Survival and effectiveness of leflunomide compared with methotrexate and sulfasalazine in rheumatoid arthritis: a matched observational study

Abstract

Objective: To determine the survival and clinical effectiveness of leflunomide (LEF) compared with methotrexate (MTX) and sulfasalazine (SSZ) for RA in an observational study.

Methods: An observational database of 1088 patients and 5141 patient years of DMARD treatment (2680 courses) from two academic hospitals was filtered for treatment with LEF, MTX, and SSZ. LEF treatment groups were matched for patients' age, baseline ESR, number of previous DMARDs, and hospital cohort with MTX and SSZ treatment groups. For these treatments, Kaplan-Meier analyses of time until the drug was discontinued (drug "survival"), and the effectiveness and safety of continuation of treatment, were performed. The change in disease activity markers (CRP, ESR) was compared between the groups.

Results: The median dose during the study increased from 10 to 15 mg MTX/week and from 1.5 to 2.0 g SSZ/day. Matched survival analysis showed better retention rates for MTX (mean (SEM) survival 28 (1) months) than for LEF (20 (1) months; p=0.001), whereas retention rates of SSZ (23 (1) months) were similar to those of LEF (p=NS). Treatments were stopped earlier because of adverse events (AEs, 3 months) than because of ineffectiveness (IE, 10 months; p<0.001). LEF and MTX were less likely to be stopped because of AEs than SSZ. LEF courses were stopped earlier for AEs (p<0.001) than MTX.

Conclusions: Current dosing strategies should be re-evaluated, and coping strategies for common AEs should be investigated. This will be necessary to achieve better drug retention of LEF. At present, MTX continues to be the most effective drug in clinical practice.

Figure 1

Population studied, characteristics, and flow chart of analysis. The study group comprised a consecutive inception cohort of 1088 patients with RA, who had undergone 2680 DMARD courses.

Figure 2

Cumulative drug retentions (%) of LEF, MTX, and SSZ. The period to the right of the reference line (at 30 months) indicates an unduly small number of patients left at risk (<10%). (A) All treatments started after 1999, mean survivals (SE) of the drugs (months)—MTX (n=209): 26 (1); SSZ (n=70): 23 (2); LEF (n=168): 20 (1). Breslow test: p=0.03 for MTX v LEF, otherwise: p=NS. (B) Matched analysis: LEF (n=168) v MTX (n=168)—MTX: 28 (1) months; LEF: 20 (1) months. Breslow test: p=0.001. (C) Matched analysis: LEF (n=159) v SSZ (n=159)—LEF: 20 (1) months; SSZ: 23 (1) months. Breslow test: p=NS.

Figure 3

Timing of DMARD discontinuation due to (A) inefficiency (A) and (B) AEs. Only treatments with an event during the first 42 months are displayed (n). The median timing of discontinuations (for all DMARDs) due to inefficiency (A) was 10 months (quartiles: 6; 18 months); and due to adverse events (B) was: median 3 months (1; 10 months) (p<0.001, Kolmogorov-Smirnov test for equality of distributions). Between-drug comparison for timing of inefficiency (A): p=NS; and for timing of adverse events (B): p<0.001 (earlier timing of events while receiving LEF and SSZ than while receiving MTX treatment).

Figure 4

Survival of treatment effectiveness. Cumulative drug retentions (%) of LEF, MTX, and SSZ, when only discontinuation due to inefficiency was analysed, assuming permanent safety otherwise (that is, censoring at time of occurrence). The period to the right of the reference line (30 months) indicates an unduly small number of patients left at risk (<10%). Note the discontinuous y axis. (A) All treatments after 1999, mean survivals (SE) of the drugs (months)—MTX (n=209): 33 (1); SSZ (n=70): 30 (2); and LEF (n=168): 26 (1). Breslow test: p=0.04 for MTX v LEF, otherwise: p=NS. (B) Matched analysis of LEF v MTX (n=168)—LEF: 26 (1) v MTX: 32 (1); Breslow test: p=0.04. (C) Matched analysis of LEF v SSZ (n=159)—LEF: 26 (1) v SSZ: 30 (1); Breslow test: p=NS.

Figure 5

survival of treatment safety. Cumulative drug retentions (%) of LEF, MTX, and SSZ, while only discontinuation due to adverse events was analysed, assuming permanent effectiveness otherwise (that is, censoring at time of occurrence). The period to the right of the reference line (30 months) indicates an unduly small number of patients left at risk (<10%). Note the discontinuous y axis. (A) All treatments after 1999, mean survivals (SE) of the drugs (months)—MTX (n=209): 37 (1); SSZ (n=70): 36 (2); and LEF (n=168): 30 (1). Breslow test: p=0.01 for MTX v LEF, otherwise: p=NS. (B) Matched analysis of LEF v MTX (n=168)—LEF: 30 (1) v MTX: 34 (1); Breslow test: p=NS. (C) Matched analysis of LEF v SSZ (n=159)—LEF: 31 (1) v SSZ: 33 (1); Breslow test: p=NS.