Elsevier

The Lancet

Volume 364, Issue 9430, 17–23 July 2004, Pages 263-269
The Lancet

Articles
Effect of a treatment strategy of tight control for rheumatoid arthritis (the TICORA study): a single-blind randomised controlled trial

https://doi.org/10.1016/S0140-6736(04)16676-2Get rights and content

Summary

Background

Present treatment strategies for rheumatoid arthritis include use of disease-modifying antirheumatic drugs, but a minority of patients achieve a good response. We aimed to test the hypothesis that an improved outcome can be achieved by employing a strategy of intensive outpatient management of patients with rheumatoid arthritis—for sustained, tight control of disease activity—compared with routine outpatient care.

Methods

We designed a single-blind, randomised controlled trial in two teaching hospitals. We screened 183 patients for inclusion. 111 were randomly allocated either intensive management or routine care. Primary outcome measures were mean fall in disease activity score and proportion of patients with a good response (defined as a disease activity score <2·4 and a fall in this score from baseline by >1·2). Analysis was by intention-to-treat.

Findings

One patient withdrew after randomisation and seven dropped out during the study. Mean fall in disease activity score was greater in the intensive group than in the routine group (–3·5 vs –1·9, difference 1·6 [95% CI 1·1–2·1], p<0·0001). Compared with routine care, patients treated intensively were more likely to have a good response (definition, 45/55 [82%] vs 24/55 [44%], odds ratio 5·8 [95% CI 2·4–13·9], p<0·0001) or be in remission (disease activity score <1·6; 36/55 [65%] vs 9/55 [16%], 9·7 [3·9–23·9], p<0·0001). Three patients assigned routine care and one allocated intensive management died during the study; none was judged attributable to treatment.

Interpretation

A strategy of intensive outpatient management of rheumatoid arthritis substantially improves disease activity, radiographic disease progression, physical function, and quality of life at no additional cost.

Introduction

Present treatment strategies for rheumatoid arthritis use disease-modifying antirheumatic drugs (singly or in combination) as early as possible in the disease process, because suppression of disease activity correlates with reduction in radiological joint damage.1 The challenge is to ascertain whether tight control of the inflammatory response can be achieved and sustained in a large proportion of rheumatoid arthritis patients, and to assess the effect of such management on symptoms and medium-term and long-term outcomes. To date, published trials have focused on specific drug combinations2, 3 or have used an open study design without masked assessments of disease activity.4 Moreover, in most trials, combinations of two or more disease-modifying antirheumatic drugs have been used from the outset, whereas many clinicians do not favour this approach in all patients with a new diagnosis of rheumatoid arthritis, preferring instead to step up treatment in those with disease that has proven resistant to monotherapy.5 We did a randomised controlled trial (with masked assessments) of a therapeutic strategy aiming for sustained, tight control of disease activity compared with routine outpatient care.

This study was undertaken in two National Health Service (NHS) teaching hospitals in Glasgow, UK. Between August, 1999, and April, 2001, we recruited patients aged between 18 and 75 years who had had rheumatoid arthritis for fewer than 5 years. All patients had active disease, defined by a disease activity score of more than 2·4. We excluded patients who had previously received combination disease-modifying antirheumatic drug treatment, or had relevant concurrent liver (aspartate aminotransferase >80 IU/L, alkaline phosphatase >700 IU/L), renal (creatinine >0·2 mmol/L), or haematological disease (white-cell count <4·0×109/L, platelet count <150×109/L). The local ethics committee in Western Infirmary, Glasgow, approved this protocol. All patients gave written informed consent.

The treating doctor telephoned an administrative co-ordinator, who randomly assigned patients either intensive or routine management, with randomisation software. Patients assigned to the intensive group were seen every month by the same rheumatologist (CG), and their disease activity score was calculated. This score is a validated composite of erythrocyte sedimentation rate, Ritchie articular index, joint swelling count, and patients' global assessment of disease activity.6, 7 Disease activity scores of 3·6, 2·4, and 1·6 represented high, moderate, and low disease activity, respectively. At every monthly assessment, we injected any swollen joint amenable to intra-articular steroid, unless the joint had been injected within the previous 3 months or the patient declined. We injected a maximum of three joints per assessment, up to a total dose of 120 mg triamcinolone acetonide per visit. Within the first 3 months of starting a new disease-modifying antirheumatic drug, if 120 mg of triamcinolone acetonide was not injected intra-articularly, we gave the balance by intramuscular injection if the disease activity score remained more than 2·4. At every assessment after month 3, patients with a score of more than 2·4 received an escalation of their oral treatment according to a protocol (figure 1), unless they declined or toxic effects precluded this approach. We recorded adverse events, and drug-related toxic effects were managed empirically by the rheumatologist.

Treatment for patients assigned to the routine care group was supervised in the usual rheumatology follow-up clinics, which were led by two consultant rheumatologists and included trainee rheumatologists working under supervision. These participants were reviewed every 3 months, with no formal composite measure of disease activity used in clinical decision making. Disease-modifying antirheumatic drug monotherapy was given in patients with active synovitis, and failure of treatment (because of toxic effects or lack of effect) resulted in a change to alternative monotherapy, or addition of a second or third drug at the discretion of the attending rheumatologist. Intra-articular injections of corticosteroid were given to patients assigned routine care with the same restrictions as those in the intensive group.

Every 3 months, a metrologist assessed patients from both groups. To ensure masked assessments, the metrologist knew of the aim and details of the study design, but was unaware of participants' assigned treatment groups. Before every assessment, patients were sent a letter stating that they should make no mention of their drug treatment or the identity of their doctors. Assessments were undertaken without the case record and on a different clinic day from that attended by the intensive group for their monthly visits—hence all patients (intensive and routine) attended the assessment clinic every 3 months. The assessing metrologist did not attend the other outpatient clinics, at which the intensive group attended every month, and did not take part in intensive treatment. No intra-articular injections were allowed in the month preceding the assessments.

Primary outcome measures were mean fall in disease activity score, and the proportion of patients with a good response (European League Against Rheumatism [EULAR] definition—ie, disease activity score <2·4 and a fall in score from baseline by >1·2). Secondary outcome measures consisted of the proportion of patients in remission (EULAR definition—ie, disease activity scores <1·6),8 American College of Rheumatology (ACR) response rates, and other constituents of the EULAR core measures of disease activity and outcome, namely visual analogue pain score, assessor's global assessment of disease activity, and patient's function measured by the health assessment questionnaire (at 0 and 18 months). ACR 20, 50, and 70 responses are defined as at least 20%, 50%, and 70% improvement in joint swelling and joint tenderness counts, and three of five other variables (ie, erythrocyte sedimentation rate, health assessment questionnaire, pain score, and assessors' and patients' global assessments). We used the short form-12 questionnaire (a reliable, valid, and responsive measure of health status in rheumatoid arthritis)9 to measure health-related quality of life. Two radiologists scored radiographs of hands and feet at 0 and 18 months with the van der Heijde modification of the Sharp score.10 Films were scored with the sequence of radiographs known, but the radiologists were masked to treatment groups.

We obtained data on resource use from case-note review and patients' diaries. We measured hospital resource use in terms of the number of outpatient visits, inpatient stays by specialty, and prescription costs of disease-modifying antirheumatic drugs or non-steroidal anti-inflammatory drugs. Community resource use was obtained with 1-month prospective patients' diaries at 0, 6, and 12 months. We defined community resources as visits to the family practitioner, practice nurse, or other health professionals, and blood-test monitoring. We assumed that the rate of resource use was constant for every subsequent month after completion of the previous patient's diary. Perspective of the economics evaluation was from the NHS, although we also measured travel costs by self-completed patients' questionnaires. We costed resource use with published unit-cost data.11, 12, 13, 14 For example, the unit cost for a rheumatology inpatient bed per day and an outpatient attendance was £232 and £33, respectively. We calculated incremental cost differences between each arm of the trial for patients who completed the trial. All costs have been expressed in 2001–02 prices. No discounting was applied because of the short study timeframe.

We did an intention-to-treat analysis; patients who died, were lost to follow-up, or withdrew from the trial were designated as non-responders. We assessed the mean fall in disease activity score with Student's t test and the proportion of patients with a good response with the Mantel-Haenszel procedure, with SAS version 8.02. A difference in disease activity score of 1·1 was clinically significant, and the SD of the fall in score in a previous trial15 was 0·7. With a significance of p<0·01 and a power of 95%, 21 patients per group needed to be analysed. However, to detect an increase in the number of good responders from 40% to 70%, for a trial of 95% power at p<0·01, 53 patients per group would need to be analysed. To assess interobserver variability, we calculated the correlation between the two radiologists' scores; median change in joint-space narrowing score, erosion score, and total Sharp score was measured by Mann-Whitney non-parametric analysis. We used non-parametric bootstrapping techniques to calculate 95% CIs around cost differences with the percentile method. Sensitivity analysis was used to vary the unit-cost data.

The study was funded by the Chief Scientist's Office, Scottish Executive, and provided feedback on study design, but had no role in data collection, data analysis, data interpretation, writing of the report, or decision to submit the paper for publication.

Section snippets

Results

We screened 183 patients for the study; after exclusions, 55 were assigned intensive management and 55 routine care (figure 2). Baseline characteristics and measures of disease activity in the two groups were similar, although patients randomly assigned to the intensive group had slightly higher erythrocyte sedimentation rates and C-reactive protein concentrations, but slightly less radiological damage, than those in the routine group (table 1). Differences in baseline measures of disease

Discussion

Our evidence lends support to the hypothesis that tight control of rheumatoid arthritis can be achieved in most patients with early rheumatoid arthritis, with a strategy of intensive treatment. It is noteworthy that, in this era of targeted biological therapies, this tight control was achieved with standard disease-modifying antirheumatic drugs without the use of anti-tumour necrosis factor (TNF) treatment.

The strategy used in this trial was multifaceted—treatment consisted of the frequent,

References (20)

There are more references available in the full text version of this article.

Cited by (1276)

View all citing articles on Scopus
View full text