Abstract
Objective. The longterm use of tumor necrosis factor (TNF)-α blockers is limited by the formation of neutralizing antibodies. To the best of our knowledge, immunogenicity in psoriatic arthritis (PsA) has not been investigated in depth. Our objective was to evaluate the prevalence and significance of TNF-α blocker immunogenicity in PsA.
Methods. Consecutive patients with PsA treated with either infliximab (IFX), adalimumab (ADA), or etanercept (ETN) > 3 months participated in our cross-sectional study. Their demographic and clinical characteristics, skin and joint disease activity, and records of use of methotrexate (MTX) and other medications were collected. Drug levels (ELISA) and antidrug antibodies (ADAb; Bridging ELISA) were evaluated before the next injection or infusion.
Results. A total of 93 patients with PsA were recruited (48 receiving ADA, 24 IFX, and 21 ETN), with a mean age of 53 years (range 21–83 yrs), composed of 53% women. One-fourth of the patients were concomitantly treated with MTX. Altogether, 77% of the patients demonstrated therapeutic drug levels. High levels of ADAb were found in 29% of patients taking ADA, 21% taking IFX, and 0% taking ETN. ADAb significantly correlated with lower drug levels, higher 28-joint Disease Activity Scores, and higher global assessments. MTX use correlated significantly with a lower prevalence of ADAb.
Conclusion. Significant levels of ADAb were present in up to 29% of patients with PsA treated with ADA or IFX. ADAb clearly correlated with low therapeutic drug levels and higher disease activity variables. The use of MTX significantly decreased ADAb prevalence, and its use should be strongly considered in combination with TNF-α blocker antibodies in patients with PsA.
The management of psoriatic arthritis (PsA) has been revolutionized by the use of tumor necrosis factor (TNF)-α blockers, such as infliximab (IFX), adalimumab (ADA), etanercept (ETN), and others1. Moreover, while in the past the use of TNF-α blockers was restricted in some chronic infections, such as latent tuberculosis (TB) and hepatitis C virus (HCV) positivity, today its use is no longer limited and its longterm administration is safe with proper chemoprophylaxis (latent TB) and close monitoring (HCV), as demonstrated by several reports2,3. Despite the ability of those agents to decrease disease activity, their longterm effect is limited in some patients. Limited drug efficacy in patients with rheumatoid arthritis (RA), ankylosing spondylitis (AS), and inflammatory bowel disease (IBD) has been partially attributed to the development of immunogenicity4,5,6. The creation of antibodies against anti-TNF-α drugs has been associated with low levels of TNF-α blockers7,8, hypersensitivity reactions9,10, and decreased therapeutic response (i.e., loss or absence of effect)11,12,13. The immunogenicity phenomenon has been described against monoclonal antibodies to TNF-α (IFX and ADA)14 and rarely to the soluble receptor (ETN)11,15,16,17,18,19,20,21.
The prevalence of antidrug antibodies (ADAb) in RA is between 20–40%, and it is reportedly reduced significantly by the use of methotrexate (MTX)11,12,19,22,23,24,25,26. Similar prevalence levels, and some even higher (up to 61%) have been described in patients with IBD, especially in Crohn disease (CD), while concomitant MTX and azathioprine (AZA) were shown to attenuate immunogenicity5,27,28,29. Studies on AS reported an ADAb prevalence of 25–64%6,13,18,30,31,32. About one-third of patients with psoriasis who were treated with IFX as monotherapy developed human antichimeric antibodies that were associated with a decreased therapeutic response18,33,34,35. It has been suggested that fixed time intervals between injections might reduce immunogenicity22. In addition, the effect of a simultaneous use of MTX was evaluated on a small number of patients, and the results indicated a reduced immunogenicity and an improvement of the clinical response36.
It appears that the immunogenicity of TNF-α blockers in PsA has not been thoroughly investigated. To the best of our knowledge, the only study that directly addressed this subject included 22 patients with PsA who were treated with ADA and had an ADAb prevalence of 18% that was associated with a decreased clinical response and low drug level37. Kavanaugh, et al’s earlier study on 200 patients with PsA (the Infliximab Multinational Psoriatic Arthritis Controlled Trial 2) was a phase III randomized controlled trial that evaluated safety and efficacy in patients with PsA receiving IFX, although it did not specifically evaluate immunogenicity38. Their results demonstrated an ADAb prevalence of 15% (26% in patients not taking MTX and 3.6% in patients taking concomitant MTX, not analyzed statistically).
MTX is the main immunomodulatory drug that moderates immunogenicity18,19,36,39,40,41. AZA attenuates immunogenicity in IBD similarly to MTX18,42,43,44, but this has not been reported in rheumatologic conditions such as RA, spondyloarthropathies (SpA), and psoriasis18. Moreover, there is insufficient evidence to support the role of other disease-modifying antirheumatic drugs (sulfasalazine, leflunomide, hydroxychloroquine, and cyclosporine A) or prednisolone in moderating ADAb formation18,39. The goals of our present study were to evaluate the prevalence of TNF-α blocker immunogenicity in PsA and its correlation to drug levels, disease activity, and the effect of concomitant use of MTX.
MATERIALS AND METHODS
Study design
We conducted a cross-sectional study (transversal analyses) on 93 patients with PsA treated with IFX, ADA, or ETN. The primary endpoint was the prevalence of ADAb to TNF-α blockers in PsA. The secondary endpoints were (1) the correlation between the presence of immunogenicity and drug levels, (2) demographic and clinical features (including disease activity scores), and (3) the effect of concomitant use of MTX. Patients’ enrollment was at the trough level of the anti-TNF-α drug, i.e., just before the next administration of the anti-TNF drug. Clinical assessment was performed and blood was drawn at this timepoint.
Study participants
The study was composed of consecutive patients with PsA according to the Classification for Psoriatic Arthritis criteria. They were aged above 18 years, and were undergoing intravenous IFX 5 mg/kg treatment according to the standard protocol, subcutaneous ADA 40 mg every other week, or SC ETN 50 mg/week. Patients who had been treated with IFX, ADA, or ETN for less than 3 months or who were receiving other TNF-α antagonists were excluded from the study.
Demographic data
Clinical data included the pattern of the PsA arthropathy, duration of psoriasis and PsA, comorbid conditions, and current drugs being taken (including MTX) and their dosage. Assessment of disease activity included the 28-joint Disease Activity Score (DAS28), pain evaluation using a visual analog scale (VAS) of 100 mm, the Psoriasis Area and Severity Index (PASI), and the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI). Blood was drawn for analyses of the erythrocyte sedimentation rate, anti-TNF-α drug level, and ADAb.
Assessment of drug levels and ADAb
The drug level was measured by ELISA (Promonitor commercial kit of the Progenika Biopharma group) and expressed in microgram per milliliter (µg/ml). The IFX detection threshold was 0.035 µg/ml: levels < 0.035 µg/ml were considered non-detectable, levels 0.035–1.5 µg/ml were considered subtherapeutic, and levels > 1.5 µg/ml were considered therapeutic. Those ranges were derived from clinical studies mainly in RA (and fewer in CD)44,45, including a study done by means of the Promonitor kit on 78 patients with RA for whom there were 612 blood samples. The detection threshold of ADA was 0.024 µg/ml: levels < 0.024 µg/ml were considered undetectable, levels 0.024–0.8 µg/ml were considered subtherapeutic, and levels > 0.8 µg/ml were considered therapeutic. Those ranges are derived from clinical studies45, including 1 done by means of the Promonitor kit on 51 patients with RA for whom there were 171 blood samples. The ETN detection threshold was 0.035 µg/ml. There are no therapeutic ranges because of the lack of data: levels < 0.035 µg/ml were considered undetectable and levels > 0.035 µg/ml were considered positive responses.
ADAb and their concentrations were tested by Bridging ELISA (Promonitor). Concentrations were expressed in absorbance units per milliliter (Au/ml). The anti-IFX antibodies detection threshold was 2 Au/ml. We defined a high concentration as the value of the threshold multiplied by 14, i.e., 28 Au/ml. The anti-ADA antibodies detection threshold was 3.5 Au/ml, and a high concentration was defined as a value of the threshold multiplied by 14, i.e., 50 Au/ml. The anti-ETN antibodies detection threshold was 142 Au/ml, and a high concentration was defined as a value of the threshold multiplied by 14, i.e., 2000 Au/ml.
It is noteworthy that high concentration of ADAb (“high” ADAb) is a term mainly used for anti-ADA antibodies. The customary value of a high concentration of antidrug antibodies is above 100 Au/ml for the radioimmunoassay (RIA) test, which is a multiplier of 8.5 from the threshold of positive test, which is 12 Au/ml19,32. To date, there is no data known to us that defines high ADAb in the Bridging ELISA test. According to the statistical distribution of ADAb in our study participants, and since the threshold of Bridging ELISA test for positive anti-ADA antibodies is 3.5 Au/ml, we defined high concentration as 50 Au/ml, using a multiplier of 14. We used rougher values with a multiplication by 14 (and not by 8.5) because Bridging ELISA is less accurate than RIA46,47. Indeed, these values were arbitrarily determined, but rationally and according to our participants’ ADAb levels distribution. A recent paper by Chen, et al has shown good concordance between RIA and Bridging ELISA, although again RIA was more accurate48.
Statistical analysis
Statistical processing was done by Excel (Microsoft Corp.) and SPSS (version 14, IBM Corp.) software. We used the Pearson chi-square test, Fisher’s exact test, and Spearman rank correlation coefficient test to obtain correlations between variables. The Student t test and ANOVA were used to compare the means of variables with normal distributions. Medians of variables without normal distribution were tested using the Mann-Whitney U test and Kruskal-Wallis test. A p value < 0.05 was considered statistically significant.
RESULTS
Demographic and clinical characteristics of the participants
Between March 2011 and April 2013, 93 patients with PsA were recruited into our study. Their mean age was 53 years (range 21–83 yrs), and 53% of the subjects were women. The mean durations of the skin disease and the arthropathy were 21 years and 15 years, respectively. The most common pattern of disease was peripheral polyarthritis (48.5%), followed by combined axial and peripheral (23.5%), oligoarthritis (21.5%), and pure axial (7.5%). Forty-eight patients were taking ADA (51.6%), 24 were taking IFX (25.8%), and 21 were taking ETN (22.6%). About one-fourth of the patients (25.8%) were concomitantly taking MTX at an average weekly dosage of 13.3 mg (Table 1).
Clinical and demographic data.
Table 2 summarizes the variables of disease activity. The mean DAS28 was 3.4 (moderate disease activity), the mean VAS was 27, the mean BASDAI was 3.95 (moderate to severe), and the mean PASI was 3.5 (mild). Patients receiving ETN presented a significantly milder degree of PsA, while IFX-treated patients had a more severe disease form according to DAS28. Patients treated with ADA had higher levels on the BASDAI. The severity of skin disease (by PASI) was similar for all medication groups.
Variables of disease activity.
ADAb prevalence and drug levels
The prevalence of immunogenicity in the whole group was 33.3% (Table 3). One-fifth of the patients had high concentrations of antibodies. None of the members of the ETN group had ADAb. Analysis of the levels of antibodies in the ADA group yielded 2 groups of patients: 54% of the entire group developed ADAb, but only 29% had high ADAb concentrations. ADAb to IFX was found in 21% of the IFX-treated patients, and all of them had high concentrations of antibodies. All of the patients taking ETN demonstrated therapeutic drug levels compared with 79% of the ADA-treated patients and 54% of the IFX-treated patients.
Drug levels and antibodies prevalence. Values are %.
Correlations between immunogenicity and drug levels
The presence of ADAb and high ADAb concentrations significantly correlated with nontherapeutic drug levels: 45% of ADAb-positive patients and 63% of high ADAb patients had nontherapeutic drug levels comparing to only 11% of patients without ADAb. Patients with ADAb (high and low titers) concentrations had a significantly higher DAS28 (3.9 and 4.1, respectively, compared with a score of only 3.2 in patients without ADAb) and a higher VAS (36 and 42, respectively, compared with a score of only 23). There was a trend toward a higher PASI, younger age, and male sex in the ADAb-positive patients; those values did not reach a level of significance (Table 4 and Table 5). The DAS28, VAS, and PASI scores were lower in patients with therapeutic drug levels compared to patients with nontherapeutic drug levels (3.3 vs 3.8, 25 vs 35, and 2.7 vs 6, respectively), but those values did not reach a level of statistical significance.
Correlation between demographic and clinical data and the presence of antibodies.
Correlation between demographic and clinical data to antibodies concentration levels.
The effect of MTX on immunogenicity
One-fourth of the study subjects were treated concomitantly with MTX at a mean dosage of 13.3 mg/week. The MTX-treated patients were significantly older (59 yrs vs 51.5 yrs, p = 0.024). The MTX-treated patients had significantly fewer ADAb (i.e., 16.7% vs 39.1%, p = 0.049; Table 6). Moreover, fewer MTX-treated patients had high ADAb concentration (16.7% compared to 21.7% for the non-MTX patients). There was a nonsignificant trend toward a higher proportion of patients with therapeutic drug levels among the MTX-treated patients (92% vs 72.5% for the non-MTX patients). Interestingly, the MTX-treated patients had a higher DAS28 than the non-MTX patients (4 vs 3.2, p = 0.02).
Correlations between concomitant MTX use and immunogenicity, and drug level and demographic and clinical data.
DISCUSSION
TNF blockers have had an enormous effect on the treatment, course of disease, and prognosis of SpA, including PsA. Indeed, these biologic drugs represent a breakthrough in the management of these conditions. However, the use of these agents has been limited by the phenomenon of immunogenicity, which might induce low drug levels and decreased clinical response in RA, AS, and IBD.
The results of our cross-sectional study have demonstrated that the phenomenon of immunogenicity is limited to IFX and ADA in patients with PsA. The results showed that it reaches a level of up to 50% in ADA-treated patients (29% with high levels), 21% in IFX-treated patients, and none in ETN-treated patients. A clear correlation was found between the presence of immunogenicity, drug levels, and clinical response. The use of MTX significantly decreased the proportion of ADAb.
Our results are similar to the prevalence of immunogenicity in RA11,22,23. As others had found in RA11,12, we observed a correlation between ADAb and low drug levels and decreased clinical response. Interestingly, the effect of immunogenicity in our study was more prominent in active peripheral arthritis, as reflected by DAS28, than in axial arthropathy, as reflected by BASDAI.
Another important finding in our study was the clearly reduced immunogenicity in patients who were also receiving MTX, as had been shown in RA22,23. While the concomitant use of MTX with monoclonal antibodies is indicated in RA, there are no similarly clear recommendations in PsA. Our results suggest that the use of MTX should be strongly considered in addition to monoclonal antibodies.
Noteworthy, the patients taking ADA demonstrated an immunogenicity of 54%, but only half of them (29%) showed high concentrations of antibodies. This finding suggests that a variety of levels might be found that might affect the significance of immunogenicity differently among patients producing ADAb.
Most of the known literature about immunogenicity of SpA is in the setting of AS, where its prevalence is about 25%, with correlations between monoclonal ADAb and low drug levels, decreased clinical response, and a high rate of injection-induced hypersensitivity reactions6,13,31. A recent trial by Kneepkens, et al demonstrated an ADAb prevalence of 27% with a correlation to decreased clinical response in patients treated with ADA32. Moreover, those authors found, as we did, a correlation between high ADAb concentrations and decreased clinical response in ADA-treated patients. Studies of immunogenicity in patients with psoriasis demonstrated that ADAb is correlated to IFX and ADA, but not to ETN. Its prevalence was up to one-third, and there was a correlation to decreased drug levels and decreased therapeutic response33,34,35,42,49. The concomitant use of MTX reduced ADAb prevalence in SpA and psoriasis, but unlike RA, it is not clear whether it improved anti-TNF efficacy18. On the other hand, attenuating anti-TNF immunogenicity may be even more crucial because there are fewer alternative classes of biologics to use in the event of treatment failure.
Our study has several limitations. It is cross-sectional and based on a relatively small number of patients. The evaluation of ADAb was performed using Bridging ELISA, which is reliable, but considered to be less accurate than RIA46,47,48. Likewise, the evaluation of disease activity was based on DAS28, which might not be ideal in PsA.
Similar to RA, the phenomenon of immunogenicity is present in PsA as well, with a prevalence of up to 50%. Immunogenicity clearly affected the therapeutic effect of the monoclonal antibodies and seemed to be attenuated by the concomitant use of MTX. Large-scale studies are warranted to confirm our observations.
- Accepted for publication September 8, 2014.
REFERENCES
In this issue
