Triple therapy in early active rheumatoid arthritis: A randomized, single-blind, controlled trial comparing step-up and parallel treatment strategies

Abstract

Objective

The Tight Control of Rheumatoid Arthritis study previously demonstrated that an intensive step-up disease-modifying antirheumatic drug (DMARD) treatment strategy targeting persistent disease activity was superior to routine care in the management of early rheumatoid arthritis (RA). We undertook this study to test the hypothesis that early parallel triple therapy achieves better outcomes than step-up therapy within an intensive disease management regimen.

Methods

Ninety-six patients with early RA (mean disease duration 11.5 months) were randomized to receive step-up therapy (sulfasalazine [SSZ] monotherapy, then after 3 months, methotrexate [MTX] was added, and when the maximum tolerated dosage of MTX was reached, hydroxychloroquine [HCQ] was added) or parallel triple therapy (SSZ/MTX/HCQ). All patients were assessed monthly for 12 months. If their disease activity score in 28 joints (DAS28) was ≥3.2, the dosage of DMARDs was increased according to protocol, and swollen joints were injected with triamcinolone acetonide (maximum dosage 80 mg per month). A metrologist who was blinded to the treatment allocation performed assessments every 3 months. The primary outcome measure was the mean decrease in the DAS28 score at 12 months.

Results

Both groups showed substantial improvements in disease activity and functional outcome. At 12 months, the mean decrease in the DAS28 score was −4.0 (step-up therapy group) versus −3.3 (parallel therapy group) (P = 0.163). No significant differences in the percentages of patients with DAS28 remission (step-up therapy group 45% versus parallel triple therapy group 33%), DAS28 good response (60% versus 41%, respectively), or American College of Rheumatology criteria for 20% improvement (ACR20) (77% versus 76%, respectively), ACR50 (60% versus 51%, respectively), or ACR70 (30% versus 20%, respectively) responses were seen. Radiologic progression was similar in both groups.

Conclusion

This study confirms that highly effective control of disease activity can be achieved using conventional DMARDs as part of an intensive disease management strategy. Within this setting, step-up therapy is at least as effective as parallel triple therapy.

The management of early rheumatoid arthritis (RA) has evolved rapidly over the last decade. Disease-modifying antirheumatic drugs (DMARDs) are used early in RA treatment with the aim of controlling synovitis and reducing radiologic progression. Randomized controlled trials have established that combinations of conventional DMARDs are more effective in controlling disease activity (1–4) and that this can have lasting implications on the rate of radiologic progression (3, 4). The Tight Control of Rheumatoid Arthritis (TICORA) trial (5) also demonstrated that intensive management of early RA, consisting of regular assessment and targeting of persistent disease activity by intensifying therapy in patients with a Disease Activity Score (DAS) of >2.4, results in superior outcomes when compared with routine outpatient management (6). Using the TICORA study regimen, the vast majority of patients improve and a majority of patients enter low disease activity states (or clinical remission).

However, a great deal of uncertainty remains. Can the results of the TICORA study be reproduced? In the TICORA study, more than half of the patients ended the trial using combination DMARD therapy. Was this the reason for the superiority of the strategy? The clinical results of the TICORA study were exceptional, but the reductions in radiologic progression were modest by comparison. Is this because it took time for optimal disease activity control to occur, and would more rapid control of inflammation by using the combination DMARD therapy earlier result in still better results? This study was undertaken to investigate whether the early use of parallel triple therapy (with methotrexate [MTX], sulfasalazine [SSZ], and hydroxychloroquine [HCQ]) within an intensive management strategy would be more effective than step-up therapy.

PATIENTS AND METHODS

Study participants.

The study was undertaken in 3 National Health Service teaching hospitals in Glasgow, UK. Between February 2003 and March 2005, patients between 18 and 80 years of age who were newly diagnosed as having RA were assessed. Patients with symptom duration of <5 years who had active RA, as defined by a Disease Activity Score in 28 joints (DAS28) of >5.1, and who had not previously been treated with DMARDs (other than HCQ) were allowed to participate in the trial (7). Patients with relevant concurrent liver (aspartate aminotransferase >80 IU/liter or alkaline phosphatase >700 IU/liter), renal (serum creatinine >130 μmoles/liter), hematologic (total white blood cell count <4 × 10⁹/liter, platelet count <150 × 10⁹/liter), or severe respiratory disease, or who were pregnant, planned to become pregnant, or were unwilling to use effective contraception were excluded. The protocol was approved by the Local Ethics Committee, and all patients were given written information about the trial. All patients provided written informed consent.

Procedures.

The treating doctor telephoned an administrative coordinator who randomly assigned patients either to the parallel triple therapy group or to the step-up therapy group using randomization software. All patients were seen monthly by the same rheumatologist (SAS). At each monthly visit, their DAS28 score was measured. This is a validated composite measure of disease activity based on the erythrocyte sedimentation rate (ESR), 28-joint swollen joint count, 28-joint tender joint count, and patient's global assessment of disease activity (using a 100-mm visual analog scale) (7). DAS28 scores of 5.1, 3.2, and 2.6 represented high, moderate, and low disease activity, respectively. At each monthly visit, any swollen joint amenable to intraarticular injection was injected with a steroid (20–40 mg triamcinolone acetonide) unless the joint had been injected within the previous 3 months or if the patient declined. A maximum total dosage of 80 mg of triamcinolone was injected per visit. Within the first 3 months of starting a new DMARD, if 80 mg of triamcinolone was not injected intraarticularly, the balance was provided by intramuscular injection if the DAS28 score remained ≥3.2. At each assessment, patients with a DAS28 score ≥3.2 had their DMARD therapy increased according to a protocol, unless they declined or adverse effects precluded this approach. Adverse events and drug-related toxicity were recorded, and these were managed by the rheumatologist according to local strategies.

DMARD protocols.

Patients assigned to parallel triple therapy were initially prescribed MTX (7.5 mg/week), SSZ (500 mg twice a day), and HCQ (200 mg daily). The MTX dosage was escalated each month in 2.5–5/mg increments until good disease control was obtained (up to a maximum of 25 mg/week) or side effects occurred. Thereafter, if good disease control was not achieved (i.e., if the DAS28 score remained ≥3.2), the dosage of SSZ was increased to 40 mg/kg/day in divided doses, and subsequently, if disease control remained inadequate, the dosage of HCQ was increased to 400 mg/day (Figure 1).

Figure 1. Protocol for increasing disease-modifying antirheumatic drugs (DMARDs) in the step-up therapy and parallel triple therapy groups. bd = twice a day; DAS = Disease Activity Score; TNF = tumor necrosis factor.

Patients assigned to step-up therapy were prescribed SSZ (target dosage 40 mg/kg/day in divided doses). After 3 months, if the DAS28 score remained ≥3.2, MTX was added (7.5 mg/week, increased monthly to a maximum of 25 mg/week, as above). After the maximum tolerated dose of MTX was reached, 400 mg/day of HCQ was added in patients with persistent disease activity.

In all patients, MTX was co-prescribed with 5 mg/week folic acid 4 days after MTX dosing. In both groups, if patients had persistent disease activity despite maximal drug therapy or drug-related toxicity, then alternative DMARDs or biologic agents could be used singly or in combination in order to control disease activity.

Assessments.

All patients were independently assessed at baseline and every 3 months for 1 year by the same metrologist (AS), who was blinded to treatment allocation. No intraarticular or intramuscular steroid injections were allowed in the 4 weeks preceding an assessment. Data were collected on the European League Against Rheumatism (EULAR) core measures of disease activity and outcome (28-joint swollen joint count, 28-joint tender joint count, ESR, patient's global assessment of disease activity, physician's global assessment of disease activity, pain score, and Health Assessment Questionnaire) (8, 9) and health-related quality of life (Short Form 12 health survey) (10). The primary outcome measure was the mean decrease in the DAS28 score at 12 months. Secondary outcome measures consisted of the proportion of patients with a EULAR good response (as defined by a final DAS28 score <3.2), the proportion of patients in remission (DAS28 score <2.6), and the proportions of patients who met the American College of Rheumatology criteria for a 20% response (ACR20), a 50% response (ACR50), and a 70% response (ACR70). Plain radiographs of the hands and feet were obtained at baseline and at 12 months. Two radiologists (RV and WK) who were blinded to treatment allocation independently scored the films using the modified Sharp/van der Heijde score (11). The radiographs were scored in pairs, with the chronology of the radiographs known.

Statistical analysis.

Intent-to-treat analysis was performed. Patients who died, were lost to followup, or who withdrew from the trial were designated as nonresponders. The mean decrease in the DAS28 score was assessed using Student's t-test, and the proportions of patients with a EULAR good response, EULAR remission, and ACR20, ACR50, and ACR70 responses were assessed by the Mantel-Haenszel test, using SAS v8.02 software (SAS Institute, Cary, NC). A difference in the DAS28 score of 1.1 was clinically significant, and the SD of the decrease in the DAS28 score in the patients was 1.46. Using a significance of P < 0.05, and a power of 90%, 37 patients per group needed to be analyzed. The correlation between the 2 radiologists' scores was computed to assess interobserver variability. The scores were then averaged, and the mean change in the joint space narrowing score, the erosion score, and the total Sharp score was analyzed.

RESULTS

Patients with newly diagnosed RA (n = 263) were screened for the study; after exclusions, 47 patients were randomized to receive step-up therapy and 49 to receive parallel triple therapy (Figure 2). Baseline characteristics and measures of disease activity in the 2 groups were similar, although patients in the step-up therapy group had higher baseline ESR and C-reactive protein (CRP) levels (Table 1). These differences were not statistically significant.

Figure 2. Randomization of the study patients. RA = rheumatoid arthritis; DAS28 = Disease Activity Score in 28 joints; DMARDs = disease-modifying antirheumatic drugs.

Table 1. Baseline characteristics of the study participants*

	Step-up therapy (n = 47)	Parallel triple therapy (n = 49)
* Except where indicated otherwise, values are the mean ± SD. DAS28 = Disease Activity Score in 28 joints; VAS = visual analog scale; CRP = C-reactive protein; ESR = erythrocyte sedimentation rate; HAQ = Health Assessment Questionnaire; SF-12 = Short Form 12.
Sex, % female	79	76
Age, years	55 ± 11	55 ± 15
Rheumatoid factor positive, %	72	69
Disease duration, months	13 ± 12	10 ± 9
DAS28	6.9 ± 0.9	6.8 ± 0.9
Swollen joint count (range 0–28)	13 ± 5	12 ± 4
Tender joint count (range 0–28)	18 ± 6	18 ± 6
Pain score (range 0–100-mm VAS)	65 ± 22	71 ± 26
Patient's global assessment, (range 0–100-mm VAS)	77 ± 18	82 ± 18
Physician's global assessment (range 1–5)	3.6 ± 0.6	3.4 ± 0.6
CRP, mg/liter	58 ± 53	38 ± 41
ESR, mm/hour	45 ± 30	36 ± 24
HAQ score (range 0–3)	2.0 ± 0.7	1.9 ± 0.7
SF-12 health survey score	28 ± 7	27 ± 8
Erosions at baseline, %	85	84
Total Sharp score, median (range)	42.5 (0–97)	38.5 (0–109)
Total Sharp score	44 ± 21	39 ± 22

Clinical responses.

After 12 months of followup, both groups demonstrated substantial improvements in the mean DAS28 score from baseline (Figure 3). There was a trend toward greater improvement in the step-up therapy group compared with the parallel triple therapy group (mean change in DAS28 score −4.0 versus −3.3), but this was not statistically significant (P = 0.163). Table 2 shows the proportions of patients who achieved EULAR good response, EULAR remission, and ACR20, ACR50, and ACR70 responses. At 12 months, more patients in the step-up therapy group achieved EULAR good responses (60%) and EULAR remission (45%) compared with patients in the parallel triple therapy group (41% and 33%, respectively), and a similar pattern was seen with ACR50 and ACR70 responses. However, these differences failed to reach statistical significance. Improvements were seen in both groups in all disease activity variables, as well as in physical function and quality of life, but there were no significant differences between groups (Table 3).

Figure 3. Mean change in the Disease Activity Score in 28 joints (DAS28) from baseline.

Table 2. Percentage of patients in whom response had been achieved at the 12-month assessment*

	Step-up therapy (n = 47)	Parallel triple therapy (n = 49)	OR (95% CI)
* Intent-to-treat analysis of all patients randomized. There were no significant differences in the percentage of responders between the 2 treatment groups. OR = odds ratio; 95% CI = 95% confidence interval; EULAR = European League Against Rheumatism (response based on Disease Activity Score in 28 joints); ACR20 = American College of Rheumatology criteria for a 20% response.
EULAR good response	60	41	0.47 (0.2, 1.1)
EULAR remission	45	33	0.6 (0.3, 1.4)
ACR20	77	76	0.9 (0.3, 1.6)
ACR50	60	51	0.7 (0.3, 1.6)
ACR70	30	20	0.6 (0.2, 1.5)

Table 3. Changes in disease activity, radiographic damage, physical function, and quality of life between baseline and 12 months*

	Step-up therapy (n = 44)	Parallel triple therapy (n = 47)	Difference (95% confidence interval)
* Values are the mean ± SD. There were no significant differences in the mean change in any parameter between treatment groups. See Table 1 for definitions.
DAS28	−4.0 ± 1.8	−3.3 ± 1.6	0.5 (−0.2, 1.1)
Swollen joint count (range 0–28)	−12 ± 5	−11 ± 4	0.3 (−0.5, 1)
Tender joint count (range 0–28)	−15 ± 8	−12 ± 7	2 (−0.1, 5)
Pain score (range 0–100-mm VAS)	−42 ± 32	−43 ± 34	−1 (−13, 10)
Patient's global assessment (range 0–100-mm VAS)	−51 ± 33	−52 ± 31	−4 (−7, 16)
Physician's global assessment (range 1–5)	−1.7 ± 0.8	−1.4 ± 0.8	0.2 (−0.1, 0.5)
CRP, mg/liter	−43 ± 51	−26 ± 39	5 (−1, 12)
ESR, mm/hour	−27 ± 27	−17 ± 25	2 (−5, 9)
HAQ score (range 0–3)	−0.9 ± 0.7	−0.8 ± 0.7	0.1 (−0.2, 0.4)
SF-12 health survey score	10 ± 11	9 ± 13	0.9 (−4.2, 6.0)
Erosion score	1.1 ± 1.8	1.7 ± 2.4	−0.6 (−1.5, 0.3)
Joint space narrowing score	4.9 ± 3.9	4.8 ± 6.4	0 (−2, 2)
Total Sharp score	6.0 ± 5.3	6.6 ± 7.0	−0.6 (−3, 2)

Radiologic progression.

Eighty-nine sets of radiographs were scored at baseline and at 12 months (43 in the step-up therapy group, 46 in the parallel triple therapy group); 4 patients were lost to followup, 1 patient died, and baseline or followup films were lost in 2 patients. There was no difference between the groups in radiologic progression over 12 months (Table 3). The correlation between the 2 radiologists' scoring of the change in total Sharp score was 0.95 (P < 0.01).

Drug regimens.

Eighty-eight percent of the patients originally randomized to receive parallel triple therapy finished the trial taking the following combination DMARD therapies: SSZ/MTX/HCQ (35 patients), MTX/HCQ (5 patients), SSZ/HCQ (2 patients), or SSZ/MTX (1 patient). Four patients from the parallel triple therapy group ended the trial taking MTX (2 patients) or sodium aurothiomalate (2 patients) monotherapy. In contrast, 49% of the step-up therapy group were taking SSZ (12 patients), MTX (9 patients), HCQ (1 patient), or sodium aurothiomalate (1 patient) monotherapy at 12 months. Forty percent of the step-up therapy group ended the trial taking the following combination DMARD therapies: SSZ/MTX (13 patients), MTX/HCQ (4 patients), SSZ/HCQ (1 patient), or SSZ/MTX/HCQ (1 patient). At study completion, 2 patients in the step-up therapy group were temporarily receiving no DMARD therapy because of adverse events. No patients received biologic treatments during the 12-month followup period.

Patients in the step-up therapy group received higher dosages of SSZ than those in the parallel triple therapy group (mean dosage 2.7 gm/day versus 1.2 gm/day) and higher dosages of HCQ (mean dosage 400 mg/day versus 238 mg/day), but similar dosages of MTX were given to both groups (mean dosage 14.6 mg/week versus 16 mg/week). Three patients (1 in the step-up therapy group and 2 in the parallel triple therapy group) were taking oral prednisolone (mean dosage 14.2 mg/day) prior to entering the trial; however, only 1 patient continued to take prednisolone at the end of the trial (5 mg/day). Patients in both groups received similar amounts of intraarticular and intramuscular triamcinolone acetonide during the study. The mean dosage of triamcinolone acetonide in the step-up therapy group was 32.6 mg/month, with 29.4 mg/month given in the parallel triple therapy group. The difference in the mean triamcinolone dosage between the 2 groups was equivalent to only 4 mg of prednisolone per month. In both groups, more triamcinolone was used early in the study; 57% of the total dose of triamcinolone given over 12 months was administered in the first 3 months of treatment.

Adverse events.

Members of both groups reported adverse events with similar frequency, and the nature of toxicity reported was as would be expected for the DMARDs used in the study. A total of 135 adverse events were reported in the step-up therapy group (48 gastrointestinal, 6 abnormal liver function tests, 27 infective, 16 mucocutaneous, 8 hematologic, 13 neurologic, and 17 other events). There were 141 adverse events reported in the parallel triple therapy group (52 gastrointestinal, 5 abnormal findings on liver function tests, 29 infective, 19 mucocutaneous, 8 hematologic, 6 neurologic, and 22 others). Adverse events resulted in 18 drug withdrawals in the step-up therapy group (SSZ withdrawn in 17 patients and MTX in 1 patient) and 15 drug withdrawals in the parallel therapy group (SSZ withdrawn in 8 patients, MTX withdrawn in 2 patients, HCQ withdrawn in 3 patients, and SSZ/MTX/HCQ withdrawn in 2 patients). One patient received prolonged hospital treatment during the trial following a spontaneous rupture of the esophagus that occurred at 6 weeks of followup. This patient had been assigned to the parallel triple therapy group and had no gastrointestinal symptoms prior to this event. No cause of the esophageal rupture was identified at the time of surgery. One patient died during the study followup. This patient had been assigned to the parallel triple therapy group and, at 11 months of followup, died of metastatic renal carcinoma thought to be unrelated to her study treatment.

DISCUSSION

The findings of this study confirm that substantial improvements in disease activity can be achieved using conventional DMARDs as part of an intensive disease management strategy in early active RA. Using this strategy of regular objective measurement of disease activity with the aim of targeting persistent disease, more than half of the patients studied achieved EULAR good control or ACR50 responses. Although anti–tumor necrosis factor (anti-TNF) agents could be instituted in those patients who failed to respond to or did not tolerate maximum doses of parallel triple therapy, no patient was given anti-TNF therapy during the 12-month followup period. Despite the use of conventional DMARDs, clinical response rates seen in our patients were similar to those reported in trials of etanercept (12), infliximab (13), and adalimumab (14) in early RA.

The primary objective of the trial was to compare the efficacy of a step-up DMARD regimen with a parallel triple-DMARD regimen. In the TICORA study, 51% of patients who received intensive treatment had to step up to triple therapy to achieve good control during the 18-month followup period. This raises the possibility that patients might benefit from earlier introduction of parallel triple therapy with the option of stepping down if good control was achieved and maintained. This study shows, however, that within the setting of an intensive management strategy, step-up therapy is at least as effective as parallel triple therapy in patients with high levels of disease activity. Twenty-six percent of patients who were screened did not meet the inclusion criteria because of low DAS28 scores, and the study leaves unanswered the question of how best to treat patients with mild disease.

Although there have been several studies of combination DMARDs in various strategies (step-up and parallel) in established RA, relatively few studies have been performed in the early stages of the disease. The Finnish Rheumatoid Arthritis Combination Therapy (FIN-RACo) study (3) is the only other published clinical trial of parallel triple therapy as the initial therapy in early RA. In that trial, the combination group had better clinical outcomes after 2 years of followup compared with the monotherapy (SSZ) group. However, all patients in the combination group also received low dosages of oral prednisolone (5 mg/day), and the study was open, with no blinded assessments of clinical outcome. Radiologic assessments were blinded, however, and it is noteworthy that the combination group had significantly less radiologic progression at 2 years. Radiologic benefit was still evident after 5 years of followup (15), and there was less work disability in the combination group over 5 years (16).

The Combinatietherapie Bij Reumatoïde Artritis (COBRA) study (4) used a step-down strategy, starting with high-dose oral prednisolone, MTX, and SSZ, and tapering to SSZ monotherapy by 40 weeks of followup. The control group received SSZ monotherapy. The combination group achieved a more rapid and greater clinical response compared with the monotherapy group, but this clinical benefit was lost when treatment in the combination arm was stepped down. However, there was less radiologic progression in the combination group, and although the rate of progression increased to that of the monotherapy group once combination therapy had been tapered, long-term followup showed sustained radiologic benefit in the combination arm (17).

Both the COBRA and FIN-RACo studies support the hypothesis that aggressive disease control in the early phase of RA results in long-term radiologic benefit that may translate into better functional outcome. Similarly, the TICORA study demonstrated less radiologic progression in patients treated with an intensive, aggressive step-up strategy compared with routine care. The Behandelstrategieën voor Reumatoide Artritis (BeSt) study (18) compared 4 treatment strategies in early RA: sequential monotherapy, step-up therapy, COBRA-type combination therapy, and MTX/infliximab combination therapy. Although clinical responses were most rapid in the COBRA-type combination and MTX/infliximab combination groups, at 2 years there were no significant differences in clinical response between these 2 groups and the sequential monotherapy and step-up therapy groups. There was significantly less radiologic progression in the initial combination arms. This is as would be expected given the now well-established effects of anti-TNF agents on radiologic progression (19), and it confirms the effect of the COBRA regimen previously demonstrated by Boers et al (4).

Of particular interest in the BeSt study, step-up therapy did not result in better outcomes than did sequential monotherapy. This was not consistent with the findings of the TICORA study, but important differences in treatment strategy exist. Although both studies used a target of DAS-based good control to increase treatment, patients in the BeSt study only had increases in treatment every 3 months as opposed to the monthly assessments and treatment changes used in the intensive TICORA study regimen. Additionally, in the TICORA study, rapid early improvement in disease activity and function were achieved using liberal intraarticular steroid injections, although the mean dosage of steroid used in the first 3 months was only 2.7 mg/day of prednisolone equivalent. These 2 differences between BeSt step-up therapy and TICORA step-up therapy may explain why the intensive TICORA study regimen achieved such superior results.

Because all published studies to date have demonstrated the superiority of triple therapy, our finding that step-up therapy is at least as effective as parallel triple therapy in early RA is unexpected. In the TICORA study, the intensive step-up strategy resulted in exceptional clinical results, and it simply may not be possible to improve upon this with conventional DMARD therapies. Our step-up strategy is multifaceted, and we would argue that regular objective assessment, targeted treatment, and early restoration of function using intraarticular steroid injections are important aspects of that strategy. Previously published studies have compared triple therapy with single- or dual-DMARD therapy, but not within the setting of an intensive management regimen. One other possible explanation is that the dosages of MTX, SSZ, or HCQ used in our triple therapy regimen were too low. In particular, the mean dosage of SSZ used in the step-up therapy group was significantly higher than that used in the parallel triple therapy group. Better clinical outcomes might have been achieved if higher dosages of SSZ had been provided, but the dosage of SSZ used was the same as that used in the FIN-RACo study and the O'Dell study (1), both of which favored triple therapy.

Although substantial improvements in the mean DAS28 score were observed in this study, the ACR20, ACR50, ACR70, and EULAR response rates were not as high as those seen in the TICORA study. Direct comparisons of different study cohorts should always be undertaken with caution; however, there were some differences in study design that may be relevant. First, the TICORA study found outcomes after 18 months of intensive treatment as opposed to the 12-month study end point in this trial. Patients in the TICORA study continued to accrue benefit between 12 and 18 months, and it is possible that greater improvement would have been achieved had our study continued for 18 months. Additionally, the step-up protocol for our study was not the same as the step-up protocol for the TICORA study. In the TICORA study, if good control was not achieved after 3 months of SSZ treatment, the next step was triple therapy. In our study, MTX was added to SSZ initially, and HCQ was added only if the patient failed to achieve good control with the MTX/SSZ combination. This ensured that the step-up therapy group received a true step-up regimen, allowing us to compare step-up therapy with parallel triple therapy rather than measuring the effect of delaying the introduction of triple therapy by 3 months.

Because clinical outcomes were not significantly different between the 2 treatment groups, it is predictable that there was no difference in radiologic outcome. Unfortunately, radiologic progression occurred in both groups, and despite our best use of conventional DMARDs, it is unlikely that we will ever achieve the reduction of radiologic progression seen with anti-TNF agents. The majority of our study cohort (85%) had erosions at presentation, and it may be that less progression would have occurred with earlier treatment. Although the majority of patients who entered this study were seen immediately after referral by their general practitioner, the average disease duration was still >10 months from symptom onset. The lag time to the introduction of DMARDs has decreased considerably over the years (20), but clearly, we are still not achieving the target of truly early treatment. We must continue to invest more into educating the public, primary care providers, and those who control health care resources about the importance of early recognition and treatment of RA.

In conclusion, using an intensive management regimen with conventional DMARDs in early RA results in substantial improvements in clinical disease activity and functional outcome; within this setting, step-up therapy is at least as effective as parallel triple therapy, with 45% of patients achieving remission after 12 months. However, despite optimal use of conventional DMARDs, a significant proportion of patients fail to achieve good response, and radiologic progression may still occur despite clinically low disease activity. Anti-TNF agents may be able to dramatically slow or even arrest radiologic progression, but the routine use of anti-TNF agents in early disease is prohibitively expensive for most health care providers. Further controlled trials are needed to evaluate the role of anti-TNF therapy in early RA in comparison with or as part of the TICORA study–type intensive management strategy.

AUTHOR CONTRIBUTIONS

Dr. Saunders 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. Capell, Porter.

Acquisition of data. Saunders, Capell, Stirling, Vallance, Kincaid.

Analysis and interpretation of data. Saunders, Capell, Vallance, Kincaid, McMahon, Porter.

Manuscript preparation. Saunders, Capell, Stirling, Porter.

Statistical analysis. McMahon.

Acknowledgements

The authors thank Drs. R. Madhok, Prof. R. D. Sturrock, Prof. I. B. McInnes, Dr. M. Field, Dr. J. A. Hunter, Dr. M. M. Gordon, Dr. A. McEntegart, Dr. H. Wilson, and Dr. P. E. McGill for allowing us to study their patients. We also would like to thank A. Tierney and R. Hampson for assistance with randomization.