Dr. Heiberg has received speaking fees and honoraria (less than $10,000) from Abbott and consultancy fees (less than $10,000) from Wyeth. Dr. Koldingsnes has received speaking fees (less than $10,000 each) from Schering-Plough, Wyeth, and Abbott. Dr. Mikkelsen has received honoraria (less than $10,000 each) from Roche and Bristol-Myers Squibb. Dr. Kvien has received consultancies and speaking fees (less than $10,000 each) from Abbott, Schering-Plough, and Wyeth.
The comparative one-year performance of anti–tumor necrosis factor α drugs in patients with rheumatoid arthritis, psoriatic arthritis, and ankylosing spondylitis: Results from a longitudinal, observational, multicenter study
Version of Record online: 31 JAN 2008
Copyright © 2008 by the American College of Rheumatology
Arthritis Care & Research
Volume 59, Issue 2, pages 234–240, 15 February 2008
How to Cite
Heiberg, M. S., Koldingsnes, W., Mikkelsen, K., Rødevand, E., Kaufmann, C., Mowinckel, P. and Kvien, T. K. (2008), The comparative one-year performance of anti–tumor necrosis factor α drugs in patients with rheumatoid arthritis, psoriatic arthritis, and ankylosing spondylitis: Results from a longitudinal, observational, multicenter study. Arthritis & Rheumatism, 59: 234–240. doi: 10.1002/art.23333
- Issue online: 31 JAN 2008
- Version of Record online: 31 JAN 2008
- Manuscript Accepted: 3 JUL 2007
- Manuscript Received: 27 MAR 2007
- The Norwegian Women's Public Health Association and Eastern Norway Regional Health Authority
- Norwegian Directorate for Health and Social Affairs
To compare the 1-year retention rates of anti–tumor necrosis factor α (anti-TNFα) medications in patients with rheumatoid arthritis (RA), psoriatic arthritis (PsA), and ankylosing spondylitis (AS) with complementary analyses of the effect on health status.
Our analyses comprised 847, 172, and 249 anti-TNFα treatment courses in patients with RA, PsA, and AS, respectively. Crude drug survival was compared and hazard ratios (HRs) for treatment termination were calculated with adjustments for age, sex, investigator's global assessment, and concomitant methotrexate (MTX). Adjusted changes in health-related quality of life (HRQOL) were compared among the groups.
Unadjusted 1-year retention rates were 65.4%, 77.3%, and 77.5% in the RA, PsA, and AS groups, respectively. The adjusted HRs for treatment termination were 0.76 (95% confidence interval [95% CI] 0.53–1.07) for PsA versus RA and 0.66 (95% CI 0.47–0.92) for AS versus RA. High baseline disease activity and female sex were significantly associated with premature treatment termination, whereas concomitant MTX was associated with better drug survival. However, the impact of MTX was apparent for RA and PsA, but not for AS in stratified analyses. The improvements in HRQOL were superior in patients with PsA and AS compared with RA.
Our results suggest that survival of anti-TNFα treatment is superior in AS and PsA patients compared with RA patients. Larger improvements in HRQOL in patients with spondylarthritides may contribute to the differences in drug survival. Concomitant MTX was associated with better retention rates in RA and PsA patients, but not AS patients.
The introduction of biologic therapies has changed the treatment strategies for inflammatory arthritides, in particular ankylosing spondylitis (AS), for which therapeutic options were previously scarce (1). Three different anti–tumor necrosis factor α (anti-TNFα) drugs (infliximab, etanercept, and adalimumab) are approved for patients with rheumatoid arthritis (RA), AS, and psoriatic arthritis (PsA). Their efficacy has been demonstrated in several randomized clinical trials (RCTs) for each diagnostic group separately (2–10), but less is known about the comparative effectiveness across the different diagnoses. Drug survival is considered a surrogate marker for effect (11). However, other factors, such as available treatment options, concomitant medication, patient characteristics, and tight followup, may influence the retention rates. The comparative drug survival between diagnostic groups has been explored previously in a few observational studies (12, 13), but ours is the first study to provide data on drug survival with concurrent analyses of effects on health status. We analyzed data from a Norwegian register of disease-modifying antirheumatic drug (DMARD) prescriptions (including biologic therapies) for patients with inflammatory arthropathies. Our goal was to compare the 1-year retention rates in patients with RA, PsA, and AS who received anti-TNFα treatment. As an additional measure of effectiveness, we explored the comparative change in health-related quality of life (HRQOL) between the diagnostic groups.
PATIENTS AND METHODS
The Norwegian DMARD (NOR-DMARD) register was established in December 2000 (14). Five Norwegian rheumatology departments consecutively included all patients with inflammatory arthritides who started a DMARD regimen. Patients were registered as a new case when they switched to another DMARD regimen, which also included adding, for example, an anti-TNFα drug to methotrexate (MTX) monotherapy. The study design was a phase IV, multicenter, longitudinal, observational study. We managed to enroll ∼85% of the patients who started with DMARD therapy into the register. The remaining 15% were either missed for inclusion, refused enrollment, were excluded due to language barriers or inclusion in RCTs, etc. By August 2006, a total of 5,811 cases were enrolled in the NOR-DMARD register.
Patients were eligible for the present analyses if they had been diagnosed with RA, PsA, or AS by the treating rheumatologist (i.e., they were given the diagnoses M05.8, M05.9, M06.0, L40.5 plus M07.0, M07.2, M07.3, or M45 according to the World Health Organization International Statistical Classification of Diseases and Related Health Problems, Tenth Revision) and received an anti-TNFα drug, with or without concomitant MTX. A total of 1,019 patients (645 RA, 150 PsA, 224 AS) received a total of 1,268 anti-TNFα treatment courses (847 RA, 172 PsA, 249 AS) that were included in the analyses. Two or more treatment courses were registered in 153 (24%) of the RA patients, 18 (12%) of the PsA patients, and 22 (10%) of the AS patients. The most common dosing regimen for infliximab was 3–5 mg/kg at baseline, after 2 and 6 weeks, and then every 8 weeks for patients with RA and PsA and 5 mg/kg at baseline, after 2 and 6 weeks, and then every 6 weeks for patients with AS. The dosing regimen for etanercept was 25 mg twice weekly and for adalimumab, 40 mg every 2 weeks. The patients gave written informed consent before participation. The study was evaluated by the regional ethical committee, and the storage of data was approved by the Data Inspectorate.
Demographic variables are recorded for all patients in the NOR-DMARD register at baseline. Furthermore, patients are assessed at baseline and after 3, 6, and 12 months and then yearly with measures of disease activity and health status that are applicable not only in RA but also in other inflammatory arthritides. All treatment terminations and reasons for withdrawal are recorded. Reasons for withdrawal are classified as lack of efficacy, adverse event (AE), both lack of efficacy and AE, patient's preference, remission, or other reason/unknown. In the present analyses, the combination of lack of efficacy and AE was classified as AE.
The Medical Outcomes Study 36-item Short Form Health Survey (SF-36) (15) is a commonly used health status measure. It contains 36 questions measuring health across 8 different dimensions: physical functioning, role limitations due to physical health problems, bodily pain, vitality, social functioning, role limitations due to emotional problems, mental health, and general health. A score is computed within each dimension with a value from 0 (worst possible health state) to 100 (best possible health state). The SF-36 is a generic measure, which therefore can be used to assess health status across different diagnoses. The investigator's global assessment of disease activity, measured on a 100-mm visual analog scale (0 = not active, 100 = extremely active), was used as a generic measure of disease activity.
The crude drug retention rates were explored with Kaplan-Meier analyses using the log rank test to compare the survival curves. The hazard ratios (HRs) for treatment withdrawal were assessed with adjustments for age, sex, investigator's global assessment of disease activity, and concomitant MTX use in Cox regression analyses. The model was tested for proportionality. We also explored the impact of disease duration, number of previous DMARDs, and type of anti-TNFα drugs by using these variables as covariates in the Cox regression analyses. The 6- and 12-month changes in SF-36 scores were assessed using analyses of covariance (16) with adjustments for age, sex, concomitant MTX use, and the baseline value of the dependent variable. Changes in SF-36 scores were examined with a last observation carried forward (LOCF) approach when values were missing, with at least 1 followup examination required. Some patients had not been included in the register for a sufficient time to have a followup examination, whereas others had terminated treatment prior to the 3-month assessment. Therefore, data for the LOCF analyses were not available in 13%, 12%, and 17% of the RA, PsA, and AS cases, respectively. Robustness analyses were performed including only the patients who had never previously taken anti-TNFα agents. A significance level of 5% was used in all analyses. No correction for multiple comparisons was performed. Statistical analyses were performed with SPSS software, version 14.0 (SPSS, Chicago, IL).
A total of 1,268 treatment courses were eligible for the analyses (847 anti-TNFα regimens in RA patients, 172 in PsA patients, and 249 in AS patients). A larger proportion of RA patients were women, and RA patients were older than patients in the PsA and AS groups. More RA patients had previously been exposed to anti-TNFα drugs. Among the RA patients, 77% were positive for rheumatoid factor (RF) and 73% had erosive disease. Of the PsA and AS patients, 87% and 25%, respectively, had peripheral arthritis. Group characteristics and baseline values are shown in Table 1.
|Characteristic||RA (n = 847)||PsA (n = 172)||AS (n = 249)|
|Age, years||52.8 ± 13.5||45.7 ± 10.8||43.4 ± 10.5|
|Female sex, %||74.3||36.6||26.1|
|Disease duration, years||11.4 ± 9.3||12.1 ± 9.3||14.0 ± 10.2|
|No. previous DMARDs||3.9 ± 2.4||2.6 ± 1.6||1.1 ± 1.2|
|VAS investigator's global||48.3 ± 19.9||43.9 ± 19.3||39.5 ± 37.0|
|Concomitant MTX, %||66.7||68.0||30.5|
|Anti-TNFα drug, no. (%)|
|Infliximab||214 (25.2)||48 (27.9)||113 (45.4)|
|Etanercept||345 (40.7)||96 (55.8)||122 (49.0)|
|Adalimumab||288 (34.1)||28 (16.3)||14 (5.6)|
|Anti-TNFα naive, %||64||73||78|
The crude 1-year survival rates of anti-TNFα drugs were 65.4%, 77.3%, and 77.5% among patients with RA, PsA, and AS, respectively (P = 0.003 for RA versus PsA and P = 0.001 for RA versus AS) (Figure 1). The withdrawal rates for the different anti-TNFα drugs are shown in Table 2. After adjusting for age, sex, investigator's global assessment, and concomitant MTX, the respective HRs (95% confidence interval [95% CI]) for treatment termination in PsA and AS patients versus RA patients were 0.76 (0.53–1.07) and 0.66 (0.47–0.92). Female sex and higher baseline disease activity, as measured by the investigator's global assessment, were associated with a higher risk of treatment termination, whereas concomitant MTX was associated with a lower risk of treatment termination (Table 3). The association between baseline disease activity and drug survival was consistent across the diagnoses and the different anti-TNFα drugs.
|HR (95% CI)||P|
|VAS investigator's global†||1.06 (1.001–1.13)||0.01|
|Concomitant MTX||0.53 (0.43–0.65)||<0.001|
We also explored the impact of disease duration and number of previous DMARDs by introducing these variables as covariates in the Cox regression analysis. They were not statistically significant covariates and they did not influence the model. There were also no interactions between disease duration and diagnosis or type of anti-TNFα drug. There was a statistically significant difference between the 3 anti-TNFα drugs in favor of etanercept versus infliximab and adalimumab regarding overall drug survival. However, etanercept was prescribed more often as the first anti-TNFα drug than the other 2. We also performed a subanalysis including only patients who had never previously taken anti-TNFα drugs. There was no longer a difference between the 3 drugs, otherwise the results remained similar. In a subanalysis including only the RA patients, we found no difference in drug survival between the RF-positive and RF-negative patients.
The combination of an anti-TNFα drug and MTX was highly associated with a lower risk of treatment termination compared with anti-TNFα monotherapy, with an HR (95% CI) of 0.53 (0.43–0.65). We also compared the survival of monotherapy and combination therapy using Kaplan-Meier plots, stratified for diagnoses (Figure 2). A statistically significant difference in favor of combination therapy was observed for both RA patients (P < 0.001) and PsA patients (P = 0.02), whereas no difference was observed between monotherapy and combination therapy for AS patients (P = 0.29).
The reasons for discontinuing treatment among patients with RA, PsA, and AS, respectively, were distributed as follows: lack of efficacy in 39.0%, 17.9%, and 36.4%; AEs in 49.3%, 69.2%, and 43.6%; patient's request in 4.8%, 5.1%, and 1.8%; and other reasons/unknown in 6.9%, 7.7%, and 18.2%. No combination treatment courses were discontinued due to lack of efficacy among patients with PsA.
The SF-36 scores at baseline were similar across the groups, except that the mean physical function score and role physical score were higher in the AS group compared with the 2 other groups (results not shown). The adjusted changes were numerically superior in PsA and AS patients versus RA patients across all 8 SF-36 dimensions at both 6 and 12 months (Figure 3), although the changes were not statistically significant for the mental and social dimensions.
The anti-TNFα drugs represent a new biologic treatment era for patients with rheumatic conditions, and the efficacy has been demonstrated through several RCTs (2–10). However, it is also important to document the effectiveness of these costly drugs, i.e., how well they perform under real-life conditions, as strict inclusion criteria and short duration often limit the external validity of results from RCTs (17, 18). A longitudinal, observational study is the preferred design for studying effectiveness (19).
In the present observational study, we assessed the survival of treatment with anti-TNFα drugs in a large number of patients with RA, PsA, and AS, and complementary analyses of the effect on health status were performed. The crude Kaplan-Meier analyses showed that a larger proportion of anti-TNFα treatment courses were terminated prematurely in RA patients compared with PsA patients and AS patients during the first year. This finding corresponds with the results from other observational studies. Carmona and Gomez-Reino analyzed data from the Spanish registry Base de Datos de Productos Biológicos de la Sociedad Española de Reumatología (BIOBADASER) and found an HR for treatment discontinuation of 0.66 for spondylarthritides versus RA after adjustment for age, sex, and use of infliximab (12). In a retrospective study, Duclos and colleagues found an HR of 1.60 for continuing anti-TNFα treatment for spondylarthritis versus RA (13).
Although drug survival is regarded as a surrogate marker for efficacy, other factors may influence the retention rates, e.g., the access to alternative treatments. Anti-TNFα treatment is the only effective treatment for AS patients with axial disease for whom nonsteroidal antiinflammatory drugs are insufficient (1), whereas several other treatment alternatives are available for patients with PsA and RA. Thus, the comparative retention rates between these diagnostic groups must be interpreted with caution.
The present study is the first to combine analyses of drug survival with assessments of comparative changes in HRQOL between patients with RA, PsA, and AS. As shown in Figure 3, the changes in SF-36 scores were numerically superior in PsA and AS patients compared with RA patients across all SF-36 dimensions, although these changes were statistically significant primarily for the physical dimensions. A ceiling effect for the mental dimensions may have reduced the chance of detecting group differences, as the baseline values for social functioning, mental health, and role mental were high (score range 50–70), and therefore limited the potential for improvement. We have also previously demonstrated that HRQOL improved more in patients with AS receiving anti-TNFα drugs than in those with RA receiving these drugs (20), but the increased health benefit in patients with PsA versus those with RA is a novel finding and supports the idea that superior effectiveness contributes to the longer retention rates in spondylarthritis patients. There was also a tendency of larger improvements in AS patients compared with PsA patients, although the difference was statistically significant only for the SF-36 physical function score at the 6-month assessment.
The overall impression from our results is that a gradient is observed for both retention rates and health benefit, with a relatively larger benefit in patients with pure axial manifestations versus peripheral arthritides. As in our study, Carmona and Gomez-Reino also found that the difference in drug survival between the PsA subgroup and RA was not statistically significant after adjustment for other important factors, whereas there was still a statistically significant difference between RA and AS (12).
In the Cox regression analyses, we found that female sex and high baseline disease activity were associated with a higher risk of treatment termination, whereas concomitant MTX was strongly associated with better drug survival. This finding contradicts the finding from the French retrospective analysis that showed a significantly better survival of anti-TNFα treatment without concomitant DMARDs (13), whereas the impact of concomitant MTX use was not addressed in the Spanish study (12). RCTs and observational studies have demonstrated that anti-TNFα drugs, in combination with MTX, have a superior effect on disease activity compared with anti-TNFα monotherapy in patients with RA (6, 21–24), whereas the role of concomitant MTX is less clear for the spondylarthritides. It is known that anti-TNFα therapy, especially infliximab, may induce antibody formation against the drug in patients with RA, and that concomitant MTX can prevent this antibody formation (25). Less is known about antibody formation against the anti-TNFα drugs in patients with PsA and AS, although a Dutch study recently demonstrated that clinical response to infliximab in patients with AS is also correlated to antiinfliximab antibody formation (26). Whether this can be prevented by concomitant immunosuppressive therapy is not known. Unfortunately, antibodies in serum are not measured in the NOR-DMARD register and frozen sera are not available. In the stratified Kaplan-Meier analysis, we found a statistically significant difference in retention rates between combination therapy and monotherapy for both RA and PsA patients, whereas the retention rates were similar in AS patients. Interestingly, no combination treatment courses were prematurely terminated due to lack of efficacy in patients with PsA, whereas lack of efficacy was the reason for ∼40% of the monotherapy discontinuations. Kristensen et al also found that lack of concomitant MTX was a predictor of premature treatment termination in patients with PsA (27). Previous RCTs of anti-TNFα drugs in patients with PsA have failed to demonstrate superior clinical improvement in patients receiving concomitant MTX versus monotherapy (2, 8, 9).
Duclos et al found significantly better survival of the first anti-TNFα treatment course compared with the second or third treatment course (13), whereas in a large UK cohort study, Hyrich et al recently demonstrated that continuation rates were high in patients with RA who switched to a second anti-TNFα drug (28). We found a tendency toward longer retention rates for the first course of anti-TNFα treatment, but the association was not statistically significant. Nevertheless, we performed robustness analyses, including only the patients who had never previously received anti-TNFα drugs, and our results remained similar. Like Duclos et al, we found no significant difference between the 3 anti-TNFα drugs, whereas other groups have found significant differences between the 3 drugs in observational studies (12, 22). However, caution should be applied in the interpretation of treatment comparisons in a nonrandomized context due to the possibility of confounding by indication.
In summary, we found that the 1-year survival of anti-TNFα drugs was superior in patients with AS and PsA compared with those with RA, and the gradient for the comparative improvement in HRQOL was similar to the gradient regarding retention rates across the diagnostic groups. Concomitant MTX was associated with better retention rates in patients with RA and PsA, but not in patients with AS. These results, taken together with findings from other observational studies, support that the effectiveness of anti-TNFα treatment is graded across the diseases, with the relatively lowest benefit in patients with RA.
Dr. Heiberg had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study design. Mikkelsen, Kvien.
Acquisition of data. Heiberg, Koldingsnes, Mikkelsen, Rødevand, Kaufmann, Kvien.
Analysis and interpretation of data. Heiberg, Koldingsnes, Mowinckel, Kvien.
Manuscript preparation. Heiberg, Koldingsnes, Rødevand, Mowinckel, Kvien.
Statistical analysis. Heiberg, Mowinckel.
- 3Long-term efficacy and safety of infliximab in the treatment of ankylosing spondylitis: an open, observational, extension study of a three-month, randomized, placebo-controlled trial. Arthritis Rheum 2003; 48: 2224–33., , , , , , et al.
- 5Adalimumab, a fully human anti tumor necrosis factor-alpha monoclonal antibody, and concomitant standard antirheumatic therapy for the treatment of rheumatoid arthritis: results of STAR (Safety Trial of Adalimumab in Rheumatoid Arthritis). J Rheumatol 2003; 30: 2563–71., , , , , , et al.
- 16Applied regression analysis and other multivariable methods. Boston: PWS-Kent Publishing Company; 1988..
- 20The comparative effectiveness of tumor necrosis factor–blocking agents in patients with rheumatoid arthritis and patients with ankylosing spondylitis: a six-month, longitudinal, observational, multicenter study. Arthritis Rheum 2005; 52: 2506–12., , , , , , et al.
- 21The PREMIER study: a multicenter, randomized, double-blind clinical trial of combination therapy with adalimumab plus methotrexate versus methotrexate alone or adalimumab alone in patients with early, aggressive rheumatoid arthritis who had not had previous methotrexate treatment. Arthritis Rheum 2005; 54: 26–37., , , , , , et al.
- 27Predictors for terminating therapy of etanercept and infliximab in psoriatic arthritis patients [abstract]. Arthritis Rheum 2006; 54: S721–2., , , , , .
- 28British Society for Rheumatology Biologics Register. Outcomes after switching from one anti-tumor necrosis factor α agent to a second anti-tumor necrosis factor α agent in patients with rheumatoid arthritis: results from a large UK national cohort study. Arthritis Rheum 2007; 56: 13–20., , , , , for the