Annals ofthe Rheumatic Diseases 1995; 54: 881-885 881 Survival analysis of disease modifying Ann Rheum Dis: first published as 10.1136/ard.54.11.881 on 1 November 1995. Downloaded from antirheumatic drugs in Spanish rheumatoid arthritis patients Jose De La Mata, Francisco J Blanco, Juan J Gomez-Reino Abstract control disease activity.A" Current available Objectives-To evaluate the duration of data show that disease modifying anti- treatment and the reasons for dis- rheumatic drugs (DMARDs) are useful in continuing therapy with disase modifying short term studies,"1 but they are seldom antirheumatic drugs in Spanish rheuma- continued for long periods, as a result of lack toid arthritis patients. of efficacy, or toxicity. Clinical trials have been Methods-An observational study was unable to show notable advantages of one made of 629 patients with rheumatoid DMARD over another,'2 therefore choice of a arthritis treated with disease modifying DMARD should be based on maximal efficacy antirheumatic drugs between 1979 and with minimum toxicity. In this regard, most 1991. The outcomes (treatment termination authors have stressed the need for long term because oftoxicity and lack ofresponse) of comparative studies;'3'5 however, such studies 991 treatment starts with intramuscular under 'controlled' conditions are impractical gold salts, D-penicillamine, azathioprine, because of high costs and logistic complexity. and methotrexate were subjected to sur- As an alternative, community based studies vival analysis. Cumulative probability of that analyse continuation on treatment with continuation of each drug (drug survival) different DMARDs and use long follow ups was calculated by the Kaplan-Meier and large numbers of patients are a common method and comparison between the sur- and useful approach to monitoring long term vival curve of each was made by log rank treatment of RA.1-'9 testing. To assess the long term effectiveness of Results-Median drug survival (95% con- DMARDs in our population, we conducted an fidence interval) was 51 (25-76.9) months observational study of a large sample of for methotrexate, 39 9 (19-9-48.2) months Spanish RA patients followed for 12 years in for azathioprine, 34*9 (29.4-41.4) months for a single rheumatology unit. We compared the gold salts, and 16-4 (13-9-21) months for survival (that is, continuation as treatment) http://ard.bmj.com/ D-penicillamine. The highest cumulative of most DMARDs currently used in our probability of drug survival at five years community, and analysed the reasons for their was for methotrexate (450/); that at 10 discontinuation. To simplify terminology in years was for gold salts (15%). Up to 600/o this text, phrases such as 'drug survival' are ofthe patients discontinued D-penicillamine used to imply 'cumulative probability of the in the first two years. Lack ofresponse was patient continuing to take the drug'. the major limiting factor for all drugs on October 1, 2021 by guest. Protected copyright. except D-penicillamine, for which it was toxicity. D-Penicillamine was associated Patients and methods with a greater rate of discontinuations PATIENTS AND DOSING SCHEDULES because oftoxicity in women and patients In this open observational study, we reviewed older than 65. Previous disease modifying the records of all RA patients treated con- antirheumatic drug administration did secutively in our institution with gold salts, not influence current drug survival. D-penicillamine, azathioprine, and metho- Conclusion-Overall, gold salts remain trexate from 1979 to 1991 (treatments with useful for the treatment of rheumatoid methotrexate started in 1986). By December arthritis over long periods of time in 1991, 629 patients with definite or classical RA the population studied. Because of the according to 1958 American Rheumatism Rheumatology and Research Units, high rate of continuation of treatment Association criteria received 991 treatment Hospital Universitario (survival) and the optimal efficacy and starts as follows: 535 treatments with gold Doce de Octubre, toxicity profiles observed with metho- salts, 178 with D-penicillamine, 126 with Madrid, Spain J De La Mata trexate after five years of treatment, it azathioprine, and 152 with methotrexate. F J Blanco should be the drug of first choice for Short breaks in a course ofDMARD treatment J J G6mez-Reino second line treatment ofthese RA patients. (usually four weeks or less) were counted as Correspondence to: time taking the drug; only 'new starts', not Dr J J G6mez-Reino, Rheumatology, (Ann Rheum Dis 1995; 54: 881-885) those restarts, were included in the study Hospital 'Doce de Octubre', analysis. The DMARD used was chosen Carretera de Andalucia, 28041 Madrid, Spain. Rheumatoid arthritis (RA) is a chronic pro- according to the physician's preference. When Accepted for publication gressive disease that requires long term treat- treatment was not effective or could not 25 July 1995 ment. The major concern in management is to be continued because of toxicity, another 882 De La Mata, Blanco, G6mez-Raino Table 1 Patient characteristics at the start oftreatment with disease modifying analyses, we used termination of treatment antirheumatic drugs because of toxicity or lack of response as out- Ann Rheum Dis: first published as 10.1136/ard.54.11.881 on 1 November 1995. Downloaded from GS D-PN AZA MTXS comes. Differences between drug survival curves were analysed by log rank method, and Number of starts 535 178 126 152 Observation period (patient years) 1405 360 282 272 median survival time was determined by Sexratio (F:M) 3:1 4:1 4:1 5-6:1* Kaplan-Meier analysis. Comparisons among <65 years (%/6) 65 63 60 77* groups for continuous variables were made by GS = Gold salts; D-PN = D-penicillamine; AZA = azathioprine; MTX = methotrexate. t test. Categorical variables were analysed by §MTX was started in 1986. *p < 0-05 compared with GS, D-PN, or AZA (x2)- the x2 test. Statistical significance was assumed for values of p < 0-05. DMARD was substituted, with one month free from second line medication. Results Gold salts were started at a dose of 50 mg/ PATIENTS week, decreasing to a minimum of 25 mg During the study period, DMARDs were monthly ifthe disease was controlled clinically; started 991 times in 629 RA patients (table 1). D-penicillamine was started at 250 mg daily, No differences in demographic characteristics with three monthly increments of 125 mg, to of the patients or pattern of drug use were a maximum of 750 mg daily; azathioprine was observed. More than 75% of the patients were prescribed at 75 mg daily, with monthly treated concomitantly with non-steroidal anti- increments of 25 mg to a maximum of 150 mg inflammatory drugs (NSAIDs); prednisone daily; and methotrexate was administered at was prescribed in fewer than 15% of these, 7-5 mg weekly, with increments of 2-5 mg to at doses generally less than 10 mg/day. a maximum of 15 mg weekly. No statistical differences related to gender or Clinical data for the study were obtained age were found in the use of gold salts, every three to four months during clinic visits. D-penicillamine and azathioprine, but metho- The use of DMARDs was recorded using trexate was prescribed more frequently for specific data collection forms, each of which women and for patients younger than 65 years noted date of start, maximal dose admin- (X2,p <0-05) (table 1). istered, date and reason for discontinuation, and side effects. The latter were monitored periodically for each drug. Complete blood cell SURVIVAL ANALYSIS counts and urine analysis were performed at six Figure 1A represents the cumulative prob- week intervals in patients receiving gold salts or ability of survival for the four drugs, taking D-penicillamine. Blood counts were obtained drug discontinuation as the end point. Drug monthly in patients receiving azathioprine. survival for D-pencillamine was significantly Complete blood cell counts and serum levels lower than that of the other DMARDs when of transaminases, alkaline phosphatase, bili- survival curves were compared by log rank test rubin, albumin, and creatinine were performed (p < 0-01). Median survivals indicated that http://ard.bmj.com/ every month in patients taking methotrexate. more than 50% of the patients stopped taking D-penicillamine after only 18 months of treat- ment (table 2). In contrast, 50% of patients CRITERIA FOR DISCONTINUATION OF taking methotrexate remained on the treat- TREATMENT ment after more than 50 months (table 2). Reasons for termination of treatment were After 34 months of treatment, 50% of patients classified as associated toxicity, lack of receiving gold salts or azathioprine were still response, or miscellaneous. taking the drug (fig lA). on October 1, 2021 by guest. Protected copyright. Associated toxicity comprised any adverse effect that required discontinuation of treat- ment. Results were expressed as number of REASON FOR DMARD DISCONTINUATION events per 1000 patient years of observation. We recorded 573 discontinuations after a Lack of response was defined as the absence total of 991 starts of DMARD treatment. of clinical benefit at any time of follow up after In most cases, gold salts, azathioprine and a miniimum of six months receiving treatment, methotrexate were discontinued because of or disease relapses not effectively controlled by lack of response. Among patients receiving adjustment of the dose of DMARD. D-penicillamine, there were as many dis- Miscellaneous causes included loss to follow continuations because of toxicity as there were up, lack of compliance, or death for unrelated discontinuations because of lack of response reasons while the patients was receiving (46% each) (table 3). No differences between DMARD treatment. the four drugs were found when survival curves were compared taking lack of response as the STATISTICAL ANALYSES Table 2 Median survival oftreatment with disease Survival analysis for the four DMARDs was modifying antirheumatic drugs (DMARD) calculated by the Kaplan-Meier method and DMARD Median 95% Confidence interval expressed as cumulative probability of drug (mondh) survival.