G. Poór for the Leflunomide Multinational Study Group and V. Strand 1

Rheumatology 2004;43:744 749 Advance Access publication 16 March 2004 Efficacy and safety of leflunomide 10 mg versus 20 mg once daily in patients with active rheumatoid arthritis: multinational double-blind, randomized trial G. Poór for the Leflunomide Multinational Study Group and V. Strand 1 doi:10.1093/rheumatology/keh168 Objective. To compare the efficacy and safety profile of two daily maintenance doses of leflunomide, 10 mg and 20 mg, for the treatment of active rheumatoid arthritis (RA). Methods. In this multinational, randomized, double-blind, parallel-group study, 402 RA patients were randomized equally to receive daily doses of (n 5 202; loading dose on day 3, 100 mg) or (n 5 200; loading dose on day 1 3, 100 mg) for 24 weeks. The study was designed to demonstrate non-inferiority of the efficacy of 10 mg compared with 20 mg by calculating 95% confidence intervals for differences in changes in tender joint count (TJC), swollen joint count (SJC) and Health Assessment Questionnaire Disability Index (HAQ DI), comparing these confidence intervals with predefined bounds. Results. In the intent-to-treat population, mean improvements at the end-point in the 10 and 20 mg groups respectively were: TJC, 7.57 and 8.89 (P 5 0.061); SJC, 6.38 and 6.96 (P 5 0.304); and HAQ DI, 0 0.37 and 0 0.49 (P 5 0.095). By American College of Rheumatology (ACR) 20% criteria, response rates were 49.8 and 56.6% respectively (P 5 0.1724). Adverse events (AEs) resulting in treatment withdrawal were higher in the 10 mg (15.3%) than in the 20 mg treatment group (12.0%), as were serious adverse events (SAEs): 12.9 vs 10.0%. Conclusions. This study rejected the hypothesis of non-inferiority of 10 mg compared with 20 mg daily maintenance doses of leflunomide. More AEs resulting in treatment discontinuation and SAEs in patients receiving daily also support a better efficacy profile for the 20 mg daily dose. KEY WORDS: Leflunomide, Randomized clinical trial, Rheumatoid arthritis, Non-inferiority study. In the last decade, the standard approach to the therapy of rheumatoid arthritis (RA) has been substantially revised. The benefit of early intervention to prevent disease progression and joint damage is well established and emphasized in current guidelines [1]. Disease-modifying anti-rheumatic drugs (DMARDs) are used in the very early course of RA, as their benefit/risk profile has been shown to compare favourably with the risk of the disease itself [2], and even against conventional nonsteroidal anti-rheumatic drugs (NSAIDs) [3]. However, long-term maintenance of patients on DMARDs has been proved problematic, as the response to individual drugs may taper off or efficacy be compromised by loss of tolerability [4]. Thus, flexible dose regimens might offer benefit for patients. The isoxazole derivative leflunomide has a unique mechanism of action, differing from other DMARDs, and was introduced for the treatment of adult RA in 1998 in the USA [5]. Leflunomide is a prodrug, which is rapidly converted to its active metabolite A771726 in the gut wall, plasma and liver. It has both antiinflammatory and immunomodulatory properties by inhibiting the de novo synthesis of pyrimidine nucleotides in immune response cells [6, 7]. Three 6- and 12-month randomized placebo-controlled studies demonstrated the efficacy and safety of leflunomide (20 mg per day) and were used as pivotal studies for registration [8 11]. The phase II and III clinical trial programme included comparisons of leflunomide 20 mg per day with methotrexate and sulphasalazine and leflunomide 10 20 mg per day in combination with methotrexate [11] or in combination with sulphasalazine, respectively [9]. In these studies the 20 mg dose was established as the optimal dose for the majority of RA patients. However, if dosing at 20 mg per day was not well tolerated, reduction to 10 mg daily was permitted. This occurred in 10% of all patients in the phase III trials. In an effort to determine the lowest effective daily maintenance dose, daily doses of 5, 10 and 25 mg of leflunomide were studied in a placebo-controlled, dose-finding phase II protocol in 402 patients with active RA [12]. The efficacy of the 5 mg dose was not different from placebo. The 10 and 25 mg doses were equivalent and statistically significantly superior to placebo by American College of Rheumatology 20% (ACR20) criteria. To establish the daily maintenance dose providing optimal control of disease activity and least treatment-associated toxicity, the present study was designed to compare efficacy and safety of daily maintenance doses of leflunomide 10 mg vs 20 mg. Methods Subjects Male or female patients, aged between 18 and 75 yr, were eligible for enrolment. All subjects recruited for this were required to have a diagnosis of RA according to ACR criteria [13] and were included if they had active disease by six or more tender joints, six or more swollen joints, investigator and patient global assessments National Institute of Rheumatology and Physiotherapy, Budapest, Hungary and 1 Division of Immunology, Stanford University, Palo Alto, CA, USA. Submitted 10 July 2003; revised version accepted 4 February 2004. Correspondence to: G. Poór, National Institute of Rheumatology and Physiotherapy, Frankel Leo u. 38 40, 1023 Budapest, Hungary. E-mail: orfireum@axelero.hu 744 Rheumatology Vol. 43 No. 6 ß British Society for Rheumatology 2004; all rights reserved

Leflunomide 10 vs 20 mg non-inferiority study 745 of not better than fair and either CRP >2.0 mg/dl or ESR >28 mm/h. Daily doses of NSAIDs and oral corticosteroids (10 mg prednisolone or the corticosteroid equivalent) could not be changed for at least 4 weeks before entering the study (visit 1). Previous therapy at any time with leflunomide was not allowed. Oral corticosteroids exceeding a prednisolone equivalent of 10 mg/day, parenteral or intra-articular glucocorticoid injection or therapy with any other DMARD were permitted only if discontinued at least 4 weeks before inclusion in the study. Subjects with chronic medical conditions which, in the investigator s opinion, would have put the patient at risk to participate in the study were excluded. Women of childbearing age were required to use adequate contraception and were monitored monthly. Male patients consented to practice contraception during the study. Study design This was a multinational, randomized, double-blind, parallelgroup study to compare the efficacy/safety profile of 10 vs 20 mg daily maintenance doses of leflunomide, performed in 62 centres in 12 countries in Europe. It was approved by the institutional review boards of each participating institution, and written informed consent was obtained from all subjects in the study. A sequence of four-digit randomization numbers was assigned to each centre and study medication labelled accordingly. At the baseline visit, eligible subjects were randomized to receive daily doses of 10 or 20 mg of leflunomide in a ratio of 1:1 in blocks of six subjects. Following a 1-week screening phase (visits 1 and 0), eligible subjects received 24-week double-blind treatment (visits 1 5 at weeks 2, 4, 8, 16 and 22), followed by a 4-week, treatmentfree observation phase (visit 6). Haematology values were monitored every 2 weeks and alanine aminotransferase (ALT) values every 4 weeks. Additional visits for blood sampling were therefore scheduled at weeks 6, 10, 12 and 14 and weeks 18, 20 and 22. Patients in the 10 mg group received a 100 mg loading dose on days 1, 2 and 3 followed by one 10 mg tablet/day from day 4 onwards; those in the 20 mg group received 100 mg on days 1, 2 and 3 followed by one 20 mg tablet/day from day 4 onwards. Dose adjustments were not allowed during the 24 weeks of protocol treatment. Subjects interrupting study medication for more than 15 days were withdrawn from the study; interruptions exceeding a total of 7 days were not allowed during the last 4 weeks of protocol treatment. Efficacy and safety measures The primary efficacy end-points included tender joint count (TJC) and swollen joint count (SJC) based on a 28-joint count, and physical function assessed by the Health Assessment Questionnaire Disability Index (HAQ DI) [14, 15] after 6 months of treatment. Secondary efficacy variables included ACR 20%, ACR 50% and ACR 70% response rates [16]. To be considered an ACR 20% responder, a subject had to have a 20% improvement from baseline to end-point in both TJC and SJC and three or more of the following five measures: investigator global assessment of disease activity, patient global assessment of disease activity, pain intensity, HAQ DI and CRP. ESR was to be used for subjects with missing CRP values at baseline or end-point. Physician and patient global assessments of disease activity used a Likert five-point scale from 1 (very good) to 5 (very poor). Investigator global assessments were to be scored independently from patient assessments. Patient assessments of pain intensity used a 100 mm visual analogue scale. Patients were also queried about the duration of morning stiffness. Baseline daily corticosteroid doses were calculated based on the average dose over 4 weeks prior to intake of study medication (baseline dosage), and compared with the average daily dose over the last 4 weeks of study medication administration (end-point dosage). Safety assessments included complete physical examinations, including body weight data at visits 1 and 5, reports of adverse events (AEs) and serious adverse events (SAEs), monitoring of laboratory parameters and blood pressure at visits 1 to 5. Laboratory tests included haematology measurements, blood chemistry and pregnancy tests and urinalyses. Compliance was assessed by tablet counts and query of patients regarding study medication intake. Statistical analysis The primary efficacy analysis was designed to test three variables [TJC and SJV, HAQ disability index (DI)] for non-inferiority of 10 vs 20 mg daily doses of leflunomide using 95% confidence intervals (CIs). The 95% CIs were calculated based on differences in adjusted mean changes between treatment groups according to analyses of covariance (ANCOVA), with treatment, treatment centres and disease duration as factors in the model and with baseline values as covariates. Results were pooled across study centres by region. Six pools were defined by country and region for ANOVA analysis as follows: Pool 1, Czech Republic (centres 1 4; 86 90), Slovakia; Pool 2, Germany, Switzerland (centres 16 29, 81 84); Pool 3, Denmark, Finland, Sweden (centres 6 14, 31 35, 71 75); Pool 4, Hungary (centres 41 44); Pool 5, France, Italy (centres 36 40, 51 60); Pool 6, Ireland, Netherlands (centres 46 47, 61 67). Efficacy results are expressed as mean±s.d. unless otherwise stated. The 10 mg dose was defined to be not inferior to 20 mg daily if the lower bound of the 95% CI for the difference between treatment groups was fully to the right of 2 for both TJC and SJC and fully to the right of 0.2 points for HAQ DI. As the primary objective was to compare the overall efficacy/safety profile of 10 vs 20 mg daily doses of leflunomide, sample sizes were calculated to ensure that 400 patients would be evaluable for safety analysis. When a non-inferiority comparison is sought, definition of a clinically meaningful difference,, between treatment groups is of great importance. The limit for non-inferiority was defined not to exceed 50% of the known superiority between leflunomide and placebo, based on published randomized controlled trials. Based on data from a 6-month phase III trial comparing leflunomide and sulphasalazine with placebo in a comparable patient population [9], as well as a phase II trial comparing two efficacious doses of leflunomide with placebo, this corresponded to differences of 2 in both TJC and SJC and > 0.22 in HAQ DI. Differences in HAQ DI correspond to published definitions of minimum clinically important differences [17]. For each primary efficacy variable, a power of at least 80, 92 and 99% respectively was calculated to demonstrate non-inferiority. Given the strong correlation between these predefined efficacy variables, assessing all three simultaneously was expected to yield an overall power exceeding 71%. The intention to treat (ITT) population consisted of all randomized subjects who received at least one dose of study medication. The primary efficacy analyses were based on the ITT population. Analyses of secondary efficacy variables of RA (pain intensity assessment, CRP, ESR, RF, and morning stiffness) are presented as mean changes from baseline to end-point (using the same ANCOVA model as for the primary variables) by treatment group, with 95% CI for differences between treatment groups (adjusted mean changes). Changes in global assessments were analysed using the Mantel Haenszel test with stratification by treatment group. ACR responder status used the ITT population in a last observation carried forward analysis with the 2 test and logistic regression analysis including treatment, treatment group, and duration of RA (2 yr, >2 yr) as factors. Changes in daily doses of

746 G. Poo r and V. Strand NSAIDs and corticosteroids between baseline and end-point due to RA were analysed using the Wilcoxon Mann Whitney test. Results Patient characteristics and disposition Of 464 subjects screened for eligibility, 402 (86.6%) were randomized to receive protocol treatment. In 202 patients randomized to receive 10 mg, 50 (24.8%) were withdrawn prematurely compared with 36 (18.0%) of 200 randomized to receive 20 mg. Figure 1 presents the treatment disposition and reasons for drop-outs. Drop-outs were considered non-responders if response criteria were not fulfilled for the respective outcome measure at the last available assessment of efficacy. At baseline, demographic and disease characteristics were not different between treatment groups (Table 1). The majority were women (83%); mean disease duration was 9.6 yr [2 yr in 14 (3.5%), 2 10 yr in 28 (7.0%) and >10 yr in 351 (87.3%)]; all had failed a mean of three DMARDs. Efficacy Baseline values of the three primary variables were similar in both treatment groups. At the end-point, reductions in TJC were 7.57 in the 10 mg group and 8.89 in the 20 mg group (P ¼ 0.0617); reductions in SJC were 6.38 in the 10 mg group and 6.96 in the 20 mg group (P ¼ 0.3041), and improvement in HAQ DI was 0.37 in the 10 mg group and 0.49 in the 20 mg group (P ¼ 0.0946). By predefined statistical analyses, criteria for non-inferiority were satisfied only by changes in SJC, not in TJC or HAQ DI. Therefore, non-inferiority of the 10 mg daily dose of leflunomide compared with 20 mg was not demonstrated. Table 2 summarizes mean changes from baseline in primary and secondary efficacy variables for both treatment groups. With the exception of ESR, all secondary outcome measures demonstrated numerically better improvement in patients receiving 20 mg compared with 10 mg daily doses of leflunomide. ACR 20% response rates were 49.8 and 56.6% in the 10 mg and 20 mg dose groups, respectively. As observed in the phase III trials, ACR 20% responses approached maximum levels after 3 months of protocol treatment in both groups, while ACR 50% and ACR 70% responses continued to improve during this 6-month study. Maximal decreases in CRP and ESR values were evident after 4 weeks of active treatment. The time courses of all efficacy variables (TJC, SJC, HAQ DI, ACR 50%, ACR 70% responses, investigator and patient global assessments, duration of morning stiffness and pain assessments) paralleled ACR 20% responses. Approximately 60% of patients were receiving concomitant corticosteroid treatment, in doses less than or equal to the equivalent daily dose of prednisone. In the treatment group receiving daily, 22.3% were prescribed higher daily corticosteroid doses compared with 6.6% in the 20 mg treatment group. Correspondingly, 26.2% of patients in the 20 mg dose group discontinued corticosteroid treatment, compared with 26.2% in the 10 mg dose group; this difference was statistically significant (P ¼ 0.0026). Changes in concomitant corticosteroid treatment between baseline and end-point are presented in Table 3. Safety Adverse events. AEs were reported in 91 (45.0%) receiving 10 mg and in 101 (50.5%) receiving 20 mg/day leflunomide. Treatment-emergent AEs are presented in Table 4. More AEs Screened Screened 464 464 Withdrawn during screening 62 Randomized 402 402 10 10 mg mg leflunomide 202 202 20 20 mg mg leflunomide 200 200 Withdrawn 50 Adverse event 32 Lack of efficacy 11 Other 7 Withdrawn 36 Adverse event 24 Lack of efficacy 6 Other 6 Completed treatment phase phase 152* Completed treatment phase phase 164 164 FIG. 1. Study flow chart. Explanation for withdrawals summarized as Other in the 10 mg group: one subject was discontinued for poor compliance; one subject did not wish to continue; one subject had a protocol deviation; three subjects did not meet inclusion criteria; one subject unknown. Explanation for withdrawals summarized as Other in the 20 mg group: one subject was discontinued for poor compliance; two were lost to follow up; two had a protocol deviation; one did not meet inclusion criteria.

Leflunomide 10 vs 20 mg non-inferiority study 747 TABLE 1. Disease characteristics and demographics No. (%) subjects in treatment group (n ¼ 202) (n ¼ 200) Total (n ¼ 402) Women (n, %) 163 80.7% 172 86.0% 335 83.3% Mean age (yr) 55.6±11.9 55.4±11.1 55.5±11.5 Mean no. of DMARDs failed 3.1±1.7 2.9±1.8 3.0±1.8 Mean disease duration (months) 111.4±108.7 118.8±117.54 115.1±113.1 Patients on steroids at baseline (n, %) 121 59.9% 122 61.0% 243 60.4% Patients on NSAIDs at baseline (n, %) 146 72.3% 131 65.5% 277 68.9% ACR functional class I(n, %) 16 7.9% 13 6.5% 29 7.2% II (n, %) 102 50.5% 97 48.5% 199 49.5% III (n, %) 84 41.6% 89 44.5% 173 43.0% IV (n, %) 0 1 0.5% 1 0.2% TABLE 2. Changes in primary and secondary clinical efficacy end-points in the ITT population Mean (S.D.) Comparison (20 10 mg) Variable/statistic* n ¼ 200) (n ¼ 197) P # 95% CI y TJC 7.57 (7.92) 8.89 (8.0) 0.0617 2.69, 0.06 SJC 6.38 (6.22) 6.96 (6.46) 0.3041 1.67, 0.52 HAQ DI** 0.37 (0.61) 0.49 (0.72) 0.0946 0.24, 0.02 ACR20 100/201 (50) 112/198 (57) 0.1724 2.96, 16.58 ACR50 41/201 (20) (51/198) (26) 0.2035 2.89, 13.61 ACR70 (14/201) (7) (19/198) (10) 0.3394 2.77, 8.04 The following clinically meaningful differences between the treatment groups were defined for the three primary end-points: TJC, 2 joints; SJV, 2 joints; HAQ DI, 0.2 points. Numbers in brackets display the baseline minus end-point comparison between treatment groups. *Changes were estimated from the ANCOVA model adjusted for investigator pool and duration of RA. **, n ¼ 147;, n ¼ 144. y Lower and upper bounds of the 95% CIs for the difference in mean changes between treatment groups. # P values were derived from the 2 test for differences in responder rates (24-week end-point) between treatment groups. TABLE 3. Changes in concomitant NSAIDs and corticosteroid medication between baseline and end-point 10 mg leflunomide 20 mg leflunomide P value for comparison* n (%) n (%) (20 10 mg) Total subjects with 121 (100.0) 122 (100.0) corticosteroids Increase 27 (22.3) 8 (6.6) No change 72 (59.5) 82 (67.2) 0.0026 Decrease or stopped 22 (18.2) 32 (26.2) *P values are derived from the Wilcoxon Mann Whitney test for differences in changes in NSAID/corticosteroid consumption between treatment groups. resulting in treatment discontinuation were reported in patients receiving 10 compared with 20 mg daily: 31 patients (15.3%) vs 24 patients (12.0%), including worsening RA, anaemia, diarrhoea and rash. AEs resulting in protocol withdrawal are presented in Table 5. SAEs were reported more frequently in patients receiving 10 vs 20 mg daily doses of leflunomide. SAEs reported in more than two subjects in each treatment group included active RA [eight subjects (4.0%) receiving 10 mg vs five subjects (2.5%) receiving 20 mg] and anaemia [one subject (0.5%) in the 10 mg group vs four subjects (2.0%) in the 20 mg group]. Laboratory examinations, blood pressure and body weight. The majority of subjects had normal liver enzyme levels at both baseline and end-point. There were no differences between treatment groups in the number of clinically noteworthy elevations in liver enzymes observed. Treatment-associated changes in ALT and aspartate aminotransferase (AST) values for each treatment group are presented in Table 6. Mean systolic/diastolic blood pressure increased between baseline and the study end-point by 1.5/1.6 mmhg in the 10 mg group and 5.0/2.2 mmhg in the 20 mg group. Treatment-associated mean decreases in weight of 0.7±3.05 kg in patients receiving 10 mg daily and 1.0±2.96 kg in the 20 mg group were observed. Discussion The present study offers important information regarding the efficacy and safety profile of leflunomide for the treatment of patients with active RA, comparing for the first time the recommended 20 mg with 10 mg daily maintenance dose. In clinical practice, if the 20 mg dose is not well tolerated, dose reduction to 10 mg daily is recommended. Methodology This comparison was performed using a non-inferiority trial design [18]. To demonstrate that a new or experimental therapy is at least as good as standard therapy in a randomized controlled trial, statistical testing of treatment results must rule out clinical

748 G. Poo r and V. Strand TABLE 4. Treatment-emergent AEs regardless of relatedness occurring in at least 5% of subjects in each treatment group No. (%) subjects with treatment-emergent AEs AE (n ¼ 202) (n ¼ 200) Total (n ¼ 402) Total subjects with events 91 (45.0) 101 (50.5) 192 (47.8) Diarrhoea 18 (8.9) 24 (12.0) 42 (10.4) Alopecia 13 (6.4) 20 (10.0) 33 (8.2) Nausea 9 (4.5) 18 (9.0) 27 (6.7) Hypertension 12 (5.9) 10 (5.0) 22 (5.5) TABLE 5. AEs leading to discontinuation of study medication No. (%) subjects with treatment-emergent AEs AE (n ¼ 202) (n ¼ 200) Total (n ¼ 402) Total subjects with events leading 31 (15.3) 24 (12.0) 55 (13.7) to discontinuation RA 9 (4.5) 3 (1.5) 12 (3.0) Anaemia 1 (0.5) 3 (1.5) 4 (1.0) Diarrhoea 1 (0.5) 3 (1.5) 4 (1.0) Rash 0 3 (1.5) 3 (0.7) TABLE 6. Summary of transaminase elevations (% of patients) Value at end-point Liver test Group Maximum value in study n 1.2 ULN 1.2 to 2 ULN 2to 3 ULN >3 ULN AST 10 mg 1.2 to 2 ULN 8 3 5 2to3 ULN 3 2 1 >3 ULN 20 mg 1.2 to 2 ULN 14 10 4 2to3 ULN 4 1 1 2 >3 ULN 2 1 1 ALT 10 mg 1.2 to 2 ULN 17 11 6 2to3 ULN 1 1 >3 ULN 4 2 2 20 mg 1.2 to 2 ULN 17 13 4 2to3 ULN 6 1 1 4 >3 ULN 4 3 1 SGGT 10 mg 1.2 to 2 ULN 13 5 8 2to3 ULN 3 3 >3 ULN 2 1 1 20 mg 1.2 to 2 ULN 19 11 8 2to3 ULN 4 1 2 1 >3 ULN 7 3 1 3 SGGT ¼ serum -glutamyl transferase. inferiority with a high probability [19]. Appropriate interpretation of results derived from this trial design requires availability of data from previous randomized controlled trials to establish historical evidence of clinical efficacy and a positive risk/benefit profile. If possible, efforts should be made to assure that subsequent study populations are, in fact, similar to those enrolled in previous trials which initially established efficacy of the active treatment [20]. These criteria were closely observed and checked in the present study, which is indeed very similar in design and study population to the comparative study of Mladenovic et al. [12]. Efficacy Both leflunomide treatment regimens had benefit on the primary efficacy end-points, reducing TJC and SJC and improving physical function, respectively. However, non-inferiority criteria were satisfied only for SJC and therefore non-inferiority of the 10 mg compared with the 20 mg daily dose was rejected. The primary efficacy results were supported by the analysis of the secondary end-points, ACR criteria and all components except ESR. A predefined and clinically relevant analysis demonstrated a difference favouring the 20 mg dose in the number of patients with reduced or discontinued concomitant steroid consumption, which had been shown in previous studies but was statistically not different from methotrexate. However, baseline corticosteroid doses varied between the 10 and 20 mg treatment groups, which may have resulted in longitudinal bias. There were more treatment discontinuations in the 10 mg group due to withdrawals resulting from AEs (15.8 vs 12.0%) and lack of efficacy (5.4 vs 3.0%). AEs attributed to active RA were the most frequent reason for discontinuing study medication in the 10 mg treatment group, which can also be interpreted as a lack of efficacy.

Leflunomide 10 vs 20 mg non-inferiority study 749 Reported improvements in primary and secondary efficacy endpoints in this trial are in agreement with phase II and III trials following 6 [9, 12], 12 [10] and 24 months [11, 21] of treatment with leflunomide. Remarkably, the leflunomide treatment effect reflected by the ACR 20% response in the 20 mg dose group in this trial, 57%, was similar to those reported in the phase III trials: 55% [9], 52% [10] and 51% [11]. This strongly supports the validity and reproducibility of the anti-rheumatic efficacy of this daily dose. The 6-month results of the placebo-controlled comparison of leflunomide with sulphasalazine performed by Smolen et al. [9] are comparable with the data from this study: improvements in tender and swollen joint counts in the leflunomide treatment group ( 9.7, 7.2 and ACR 20 response 55%, respectively) are in agreement with the results of the present study ( 8.9, 7.0, ACR 20 response 56.6%, respectively). Results following 12 and 24 months of treatment with reported in studies performed by Emery et al. [11], Scott et al. [21], Strand et al. [10] and Cohen et al. [22] confirm comparable efficacy. Safety The safety profile of leflunomide in this study was comparable with that reported in other studies [23]. The percentages of patients with AEs were similar in the 10 and 20 mg groups (66 vs 68%) but were lower than reported in a previously published phase III study with similar patient numbers [9]: 84, 79 and 84% of patients with treatment-emergent AEs received leflunomide, placebo and sulphasalazine, respectively. In the 20 mg group in the present study the frequencies of the most commonly reported AEs [diarrhoea (16%), nausea (10%), alopecia (10%) and headache (6%)] were lower than in the other phase III trials [9]. The frequency of clinically noteworthy liver enzyme elevations [>3 times the upper limit of normal range (ULN): 0 and 0.5% for AST and 1 and 0% for ALT in the 10 and 20 mg groups, respectively] were lower than in the studies performed by Strand et al. [10] (4.4% ALT >3 ULN) and Emery et al. [11] (2.4% ALT >3 ULN). In conclusion, this non-inferiority study confirms the recommendations of the present product labelling with a recommended dose of 10 20 mg once daily. In clinical practice, if the 20 mg dose is not well tolerated, dose reduction to 10 mg daily can be considered. Rheumatology Acknowledgements This study was funded by Aventis Pharma, Paris. We thank all participating centres and their staff for their cooperation and commitment, without which the study would not have been possible. VS is a consultant in clinical and regulatory affairs to many sponsors, and has served in this capacity with Aventis. She holds no stock. References Key messages A daily maintenance dose of 20 mg leflunomide appears to be more effective than. is safe and recommended for the treatment of rheumatoid arthritis. 1. American College of Rheumatology ad hoc committee on clinical guidelines. Guidelines for the management of rheumatoid arthritis. Arthritis Rheum 1996;1996:713 22. 2. Pincus T, Callahan L. What is the natural history of rheumatoid arthritis? Rheum Dis Clin North Am 1993;19:123 51. 3. Singh G. Recent considerations in nonsteroidal anti-inflammatory drug gastropathy. Am J Med 1998;105:31S 38S. 4. Felson DT, Anderson JJ, Meenan RF. Use of short-term efficacy/ toxicity tradeoffs to select second-line drugs in rheumatoid arthritis. A metaanalysis of published clinical trials. Arthritis Rheum 1992;35: 1117 25. 5. Prakash A, Jarvis B. Leflunomide: a review of its use in active rheumatoid arthritis. Drugs 1999;58:1137 64. 6. Alldred A, Emery P. Leflunomide: a novel DMARD for the treatment of rheumatoid arthritis. Expert Opin Pharmacother 2001;2:125 37. 7. Fox RI. Mechanism of action of leflunomide in rheumatoid arthritis. J Rheumatol Suppl 1998;53:20 6. 8. Sharp JT, Strand V, Leung H, Hurley F, Loew-Friedrich I. Treatment with leflunomide slows radiographic progression of rheumatoid arthritis: results from three randomized controlled trials of leflunomide in patients with active rheumatoid arthritis. Leflunomide Rheumatoid Arthritis Investigators Group. Arthritis Rheum 2000; 43:495 505. 9. Smolen JS, Kalden JR, Scott DL et al. Efficacy and safety of leflunomide compared with placebo and sulphasalazine in active rheumatoid arthritis: a double-blind, randomised, multicenter trial. European Leflunomide Study Group. Lancet 1999;353:259 66. 10. Strand V, Cohen S, Schiff M et al. Treatment of active rheumatoid arthritis with leflunomide compared with placebo and methotrexate. Leflunomide Rheumatoid Arthritis Investigators Group. Arch Intern Med 1999;159:2542 50. 11. Emery P, Breedveld FC, Lemmel EM et al. A comparison of the efficacy and safety of leflunomide and methotrexate for the treatment of rheumatoid arthritis. Rheumatology 2000;39:655 65. 12. Mladenovic V, Domljan Z, Rozman B et al. Safety and effectiveness of leflunomide in the treatment of patients with active rheumatoid arthritis. Results of a randomized, placebo-controlled, phase II study. Arthritis Rheum 1995;38:1595 603. 13. Arnett FC, Edworthy SM, Bloch DA et al. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 1988;31:315 24. 14. Smolen JS, Breedveld FC, Eberl G et al. Validity and reliability of the twenty-eight-joint count for the assessment of rheumatoid arthritis activity. Arthritis Rheum 1995;38:38 43. 15. Fries JF, Spitz P, Kraines RG, Holman HR. Measurement of patient outcome in arthritis. Arthritis Rheum 1980;23:137 45. 16. Felson D, Anderson J, Boers M et al. American College of Rheumatology. Preliminary definition of improvement in rheumatoid arthritis. Arthritis Rheum 1995;38:727 35. 17. Ortiz Z, Shea B, Garcia Dieguez M et al. The responsiveness of generic quality of life instruments in rheumatic diseases. A systematic review of randomized controlled trials. J Rheumatol 1999;26:210 6. 18. Committee for Proprietary Medicinal Products. Points to consider on switching between superiority and non-inferiority. Br J Clin Pharmacol 2001;52:223 8. 19. Laster LL, Johnson MF. Non-inferiority trials: the at least as good as criterion. Stat Med 2003;22:187 200. 20. ICH Expert Working Group. ICH harmonised tripartite guideline. Choice of control group and related issues in clinical trials. Federal Register, 2000;66:24390 1. 21. Scott DL, Smolen JS, Kalden JR et al. Treatment of active rheumatoid arthritis with leflunomide: two year follow up of a double blind, placebo controlled trial versus sulfasalazine. Ann Rheum Dis 2001;60:913 23. 22. Cohen S, Cannon GW, Schiff M et al. Two-year, blinded, randomized, controlled trial of treatment of active rheumatoid arthritis with leflunomide compared with methotrexate. Utilization of Leflunomide in the Treatment of Rheumatoid Arthritis Trial Investigator Group. Arthritis Rheum 2001;44:1984 92. 23. Osiri M, Shea B, Robinson V et al. Leflunomide for treating rheumatoid arthritis (Cochrane Review). Cochrane Database Syst Rev 2003:CD002047. [13.5.2004 6:55am] [REVISE PROOFS] [744 749] [Page No. 749] F:/OUP UK/RHE (ke)/rhe43-6/keh168.3d (RHE) Paper: keh168