LOW-DOSE METHOTREXATE WITH LEUCOVORIN (FOLINIC ACID) IN THE MANAGEMENT OF RHEUMATOID ARTHRITIS

Transcription

1 795 LOW-DOSE METHOTREXATE WITH LEUCOVORIN (FOLINIC ACID) IN THE MANAGEMENT OF RHEUMATOID ARTHRITIS Results of a Multicenter Randomized, Double-Blind, Placebo-Controlled Trial JEFFREY B. SHIROKY, CAROLYN NEVILLE, JOHN M. ESDAILE, DENIS CHOQUETTE, MICHEL ZUMMER, MARK HAZELTINE, VIVIAN BYKERK, MOHAMED KANJI, ANNE ST-PIERRE, LISE ROBIDOUX, and LUCIENNE BOURQUE Objective. To determine whether the side effects of methotrexate can be decreased by the concurrent use of leucovorin, without affecting the efficacy of the methotrexate. Methods. We conducted a multicenter randomized, double-blind, placebo-controlled trial of leucovorin administration, mg orally, to be given 24 hours after the single, weekly, oral dose of methotrexate. Every 3 weeks for 52 weeks, patients were evaluated for rheumatic disease activity and side effects. Dosage adjustments for both methotrexate and leucovorin were made as needed, according to a defined protocol. The primary outcome evaluated was the frequency of study withdrawals because of side effects and/or inefficacy. Secondary outcomes evaluated included the frequency of side effects and the relative efficacy of methotrexate in the leucovorin and placebo treatment s. Results. Ninety-two evaluable patients were analyzed (44 took leucovorin and 48 placebo). Twenty-two patients withdrew early because of side effects unrespon- From the Montreal General Hospital, McGill University, and the Hbpital Notre-Dame, the HBpital Maisonneuve-Rosemont, and the HGpital St-Luc, Universitk de Montreal; the Credit Valley Hospital, Mississauga, Ontario; and the Hbpital Charles Lemoyne, Greenfield Park, Quebec, Canada. Supported by a grant from the Arthritis Society of Canada. Dr. Shiroky is a Research Scholar of the Arthritis Society. Dr. Esdaile is a Senior Chercheur-Boursier of the Fonds de la Recherche en Santt du QuCbec. Jeffrey B. Shiroky, MD; Carolyn Neville, RN; John M. Esdaile, MD, MPH; Denis Choquette, MD; Michel Zummer, MD; Mark Hazeltine, MD; Vivian Bykerk, MD; Mohamed Kanji, MD; Anne St-Pierre, MD; Lise Robidoux, RN; Lucienne Bourque, RN. Address reprint requests to Jeffrey B. Shiroky, MD, Division of Rheumatology, Montreal General Hospital, 65 Cedar Avenue, Montreal, Quebec H3G A4, Canada. Submitted for publication October 2, 992; accepted in revised form December 6, 992. sive to our protocol, and because of inefficacy; 7 had been taking placebo and 6 had been taking leucovorin (35% versus 4%, P <.2). The number of visits during which side effects were reported was reduced by almost 5% in the leucovorin treatment (P <.). There were significant reductions in the frequencies of all common side effects. At 52 weeks, disease activity was similar in both patient s. Conclusion. The methotrexateleucovorin protocol used significantly reduces common side effects of methotrexate therapy without significantly altering efficacy. Methotrexate is considered among the most effective treatments for rheumatoid arthritis (,2). Although serious toxicity is rarely seen with the current regimen of weekly low-dose pulses, side effects account for most of the unsuccessful courses of methotrexate. Gastrointestinal symptoms, mucosal ulcers, and asymptomatic elevations in hepatic transaminase levels are the major reasons for dose reduction or discontinuation of therapy (3-6). The long-term risks of low-dose weekly methotrexate administration appear to be low, although there is justifiable concern about the potential for clinically significant chronic liver disease resulting from hepatotoxicity. While elevations in the serum hepatic transaminase levels are commonly seen, only rare cases of clinically significant liver disease have been noted in rheumatoid arthritis patients taking methotrexate over long periods (7). Prospective studies of liver biopsy tissues have identified a minority of patients with slowly progressing hepatic fibrosis without clinically apparent sequelae (8,9). Two studies Arthritis and Rheumatism, Vol. 36, No. 6 (June 993)

2 796 SHIROKY ET AL have linked this progressive fibrosis with periodic elevations in serum hepatic transaminase levels (8,9). Methotrexate, a folate analog, competitively inhibits the cellular enzyme, dihydrofolate reductase, which is involved in the essential step of reducing folate to its biologically active form. This results in a depletion of the intracellular stores of activated folate and a subsequent disruption of cellular metabolism (). Leucovorin (folinic acid) is a synthetic form of reduced folate and is designed to bypass the metabolic block. It is currently used as the antidote for severe methotrexate toxicity and prophylactically to reduce the incidence and severity of adverse reactions when intravenous infusions of large doses of methotrexate (.5- gm) are employed to treat certain cancers (). Based on the ability of leucovorin to limit adverse events, we developed a protocol for methotrexate therapy that incorporated small doses of leucovorin to be given in conjunction with the standard low dose of methotrexate for rheumatoid arthritis. Pilot work showed promising results ( ). We report here the results of a multicenter randomized, doubleblind, placebo-controlled trial in which we sought to determine whether our leucovorin protocol for lowdose methotrexate therapy reduces the frequency of methotrexate side effects and does so without affecting efficacy. PATIENTS AND METHODS Null hypothesis. The addition of leucovorin to methotrexate therapy currently used in the management of rheumatoid arthritis will not reduce the frequency of patients who are withdrawn from methotrexate treatment due to a lack of efficacy or due to an adverse reaction during a 52-week treatment period. Patients. Criteria for study inclusion were as follows: ) age 8-7 years; 2) rheumatoid arthritis, as defined by the 987 revised criteria of the American Rheumatism Association (now the American College of Rheumatology) (2); 3) active arthritis, defined as at least joints with active disease, i.e., discernible swelling, or pain with palpation and/or passive range of motion, and either at least 45 minutes of morning stiffness or an erythrocyte sedimentation rate (ESR) of at least 3 mm/hour; 4) unsuccessful treatment with aspirin or other nonsteroidal antiinflammatory drugs (NSAIDs) and at least second-line agent other than methotrexate (hydroxychloroquine, gold salts, D-penicillamine, sulfasalazine, azathioprine, or cyclophosphamide); 5) and daily dosage of corticosteroid 5 mg of prednisone or equivalent, without having been changed for at least month prior to entry. Exclusion criteria were as follows: ) prior use of methotrexate for arthritis; 2) women who were pregnant or breast feeding, or patients who were not surgically sterile or not practicing a reliable form of birth control for at least 3 months prior to study entry; 3) renal insufficiency, defined as a serum creatinine level >25 pmolesfliter (.4 mg/dl), a creatinine clearance rate <.O mvsecond (6 ml/minute), or proteinuria >.3 gm/dl; 4) known liver abnormalities, defined as transaminase or bilirubin levels persistently higher than the normal range, hepatitis within the past 2 months, or hepatitis B surface antigen identified on screening test; 5) consumption of more than 7 ounces of ethanol weekly; 6) major complicating illnesses, including diabetes mellitus requiring treatment, nondermatologic malignancies, malignant melanoma, or chronic infections; 7) white blood cell (WBC) count <4. X lo fliter, hemoglobin <9 gm/liter, or platelet count of. X Aiter; 8) history of major gastrointestinal bleeding, active peptic ulcers, or inflammatory colitis in the preceding 2 months; 9) major surgery in the preceding 2 months; ) treatment with any of the following: folate, warfarin, allopunnol, or sulfonamides (including oral hypoglycemics, antibiotics, and diuretics); ) any change in medications, including use of intraarticular corticosteroid injections, within month of entry; 2) any use of secondline agents within month of entry; 3) Steinbrocker functional class IV (3); and 4) history of poor compliance with therapeutic regimens. All patients provided uniform written informed consent approved by the internal review boards of all participating centers. Study medication. Methotrexute. All patients were initially treated with 7.5 mg of oral methotrexate to be taken as a single dose in the morning once each week. If, in the opinion of the physician (clinical investigator), there was no improvement, the dose could be increased every 6 weeks, by a minimum of 2.5 mg and a maximum of 7.5 mg, to a dosage of 5 mg/week, and then by 5 mg/week (every 6 weeks), up to a maximum dosage of 3 mg/week. Patients were given a 3-month supply of 2.5-mg tablets of methotrexate (Cyanamid Canada Ltd., Markham, Ontario, Canada) and were told to bring the pills to each followup visit (every 3 weeks). Compliance was assessed by pill count. Leucovorin. Each patient was given a supply of 5-mg tablets of either leucovorin or placebo (provided by Cyanamid Canada Ltd.). The placebo tablets were identical to the leucovorin tablets and both were tasteless. Study drug assignment was based on randomization, using a table of random numbers and stratified by center. The randomization schedule was given to the pharmacy staff of the Montreal General Hospital, where all supplies of leucovorin and placebo were prepared using identical labeling and bottling. Patients were instructed to take the leucovorin (or placebo) as a single dose 24 hours after taking the methotrexate. The initial dose was 2.5 mg (% tablet). When the methotrexate dose was increased to 5 mg or greater, the leucovorin dose was increased to 5 mg (I tablet); doses greater than 5 mg were not permitted. In the event of an adverse reaction to the methotrexate, the leucovorin (or placebo) could be increased to 5 mg as described below. A 3-month supply of leucovorin/placebo pills was given to the patients and they were instructed to bring them to each visit so the pills could be counted, for assessment of compliance.

3 LEUCOVORIN THERAPY WITH METHOTREXATE 797 Other, non-study, medications. All second-line agents were discontinued at least month before study entry. Oral prednisone (or equivalent) dosages were mg/day and stable for at least month before study entry. Aspirin and other NSAIDs were permitted throughout the study, at dosages established month prior to study entry. The dosage of neither the corticosteroid nor the NSAID could be altered during the 52-week study period. Acetaminophen was permitted as needed. One intraarticular injection of corticosteroid was permitted during the entire study period. Patient evaluation. All patients had a baseline evaluation which consisted of a complete history and physical examination, complete blood cell count (CBC) with differential count, blood biochemistry profile consisting of glucose, urea, serum creatinine, aspartate aminotransferase (AST), alanine aminotransferase (ALT), gamma glutamyl transferase (GGT), alkaline phosphatase, total bilirubin, total protein, and serum albumin levels, urinalysis, and a 24-hour urine collection for creatinine clearance and protein excretion analyses. Rheumatoid factor by nephelometry, C3, C4, serum immunoglobulins, and an ESR were obtained. Patients were seen every 3 weeks for a total of 52 weeks. At each visit, patients completed a standardized written questionnaire regarding side effects and were further interviewed by the center s research nurse regarding any new symptoms. When new symptoms or an adverse event occurred, the patient was assessed by the physician. A CBC with differential count, urea, serum creatinine, AST, ALT, alkaline phosphatase, GGT, and total bilirubin levels were obtained at each visit. At 52 weeks or at withdrawal from the study a final 24-hour urine collection was evaluated for creatinine clearance and protein levels. Clinical assessments performed every 6 weeks (every other visit) were: ) joint tenderness on pressure and/or pain on motion, scoring 6 diarthrodial (synovial) joints, on a scale of -3, where = none, = positive response on questioning, 2 = spontaneous response elicited, 3 = withdrawal by patient on examination; 2) joint swelling, scoring 58 joints, on a scale of -3, where = none, = detectable synovial thickening without loss of bony contours, 2 = loss of distinctions of bony contours, 3 = bulging synovial proliferation with cystic characteristics; 3) physician s global assessment of disease activity, graded on a scale of -5, where = asymptomatic, 2 = mild, 3 = moderate, 4 = severe, 5 = very severe. Clinical assessments performed every 3 weeks (each visit) were: ) patient s global assessment of disease activity, graded on a scale of -5, where = asymptomatic, 2 = mild, 3 = moderate, 4 = severe, 5 = very severe; 2) grip strength (expressed in mm Hg) using a standard sphygmomanometer with a standard grip bag (4); 3) morning stiffness (expressed in minutes); 4) 5-foot walking time (expressed in seconds). The Stanford Health Assessment Questionnaire (HAQ) was completed by all patients at baseline and at 52 weeks or at the time of study withdrawal. The HAQ served as an index of disability (5). Adverse reactions. All new symptoms reported were recorded. Adverse events were symptoms thought to be related to the methotrexate therapy. All adverse events were classified by the clinical investigator as being mild, moderate, or severe. Classification was based on the subjective assessment of both the patient and the investigator if these events were not the findings of a laboratory test (e.g., gastrointestinal [GI] disturbances, mouth ulcers, headaches). When adverse events were findings of a laboratory test, then the following specific definitions for mild, moderate, and severe were given: for hematologic events, mild = WBC count < x 9/liter and/or platelet count x 9/liter; moderate = WBC count x 9/liter and/or platelet count x fliter; severe = WBC count c2.5 x 9/liter and/or a platelet count 4. x 9/liter. Hepatic reactions were classified as mild = either an AST or ALT elevation above the upper limit of normal, but less than twice the upper limit of normal; moderate = an AST or ALT greater than twice the upper limit of normal, but less than 5 times the upper limit of normal; severe = greater than 5 times the upper limit of normal. All severe reactions resulted in study withdrawal, as did all pulmonary and allergic reactions. All other adverse reactions had to be present for 2 consecutive visits, 3 weeks apart, prior to modifications of the study medications. Mild reactions were handled initially by reducing the methotrexate dose by 2.5 mg. Once the reaction resolved, further increases in the methotrexate were permitted if deemed necessary by the clinical investigator. Moderate reactions were handled by withholding the methotrexate until resolution of the reaction, then resuming the methotrexate at doses 5. mg lower than previously prescribed. If moderate symptoms recurred at the reduced dose, or when the investigator deemed it necessary to increase the methotrexate, then the leucovorin dose was increased to 5 mg/week (if the methotrexate dose was <I5 mg weekly). If, after these initial procedures, the adverse reaction persisted over the 3 weeks to the next visit, then the adverse reaction was handled in an escalating manner. As an example, a persistent, mild adverse reaction would then be handled as a moderate one, and if persisting, would then be handled as a severe one (by withdrawing the patient from study). Similarly, a persistent, moderate reaction would then be handled as a severe one (withdrawal from study). Data handling and analysis. The primary outcome for this study was the frequency of withdrawals from study due to either adverse effects or inefficacy. Based on a literature review indicating a 3% withdrawal at 2 months, 96 patients (48 per ) were needed in order to detect a 67% reduction in the withdrawal rate that could be attributed to the methotrexate therapy, based on (Y =.5 (-tailed) and p =.2. For the purposes of this analysis, inefficacy was defined as failure to improve by 5% in at least of the 8 clinical variables assessed (patient s global assessment, physician s global assessment, disability index, joint tenderness index, joint swelling index, morning stiffness, grip strength, 5-foot walking time). Secondary outcomes analyzed included the efficacy of methotrexate in those completing the full 52-week study, and the number of visits at which adverse reactions attributed to methotrexate were recorded. All patients who withdrew within 8 weeks of enrollment because of either an enrollment violation or noncompliance were excluded from analysis. Patients withdrawing early because of an unrelated medical illness were included in the baseline analysis (not the 52-week analysis of efficacy). These patients were included in the analysis of the

4 SHIROKY ET AL number of visits at which adverse reactions were reported, up to the time of their study withdrawal. Baseline variables were contrasted, using Pearson chi-square, Student s t-test, Mann-Whitney test, or Kruskal- Wallis test where appropriate, for all those enrolled, and then again for those completing 52 weeks of therapy. Clinical variables at 52 weeks were contrasted in a similar manner for those completing the study (efficacy analysis). In addition, the clinical variables were further contrasted at 52 weeks with respect to change since baseline, using repeatedmeasures analysis of covariance. The primary outcome of withdrawals due to adverse events, with and without those due to inefficacy, was analyzed by Pearson s chi-square. RESULTS Population enrolled. Ninety-seven patients were entered into the study. Five patients ( taking placebo and 4 taking leucovorin) were excluded. Two of these 5 patients were excluded because of enrollment criteria violations ( had active peptic ulcer disease and had proteinuria). Three other patients were withdrawn early because of noncompliance. The remaining 92 patients (44 taking leucovorin Table. Baseline demographic and clinical variables for all 92 patients enrolled* Leucovorin Placebo Characteristic (n = 44) (n = 48) P Age (years) 53.3 f f.5.94 Sex Males 4 - Females Disease duration (years).7 -C t Disease activity Patient s global 2.3 t & assessment Physician s global 2.2 f f assessment Steinbrocker functional 2.4 f f.54.7 class HAQ disability index.26 f &.67.6 Joint tenderness index 38.3 IT t Joint swelling index 33.2 IT f Morning stiffness (minutes) 74 t f Grip strength, both hands 48 f t (mm Hg) 5-ft walking time 4. f f (seconds) Erythrocyte sedimentation 34. f f 4..5 rate (mdhour) Hemoglobin (gditer) 27.2 f f Mean corpuscular volume 84.5 f f (fl) * Values are the mean 2 SD. Range of possible scores (asymptomatic to severe): global assessments -4; Steinbrocker functional class -4; Health Assessment Questionnaire (HAQ) disability index -3. See Patients and Methods for details. Table 2. Baseline demographic and clinical variables for patients who completed 52 weeks of therapy* Leucovorin Placebo Characteristic (n = 35) (n = 29) P 53. IT t..9 Age (years) Sex Males Females Disease duration (years) Disease activity Patient s global assessment Physician s global assessment Steinbrocker functional class HAQ disability index Joint tenderness index Joint swelling index Morning st8ness (minutes) Grip strength, both hands (mm Hg) 5-ft walking time (seconds) Erythrocyte sedimentation rate (mmhour) Hemoglobin (gm/liter) Mean corpuscular volume (fl) * See Table for details. I f t t f f f f f f f f t f f f.5.6 f t f f ? t f t f and 48 taking placebo) were analyzed for pertinent demographic, clinical, and laboratory variables at baseline (Table ). There were no significant differences noted in the 5 variables analyzed, although patients in the placebo-treated tended to have a shorter disease duration and higher disability index. We also examined these baseline data for all who completed the 52 weeks of study (35 taking leucovorin and 29 taking placebo); there were no significant differences noted at baseline (Table 2). No differences were noted in any of the other laboratory values examined (data not shown). Withdrawals from study. Of the 92 patients evaluated at baseline, 64 were able to complete the full 52-week study period. Five patients withdrew early due to other medical problems. Two patients from each study had active NSAID-induced gastric ulcers, and patient (leucovorin ) had a clinically significant hemorrhagic ovarian cyst. Twenty-three out of 92 patients withdrew early, 22 for side effects failing to resolve with our protocol and due to a lack of efficacy. Seventeen of the 22 withdrawals for side effects were patients receiving placebo, and the other 5 were receiving leucovorin (P

5 LEUCOVORIN THERAPY WITH METHOTREXATE 799 Table 3. Side effects resulting in withdrawal from study* Side effects GI only Oral ulcers only Transaminase elevations only GI, oral ulcers GI, transaminase elevations Oral ulcers, transaminase elevations GI, oral ulcers, transaminase elevations Angioedema Leucovorin (n = 5) Placebo (n = 7) * Values are the number of patients. The angioedema was not proven to be due to leucovorin. GI = gastrointestinal. <.). One of the 5 leucovorin-treated patients developed non-life-threatening angioedema, presumably related to either the methotrexate or the leucovorin (Table 3). Fourteen of the 7 placebo-treated patients had a resolution of the side effects when, as per the protocol, leucovorin was added; these patients continued taking methotrexate after withdrawal from study. The other 3 patients chose not to continue methotrexate. Only patient failed to improve by 5% in at least variable, despite achieving a weekly dose of 3 mg of methotrexate. This patient, who was in the leucovorin-treatment, withdrew due to lack of efficacy. When this patient was included in the analysis of withdrawals related to methotrexate therapy (now 7 of 48 placebo versus 6 of 44 leucovorin), the results remained statistically significant (P <.2). We compared mean corpuscular volumes (MCV) in those who withdrew due to adverse reactions and those who completed the study. No differences were noted either at study entry or study completion, although the MCV tended to rise in both s over 52 weeks (baseline mean f SD versus 84.9 * 6.7 and study end 89.3? 6.9 versus 9. & 6.4, withdrew due to side effects versus completed). Similarly, those withdrawing due to side effects did not differ in age, disease duration, renal function, liver transaminase levels, or disease activity measures at baseline (data not shown). The MCV increased in both study s, but less so in the leucovorin-treated. Adverse events. Other than the non-lifethreatening angioedema, no serious toxicity was observed. No significant hematologic abnormalities were seen. One leucovorin-treated patient had mild leukopenia (3.4 X lo /liter) on only evaluation. The previous WBC count was 4.2 x lliter and the subsequent count was 4.7 x lo Aiter, without a dosage modification, which suggests the abnormal value may have been a laboratory error. As expected from previous studies, most toxicity was mild, involving oral ulcers, GI symptoms, or transaminase elevations. Most patients who withdrew due to side effects had combinations of side effects (Table 3). To examine the nature, frequency, and severity of side effects. we tabulated the number of visits Table 4. Number of visits at which adverse events were reported* Classification Total no. Total no. (%)., at which one or more adverse Mild Moderate Severe Total of visits events reported Transaminase elevations Leucovorin Placebo Gastrointestinal symptoms 3 93 Leucovorin Placebo Oral ulcers Leucovorin Placebo Visit totals Leucovorin I (7.3)t Placebo I (32.2) * Adverse events were symptoms thought to be related to the methotrexate (see Patients and Methods for details). t P <. versus placebo, calculated from the actual number of visits with one or more adverse events reported.

6 SHIROKY ET AL Table 5. Other adverse reactions reported* Alopecia (thinning) Fatigue Dry cough Dizziness Headaches Bad taste Rashes Vaginal sores Increased perspiration Lip swelling (not angioedema) Angioedema Zoster * Values are the number of patients. Leucovorin Placebo during which a side effect was reported, and then we categorized the side effects by reaction type and severity. This allowed inclusion of transient events, some of which were responsive to the protocol for modifying therapy. For this analysis, the total number of completed visits for all patients in each was used as the denominator (maximum of 7 followup visits over the 52-week study; with 4 more patients in the placebo-treated, there would be 68 more visits if all completed the study). It was established a priori to include in the numerator those adverse reactions thought to be due to the antifolate properties of the methotrexate (hepatic, mucosal, gastrointestinal, hematologic). As can be seen in Table 4, the reporting of all common side effects was substantially lower in the leucovorin. There were almost 5% fewer reports of side effects in the leucovorin-treated (7.3% versus 32.2%; P C.). The reporting of transaminase elevations was 6% lower, oral ulcers 48% lower, and GI symptoms 4% lower. Reports of moderate and severe GI symptoms were 63% fewer. Other, less frequent, symptoms are shown in Table 5. The relationships either to methotrexate or to its antifolate properties are uncertain, and thus any response to leucovorin therapy is also uncertain. No pulmonary toxicity was seen. Efficacy. The mean scores on the 8 clinical variables at 52 weeks of therapy are presented in Table 6. Comparing means at 52 weeks did not demonstrate any significant differences between s. Repeatedmeasures analysis of variance, controlling for the baseline values of each variable, did not demonstrate any significant difference at 52 weeks, except for the HAQ disability index, which had a higher degree of change in the placebo-treated. The mean weekly dose of methotrexate at week 52 was slightly higher in the leucovorin-treated than in the placebo-treated (3.6 mg [range and 2. mg [range 7.5-3, respectively), but this did not achieve statistical significance (P =.22). Seven of 44 patients taking leucovorin versus 2 of 48 patients taking placebo received a single intraarticular dose of corticosteroids during the study period (P not significant). We also compared the mean weekly dose of methotrexate in those withdrawing due to side effects in both s. This was also not statistically Table 6. Week-52 clinical variables compared with baseline data for patients who completed the study* Disease activity Patient s global assessment Physician s global assessment HAQ disability index Joint tenderness index Joint swelling index Morning stiffness (minutes) Grip strength, both hands (mm Hg) 5-ft walking time (seconds) Erythrocyte sedimentation rate (mdhour) Hemoglobin (gmaiter) Mean corpuscular volume (fl) * See Table for details. Data at week 52 Leucovorin Placebo Difference compared with baseline Leucovorin Placebo (n = 35) (n = 29) P DOUP P f f f f.92.2 t f.8. f f f f f f t t t f f f f t f f f f f f f

7 LEUCOVORIN THERAPY WITH METHOTREXATE 8 significant (leucovorin. mg [range versus placebo.3 mg [range 5-2]; P =.88). DISCUSSION This study was designed to determine whether a leucovorin protocol could be developed which could reduce the frequency and severity of adverse reactions to methotrexate without interfering with its effectiveness in a clinically significant manner. Our study demonstrates that small doses of leucovorin, when begun simultaneously with methotrexate, reduce the frequency of side effects related to the methotrexate therapy for rheumatoid arthritis. Patients who withdrew from the protocol because of side effects were those who had either the most severe or the most persistent symptoms, which failed to respond to standard clinical adjustments of the methotrexate dosage (reducing and withholding). Seventeen of the 48 patients in the placebo-treated (35%) and 5 of the 44 patients in the leucovorin-treated ( %) had to withdraw because of side effects (P <.). Most of these patients had persistent, rather than severe, side effects. Severe side effects were uncommon (Table 4). Most of those who withdrew had more than one of the common side effects (GI, hepatic, mucosal) (Table 3). It should be emphasized that it was the failure of the symptoms or laboratory abnormalities to resolve with our protocol, rather than the severity of these abnormalities, that resulted in most of the study withdrawals. This outcome was chosen solely to measure the effectiveness of this leucovorin protocol in controlling methotrexate side effects. In clinical practice, some of these patients may well choose to continue taking the same dosage of methotrexate and just tolerate the side effects. This will not be true for all patients; indeed, faced with such mild, persistent side effects, the patient may be less compliant or completely discontinue the methotrexate because of apprehension. The patient s sense of well-being, as well as willingness to increase the dosage of methotrexate to a more clinically effective level, may also be influenced by mild persistent adverse reactions. The reduction in serum transaminase elevations confirms the results of our pilot study (). In the present study, the frequency of transaminase elevations was reduced by 6% (Table 4). Among those who withdrew due to side effects, elevations in transaminase values were observed in placebo-treated patients and l leucovorin-treated patient. Two studies have suggested that elevations in transaminases occur more frequently in patients who are later found to have progressive histopathologic changes in the liver (8,9). These observations, coupled with the observation by Kremer et a of folate depletion of the liver with methotrexate treatment (6), suggest, although circumstantially, that transaminase elevations are related to hepatic folate depletion and that leucovorin therapy may be hepatoprotective. Other studies have failed to find significant reductions in side effects with the use of leucovorin (7,8). This can be explained by small sample sizes, short observation periods, and study populations of patients who had been taking methotrexate for an extensive period of time before study entry. In one study, the patients had been taking methotrexate a mean of 26 months prior to entry and then, in a randomized crossover design, were evaluated for side effects over two 4-week intervals without a washout period (7). The second study enrolled patients who had been on a regimen of methotrexate for a mean of 9 months (8). Thus, these results, at best, could only be generalized to patients who had tolerated methotrexate therapy for a fair length of time. Two other studies, one of 4 weeks and the other of 2 weeks duration, suggested that leucovorin blocks the effectiveness of methotrexate (9,2). This is to be anticipated, given that leucovorin is an antidote for methotrexate. In both studies, the leucovorin was administered within hours of the methotrexate, at doses considerably higher than the doses of methotrexate (45 mg of leucovorin versus mg of methotrexate and 5 mg of leucovorin versus a mean of 7.9 mg of methotrexate). However, the considerable experience of oncologists suggests that leucovorin should not be given in doses higher than those of methotrexate, and not so soon after the methotrexate dose. In our previously reported studies of high-dose methotrexate for rheumatoid arthritis (2,22), we found clinical improvement with minimal toxicity when the leucovorin was given in doses lower than the methotrexate and beginning 24 hours following the methotrexate. This study confirms that with appropriate dosages and timing of leucovorin, the side effects of methotrexate therapy can be limited, with no, or at the very worst, minimal, impairment of efficacy. The current study evaluated patients for 52 weeks, which allowed us to determine whether leucovorin therapy might result in a late-onset loss of efficacy due to potential cellular accumulation. No such loss of efficacy was observed. Results of clinical assessments at

8 SHIROKY ET AL the end of 52 weeks were similar in both study s. With the exception of the HAQ disability index, all other variables seemed to improve to a similar degree, in both study s. The different levels of improvement in the HAQ score is of uncertain clinical significance. Both s had equivalent disability scores at week 52, and it is possible that a score of.8 is the best possible score such a cohort of patients with chronic arthritis could achieve (rather than a score of ). The mean dose of methotrexate in the leucovorintreated was.6 mg higher than in the placebotreated, although the range of doses was the same. The clinical response to methotrexate was similar in both treatment s, but the methotrexate was better tolerated in the leucovorin-treated. If the slight difference in methotrexate dosage reflects interference of leucovorin with the action of methotrexate, it is of doubtful clinical significance. Other investigators have evaluated the effects of daily folate supplementation and found it to lessen methotrexate side effects without interfering with efficacy (23). It is difficult to compare that study with ours, however. That study was smaller (32 patients), of shorter duration (6 months), used a lower dose of methotrexate (median 7.5 mg, maximum 5 mg), and used a different dosing schedule (3 dosedweek, 2 hours between doses). The major conclusion was based on a lower mean toxicity score in the folatetreated, virtually all of which could be accounted for by only 4 patients. The folate study was too small to evaluate the effects of folate on transaminase elevations. Differences in the cost of folate supplementation and leucovorin supplementation favor the use of folate. In Quebec, month of folate supplementation would cost the patient approximately $4.5 (Canadian) compared with $3. or $26. per month for leucovorin, depending on whether the patient required /2 or tablet of leucovorin weekly. Although such costs may seem slight compared with the overall monthly costs of care for a patient with rheumatoid arthritis, for some patients, the differences in cost would be important (23). A study comparing folate and leucovorin therapy in methotrexate-treated patients seems in order. We believe our study better reflects how methotrexate is currently used than does the regimen employed in the folate study. Although clinical trials have largely used 7.5 or 5 mg of methotrexate, more flexible dosing with a higher maximum dose (25-3 mg) appears to be more widely used in clinical practice (3,24,25). It cannot yet be recommended that all patients who begin methotrexate therapy be given leucovorin or folate concomitantly. No such costbenefit analysis has been done. Our results could support the long-term use of leucovorin in those who would not otherwise be able to continue methotrexate because of drug intolerance, or in those with persistent or recurrent symptoms that make optimal dosing difficult. This is further supported by our observations that the administration of leucovorin to those in the placebo-treated who withdrew because of side effects successfully resolved these side effects. The cost of prophylaxis as well as the treatment of side effects with leucovorin in this setting may be more than offset by the patient s increased sense of well-being, by the possibility of achieving a more effective methotrexate dose, and by the potential reduction in the number of visits to the physician and associated investigative tests that would result from persistent side effects or arthritic symptoms. In conclusion, this study confirms that leucovorin (folinic acid) can be given at doses of mg weekly, as a single oral dose, 24 hours after a single weekly dose of methotrexate for rheumatoid arthritis and reduce the side effects without interfering with the efficacy of methotrexate in a clinically significant manner. In the absence of any serious cytopenias or pneumonitis, we cannot confirm whether this protocol prevents these adverse events. ACKNOWLEDGMENT The authors wish to express their gratitude to C. Kriticos for her devoted efforts in assisting in the preparation of the manuscript. REFERENCES Felson DT, Anderson JJ, Meenan RF: The comparative efficacy and toxicity of second-line drugs in rheumatoid arthritis: results of two metaanalyses. Arthritis Rheum 33:449-46, 99 Felson DT, Anderson JJ, Meenan RF: Comparing second line drugs for RA: efficacyltoxicity tradeoffs (abstract). Arthritis Rheum 34 (suppl 9):S53, 99 Weinstein A, Marlowe S, Korn J, Farouhar F: Lowdose methotrexate treatment of rheumatoid arthritis: long-term observations. Am J Med 79:33-337, 985 Weinblatt ME, Weissman BN, Holdsworth DE, Fraser PA, Maier AL, Falchuk KR, Coblyn JS: Long-term prospective trial of rnethotrexate in the treatment of rheumatoid arthritis: 84-month update. Arthritis Rheum 35:29-37, 992

9 LEUCOVORIN THERAPY WITH METHOTREXATE Fehlauer CS, Carson CW, Cannon GW, Ward JR, Samuelson CO, Williams HJ, Clegg DO: Methotrexate therapy in rheumatoid arthritis: 2-year retrospective followup study. J Rheumatol 6:37-32, Furst DE, Erikson N, Clute L, Koehnke R, Burmeister LF, Kohler JA: Adverse experience with methotrexate during 76 weeks of a longterm prospective trial in patients with rheumatoid arthritis. J Rheumatol 7: , Weinblatt M, Walker AM, Funch D, Dreyer NA, Alarcon G, Klippel J, Kremer J, Lee R, Tolman K: Serious liver disease in methotrexate-treated rheumatoid arthritis patients (abstract). Arthritis Rheum 34 (suppl9):s49, Phillips CA, Cera PJ, Mangan TF, Newman ED: Clinical liver disease in patients with rheumatoid arthritis taking methotrexate. J Rheumatol 9: , Kremer JM, Lee RG, Tolman KG: Liver histology in rheumatoid arthritis patients receiving long-term methotrexate therapy. Arthritis Rheum 32: 2-27, 989. Jolivet J, Cowan KH, Curt GA, Clendeninn NJ, Chabner BA: The pharmacology and clinical use of methotrexate. N Engl J Med 39: 94-4, 983. Shiroky J, Neville C: Leucovorin for methotrexate induced elevations in liver transaminases (abstract). Arthritis Rheum 33 (suppl 5):R34, Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, Healey LA, Kaplan SR, Liang MH, Luthra HS, Medsger TA Jr, Mitchell DM, Neustadt DH, Pinals RS, Schaller JG, Sharp JT, Wilder RL, Hunder GG: The American Rheumatism Association 987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 3:35-324, Steinbrocker, Traeger CH, Batterman RC: Therapeutic criteria in rheumatoid arthritis. JAMA 4: , The Cooperating Clinics Committee of the American Rheumatism Association: A seven-day variability study of 499 patients with peripheral rheumatoid arthritis. Arthritis Rheum 8:32-334, Fries JF, Spitz P, Kraines RG, Holman HR: Measure- ment of patient outcome in arthritis. Arthritis Rheum 23: 37-45, Kremer JM, Galivan J, Streckfuss A, Kamen B: Methotrexate metabolism analysis in blood and liver of rheumatoid arthritis patients: association with hepatic folate deficiency and formation of polyglutamates. Arthritis Rheum 29S32-835, Hanrahan PS, Russell AS: Concurrent use of folinic acid and methotrexate in rheumatoid arthritis. J Rheumatol 5:78-8, Buckley LM, Vacek PM, Cooper SM: Administration of folinic acid after low dose methotrexate in patients with rheumatoid arthritis. J Rheumatol 7: 58-6, Tishler M, Caspi D, Fishel B, Yaron M: The effects of leucovorin (folinic acid) on methotrexate therapy in rheumatoid arthritis patients. Arthritis Rheum 3 :96-98, Joyce DA, Will RK, Hoffman DM, Laing B, Blackbourn SJ: Exacerbation of rheumatoid arthritis in patients treated with methotrexate after administration of folinic acid. Ann Rheum Dis 5:93-94, Shiroky J, Allegra C, Inghirami G, Chabner B, Yarboro C, Klippel JH: High dose intravenous methotrexate with leucovorin rescue in rheumatoid arthritis. J Rheumatol 5:25-255, Shiroky JB, Neville C, Skelton JD: High dose intravenous methotrexate for refractory rheumatoid arthritis. J Rheumatol 9:247-25, Morgan SL, Baggott JE, Vaughn WH, Young PK, Austin JV, Krumdieck CL, Alarc6n GS: The effect of folic acid supplementation on the toxicity of low-dose methotrexate in patients with rheumatoid arthritis. Arthritis Rheum 339-8, Hanrahan PS, Scrivens GA, Russell AS: Prospective long term follow-up of methotrexate therapy in rheumatoid arthritis: toxicity, efficacy, and radiological progression. Br J Rheumatol28:47-53, Kremer JM, Phelps CT: Long-term prospective study of the use of methotrexate in the treatment of rheumatoid arthritis: update after a mean of 9 months. Arthritis Rheum 35:38-45, 992