HLA-B27 antigen and signs of sacroiliitis at magnetic resonance imaging (MRI). The non-radiographic axial SpA can progress towards AS within a couple of years. The clinical symptoms of axial SpA include inflammatory back pain during at least 3 months with an age at onset before 45 years and at least one of the other “SpA-features”: arthritis, enthesitis (heel), uveitis, dactylitis, psoriasis, Crohn’s disease/ulcerative colitis, good response to non-steroidal anti-inflammatory drugs (NSAIDs), family history of SpA and elevated C-reactive protein (CRP) . Peripheral SpA requires peripheral arthritis compatible with SpA (usually asymmetric and/or predominant involvement of the lower limb), enthesitis, or dactylitis and at least one of the other SpA features, as mentioned before .
Exercises and physical therapy are very important in SpA, in addition to treatment with NSAIDs. The efficacy of most disease-modifying anti-rheumatic drugs (DMARDs) in SpA is rather disappointing, especially compared to rheumatoid arthritis (RA) . Sulfasalazine is effective in peripheral SpA and shows some efficacy on spinal symptoms as well, but not on enthesitis .
Tumor necrosis factor (TNF) inhibitors however give substantial improvement to SpA. Up to now, large placebo-controlled trials have demonstrated the efficacy of infliximab, etanercept, adalimumab, golimumab, and certolizumab pegol in SpA patients. Anti-TNF therapy is indicated in cases of persistent high disease activity (Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) > 4) and insufficient response to NSAIDs or sulfasalazine in case of peripheral arthritis . Patients with non-radiographic axial SpA also seem to benefit from treatment with TNF blockers . Next to the improvement of axial symptoms in SpA patients, TNF blockers are very effective in the treatment of peripheral arthritis and enthesitis .
An important concern is that the current guidelines for the treatment of AS and RA indicate that anti-TNF therapy need to be used for a long period of time. Recent publications have shown a maintained efficacy for 8 years with infliximab , 7 years with etanercept , 5 years with adalimumab , and 3 years with golimumab in AS. Long-term use of infliximab , etanercept , adalimumab , and golimumab in RA are also published. The new TNF inhibitor certolizumab pegol has also shown good initial results in AS and RA .
In this article, we will discuss the most common and important adverse effects of the TNF inhibitors comparing its presentation in RA and AS, as well as the paradoxical effects related to uveitis, psoriasis, and inflammatory bowel disease (IBD).
TNF inhibitors survival in SpA and RA
Before discussing their adverse effects in SpA and RA, it is important to emphasize that the analysis of the published literature about retention rate of TNF inhibitors indicates that it is higher in SpA compared to RA, according to data from important national biologics registries.
The Spanish Register of Biologics (BIOBADASER) analyzed drug survival and adverse effects in 1524 patients with SpA compared to 4006 patients with RA. Drug survival was significantly greater in AS compared with RA, and this difference was even larger with prolonged exposure: 0.88 versus 0.83 at 1 year, 0.82 versus 0.72 at 2 years, and 0.76 versus 0.65 at 3 years. Similar results were observed with the other SpA compared with RA; meanwhile, there was no significant difference between the different SpAs. The hazard ratio (HR) for discontinuation was 0.66 compared with RA, and side effects were significantly less frequent in SpA (17%) compared with RA (26%) .
A retrospective analysis of 770 French patients (57.1% with RA and 37.7% with SpA) using TNF inhibitors in the daily practice showed that the retention rate was significantly longer for patients with SpA compared to RA (HR = 1.6) .
In the Norwegian Disease Modifying Anti-Rheumatic Drug (NOR-DMARD) register, the 1-year retention rates of TNF inhibitors in 1268 treatment courses (847 in RA, 172 in psoriatic arthritis (PsA), and 249 in AS) were 65.7% in RA, 77.3% in PsA, and 77.5% in AS. The adjusted HR for treatment termination was 0.66 for AS versus RA .
In the Czech National Register (ATTRA), the long-term analysis of 310 patients showed that drug survival was also better in the AS patients compared with RA: 84% versus 78% after 1 year, and 72% versus 49% after 3 years of treatment .
The analysis of drug retention in a large Swedish cohort of 9612 patients with RA, 1417 with PsA, and 1652 with other SpA, revealed that drug survival of the first TNF inhibitor in the first year was higher in SpA compared with RA and PsA .
In the Italian Monitor Net database, the analysis of drug survival of 2640 patients with RA and 1220 with SpA who started their first TNF inhibitor showed that patients with SpA, especially AS, had a lower risk of drug discontinuation (adjusted HR = 0.81) in a median follow-up of 17 months. The drug survival was significantly higher in AS patients compared to RA: 0.87 versus 0.82 in 1 year, 0.73 versus 0.71 at 2 years, and 0.69 versus 0.62 at 3 years .
The drug survival of 842 AS patients using TNF inhibitors in the Danish Biologics Register (DANBIO) was 74% in the first year and 63% in the second year. Variables at baseline associated with longer drug survival were male gender, CRP > 14, and low visual analog scale (VAS) for fatigue .
Although anti-TNF therapy can cause rapid and sustained clinical benefit in patients with RA and SpA, its abrupt discontinuation after a long-term use can cause disease relapse in a significant number of patients in 12 months. A classical example in axial SpA is the German study where 42 AS patients, considered good respondents to infliximab, had their medication stopped after 3 years of continuous use; after 12 months, 41 (97.6%) of the 42 patients presented relapse . Submitted to reinfusion of infliximab after the relapse, 40 patients referred good clinical response . Another study considered that relapse is faster in patients with peripheral SpA. A group of 26 non-AS and non-PsA patients participating in a randomized clinical trial with adalimumab in SpA had their medication stopped after 12 or 24 weeks. After discontinuation, relapse was referred by 19 patients (73.1%) in a mean period of 10 weeks . Nevertheless, a Dutch study analyzing baseline predictors of response and discontinuation in 220 AS patients showed that peripheral arthritis, among other factors (younger age, male gender, higher Ankylosing Spondylitis Disease Activity Score (ASDAS), higher BASDAI, and higher levels of erythrocyte sedimentation rate (ESR) and CRP), was considered a baseline predictor of good response to TNF inhibitors . So, a fast relapse upon discontinuation of TNF inhibitors may be considered a general SpA feature. This might be explained by the fact that conventional DMARDs are not frequently used in AS, while several other drugs (especially corticosteroids and methotrexate) are commonly prescribed in RA beyond use of TNF inhibitors.
- •
Drug survival for TNF inhibitors is higher in SpA compared to RA.
- •
This difference is larger with prolonged exposure.
- •
A fast relapse upon discontinuation of TNF inhibitors is commonly observed in SpA.
TNF inhibitors survival in SpA and RA
Before discussing their adverse effects in SpA and RA, it is important to emphasize that the analysis of the published literature about retention rate of TNF inhibitors indicates that it is higher in SpA compared to RA, according to data from important national biologics registries.
The Spanish Register of Biologics (BIOBADASER) analyzed drug survival and adverse effects in 1524 patients with SpA compared to 4006 patients with RA. Drug survival was significantly greater in AS compared with RA, and this difference was even larger with prolonged exposure: 0.88 versus 0.83 at 1 year, 0.82 versus 0.72 at 2 years, and 0.76 versus 0.65 at 3 years. Similar results were observed with the other SpA compared with RA; meanwhile, there was no significant difference between the different SpAs. The hazard ratio (HR) for discontinuation was 0.66 compared with RA, and side effects were significantly less frequent in SpA (17%) compared with RA (26%) .
A retrospective analysis of 770 French patients (57.1% with RA and 37.7% with SpA) using TNF inhibitors in the daily practice showed that the retention rate was significantly longer for patients with SpA compared to RA (HR = 1.6) .
In the Norwegian Disease Modifying Anti-Rheumatic Drug (NOR-DMARD) register, the 1-year retention rates of TNF inhibitors in 1268 treatment courses (847 in RA, 172 in psoriatic arthritis (PsA), and 249 in AS) were 65.7% in RA, 77.3% in PsA, and 77.5% in AS. The adjusted HR for treatment termination was 0.66 for AS versus RA .
In the Czech National Register (ATTRA), the long-term analysis of 310 patients showed that drug survival was also better in the AS patients compared with RA: 84% versus 78% after 1 year, and 72% versus 49% after 3 years of treatment .
The analysis of drug retention in a large Swedish cohort of 9612 patients with RA, 1417 with PsA, and 1652 with other SpA, revealed that drug survival of the first TNF inhibitor in the first year was higher in SpA compared with RA and PsA .
In the Italian Monitor Net database, the analysis of drug survival of 2640 patients with RA and 1220 with SpA who started their first TNF inhibitor showed that patients with SpA, especially AS, had a lower risk of drug discontinuation (adjusted HR = 0.81) in a median follow-up of 17 months. The drug survival was significantly higher in AS patients compared to RA: 0.87 versus 0.82 in 1 year, 0.73 versus 0.71 at 2 years, and 0.69 versus 0.62 at 3 years .
The drug survival of 842 AS patients using TNF inhibitors in the Danish Biologics Register (DANBIO) was 74% in the first year and 63% in the second year. Variables at baseline associated with longer drug survival were male gender, CRP > 14, and low visual analog scale (VAS) for fatigue .
Although anti-TNF therapy can cause rapid and sustained clinical benefit in patients with RA and SpA, its abrupt discontinuation after a long-term use can cause disease relapse in a significant number of patients in 12 months. A classical example in axial SpA is the German study where 42 AS patients, considered good respondents to infliximab, had their medication stopped after 3 years of continuous use; after 12 months, 41 (97.6%) of the 42 patients presented relapse . Submitted to reinfusion of infliximab after the relapse, 40 patients referred good clinical response . Another study considered that relapse is faster in patients with peripheral SpA. A group of 26 non-AS and non-PsA patients participating in a randomized clinical trial with adalimumab in SpA had their medication stopped after 12 or 24 weeks. After discontinuation, relapse was referred by 19 patients (73.1%) in a mean period of 10 weeks . Nevertheless, a Dutch study analyzing baseline predictors of response and discontinuation in 220 AS patients showed that peripheral arthritis, among other factors (younger age, male gender, higher Ankylosing Spondylitis Disease Activity Score (ASDAS), higher BASDAI, and higher levels of erythrocyte sedimentation rate (ESR) and CRP), was considered a baseline predictor of good response to TNF inhibitors . So, a fast relapse upon discontinuation of TNF inhibitors may be considered a general SpA feature. This might be explained by the fact that conventional DMARDs are not frequently used in AS, while several other drugs (especially corticosteroids and methotrexate) are commonly prescribed in RA beyond use of TNF inhibitors.
- •
Drug survival for TNF inhibitors is higher in SpA compared to RA.
- •
This difference is larger with prolonged exposure.
- •
A fast relapse upon discontinuation of TNF inhibitors is commonly observed in SpA.
Adverse effects
Although the use of TNF inhibitors has led to an efficient control of signs and symptoms of AS and RA patients, allied to a significant improvement in quality of life, most of these patients will need a long-term treatment, in many cases for indeterminate period of time, especially in AS. So, the rheumatologists must be aware of the side effects of this therapy. As there is a vast literature about this important topic, we will focus on the analysis of the information from the large national registries published in the last 5 years.
Infection
Infection is by far the most common and most important adverse effect of TNF inhibitors in the treatment of RA and AS. Although drug survival is higher in SpA patients, most of the important publications about serious infection with the use of TNF inhibitors are related to RA and describe bacterial infections.
Current data can support that the increased risk of infection is higher in the first year of treatment. The analysis of prospective observational data of the British Society for Rheumatology Biologics Register (BSRBR) found that the adjusted HR of serious infection was 1.2 during the first 6 months of treatment in a group of 11,798 patients treated with TNF inhibitors compared to 3598 patients using non-biological DMARD. This risk did not differ significantly among the three most common TNF inhibitors .
The Japanese registry for long-term safety for RA (REAL) compared the frequency of serious infection in the first year of observation in a group of 1144 patients, including 646 using TNF inhibitors and 498 patients with non-biological DMARDs. The incidence rate for serious infections was 6.42/100 patients-years (PY) for the TNF inhibitors group and 2.64/100 PY for the non-biological DMARD group. The multivariate analysis showed that TNF inhibitors represented a significant independent risk factor for serious infection (relative risk – RR = 2.37) . The same group found that the risk for serious infection, although elevated in the first year of treatment, was not elevated in the second and third years of treatment (RR = 1.38) .
The 5-year follow up of the Dutch Rheumatoid Arthritis Monitoring (DREAM) registry analyzed the risk of serious infection with different TNF inhibitors in RA: 3.86/100 PY for infliximab, 2.61/100 PY for adalimumab, and 1.66/100 PY for etanercept. These data confirmed that the risk of serious infection was lower for etanercept than that observed in the monoclonal antibodies infliximab (adjusted HR = 0.49) and adalimumab (adjusted HR = 0.55) .
A recent study analyzing risk of serious infection according to age (<65 years and ≥65 years) in 3485 RA patients in USA showed that the incidence rate of serious infection was 5.4/100 PY in the group <65 years and 16.0/100 PY for the group ≥65 years. Compared with etanercept, the adjusted HR during the first 3 months of treatment was higher for infliximab in patients <65 years (HR = 3.01), but not in those ≥65 years (HR = 0.94) .
An USA multi-institution collaboration study analyzing a large series of 10,484 RA and 3213 SpA patients did not find an increased risk of hospitalization for serious infection in a group who used TNF inhibitors compared with non-biological DMARDs. The serious infection hospitalization rates in RA were 8.16/100 PY for the TNF-inhibitors group and 7.78/100 PY for the non-biological group, with an adjusted HR = 1.07. Regarding the SpA patients, the hospitalization rates were 5.41/100 PY for the TNF-inhibitors group and 5.19/100 PY for the non-biological group, with an adjusted HR = 1.10 .
Tuberculosis
An increased incidence of tuberculosis (TB) was observed immediately after the first patients were treated with a TNF inhibitor (infliximab). Since then, this increased risk of TB led to the necessity of a prophylaxis for the patients with signs of previous contact with the disease before TNF inhibitors were started. The follow-up data from registries of biologic drugs in rheumatic diseases are very important to help the management of the long-term use of TNF inhibitors, especially in patients with latent TB. A recent review of the risk of TB after the use of TNF inhibitors for RA, AS, and PsA found 10 cases (0.21%) in 4590 patients with infliximab, nine cases (0.12%) in 7009 patients with adalimumab, and four cases (0.05%) among 7741 patients using etanercept in randomized controlled trials; 83% of the cases of TB occurred in patients with RA. Although the risk of TB was elevated with the three more common TNF inhibitors, this risk was three to four times higher in patients using monoclonal antibodies. More than 80% of the patients presented some deviation from the recommended TB prevention procedures .
The French Research Axed on Tolerance of Biotherapies (RATIO) registry collected 69 cases of TB in a group of autoimmune inflammatory diseases, including 40 patients with RA and 18 with SpA. The TNF inhibitors associated with these cases were infliximab ( n = 36), adalimumab ( n = 28), and etanercept ( n = 5) and, importantly, none of the patients had received correct prophylactic treatment. The IR of TB adjusted for sex and age for the use of TNF inhibitors was 116.7/100,000 PY, significantly higher than the IR observed in the general French population. The standardized IR was significantly higher for the monoclonal antibodies compared to the soluble receptor: infliximab – SIR 18.6; adalimumab – SIR 29.3; etanercept – SIR 1.8 . The British Biologics Register compared TB rates in 10,712 RA patients treated with TNF inhibitors (3925 with infliximab, 3913 with etanercept, and 3504 with adalimumab) and 3232 patients treated with conventional DMARDs. A significantly higher rate of TB was observed with the monoclonal antibodies infliximab (136/100,000 person-years) and adalimumab (144/100,000 person-years) compared with the soluble receptor etanercept (39/100,000 person-years). The adjusted IR was 3.1 for infliximab and 4.2 for adalimumab compared with etanercept. Regarding ethnicity, non-white patients presented a sixfold increase of TB when compared with white patients .
After the clear data showing increased risk of TB with the use of TNF inhibitors, the performance of an interferon-gamma release assay (IGRA) compared to the standard tuberculin skin test (TST) as a screening tool for latent TB was tested previously to the use of the new monoclonal antibody golimumab in patients with RA, AS, and PsA. Among the 2282 patients tested for both methods, 13.8% had at least one positive test for latent TB, including 9.4% with positive results for TST and 7% with positive results for IGRA, with 2.6% presenting positive results in both tests. The repetition of IGRA in patients whose results were initially indeterminate was efficient in more cases than it was previously reported .
The risk of TB with the use of TNF inhibitors can be significantly higher in countries where TB is endemic. A study analyzing the risk of TB in 15 Asian countries (Bangladesh, Cambodia, China, Hong Kong, India, Indonesia, Japan, Korea, Malaysia, Pakistan, Philippines, Singapore, Taiwan, Thailand, and Vietnam) revealed that it is significantly higher than the risk in Western Europe and North America. The number needed to treat (NNT) to reduce one TB event ranged from 8 to 173 for etanercept compared with adalimumab, and from 13 to 283 for etanercept instead of infliximab . In Brazil, the initial report of the Brazilian Registry of Biologics (BIOBADABRASIL), analyzing 1037 patients (69.7% with RA and 21.2% with SpA), with a total exposure to TNF inhibitors of 2101 PY, found three cases of active TB while in use of TNF inhibitors. Although a small number of patients with TB, it was higher than the estimated incidence for the Brazilian population (25−49/100,000) .
As there is a significant difference in the incidence of TB in the different countries and continents, and knowing that this incidence can vary according to the social class of the patients, it is important for the rheumatologists to be aware of the local guidelines for the detection and management of latent TB in their specific countries.
Opportunistic infection
The Safety Assessment of Biologic Therapy (SABER) study, analyzing 33,324 new users of TNF inhibitors in USA for RA, IBD, and AS/PsA/psoriasis, found 80 non-viral opportunistic infections; the main etiologies, in a percentage ≥5%, were: pneumocystosis (20%); nocardiosis/actinomycosis (15%); tuberculosis (12.5%); histoplasmosis (11.3%); non-tuberculosis mycobacteria (11.3%); salmonellosis (10%); listeriosis (5%); and legionellosis (5%). An IR of 2.7/1000 PY for TNF inhibitors compared with 1.7/1000 PY for non-biological DMARDs was obtained, allowing an adjusted HR = 1.6 for TNF inhibitors regarding opportunistic infections .
The analysis of the Spanish BIOBADASER cohort with 3301 patients predominantly using TNF inhibitors for RA (2990 patients) found a higher IR of serious infection in the non-RA group, with an IR ratio of 3.15 before adjustment and 1.96 after adjustment for age, comorbidity, and corticosteroid use, indicating that these variables should be carefully analyzed and taken into account in the study design and interpretation .
The French national registry (RATIO) analyzed all the non-tuberculosis opportunistic infections in a period of 3 years of treatment with TNF inhibitors for RA and SpA. Forty-five opportunistic infections were observed in 43 patients, predominantly with RA (26 patients). The main infections were viral (40% – herpes zoster, varicella, cytomegalovirus), bacterial (33% – listeriosis, nocardiosis, atypical mycobacteriosis, non-thypoid salmonellosis), fungal (22% – pneumocystosis, aspergillosis, cryptococcosis) and parasitic (5% – leishmaniasis); admission to the intensive care unit was necessary in 10 patients (23%), and four patients (9%) died. The risk factors for opportunistic infections were the monoclonal antibodies (OR = 17.6 for infliximab and OR = 10.0 for adalimumab compared with etanercept) and use of oral steroid >10 mg daily during the previous year (OR = 6.3) .
Herpes zoster
The reactivation of varicella zoster virus (herpes zoster (HZ) virus or “shingles”) is commonly described in short- and long-term use of TNF inhibitors. A recent publication analyzing 32,208 patients who used TNF inhibitors in USA identified an IR of HZ of 12.1/1000 PY for RA, 11.3/1000 PY for IBD, and 4.4/1000 PY for AS/PsA/psoriasis. In RA patients, the adjusted IR was similar between TNF inhibitors and non-biologic DMARD, and comparable between the three TNF inhibitors . Most published studies about the risk of HZ infection described RA cohorts. The German biologics register (RABBITT), analyzing 5040 RA patients using TNF inhibitors and conventional DMARDs, revealed that the IR of HZ infection was 11.1/1000 PY for monoclonal antibodies, 8.9/1000 PY for etanercept, and 5.6/1000 PY for the conventional DMARDs . A large cohort of patients with RA (11,881 patients using TNF inhibitors and 3673 using non-biological DMARDs) from the British Biologics Register was analyzed for the occurrence of skin and soft tissue infections. The IR for soft tissue infections was 1.6/100 PY for TNF inhibitors and 0.7/100 PY for non-biological DMARD. Regarding HZ, the IR was 1.6/100 PY for TNF inhibitors and 0.8/100 PY for non-biological DMARD .
These results are significant to support the evaluation of HZ vaccination for patients with RA using TNF inhibitors. In the first reported study, 551 (1.2%) of 44,115 patients with autoimmune diseases in USA received HZ vaccine; 6% were using TNF inhibitors at the time of vaccination. The incidence rates of HZ in vaccinated and non-vaccinated patients were similar, although the very small number of vaccinated patients did not allow definite conclusions . A Spanish study analyzed the risk of hospitalization due to HZ in a population of 4655 patients with rheumatic diseases in BIOBADASER; the estimated IR of hospitalization was 32 cases/100,000 PY, significantly higher than the expected rate for the general population (3.4 cases per 100,000 PY) .
Hepatitis
As the reactivation of hepatitis B in patients who referred previous infection was previously described, it is a medical condition which always worried the rheumatologist. A recent systematic review of the literature described reactivation of hepatitis B in 15 cases (12.3%) of 122 HBsAg-positive patients with rheumatic diseases in nine studies; antiviral prophylaxis was prescribed in 39.3% of the patients. Ten patients had detailed information (four patients with RA, four with AS, and two with PsA) and presented a favorable outcome . An analysis of the follow-up of 21 patients with past history of hepatitis B which used TNF inhibitors for the treatment of rheumatic diseases for a mean of 27 months did not show reactivation of hepatitis B . A Greek study with 14 patients with chronic hepatitis B virus (HBV) infection, 19 with HBV-vaccinated patients, and 19 with resolved HBV infection who had long-term use of TNF inhibitors associated with antiviral therapy did not find viral reactivation nor developed liver decompensation .
Regarding the safety profile of TNF inhibitors in patients with chronic hepatitis C, a review of literature has found 153 patients, with mean treatment duration of 12 months. The individual analysis showed that 91 patients had RA, 22 had psoriasis, six had IBD and 14 other chronic inflammatory diseases, most of them using etanercept. There was only one case of worsening of hepatitis C and five suspected cases . Although these results are not bad, a definite statement about the use of TNF inhibitors in the setting of chronic hepatitis C cannot be made.
Vaccines
A revision about vaccination in rheumatic patients using different biologic drugs has recently been published . Although many centers are used to vaccinate patients with RA and AS, especially against influenza and pneumococcus, before starting TNF inhibitors, there are no specific international guidelines about this subject. As a recent study showed that postvaccination antibody levels can decrease significantly 1.5 year after pneumococcal vaccination in patients with RA and AS, especially in patients using methotrexate and TNF inhibitors, early revaccination can be necessary in patients using immunosuppressive therapy . A recent Brazilian study assessing the immunogenicity of the influenza A H1N1 vaccine in RA and SpA described reduced antibody responses in SpA patients treated with monoclonal antibodies, but not with etanercept .
- •
The risk of serious infection is higher in the TNF inhibitor group compared with the non-biological group.
- •
The risk of serious infection is higher especially in the first year of treatment, in those using monoclonal antibodies and in the group older than 65 years.
- •
The risk for serious infection can decrease significantly when adjusted for age, presence of comorbidities, and corticosteroid use, especially in patients with RA.
- •
Tuberculosis prevention procedures are mandatory with the use of TNF inhibitors, and its risk is higher with monoclonal antibodies.
Cancer
A meta-analysis with an extensive review about malignancies in RA in prospective studies found that melanoma (1.79) and non-melanoma (1.45) skin cancer was the only site with increased risk of malignancies in patients using TNF inhibitors . The increased risk of lymphoma was described initially in the treatment of RA with TNF inhibitors, but it was not completely confirmed. The previous use of other immunosuppressive drugs can also affect these results, as well as the prevalence of cancer in different countries.
The Swedish Biologic Register (ARTIS) and the Swedish Cancer Register analyzed the cancer risk in a national cohort of 67,526 RA patients, including 6366 using TNF inhibitors, and compared with the general population of Sweden. After the occurrence of 240 cancers, the RR of malignancy was 1.00 in the group of patients using TNF inhibitors compared with the biologics-naïve RA patients. The RR of malignancy did not increase with increasing time since the start of TNF inhibitors nor with the cumulative duration of anti-TNF treatment in the first 6 years after the start of the drug . The same Swedish Biologic Register analyzed the risk of lymphoma in a national cohort of 66,743 RA patients, including 6604 using TNF inhibitors, and compared to a general population of 471,024 people. The RR of lymphoma was higher in patients using TNF inhibitors compared to the TNF-naïve RA patients (RR = 1.35) and to the general population (RR = 2.72). Interestingly, in a group attended between 1998 and 2006, the entire increase in the lymphoma risk started therapy between 1998 and 2001, indicating that the selection of patients for treatment can have influence in the observed risk of malignancy . The Danish Biologics Register (DANBIO) described an HR of 1.02 for overall cancer in a group of 3347 patients treated with TNF inhibitors compared with a non-biologics group. RA patients presented an increased risk for cancer overall, lymphatic–hemopoietic and non-melanoma skin cancer compared with the general population .
Recent studies have analyzed the increased IR of different skin cancers in RA. The British Biologic Register compared the follow-up of 11,881 patients using TNF inhibitors and 3629 using non-biological DMARDs regarding keratinocyte skin cancer (basal cell carcinoma and squamous cell carcinoma). The standardized incidence ratio (SIR) was 1.72 for TNF inhibitors and 1.83 for non-biological DMARDs, significantly higher than the English population . Another study from the same British cohort compared the IR of solid cancers in 11,767 patients treated with TNF inhibitors and 3249 patients using non-biological DMARDs with the UK national cancer registries. As many as 427 solid cancers were reported. No significant difference in the risk of solid cancers was observed in both groups (HR = 0.83), after adjusting for differences in baseline characteristics .
The risk of lymphoma was analyzed in a group of 8707 AS patients matched with >40,000 non-AS individuals from the Swedish population; the HR was 0.9 compared with the general population . In the analysis of the long-term safety of 23,458 patients with rheumatic autoimmune diseases (including 1648 AS patients) who used adalimumab, the overall rate for cancer was similar to the general population. The only cancer site with higher than expected incidence (in RA, psoriasis, and Crohn’s disease) was the non-melanoma skin cancer . The analysis of nine clinical trials using etanercept in AS also did not find an increased risk of malignancy .
- •
Melanoma and non-melanoma skin cancer is the only site with increased risk of malignancies in RA patients using TNF inhibitors.
- •
The increased risk of lymphoma observed in the first years of treatment with TNF inhibitors in RA may be biased by patients’ selection criteria.
- •
There is no increased risk of malignancies associated with the use of adalimumab and etanercept in AS.
Neurologic
Since TNF inhibitors were introduced in the treatment of autoimmune rheumatic diseases, the occurrence of demyelinating disorders was reported. Although TNF inhibitors might trigger the demyelinating process, many times the disease evolves independently of the discontinuation of the biologic drug.
A long enquiry among 1800 French rheumatologists and internists, searching the pattern of demyelinating disorders during therapy with TNF inhibitors found 22 cases of central nervous system (CNS) and 11 cases of peripheral nervous system (PNS) involvement; 16 patients had RA, 11 had AS, and four PsA. CNS involvement was characterized by encephalic lesions (16 cases), transverse myelitis (eight cases), and retrobulbar optic neuritis (five cases). All patients discontinued TNF inhibitors and 20 cases initially presented improvement, although five patients developed multiple sclerosis. Chronic (nine cases) or acute (two cases) inflammatory demyelinating polyneuropathy characterized PNS involvement. TNF inhibitors had to be discontinued in 10 patients and intravenous immunoglobulin was prescribed in eight patients .
A recent prospective study performed brain magnetic resonance imaging (MRI) and neurophysiological tests in 77 patients who were candidates for use of TNF inhibitors (36 RA, 24 PsA, and 17 AS). Two patients were contraindicated for the use of TNF inhibitors due to an altered brain MRI, and three other patients developed neurological disorders (brain demyelinating disorder, optic neuritis, and peripheral neuropathy) .
A retrospective analysis of a large cohort of 61,227 patients with inflammatory autoimmune diseases who had used TNF inhibitors or non-biological DMARDs found six cases of optic neuritis, three in each group. So, the rate of optic neuritis was similar in TNF inhibitors (4.5/100,000 PY) and conventional DMARDs (5.4/100,000 PY) .
Heart
It has been considered that RA and SpA present an increased morbidity and mortality associated with accelerated atherosclerosis and cardiovascular disease. From this point of view, it could be supposed that an adequate control of inflammation in RA would also reduce the incidence of cardiovascular disease . In the analysis of 6000 RA patients from the Swedish Rheumatology Register, treatment with TNF inhibitors was not associated with the risk of acute coronary syndromes (HR = 0.80) within the first years of disease . There is also evidence that the use of TNF inhibitors in AS might slow down the subclinical atherosclerosis .
Lung
Since the advent of the TNF inhibitors, the hypothesis that they could increase the risk of interstitial lung disease (ILD) was considered . A recent cohort study analyzing patients with RA, AS, PsA, and IBD from the Kaiser Permanente Northern California found ILD in 23 of 4200 (0.5%) patients who used TNF inhibitors and 15 of 5423 patients (0.3%) who used non-biological DMARDs. The higher adjusted SIR for ILD was obtained in RA patients (0.21/100 PY) and was lower in the other diseases. The adjusted HR for ILD in patients who used TNF inhibitors compared to patients who used non-biological DMARDs was 1.03, confirming the non-association between TNF inhibitors and ILD .
Kidney
Renal failure is not a common finding in patients who use TNF inhibitors. Nevertheless, 26 cases from literature search and three from a single cohort analysis reported the occurrence of biologics-induced autoimmune renal disorders; 22 patients had RA, five AS, and two PsA. Isolated autoimmune renal disorders were diagnosed in 44.8% of the cases, glomerulonephritis associated with vasculitis syndrome was present in 41.3% of the cases, and glomerulonephritis in lupus-like syndrome was reported in 13.9%. The use of etanercept was referred in 51.3% of the affected patients. In these cases, TNF inhibitors must be stopped and patients should be treated according to clinical manifestations and biopsy findings . A direct causal relationship between TNF inhibitors and renal involvement can be questioned, because renal disorders might also be due to a systemic manifestation of an autoimmune disease or a side effect of NSAIDs or DMARDs.
Mortality
A cohort study including several large health-care programs in USA analyzing mortality from all causes in autoimmune diseases found no significant difference comparing a group who used TNF inhibitors with another group using non-biological DMARDs. By the end of the follow-up, 2924 (6.3%) from the 46,424 patients had died, including 1754 (6.1%) of the 28,941 who used TNF inhibitors and 1170 (6.7%) of the 17,483 patients using conventional DMARDs. The adjusted HR for mortality in the TNF inhibitors group compared with the group with non-biologics was 0.93 for RA, 0.81 for SpA, and 1.12 for IBD, not differing between infliximab, etanercept, and adalimumab .
The Swedish Biologic Register compared the risk of death among RA patients using TNF inhibitors ( n = 8562) and a biologic-naïve group ( n = 69,981); after 113 deaths in the TNF inhibitors group and 256 in the biologic-naïve group, the RR of death following treatment with TNF inhibitors was 1.1, considered nonsignificant .
- •
The risk of demyelinating disorders is not increased in patients using TNF inhibitors for RA and AS compared with the group not using biological drugs.
- •
There is no increased risk of ILD associated with the use of TNF inhibitors in RA and AS.
- •
The risk of mortality associated with the use of TNF inhibitors in RA and AS is not increased compared with the non-biological group.
Related posts:
The concept of spondyloarthritis: Where are we now?
Genetics of psoriatic arthritis
Novel imaging modalities in spondyloarthritis
Bone formation in axial spondyloarthritis
Nonpharmacologic therapies in spondyloarthritis
Pharmacologic treatment of psoriatic arthritis and axial spondyloarthritis with traditional biologic and non-biologic DMARDs
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
