Dose escalation of parenteral methotrexate in active rheumatoid arthritis that has been unresponsive to conventional doses of methotrexate: A randomized, controlled trial




To examine whether dose escalation of intramuscular (IM) methotrexate (MTX) up to 45 mg/week improves disease control in patients who have active rheumatoid arthritis (RA) despite receiving conventional doses (15 mg/week) of IM MTX, and to obtain preliminary data on patient tolerability and adverse effects of higher doses of IM MTX.


Patients >18 years of age who had active RA, defined as a European League Against Rheumatism (EULAR) Disease Activity Score in 28 joints (DAS28) of >3.2, and who had received 15–20 mg/week of oral MTX for at least 2 months were switched (week 0) to 15 mg/week of IM MTX for 6 weeks. Patients whose DAS28 remained >3.2 at both week 4 and week 6 were randomized, in a double-blind manner, either to continue to receive 15 mg/week IM MTX with monthly placebo escalation or to receive escalating doses of IM MTX monthly up to 45 mg/week. The dose of MTX or placebo was escalated every 4 weeks if the DAS28 was >2.5. Safety assessments and determination of the DAS28 were performed every 2 weeks and monthly, respectively. Disease activity parameters from the American College of Rheumatology (ACR) core disease activity set and health status as recorded on the Medical Outcomes Study Short-Form 12 were determined at baseline (week 0) and final assessment (week 22). The primary outcome was the percentage of patients in each group achieving a target DAS28 of <3.2. Secondary outcomes comprised the percentage of patients whose DAS28 improved by >1.2, the percentage of patients achieving a 20% improvement in the ACR core disease activity measures (ACR20), and the percentage of patients achieving a good response, a moderate response, or no response in accordance with the EULAR response criteria.


Sixty-four patients were eligible for entry and were switched from oral MTX to 15 mg/week IM MTX. At baseline, the mean ± SD DAS28 was 5.6 ± 0.88; after 6 weeks of IM MTX, the DAS28 had improved by a mean of 0.42 (95% confidence interval [95% CI] 0.15–0.69). At 6 weeks, 54 patients still had a DAS28 of >3.2 and were therefore eligible for randomization. By 22 weeks, 1 patient (3.7%) in each group achieved the primary outcome of a DAS28 <3.2 (95% CI for the difference between the groups −15% to +15%). Five patients (18.5%) in each group showed an improvement of >1.2 in the DAS28 (95% CI for the difference between the groups −18% to +18%). One patient (3.7%) in each group achieved an ACR20 response, but none achieved a good response as defined by the EULAR response criteria. One patient in each group had a serious adverse reaction; minor adverse reactions were more frequently reported in the dose escalation group.


Switching from oral to parenteral MTX 15 mg/week results in a minor improvement in disease control. For patients with active RA receiving 15 mg/week IM MTX, increasing the dose up to 45 mg/week does not improve disease control. Higher doses of IM MTX were generally well tolerated and not associated with an increase in serious adverse reactions to the drug.

The efficacy of low-dose weekly methotrexate (MTX) for the treatment of patients with active rheumatoid arthritis (RA) has been demonstrated in several large randomized, controlled trials and long-term followup studies (1–4). The favorable efficacy and toxicity profile of MTX, compared with other disease-modifying antirheumatic drugs (DMARDs), has ensured its central position, both as monotherapy and as the mainstay of combination therapy for RA (5). Nevertheless, despite its rapid onset of action with initial improvement seen within weeks of treatment, a 50% improvement in key parameters is achieved in only a small proportion of patients (5). Patients who have achieved a partial response to MTX form an important group for whom there is an increasing number of therapeutic options, which differ in cost and efficacy (6).

Against this background we have examined the safety and efficacy of an alternative strategy for this group of patients, namely a switch to parenteral administration of MTX and escalation of the dose beyond conventional doses of 20–25 mg/week (7) up to 45 mg/week. There are preliminary data to support this strategy (8–10), but the efficacy of dose escalation has not been formally addressed in a rigorous prospective, randomized, double-blind, placebo-controlled study with the use of a disease activity threshold to trigger dose escalation.


Patient selection.

Patients with active RA who were attending the rheumatic diseases clinic at the University of Edinburgh's Western General Hospital were screened for eligibility. The eligibility criteria were as follows: age >18 years, a diagnosis of RA according to the American College of Rheumatology (ACR; formerly, the American Rheumatism Association) 1987 revised criteria (11), active disease defined as a European League Against Rheumatism (EULAR) Disease Activity Score in 28 joints (DAS28) of >3.2 (12), a stable dose of oral MTX at 15–20 mg/week received for at least 2 months, and previous treatment with at least 1 other DMARD besides MTX received for >4 months with >2 months at full dose. Exclusion criteria were as follows: an inability, for medical reasons, to increase the MTX dose, administration of an intraarticular steroid injection or change in oral steroid dose within the previous 4 weeks, a daily oral steroid dose >10 mg/day, and elective surgery planned within 6 months. The Lothian Research Ethics Committee granted ethical approval and all patients gave their written, informed consent.

Study design.

Phase I.

At baseline (week 0), eligible patients were switched to 15 mg/week intramuscular (IM) MTX for 6 weeks. Determination of the DAS28 was repeated at week 4 and week 6, and patients whose DAS28 remained >3.2 at both week 4 and week 6 entered the randomization phase.

Phase II.

Patients were randomized to receive either 15 mg/week IM MTX with placebo dose escalation or escalation of the IM MTX dose up to 45 mg/week. The MTX or placebo dose was escalated every 4 weeks if the DAS28 was >2.5. This threshold of a DAS28 >2.5 for escalating the dose was selected to ensure that therapy would be increased unless there had been a significant improvement in disease control. In the intervention arm, the dose of MTX was increased to 20 mg, 25 mg, 35 mg, and 45 mg/week consecutively every 4 weeks, provided that there were no medical contraindications and that the DAS28 remained above 2.5. In the placebo arm, patients were administered 15 mg MTX with the addition of carrier solution at an equal volume and in a color identical to that in the intervention arm, provided that the DAS28 was above 2.5. If any patient achieved a DAS28 of ≤2.5, the therapy was left unchanged until the next assessment. All patients received 5 mg folic acid/week within 24 hours after the MTX.

Alteration in dose or change to the route of administration of concomitant medication for arthritis was not permitted during the study. No concomitant DMARD therapy was permitted. A maximum of 2 intraarticular steroid injections of 40 mg methylprednisolone acetate (Depot-Medrone; Pharmacia, Piscataway, NJ) per patient were permitted during the course of the study but were prohibited within the final 6 weeks of the trial. Antiemetic and other therapy for adverse effects were left to the discretion of the treating physician, and the use of these treatments was recorded.

Sample size.

In the control group, a maximum placebo response of 30% of patients achieving the DAS28 target was anticipated. A minimum difference of 30% (placebo) versus 80% (active therapy) was sought. To detect this difference with 90% power required a sample size of 22 patients in each group (at a 5% significance level). Allowing for a 30% dropout rate, a total of 55 patients was required.


Computer-generated random numbers were used to allocate each treatment group. Treatment codes were kept off site but could be accessed if required by the safety monitors. Block randomization and stratification according to the use/nonuse of oral steroid was used to ensure parity of treatment allocation.


The patients, metrologists, and physicians were blinded to the treatment allocation. All of the IM MTX was constituted to an equal volume and matching color by an off-site pharmacy (Department of Pharmacy and Pharmacology, University of Bath, UK) and dispensed from a prefilled syringe. Patients who completed the protocol remained on the masked therapy until the conclusion of the study. Patients who withdrew from the study were treated at the physician's discretion, with blinding to the trial medication maintained until the conclusion of the study.

Statistical analysis.

Analysis was performed on the basis of intention to treat with last observation carried forward. Analysis of covariance adjusted for baseline level was used to estimate the mean effect of the intervention, as compared with the control, on the DAS28. Confidence limits for the differences in percentage response between the 2 groups was calculated using Newcombe's method.



Age, sex, disease duration, number of previous DMARDs used, overall duration of therapy with MTX prior to baseline (week 0) assessment, rheumatoid factor status, and the presence of erosions on radiographs of the hands, wrists, or feet were recorded for all patients.

Clinical assessments.

At each visit, a 28-joint count of the tender (TJC) and swollen (SJC) joints, the erythrocyte sedimentation rate (ESR; Westergren method), and results from a 100-mm linear general health scale (GH) were recorded. In addition, at weeks 6 and 22 (randomization and final assessment, respectively), the modified Health Assessment Questionnaire (13), the Short-Form 12 (SF-12) version of the Medical Outcomes Study (14), a visual analog scale for physician assessment of disease activity (100 mm), and a visual analog scale for patient assessment of pain (100 mm) were completed.

Disease activity, expressed as the DAS28, was calculated at baseline (week 0), after 4 weeks, at 6 weeks (randomization), and then at 4-week intervals up to week 22 (final outcome assessment). The DAS28 is a continuous composite index (range 0–10) that has been well documented and validated for the measurement of disease activity in RA (12). This index is calculated as follows: DAS28 = equation image + equation image + 0.70(lnESR) + 0.014(GH). A DAS28 >5.1 and <3.2 indicate high and low levels of disease activity, respectively. A change in the DAS28 of >0.6 constitutes a change greater than the measurement error of the instrument, while a change of >1.2 (twice the measurement error) is a clinically significant change.

Primary outcome.

The percentage of patients in each group who achieved the target DAS28 of <3.2 was considered the primary outcome measure.

Secondary outcomes.

Secondary outcome measures were the percentage of patients in each group whose DAS28 improved by >1.2, the percentage of patients who achieved a 20% improvement in the ACR core disease activity measures (ACR20 response), the percentage of patients achieving a good response, a moderate response, or no response according to the EULAR response criteria (15), the change in the ACR core set disease activity measures (16), and health status recorded using the SF-12.

Safety assessments.

A full blood cell count as well as renal and liver function tests were performed every 2 weeks from week 0 to week 22 and when clinically indicated. A baseline chest radiograph and pulmonary function tests were performed at baseline and repeated as clinically indicated.

Toxicity criteria.

Toxicity criteria, which were applied to determine whether or not the dose of MTX should be increased, comprised a reduced total white cell count, levels of neutrophils or platelets lower than the normal range, liver function test abnormalities (>2 times the upper limit of normal), or creatinine levels >25% over baseline levels. MTX was withheld in the event of intercurrent infection and was reintroduced at the discretion of the treating physician.

Serious adverse reactions.

Patients were permanently withdrawn from the study if they had a serious adverse reaction, defined as a total white blood cell count of <3.0 × 109/liter, neutrophil count of <2.0 × 109/liter, platelet count of <100 × 109/liter, abnormality on liver function tests (>3 times the upper limit of normal), or >50% rise in the creatinine level over baseline. Patients who developed a cough, shortness of breath, or fever that was unexplained by intercurrent infection and that lasted more than 4 days were also classified as having a serious adverse reaction and were permanently withdrawn. A patient who developed any other unexplained serious event, whether or not it was apparently related to the MTX, could at the discretion of the safety monitor be classified as having a serious adverse reaction and be permanently withdrawn.

Minor adverse reactions.

Adverse reactions that were considered minor comprised the presence of nausea, vomiting, diarrhea, rash, headache, dizziness, mouth ulcers or stomatitis, and hair loss. Cough, dyspnea, fever, or chills of <4 days' duration were also classified as minor adverse reactions. A structured questionnaire was administered every week to document these minor adverse reactions and any action that was taken. A minor adverse reaction was defined as a symptom that prompted the issue of a drug prescription or that was reported for 3 consecutive weeks. Independent safety monitors (EM and VD) adjudicated on issues relating to a patient's withdrawal from the study. They were also empowered to discontinue the study if there were >4 serious adverse events in the intervention arm of the trial and if the frequency of serious events reached twice that of the control arm, or if, in their professional opinion, unforeseen safety issues arose but <4 adverse events had been recorded.


Prerandomization (phase I).

Sixty-four patients were enrolled and were switched from oral MTX to 15 mg/week IM MTX (week 0). At baseline (week 0), the mean ± SD DAS28 was 5.6 ± 0.88. After 6 weeks of IM MTX, the DAS28 had improved by a mean of 0.42 (95% confidence interval [95% CI] 0.15–0.69). Fifty-four patients still had a DAS28 of >3.2 and were therefore eligible for randomization, 4 patients had achieved a DAS28 of <3.2 and were maintained on 15 mg IM MTX, and 6 patients declined to participate further (Figure 1).

Figure 1.

Randomization, assignment, and discontinuation of patients with rheumatoid arthritis in the dose escalation trial of methotrexate (MTX). DAS = Disease Activity Score in 28 joints.

Randomization (phase II).

Baseline characteristics.

The 54 patients who continued to have a DAS28 of >3.2 were randomly allocated to 1 of the 2 treatment groups. At randomization, the groups were well matched for all demographic, disease, and health status characteristics (Table 1). Both groups had a high percentage of patients with seropositive and erosive RA, with a mean duration of disease of ∼10 years. The groups were comparable with respect to the use of oral steroid, the number of previous DMARDs, and the duration of prior therapy with MTX. One-third of the patients in each group had received MTX for more than 3 years. The baseline mean DAS28 of 5.4 in each group indicates that high levels of disease activity were present despite treatment with MTX.

Table 1. Patient characteristics at randomization (week 6)*
Patient characteristicsControls (n = 27)MTX escalation (n = 27)
  • *

    MTX = methotrexate; DMARDs = disease-modifying antirheumatic drugs; DAS28 = Disease Activity Score in 28 joints; PCS12 = physical component of the Short-Form 12; MCS12 = mental component of the Short-Form 12.

 Age, mean ± SD years58 ± 1156 ± 12
 Female, no. (%)20 (74)24 (88)
 Race, no. (%) white27 (100)27 (100)
Disease status  
 Disease duration, mean ± SD years9.3 ± 5.710.0 ± 6.9
 Rheumatoid factor positive, no. (%)23 (85)23 (85)
 Erosions, no. (%)25 (92)23 (85)
Drug treatments  
 Previous DMARDs (including MTX), median number23
 Receiving prednisolone, no. (%)5 (18.5)5 (18.5)
 Dose of prednisolone, mean mg/day7.56.6
 Receiving MTX, no. (%)  
  <12 months6 (22)8 (30)
  >12<24 months8 (30)7 (26)
  >24<36 months4 (15)3 (11)
  >36 months9 (33)9 (33)
 Duration of MTX, mean ± SD months34 ± 3131 ± 22
DAS28, mean ± SD score5.4 ± 1.15.4 ± 0.97
PCS12, mean ± SD score34 ± 1029 ± 7
MCS12, mean ± SD score48 ± 1146 ± 10

Clinical outcome.

In the intervention group, the median dose of MTX achieved was 45 mg/week (range 20–45); in 21 patients (78%), the maximum dose of 45 mg/week was reached. One patient in each group (3.7%) achieved the primary outcome of a DAS28 of <3.2 (Table 2) (95% CI for the difference between intervention and control groups for achieving the primary outcome −15% to +15%). Five patients (18.5%) in each group showed a >1.2-unit improvement in their DAS28 (95% CI for the difference between the intervention and control groups achieving an improvement in DAS28 >1.2 −18% to +18%). One patient in each group (3.7%) achieved an ACR20 response (Table 2). No patients achieved a good response on the basis of the EULAR response criteria (Table 2), and there was no significant difference between the groups in either the physical or mental health components of the SF-12 score (data not shown). Moreover, there was no significant difference between the groups for the change in DAS28 or change in individual components of the ACR core disease activity set (Table 3).

Table 2. DAS28, ACR20 response, and EULAR response criteria in the study groups by week 22
OutcomeControls (n = 27)MTX escalation (n = 27)
  1. * Values are the no. (%) of patients. DAS28 = Disease Activity Score in 28 joints; ACR20 = American College of Rheumatology 20% improvement criteria; EULAR = European League Against Rheumatism; MTX = methotrexate.

Primary outcome  
 DAS28 <3.21 (3.7)1 (3.7)
Secondary outcomes  
 Improvement in DAS28 >1.25 (18.5)5 (18.5)
 ACR20 response1 (3.7)1 (3.7)
 EULAR response  
  Moderate7 (26)8 (30)
  None20 (74)19 (70)
Table 3. Change in DAS28 and American College of Rheumatology core disease activity set measures in the study groups by week 22*
Disease activity measureControls (n = 27)MTX escalation (n = 27)
  • *

    Values are the mean ± SD, with last observation carried forward for noncompleters. A negative absolute change or percentage change indicates improvement in the response criterion. No significant difference between groups for any variable (P > 0.1). The visual analog scale (VAS) for pain ranges from 0 = no pain to 100 = extreme pain. Patient and physician global assessments of disease activity range from 0 = no disease activity to 100 = extreme disease activity. The Health Assessment Questionnaire (HAQ) scale ranges from 0 = no difficulty to 3 = unable to perform activity. DAS28 = Disease Activity Score in 28 joints; MTX = methotrexate.

 Baseline5.4 ± 1.15.4 ± 0.97
 Absolute change−0.7 ± 1.3−0.5 ± 1.0
 Percentage change−13−9
Erythrocyte sedimentation rate, mm/hour  
 Baseline33 ± 19.134 ± 23.5
 Absolute change−5.4 ± 14.82.2 ± 14.3
 Percentage change−166.4
Swollen joint count (range 0–28 joints)  
 Baseline11 ± 69 ± 5
 Absolute change−2 ± 5−1 ± 4
 Percentage change−16−7
Tender joint count (range 0–28 joints)  
 Baseline9 ± 79 ± 7
 Absolute change−3 ± 7−4 ± 6
 Percentage change−34−40
Patient's global disease assessment  
 Baseline49 ± 2554 ± 18
 Absolute change−10 ± 35−12 ± 27
 Percentage change−20−22
Physician's global disease assessment  
 Baseline46 ± 2040 ± 18
 Absolute change−6.6 ± 20−3.5 ± 24
 Percentage change−14−9
Pain (VAS scale 0–100)  
 Baseline56 ± 2155 ± 18
 Absolute change−18 ± 36−9 ± 14
 Percentage change−32−16
HAQ (scale 0–3)  
 Baseline1.2 ± 0.651.4 ± 0.47
 Absolute change0.14 ± 1.0−0.05 ± 0.44
 Percentage change12−4

The frequency of intraarticular steroid injections was greater in the dose escalation group, both with respect to the number of individual patients receiving an injection (10 patients versus 16 patients in the control group) and with respect to the total number of injections administered (12 injections versus 20 injections in the control group). No injections were administered within 6 weeks of the final assessment, in accordance with the protocol.

Treatment discontinuations.

Four of 27 patients (15%) discontinued therapy in the control group; 1 of these was due to a serious adverse reaction (rise in liver enzyme levels >3 times the upper limit of normal while receiving MTX 15 mg/week) and 3 were due to inefficacy of the treatment. Five of 27 patients (18.5%) discontinued therapy in the intervention group; 1 of these was due to a serious adverse reaction (recurrent chest infections while receiving MTX 20 mg/week) and 4 were due to inefficacy of the treatment. No clinically significant hematologic toxic effects were experienced and there was no discernable trend in the hematologic parameters to suggest that toxicity may have been an issue. Minor abnormalities in the liver enzyme levels (>2 but <3 times the upper limit of normal) were documented in 1 patient in each group. Hair loss and dizziness were the only minor adverse reactions to be more frequent in the dose escalation group (Table 4).

Table 4. Reasons for withdrawals, and major and minor adverse reactions*
 Controls (n = 27)MTX escalation (n = 27)
  • *

    MTX = methotrexate; LFTs = liver function test results; WCC = white blood cell count.

 Serious adverse reaction11
Serious adverse reactions  
 LFTs >3 times normal range10
 WCC <4 × 109/liter00
 Platelets <100 × 109/liter00
 Creatinine >50% rise00
 Recurrent chest infections01
Minor adverse reactions  
 Oral ulcers44
 Hair loss04
 LFTs >2 times normal range11
 Creatinine >30% rise20


Despite worldwide clinical experience, there are limited data on the efficacy of doses of MTX above the current conventional, somewhat arbitrary, doses of 20–25 mg/week in patients with active RA. The pharmacokinetics of MTX and its likely mechanisms of action in RA are well described (17, 18). Absorption of oral doses of MTX shows considerable variation; even at moderate doses of 15–20 mg/week, absorption varies considerably between doses and individuals (19). The bioavailability of 10 mg/m2 (∼17.5 mg), relative to intravenous administration, is reported to range from 0.25 to 1.49, with a mean of 0.70 (20). At higher doses, the intestinal folate transport system saturates and the parenteral route is required to achieve higher blood levels (19). Although a direct relationship between the MTX dose and clinical response has been well established in RA over the dose range of 7.5–25 mg/week (8, 21), the efficacy of dose escalation in patients achieving a partial response to oral MTX has not been formally examined.

Two open, uncontrolled pilot studies involving patients who had achieved a partial response to conventional doses of oral MTX examined the efficacy of higher doses of intravenous MTX (40 mg/m2 and 500 mg/m2), and the results suggested some improvement in disease control (9, 10). Another placebo-controlled trial examined the efficacy and safety of oral MTX up to 20 mg/m2 in 52 patients who had previously failed treatment with either D-penicillamine or intramuscular gold but had never been exposed to MTX (8). Although improvement in disease control was noted, this might not have been solely attributable to the MTX, since the intervention included a 2-week period of inpatient care, which is in itself a very effective means of suppressing disease activity in RA. The patients in that study had not been previously exposed to MTX, the intervention group received oral rather than parenteral MTX, and the analysis of efficacy was limited to patients receiving MTX up to 22.5 mg/week. Thus, although the findings of these earlier studies suggest improved disease control with higher doses of MTX, the study designs were such that the efficacy of dose escalation was uncertain.

Our study population consisted of patients with active RA whose mean duration of disease was 9–10 years. Their previous exposure to DMARDs was between 2 and 3 drugs, and they had been receiving MTX for a median of 2.5 years at a stable dose of 15–20 mg/week for at least 2 months prior to randomization. Eighty-five percent of the patients were seropositive for rheumatoid factor and 88% had erosive disease. In each of these key features, this patient population was similar to that of the Anti–Tumor Necrosis Factor Trial in RA with Concomitant Therapy (22) and Anti–Tumor Necrosis Factor Research Study Program of the Monoclonal Antibody Adalimumab in RA (23) studies, which evaluated the addition of biologic therapies to concomitant MTX; our cohort was also similar to that of a recent study in which leflunomide was added to the MTX regimen (24). In contrast to the clinical benefits reported in these previous studies, we found no evidence that increasing the dose of IM MTX beyond 15 mg/week improved disease control, although a few patients in our study did respond to a switch from oral to parenteral administration of MTX. Thus, dose escalation of MTX is not a viable alternative treatment strategy as compared with the addition of either a conventional or biologic drug in such patients.

The strengths of this study are the close matching between the 2 groups, the low dropout rate, the careful attention to blinding of patients and observers, and the maintenance of the blinded protocol until study completion. Also, by switching all patients to parenteral MTX for a period of 6 weeks prior to randomization, it was ensured that poor disease control was not simply due to poor absorption of the drug. Use of parenteral MTX during the randomization phase ensured >95% bioavailability of MTX and 100% patient adherence. Greater numbers of withdrawals from both groups might have been expected as a result of the failure to control disease activity and the burden of weekly travel for outpatient injections. Prompt management of side effects together with patient anticipation of the benefits of dose escalation may partly explain the low discontinuation rate. The low rate of serious adverse reactions may also have influenced the rate of withdrawal.

There were no hematologic toxic effects, and abnormalities in renal or liver function were also uncommon. The trial could be criticized for its relatively small size. However, the 95% confidence limits, calculated using Newcombe's method, for the differences between the groups in the percentage achieving the primary outcome (a DAS28 <3.2) or achieving an improvement in the DAS28 of >1.2, show that the maximum clinical benefit that this trial might have missed is below 20%. The 95% confidence limits for the differences in the mean DAS28 (−0.78 to +0.48) also suggest that any difference between the groups is smaller than 1.2 DAS28 units. These data suggest that a larger, more highly powered trial is probably not justified. The lack of any appreciable benefit of escalating the dose of IM MTX was surprising and, having reviewed the methods and results, remains unexplained. It is possible that the effectiveness of MTX on inflammatory pathways is maximal at these doses; this contrasts with its use in oncology in which cytotoxic effects are sought (17).

In conclusion, in some patients who have active RA despite receiving up to 20 mg oral MTX, switching to parenteral MTX may improve disease activity. However, for the majority of patients who do not respond to switching to the parenteral route, there is no benefit in escalating the dose and an alternative strategy should be tried. It is important to note that the conclusions of our study should not be generalized to all patients with active RA. The response to dose escalation in patients with early disease or who are DMARD naive was not examined and may be different. Future studies should focus on comparison of the clinical and economic benefits of combining MTX with either conventional DMARDs or biologic agents (6).


We thank Fauldings Pharmaceuticals for the gratis supply of MTX and placebo, the Wellcome Trust Clinical Research Facility for extending the use of their staff and facilities, and our colleagues in the Rheumatic Diseases Unit for facilitating inclusion of their patients.