Power point

The Efficacy of Dexamethasone Iontophoresis for the Treatment of Rheumatoid Arthritic Knees: A Pilot Study

Linda C. Li, Roger A. Scudds, Carol S . Heck, and Manfred Harth

Objective. To examine the efficacy of dexametha- sone sodium phosphate (DEX) iontophoresis for the treatment of rheumatoid arthritis (RA) of the knee, and to obtain statistical information for a future random- ized controlled trial (RCT).

Methods. Ten subjects with RA, ranging in age from 34-75, were randomly assigned to either the experi- mental or placebo group. Iontophoresis treatments were given to both groups on days 1 , 3, and 5. Five subjects in the experimental group received a mixture of 1 ml of DEX (4 mglml) and 1 m l of injectable sterile water; those in the placebo group received 2 ml of sa- line solution. Pain on movement, at rest, and on pres- sure, active joint count, and active range of motion, were evaluated on days 1, 5, and 20. The patient’s global assessment of treatment efficacy was also as- sessed on days 5 and 20. Mann-Whitney U tests and Friedman two-way analyses of variance were per- formed for statistical analyses.

Results. Pain at rest was found to be statistically different between the two groups (P = 0.031 7). Statis- tical significance was also found over time for pain on movement within the experimental group (P = 0.0224).

Conclusion. The results suggest that DEX iontopho-

Linda C. Li, MSc, The Arthritis Society, Consultation and Therapy Service; and Roger A. Scudds, PhD, PT, Department of Physical Therapy, Carol S. Heck, PhD, PT, Department of Physical Therapy, and Manfred Harth, MD, FRCP (C), Department of Medicine, Di- vision of Rheumatology, The University of Western Ontario, Lon- don, Ontario, Canada.

Address correspondence to Roger A. Scudds, PhD, PT, Depart- ment of Physical Therapy, Wichita State University, Wichita, KS

Submitted for publication January 27, 1995; accepted in revised 67260-0043.

form November 21, 1995. 0 1996 by the American College of Rheumatology.

126

resis i s more effective than placebo in relieving pain at rest and on movement in the RA knee. Based on the study data, a total of 40 subjects will be required for an RCT of a similar nature.

Key words. Rheumatoid arthritis; Iontophoresis; Pain; Corticosteroid; Knee.

INTRODUCTION

Rheumatoid arthritis (RA) is a chronic systemic dis- ease that may gradually lead to the destruction of car- tilage and bone. It is characterized by synovitis, which results in thickening of the synovium and swelling of the joint (1). RA can affect any diarthrodial joint in humans (2). Despite the fact that it was first docu- mented in 1800, the etiology and pathology of RA re- main obscure (3).

Intraarticular corticosteroid injections (IACI) have been used for the treatment of RA since the early 1950s (4). This technique is performed by injecting the med- ication directly into the affected synovial sac. Although IACI is a relatively safe and effective adjunct to the management of RA, patients often experience a high level of pain during the treatment. Also, the invasive nature of this procedure may result in further tissue damage (5). Therefore, it is worthwhile to investigate alternative noninvasive drug delivery techniques which may provide safe, comfortable, and effective treatment.

The use of iontophoresis for the management of RA and other inflammatory conditions was first examined in the mid-1950s (6). According to K a h n (71, ionto- phoresis is defined as “the introduction of substances into the body for therapeutic purposes by means of a

0893-7524/96/$5.00

Arthritis Care and Research Dexamethasone Iontophoresis for R A Knees 127

direct current.” The possible mechanisms of iontopho- resis and its applications were recently described (8).

Iontophoresis has been used to treat a variety of in- flammatory conditions, such as chronic osteomyelitis (9), myositis ossificans (lo), lateral epicondylitis, bi- cepital tendinitis, subacromial bursitis, patellar tendi- nitis, calcaneofibular ligament strain (11), shin-splints (12), and exercise-induced muscle soreness (13). In 1988, Hasson et a1 conducted a double-blind study of a single subject who had RA of both knees to evaluate the efficacy of corticosteroid iontophoresis (14). Com- bined dexamethasone sodium phosphate (DEX) and lidocaine hydrochloride (LID) was delivered to one knee by a 4-mA direct current for 20 minutes; the other knee received only LID iontophoresis (placebo). After 24 treatments, it was found that DEX iontophoresis was more effective in improving the strength and endur- ance of quadriceps muscles as compared to the pla- cebo. A later case study by Hasson et a1 also reported that 24 treatments of DEX iontophoresis and exercise improved quadriceps strength, knee active range of mo- tion (AROM), swelling, and cardiorespiratory endur- ance in an RA patient (15). Unfortunately, it is difficult to draw firm conclusions &om these studies because of the small sample sizes and the lack of proper sta- tistical analyses. As well, neither of these studies ex- amined the effects of the treatment on pain, which is an important indicator of the severity of RA.

In order to obtain the essential background infor- mation and to determine the feasibility of a randomized control trial (RCT), a pilot study was performed over a %month period. The purpose of the study was to examine the efficacy of DEX iontophoresis for the treat- ment of RA knees. The results of this study will be used to estimate the sample size required for an RCT.

PATIENTS AND METHODS

Subjects. Ten subjects (2 males and 8 females) rang- ing in age from 34 to 75 years (mean 56.60) with a diagnosis of RA, as defined by the American College of Rheumatology (ACR; formerly, the American Rheu- matism Association) criteria (16), were enrolled in this study. All the subjects had at least one painful knee as a result of RA, which had not been relieved by the same oral medication that they had used for at least 1 month. Patients who were hypersensitive to DEX or to direct current who had received a corticosteroid injec- tion in the affected knee during the 2 months before the study, or had joint replacement surgery, a metal implant, or any surgery in the treated leg during the 3 months prior to the study were excluded. Skin sen-

sation (hot/cold, sharp/dull) of the treated leg was test- ed before patients were admitted to the study.

All the subjects were recruited from the rheumatol- ogy outpatient clinic of University Hospital and St. Joseph’s Health Centre (London, Ontario, Canada). None of these patients were receiving physiotherapy, occupational therapy, or chiropractic treatment in the month before or during the study. No adjustment to the oral medications was allowed during the study- period.

Initial contact with each potential subject was made by one of us (LCL) no later than 48 hours after receiving the referral. Informed consent was obtained on the first visit. The study protocol was approved by the Uni- versity of Western Ontario Review Board for Health Sciences Research Involving Human Subjects.

Study design and setting. A double-blind, random- ized, placebo-controlled design was employed. Partic- ipants were randomly assigned to either the experi- mental or placebo group, using a table of random num- bers. All treatments and assessments were performed by the same investigator (LCL). Subjects received treat- ment on days 1, 3, and 5. Five subjects in the exper- imental group received DEX iontophoresis, and 5 in the placebo group received sodium chloride iontopho- resis. Neither the investigators nor the subjects were aware of the group assignment.

Three assessments were performed during the 20- day study. The baseline assessment was performed on day 1, prior to the first treatment. An interim assess- ment was performed on day 5, before the third treat- ment was administered. This was to examine the short- term effects of the designated treatment. Finally, a fol- lowup assessment was performed on day 20 to deter- mine the long-term effects.

Outcome measures. Pain on movement was used as the primary outcome measure. The secondary outcome variables were pain at rest, pressurepain threshold, M O M , active joint count, and the patient’s global as- sessment (PGA) of treatment efficacy.

Equipment and measuring instrument. The Dupe1 dual-channel iontophoresis system (Empi, St. Paul, MN) was used in this study. It is a battery-operated unit, which supplies a constant direct current ranging from 1 mA to 4 mA, and a target dosage of 1-160 mA (?2) per minute on each channel. The delivery elec- trode package (Empi buffered iontophoretic delivery electrodes) consists of 2 self-adhesive, non-reusable electrodes. The delivery electrode is 1.75 X 3.75 inch- es, with an active area of 3.1 square inches. The dis- persive electrode is made of conductive gel. Each elec-

128 Li et a1 Vol. 9, No. 2, April 1996

trode measures 3.5 X 3.5 inches, with an active area of 6.9 square inches.

A Fischer pressure dolorimeter (Pain Diagnostics and Thermography, Great Neck, NY), which has high inter- and intra-rater reliability (17), was used to eval- uate pain threshold. A hand-held goniometer (Smith and Nephew Rolyan, Inc., Menomonee Falls, WI) was used to measure the AROM of the knee. Several groups of investigators (18,19) have reported that the reliability and validity of this instrument for assessing knee AROM is high.

The number of joints with active RA was determined with a self-report articular index (SRAI) (20). Subjects were provided with standardized written instructions which asked them to indicate any joints that were “tender and inflamed at present” (21). The scores were calculated by summing the weighted count of the joints.

A 10-cm visual analog scale (VAS), which has a high test-retest reliability in RA patients (22), was used for assessing pain on movement and at rest. This was in the form of a horizontal line, with the words “no pain” at one end and “worst pain imaginable” at the other end. Subjects were instructed to indicate the amount of knee pain they experienced at the time of evaluation.

The PGA of treatment efficacy was measured with a 5-point Likert scale, where 5 = much better, 3 = no change, and 1 = much worse. Subjects were required to rate the condition of their knees in comparison to the first visit.

The drug solutions used in this study were: 1) dexa- methasone sodium phosphate (4 mg/ml; Pharmasci- ence Inc., Montreal, Quebec, Canada), 2) Injectable ster- ile water (Ayerst Laboratories, Montreal, Quebec), 3) sodium chloride solution (0.9%; Baxter, Toronto, On- tario). These were prepared in an unused syringe on the same day of each treatment. A mixture of DEX (1 ml) and sterile water (1 ml) was used in the experi- mental group, while the placebo group received 2 ml of NaC1. Since both DEX and NaCl are clear, colorless, and odorless solutions, neither the investigator nor the subjects were able to distinguish between the 2 forms of treatment from the appearance of the solutions.

Assessment. The outcome measures were recorded in the following order during the baseline assessment (day 1): 1) number of active joints, 2) pain at rest, 3) pressure pain threshold, 4) AROM, and 5) pain on movement. A similar order was employed on days 5 and 20, except that PGA was also assessed on these days. PGA was evaluated before the pressure-pain threshold.

Subjects sat comfortably in a chair during the as- sessment of the number of active joints, pain at rest,

and PGA. Pressure-pain threshold was measured with the patient supine. The measurement was taken at the medial and lateral joint lines, just anterior to the medial and lateral collateral ligaments of the treated knee. Pressure was applied at a constantly increasing rate of 1 kg/second and stopped when subjects verbally in- dicated they were just beginning to feel pain. Active range of motion was assessed with subjects lying on the untreated side. The treated leg was elevated by a supporting board. Measurements were performed with the knee in an unassisted, fully extended, and fully flexed position. Pain on movement was evaluated by asking subjects to perform 5 repetitions of knee flexion and extension at their own pace while lying on the untreated side. They were then asked to rate the amount of pain experienced, using a VAS. In this po- sition, the rating of pain was not influenced by other inflamed joints, which might also be painful on active movement.

Treatment. Each subject received treatments in only one knee. Those who had bilateral involvement re- ceived treatment in the knee that was more painful and swollen at the f i s t visit. If both sides were equally involved, treatments were given to the right side only.

During the treatment, the subjects’ knees were po- sitioned in approximately 30” of flexion. The negative electrode was used as the delivery electrode for both groups. It was expected that DEX would be delivered subcutaneously in the experimental group, and that chloride ions would be introduced in the placebo group. Once delivered into the body, both chemicals would recover to their original form after combining with sodium ions.

The delivery electrode was placed on the joint line of the knee, immediately medial to the inferior medial border of the patella. The dispersive electrode was placed over the medial muscle mass of the triceps sur- ae. The centers of the electrodes were 20 cm apart. A 4-mA direct current and a current dosage of 80 mA per minute were delivered via the iontophoresis unit. Each treatment lasted 20 minutes. Subjects were allowed to continue their normal daily activity during the study period.

Data analysis. Data analyses were performed using the SPSS/PC+ (version 5.0.1) statistical package. Mann-Whitney U tests were performed to compare the average scores of the 6 outcome variables between the two study groups. The within-group effects over days for each variable (except PGA) were assessed using the Friedman two-way analysis of variance (ANOVA) (22). Multiple comparisons were calculated manually when a significant difference was observed (22). Wilcoxon

Arthritis Care and Research Dexamethasone Iontophoresis for RA Knees 129

Table 1. Characteristics of the study subjects

Experimental group Placebo group

(n = 5) (n = 5)

Sex (female/male) 510 312 Age, mean t SD years Weight, mean 2 SD kg Height, mean ? SD cm Duration of RA, mean

Additional medications used for

56.60 t 14.74 57.00 ? 17.96 70.18 2 6.37 73.88 t 26.31

165.58 t 3.42 164.78 ? 4.80

t SD years 22.00 2 18.33 14.20 ? 8.70

RA* NSAIDs 4 4 Corticosteroids 2 2 SAARDs 4 4

' RA = rheumatoid arthritis; NSAIDs = nonsteroidal antiinflammatory drugs; SAARDs = slow-acting antirheumatic drugs.

Table 3. Results of Mann-Whitney U analysis for between-group effects

Mean rank

Experi- mental Placebo

Outcome variable group group P

Pain on movement Pain at rest

Pressurepain threshold Medial aspect Lateral aspect

Active range of motion Extension Flexion

Active joint count Patient's global assessment

ment efficacy

3.80 3.40

6.80 6.00

6.60

5.20 6.80 5.60

6.80 of treat-

7.20 7.60

4.20 5.00

4.40

5.80 4.20 5.40

4.20

0.095* 0.0317f

0.222 0.691

0.310

0.841 0.222 1.000

0.222

signed-ranks tests were performed to assess the differ- ence in the PGA scores between days 5 and 20 in each group. The level of significance was set at an (Y level of 0.05.

RESULTS

Fourteen subjects were referred to the study. Three patients were asymptomatic when contacted by the in- vestigator. One subject was excluded because his med- ication was adjusted during the study period. A total of 10 subjects participated and completed the study protocol. No significant differences in the age, height, weight, and onset of symptoms were found between the experimental and placebo groups ( P > 0.05) (Table 1). Four subjects from the experimental group and 3

* P < 0.10. t P C 0.05.

from the placebo group reported bilateral involvement. Treatments were performed on the left knee in 3 sub- jects in both groups. There were no reports of burns or skin irritations as a result of the treatment.

The group means and standard deviations for each variable from the 3 assessments are presented in Table 2 . Mann-Whitney U tests were performed on the av- erage scores of the outcome variables. These were cal- culated by averaging the total scores of each variable obtained from the 3 assessments (23). A statisticaIIy significant between-group difference was found for pain at rest (P = 0.0317). No significant differences were found between the groups on any other outcome measure (Table 3).

Table 2. Mean (?SD) scores of six outcome variables measured on days 1, 5, and 20

Experimental group Placebo group

Variable* Day 1 Day 5 Day 20 Day 1 Day 5 Day 20

Pain on movement (cm) 3.31 2 2.73 0.60 t 0.65 0.71 ? 0.80 5.13 ? 2.24 4.35 2 4.01 4.69 ? 4.08 Pain at rest (cm) 0.94 t 1.19 0.20 ? 0.27 0.21 2 0.23 3.89 ? 2.76 3.81 t 3.84 3.97 t 4.15

Pressure pain threshold (kg)

Medial side 3.38 t 2.42 2.48 t 1.98 2.63 2 0.85 1.36 -t 1.07 1.39 t 0.92 1.74 2 1.27 Lateral side 1.77 t 1.02 2.06 t 1.48 2.31 2 1.03 1.77 -t 1.02 1.65 2 1.20 1.48 t 1.13

AROM (degree) 5.60 t 11.06 Extension 3.00 t 9.33 1.40 2 10.09 0.60 ? 3.78 4.60 t 11.44 3.80 2 7.50

Flexion 107.60 t 8.96 111.00 t 8.43 109.40 2 9.04 89.60 t 13.61 99.80 t 21.12 89.30 ? 24.81 Active joint count 268.40 2 59.44 234.40 2 146.99 189.00 2 132.76 213.80 t 71.15 199.00 ? 81.77 224.60 2 87.17 PGA - 3.80 2 0.84 3.60 2 0.550 - 3.20 t 1.10 2.80 2 1.48

* AROM = active range of motion, (positive reading of knee extension indicates degree of AROM lacking; negative reading indicates degree of knee hyper- extension]: PGA = patient's global assessment of treatment efficacy.

130 Li et a1 Vol. 9, No. 2 , April 1996

Table 4. Friedman two-way analysis of variance for within-group effects

Experimental group Placebo group

Mean rank Mean rank

Outcome variable Day 1 Day 5 Day 20 xz P Day 1 Day 5 Day 2 0 xz P

Pain on movement 3.00 1.40 1.60 7.60 0.0224* 2.00 1.90 2.10 0.10 0.951 Pain at rest 2.50 1.80 1.70 1.90 0.387 1.60 1.90 2.50 2.10 0.350

Pressurepain threshold Medial aspect 2.00 1.60 2.40 1.60 0.449 2.00 1.80 2.20 0.40 0.819 Lateral aspect 1.60 2.00 2.40 1.60 0.449 2.20 2.10 1.70 0.70 0.705

AROMt Extension 2.20 2.00 1.80 0.40 0.819 2.20 1.80 2.00 0.40 0.819 Flexion 1.60 2.60 1.80 2.80 0.247 1.70 2.60 1.70 2.70 0.259

No. joints with active disease 2.30 2.20 1.50 1.90 0.387 2.30 1.50 2.20 1.90 0.387

* P < 0.05. t AROM = active range of motion.

Friedman two-way ANOVAs were performed for pain on movement, at rest, and on pressure, for AROM, and number of joints with active RA in the experimental and placebo groups, respectively (Table 4). The results revealed a significant difference over time only for pain on movement within the experi- mental group (P = 0.0224). Multiple comparisons showed statistically significant differences between days 1 and 5, and between days 1 and 20. No signif- icant changes were found for PGA between days 5 and 20 ( P > 0.05).

A sample size was estimated for an RCT using the data for pain on movement. A 3-cm difference of the mean VAS scores between the experimental and the placebo groups was used as the minimal clinically significant difference. A standard deviation of 2.54 was obtained for the 10 subjects at baseline. Based on this information, the suggested sample size is 16 sub- jects per group, with an a level of 0.05 and a p level of 0.10. Taking into account that 10% of the subjects may fail the screening process and another 20% may drop out, a total of 45 subjects will be required for an RCT (24).

DISCUSSION

The main purpose of this pilot study was to examine the efficacy of iontophoresis as a local corticosteroid- delivery technique for the treatment of RA knees. The major results revealed that DEX iontophoresis was more effective in reducing pain at rest compared to placebo. Also, within the experimental group, pain on movement was found to be significantly reduced kom day 1 to day 5, and between day 1 and day 20. No

significant changes were observed in the placebo group. These findings agree with those of Coyer (6), who reported that 13 of 15 RA patients who received citrate iontophoresis in their fingers experienced pain relief, compared to only 2 of 20 subjects who had pain relief from anodal or cathodal galvanism.

This study used outcome measures that are recom- mended by the ACR for clinical studies in RA (25). Several recent studies have suggested that outcome measures, such as the patient’s assessment of pain, are sensitive to change in RA clinical trials (26-28). The present study agrees with this position because of the significant differences in pain on movement and at rest despite the small sample size. Other outcome variables, such as muscle strength and morning stiffness, were not included because of their low sensitivity to change

No statistically significant differences were found for pressure-pain threshold, AROM, active joint count, or PGA. The small sample size may be partially respon- sible for these findings. Due to the limited statistical information hom the previous iontophoresis studies, a proper sample size calculation was impossible. Thus, a sample of convenience was used in this study. How- ever, the small sample might increase the risk of pro- ducing false-negative statistical results. Also, the sen- sitivity of some variables such as pressure-pain thresh- old and AROM is questionable. These variables would require a relatively large sample in order to detect a minimal clinically significant difference. This suggests that the selection of outcome measures for the future RCT should focus on their clinical applicability and sensitivity to change.

The major drawback of this study was the small number of subjects involved. However, the impact of

(26-28).

Arthritis Care and Research Dexamethasone Iontophoresis for R A Knees 131

DEX iontophoresis on the patients’ assessment of pain indicates the potential of this technique. Further re- search in this area should afford a better understanding of its mechanisms and clinical value. Also, since one of the goals of RA treatment is to maintain functional independence, an outcome measure to evaluate this area is recommended.

IACI is often employed as an adjunctive treatment for RA. This technique allows the maximal antiinflam- matory effects to take place locally, while minimizing adverse systemic effects. However, the use of local in- jection may be hindered by the possible tissue trauma and pain caused by the invasive procedure. Iontopho- resis has been suggested as an alternative to local in- jection (14,15); however, clinical studies which doc- ument the effects of this technique are scarce. There- fore, this pilot study was conducted to assess the ef- ficacy of DEX iontophoresis for the treatment RA and the feasibility of an RCT. The results indicate that al- ternate-day iontophoresis treatment may be effective for reducing pain at rest and on movement.

It is not uncommon for RA patients to require local corticosteroid therapy for relief of symptoms. IACI is an effective and relatively safe treatment. However, other methods of drug delivery should be explored. Iontophoresis may offer a painless and effective al- ternative to the traditional drug-delivery technique. The future role of DEX iontophoresis treatment for RA will rely on further investigations through well-de- signed RCTs.

Special thanks to Mrs. Debbie Robertson for her assistance in this study.

REFERENCES

1.

2.

3.

4.

5.

6.

7.

Bernstein RM, Hughes RV Rheumatoid arthritis: current concepts. Int Rehabil Med 4:119-123, 1982 Mosby’s Medical and Nursing Dictionary, St. Louis, CV Mosby Company, 1986 Gardner DL: Articular tissues: gross and microscopic pathology. In, The Pathology of Rheumatoid Arthritis. Edited by DL Gardner. London, Edward Arnold (Pub- lishers) Ltd., 1 9 7 2 Hollander JL, Brown EM, Jessar RA, Brown CY Hydro- cortisone and cortisone injected into arthritic joints- comparative effects of and use of hydrocortisone as a local antiarthritic agent. JAMA 147:1629-1635, 1951 Gray RG, Gottlieb NL: Intra-articular corticosteroids: an updated assessment. Clin Orthop 1 7 7 2 3 5 4 6 3 , 1983 Coyer AB: Citrate iontophoresis in rheumatoid arthritis of the hands. Ann Phys Med 216-19, 1954 Kahn J: Iontophoresis. In, Principles and Practice of

Electrotherapy. Edited by J Kahn. New York, Churchill Livingstone, 1991

8. Li LC, Scudds RA: Iontophoresis: an overview of the mechanisms and clinical application. Arthritis Care Res

9. Becker RO, Spadaro JA: Treatment of orthopaedic in- 8:51-61, 1995

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

fections with electrically generated silver ions. J Bone Joint Surg [Am] 60-A:871-881, 1978 Wieder DL: Treatment of traumatic myositis ossificans with acetic acid iontophoresis. Phys Ther 72133-

Harris PR: Iontophoresis: clinical research in muscu- loskeletal inflammatory conditions. J Orthop Sports Phys Ther 4:109-112, 1982 Delacerda FG: Iontophoresis for treatment of shin- splints. J Orthop Sports Phys Ther 4:183-185, 1982 Hasson SM, Wible CL, Reich M, Barnes WS, Williams JH: Dexamethasone iontophoresis: effect on delayed muscle soreness and muscle function. Can J Sport Sci

Hasson SM, English SE, Daniels JC, Reich M: Effect of iontophoretically delivered dexamethasone on muscle performance in a rheumatoid arthritic joint: a case study. Arthritis Care Res 1:177-182, 1988 Hasson SM, Henderson GH, Daniels JC, Schieb DA: Ex- ercise training and dexamethasone iontophoresis in rheumatoid arthritis: a case study. Physiother Can 43:

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 lU, Hunder GG: The American Rheumatism Association 1987 revised cri- teria for the classification of rheumatoid arthritis. Ar- thritis Rheum 31:315-324, 1988 Delaney GA, McKee AC: Inter- and intra-rater reliability of the pressure threshold meter in measurement of my- ofascial trigger point sensitivity. Am J Phys Med Rehabil

Mitchell WS, Millar J , Sturrock RD: An evaluation of goniometry as an objective parameter for measuring joint motion. Scott Med J 20:57-59, 1975 Gogia PP, Braatz JH, Rose SJ, Norton BJ: Reliability and validity of goniometric measurements at the knee. Phys Ther 67:192-195, 1987 Stewart MW, Palmer DG, Knight RG: A self-report ar- ticular index measure of arthritic activity: investigations of reliability, validity and sensitivity. J Rheumatol 17:

Ferraz MB, Quaresma MR, Aquino LRL, Atra E, Tugwell P, Goldsmith CH: Reliability of pain scales in the as- sessment of literate and illiterate patients with rheu- matoid arthritis. J Rheumatol 17:1022-1024, 1990 Pocock SJ, editor: Clinical Trials. New York, John Wiley

Conover WJ, editor: Practical Nonparametric Statistics. Second edition. New York, John Wiley & Sons, 1980 Donna A: Design and Analysis of Clinical Trial: Epi-

137, 1992

17:8-13, 1992

11-14, 1991

72: 13 6-1 3 9, 1993

1011-1015, 1990

& Sons, 1989

132 Li et a1 Vol. 9, No. 2 , April 1996

demiology 560B, The University of Western Ontario, London, Ontario, 1993

25. Felson DT, Anderson JJ, Boers M, Bombardier C, Cher- noff M, Fried B, Furst D, Goldsmith C, Kieszak S, Light- foot R, Paulus H, Tugwell P, Weinblatt M, Widmark R, Williams HJ, Wolfe F: The American College of Rheu- matology preliminary core set of disease activity mea- sures for rheumatoid arthritis clinical trials. Arthritis Rheum 36:729-740, 1993

26. Anderson JJ, Felson DT, Meenan RF, Williams HJ:

Which traditional measures should be used in rheu- matoid arthritis clinical trials? Arthritis Rheum 32:

27. Bombardier C, Raboud J, the Auranofin Cooperating Group: A comparison of health-related quality-of-life measures for rheumatoid arthritis research. Control Clin Trials 1232438-2568, 1991

28. Gotzsche PC: Sensitivity of effect variables in rheuma- toid arthritis: a meta-analysis of 1130 placebo controlled NSAID trials. J Clin Epidemiol 13:1313-1318, 1990

1093-1099, 1989