Baseline patient characteristics
The data analysis set consisted of 3,404 patients. Baseline demographic and disease characteristics are presented in Table 1. Patients had a mean age of 54.6 years, were mostly (78%) women, exhibited moderate to high disease activity (mean DAS28 of 5.9, range 3.4–8.3), and had a mean disease duration of 11.7 years. More than 90% of patients received a diagnosis of RA more than 2 years before study enrollment, and almost half of the patients received the diagnosis more than 10 years ago.
Table 1 Baseline characteristics of patients at the time of initiation of adalimumab therapy
Almost all (95.7%) patients had received prior therapy with at least one conventional DMARD, mostly MTX (83.0%). Approximately, 30% of the patients had been previously treated with at least one biologic agent and 6.8% had received prior treatment with two or more biologic agents. The most frequently used biologic agents were etanercept (20.7%) and infliximab (13.4%). Mean treatment duration for prior biologic agents ranged from a mean of 12.6 months for anakinra to 15.3 months for infliximab. The most frequent reasons for discontinuation of biologic agents were lack of effectiveness (66.1%) and lack of tolerance (24.7%).
Regression analysis set
The regression analysis set consisted of patients with data documented for the DAS28 and FFbH at baseline and at month 12 of follow-up. A total of 779 patients withdrew from the study at month 3 or month 6; these patients did not have data at month 12 and therefore did not meet the criteria for inclusion in the regression analysis set (N = 2,625). An additional 233 patients withdrew at 12 months. Because complete data for month 12 were available for these patients, they were included among the 2,625 patients in the regression analysis set. The most frequent reasons for withdrawal in the first 12 months of therapy were lack of efficacy (538 patients) and adverse drug reactions (247 patients).
Baseline characteristics for the data analysis set and regression analysis set were similar (Table 1).
Impact of treatment on concomitant therapy
The proportions of patients receiving concomitant therapy at the time of initiation of adalimumab therapy were similar in the data analysis set and the regression analysis set (Table 1). At the time of adalimumab treatment initiation, approximately 75% of patients were receiving one or more conventional DMARDs, mostly MTX or leflunomide. In the regression analysis set, the percentage of patients with concomitant DMARD therapy decreased to 69.7% after 12 months of treatment with adalimumab. However, the percentage of patients receiving concomitant treatment with MTX (as recommended) remained stable at 54.7%. The percentage of patients taking leflunomide was reduced by about one third during the same time interval (from 20.4 to 13.9%).
In the regression analysis set, the percentage of patients receiving glucocorticoids in addition to adalimumab decreased from 85.6% at baseline to 73.9% at month 12, and the mean equivalent dose of prednisolone decreased from 8.8 to 5.8 mg/day in the same time interval. The proportion of patients with concomitant painkillers was reduced for nonsteroidal anti-inflammatory drugs (NSAIDs; from 45.8 to 37.4%), cyclooxygenase-2 (COX-2) inhibitors (from 21.2 to 13.3%), and analgesics (from 25.5 to 16.6%).
Impact of treatment on disease activity and functional capacity
In the regression analysis set, disease activity decreased (from a mean DAS28 score of 5.9–3.9) and functional activity improved (from a mean FFbH score of 59.0–68.4% points; higher FFbH scores indicate greater functional capacity) during the first 12 months of adalimumab treatment (Fig. 1; Table 2). Outcomes were similar when the 1,012 patients with documented withdrawals were included in LOCF analyses of the data analysis set (N = 3,404; Table 2). Mean DAS28 scores decreased from 5.9 to 4.2 at month 12 in the data analysis set and mean FFbH scores increased from 58.3 to 65.7% points. However, compared to the regression analysis set, mean DAS28 scores in the data analysis set were slightly higher at month 12 (4.2 vs. 3.9) and mean FFbH scores were slightly lower (65.7 vs. 68.4% points), consistent with the observation that lack of efficacy accounted for the largest proportion of withdrawals.
Table 2 Disease activity (DAS28) and functional capacity (FFbH) at baseline, month 3, and month 12 in the data analysis set (N = 3,404, last observation carried forward) and the regression analysis set (N = 2,625)
The greatest improvements in disease activity and functional capacity occurred during the first 3 months of adalimumab therapy (Fig. 1). Mean DAS28 and FFbH scores continued to improve at 6 months and were sustained throughout 12 months. The DAS28 and FFbH scores displayed a low correlation throughout the course of the study (r = −0.41 at baseline and −0.5 at month 12).
At 12 months, 19.1% of patients in the regression analysis set were in clinical remission, defined as DAS28 < 2.6 points [14], 41.9% of patients had experienced a clinically relevant functional improvement, defined as an individual FFbH increase of 11 or more percentage points from baseline [15], and 57.8% were functionally independent (FFbH ≥ 67% points) [16] (Table 2). However, 9.1% of patients experienced a clinically significant decrease in functional capacity (FFbH decrease of at least 11% points) during this time period. The proportions of patients with clinical improvements were slightly lower in the data analysis set (Table 2).
Patient and disease characteristics with an influence on therapeutic effectiveness
In the regression model for disease activity (N = 2,625), 8 of the 35 tested variables showed a significant partial correlation with changes in DAS28 at month 12 and therefore can be regarded as having an impact on effectiveness in terms of disease activity (Table 3). Among these variables, baseline DAS28 exerted the strongest influence on therapeutic gain. High DAS28 baseline values correlated with reduced DAS28 values at month 12 and thus had a positive impact on reduction in disease activity. The additional effects of all other baseline parameters were less pronounced. Other variables with a positive effect on reduced disease activity were high FFbH scores at baseline, concomitant MTX therapy, and male gender. Previous biologic treatment, older age, increased number of tender joints, and use of nonselective NSAIDs were associated with reduced therapeutic gain as assessed by DAS28.
Table 3 Parameters with significant partial correlation with DAS28 changes at month 12 in the regression analysis set
The 11 parameters that modified the increase in functional capacity at month 12 (Table 4) were only partially identical with the ones associated with reduced disease activity. The strongest influence was exerted by baseline FFbH scores; high functional capacity at baseline was associated with a less pronounced increase of FFbH. Accordingly, patients with low functional capacity at baseline were more likely to achieve functional improvements during adalimumab treatment. Variables with less pronounced effects that were associated with reduced functional improvement were disease duration, previous joint replacement surgery, older age, previous biologic therapies, and high body mass index (BMI). Employment, male gender, use of COX-2 inhibitors, ESR levels, and baseline global patient assessment had a positive influence on increases in functional capacity.
Table 4 Parameters with significant partial correlation with FFbH changes at month 12 in the regression analysis set
No significant partial correlation with the two endpoints of therapeutic outcome was found for CRP levels, subjective pain experience, subjective limitations, number of concomitant DMARD therapies, duration of previous DMARD therapies, erosive changes, or concomitant diseases. Accordingly, we conclude that these characteristics do not independently contribute to therapy success or failure.
Analysis of selected patient subgroups
Subgroup analysis was performed on selected significant parameters to illustrate how these patient and disease characteristics impact therapeutic outcome. When classified into patients with high or moderate/low disease activity (baseline DAS28 value >5.1 and ≤5.1, respectively), the group of patients with high baseline DAS28 values experienced a larger reduction in mean disease activity (from a mean DAS28 of 6.4–4.2 over 12 months) than the group with moderate/low baseline DAS28 (from a mean DAS28 of 4.4–3.2).
A similar pattern was seen for patient function. When patients were classified by baseline FFbH scores, patients with higher physical impairment (baseline FFbH score ≤50% points) improved from a mean of 34.2–50.9 points during 12 months of treatment with adalimumab, while those with less impairment (baseline FFbH score >50% points) had much lower levels of improvement (from a mean of 73.9–78.9 points). Despite the greater extent of improvement, patients with high disease activity and low functional capacity at baseline did not reach the DAS28 and FFbH levels of the groups of patients with better baseline values.
Previous biologic treatment was associated with less pronounced therapeutic success of adalimumab for both disease activity and functional capacity. When classified into subgroups by the number of previous biologic therapies, patients with 3 previous biologic therapies had higher DAS28 values at baseline compared with patients with 1 or 0 previous biologic therapies. While all subgroups showed a decrease in DAS28 values during the 12 months of adalimumab treatment, this decrease was most pronounced in patients without previous biologic treatment and least pronounced in those with 2 or 3 previous biologic therapies (Fig. 2). The effect of prior biologic therapy was also evident when analyzing functional capacity. Patients without previous biologic treatment showed the highest baseline FFbH values (60.9% points) and improved to 70.7% points at month 12. With increasing numbers of previous biologic therapies, FFbH baseline and month 12 scores declined steadily. Patients with 3 previous biologic therapies had a mean FFbH baseline value of 48.8% points and gained only 4.4% points (to 53.2% points) by month 12.