The Problem of Therapeutic Evaluation in Rheumatoid Arthritis By STANLEY L. WALLACE AND CHARLES RACAN The difficulties encountered in the eval- Le difficultates incontrate in le evaluuation of therapy for rheumatoid ar- tation de therapias pro arthritis rheumathritis are reviewed. The authors stress toide es delineate. Le autores sublinea the necessity for controlled studies, blind le necessitate de studios a controlo, de comparisons of untested agents, appro- comparationes “occultate” de Dove priate methods for measurement of thera- agentes, de appropriate methodos pro peutic response and statistical evaluation le mesuration del responsa therapeutic, e del evalutation statistic del resultatos. of results. T HE EVALUATION of therapeutic agents in rheumatoid arthritis is diffi- cult, the literature being replete with claims and counterclaims of the efficacy of many agents. So complex is this situation, as a matter of fact, that it is unusual to find a disease in which only one agent-the salicylates-has received universal acceptance even though that agent despite its obvious advantages of safety and availability is definitely wanting in effectiveness in many instances. This is in direct contrast to such other fields of medicine as the infectious diseases, hematology and cardiology, where there are many widely accepted therapeutic agents. With such difficulties, the evaluation of therapy in rheumatoid arthritis must include a clearcut distinction between symptom alleviation and effects on the course of the disease. Thus a potent analgesic such as morphine which will relieve symptoms is rarely advocated as a therapeutic remedy in rheumatoid arthritis. Attention must also be paid to the distinction between acute and long-lasting benefits. There is no cavil with the observation that suppression of almost all the manifestations of the disease may be accomplished by the administration of steroids, but this benefit may be short-lived. In a recent study,l even after as brief a period as one year, the results of steroid therapy were in no way better than those with salicylates, although the initial improvement with the latter agent was less. The duration of beneficial effects is of paramount importance in evaluation of therapy. In the older literature, most agents used in the therapy of rheumatoid arthritis were described as beneficial in about two-thirds of patients. In recent years accurate methods have been evolved for the evaluation of therapeutic agents in disease. These methods have been applied to rheumatoid arthritis in comparing cortisone to aspirin in patients with early‘ and late2 disease, and to an evaluation of chloraquin;3 however, they have not as yet been applied to the evaluation of gold therapy in this disease.q From the Department of Medicine, Cokcmbfa University College of Physiciuns and Surgeons, and the Edmard Daniels Fnulkner Arthritis Clinic oj the Presbyterian Hospitul, h’ew York, N.Y. S. L. W. is a trainee of the Notionul Institute of Arthritis and Metabolic Diseases. 20 THERAPEUTIC EVALUATION IN RHEUMATOID ARTHRITIS 21 Theoretically, the controlled clinical trial of therapeutic agents should be the only appropriate way of assessing the value of these agents in rheumatoid arthritis. The concept of the controlled clinical trial includes, however, a large number of variables that need to be considered separately. It is the purpose of this paper to discuss the various aspects of the clinical trial with particular reference to rheumatoid arthritis. The clinical evaluation of drugs can be divided into three phases: l., selection and characterization of patients to be treated; 2., technics for carrying out the study; and 3., methods of measuring the results of therapy. SELECTION AND CHARACTERIZATION OF P A m V T S To study the effect of a drug, it is necessary first to define the disease under treatment. In many studies in the literature, no criteria for selection of patients are listed. These omissions emphasize the need for more specific and generally accepted standards for the diagnosis of rheumatoid arthritis. Recently a committee of the American Rheumatism Association has listed diagnostic criteria for definite, probable, and possible rheumatoid arthritis.” These criteria are inescapably arbitrary. The presence of symmetric polyarthritis with the usually associated symptoms persisting for six weeks, and not caused by another specific disease process, would be enough to classify a given patient as definitely having rheumatoid arthritis. There are some who feel that these criteria are not sufficiently restrictive, and that it would be better to limit the diagnosis to patients with arthritis and nodules, positive agglutination tests, or widespread connective tissue lesions. The criteria of the American Rheumatism Association committee are useful in setting forth minimum standards for selection of patients in the study of therapeutic agents in rheumatoid arthritis. However, further delimitation and characterization of the patient population to be studied is desirable. The age and sex of the patient, duration of the disease prior to therapy, and results of serologic testing must be known. The presence or absence of nodules should be noted. As approximate measures of the severity of the arthritis, the classifications of progression (table 1) and of functional capacity (table 2) of the American Rheumatism Associationa are of value. Finally, some method of quantitation of the activity of idammation in the disease before therapy is needed. Lansbury‘ has reported a method for measuring rheumatoid activity, using a combination of three objective indexes (sedimentation rate, hemoglobin and the patient’s ability to compress an inflated blood pressure cuff) and three subjective indices (pain on motion, morning stiffness, and fatiguability). The latter symptoms are quantitated by determining the number of aspirin tablets taken daily by the patient for pain, the duration of morning stihess, and the number of hours of physical activity before fatigue occurs. However, such manifestations are subject to marked variation unrelated to the state of rheumatoid activity. The numbers determined by application of this method are interpolated in a standard table (table 3) and “percentages” calculated, which are averaged to give an index of rheumatoid activity. Joint inflammation is not measured directly in any of the indexes. Lansbury’s results, despite obvious sources of 22 STANLEY L. WALLACE AND CHARLES RACAN TABLE1.-Classification Stage I Roentgenologic Signs Muscle Atrophy 0 osteoporosis; sometimes no destructive changes. of Rheumatoid Progression Extra-articular Lesions (Nodules. tenovaginitis) Joint Deformity Ankylosis 0 0 0 osteoporosis; slight cartilage or subchondral bone destruction may be present. adjacent may be present 0 0 111 Osteoporosis; cartilage destruction; bone destruction. extensive may be present subluxation ulnar deviation, and/or hyperextension 0 IV same as I11 with bony ankylosis. extensive may be present same as I11 I1 .- fibrous or bony ankylosis (Reprinted from Steinbrocker, O., Tracger, C . H. and Batterman, R. C.: Therapeutic criteria in rhedmatoid arthritis. J.A.M.A. 140:659, 1949. ) TABLE2.-Classificution of Functiorml Capacity - Class I ~ _ _ Complete Ability to carry on all usual duties without handicaps I1 Adequate for nomial activities Despite handicap of discomfort or limited motion at one or more ioints 111 Limited Only to little or none of the duties of usual occupation or self care IV Incapacitated largely or wholely Bedridden or confined to wheelchair; little or no self care. (Reprinted from Steinbrocker, O., Traeger, C. H. and Batterman, R. C.: Therapeutic criteria in rheumatoid arthritis. J.A.M.A. 140359, 1949.) error, indicate a satisfactory correlation with clinical impressions of rheumatoid activity. We have recently begun to employ a modification of Lansbury’s method (table 4), which is simpler and less time consuming to use and so far, in a limited number of patients, has also been satisfactory in quantitating rheumatoid activity. This method also uses six manifestations of active inflammation in rheumatoid arthritis, substituting the presence and spread of joint inflammation (not deformity) in the patient for his ability to compress an inflated blood pressure cuff. Since weakness and fatigue are so extremely 23 THERAPEUTIC EVALUATION IN RHEUMATOID ARTHRITIS Morning Stiffness 5 10 16 20 30 45 Hrs. 1.0 1.5 2.0 2.6 3.0 3.6 4.0 4.6 5.0 5.5 6.0 6.6 7.0 7.6 8.0 .. .. .. .. .. .. .. .. .. 2 5 7 9 14 21 29 43 67 71 86 100 114 129 148 161 171 186 200 214 229 .. .. .. .. .. .. .. .. .. Fatigue Afkr 8.0 7.5 7.0 6.5 6.0 5.5 6.0 4.6 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 .. .. .. .. .. .. .. .. .. .. .. .. .. .. Aspirin D8Y Tabe. % % 0 14 29 43 61 71 86 100 114 129 143 167 171 186 200 214 229 .. .. .. .. .. .. .. .. .. .. .. .. .. .. TABLE3.-Thc I,unsbuy Systemic Zndex 0 1 2 3 4 5 6 I 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 .. .. .. .. .. 280 210 260 250 240 230 220 210 200 190 180 170 160 150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 .. .. .. 0 .. 0 6 12 19 25 31 37 44 50 56 0 0 0 0 0 81 87 6 12 19 26 31 37 44 50 66 62 94 100 106 112 119 125 131 137 144 160 166 162 169 176 69 76 81 87 94 100 106 112 119 125 131 137 144 160 62 69 76 .. ESR Anemia - 0 12 25 37 50 62 75 87 100 112 125 137 150 162 175 in7 200 212 925 237 260 262 276 287 300 812 .. .. .. 6-min. Cutler Grip Weakness mm. M% F% .. .. ow. M% 13.4 13.2 13.0 12.8 12.6 12.4 12.2 12.0 11.8 11.6 11.4 11.2 11.0 10.8 10.6 10.4 10.2 10.0 9.8 9.6 9.4 9.2 9.0 8.8 8.6 8.4 8.2 8.0 0 7 15 22 0 0 0 30 0 31 44 52 60 0 .. .. .. 0 89 0 9 19 29 33 48 67 96 67 104 111 116 126 133 141 148 166 163 170 185 193 200 76 86 95 106 114 124 133 143 153 162 111 181 190 200 .. .. .. .. 61 74 81 178 1.0 1.6 2.0 2.6 3.0 3.5 4.0 4.5 5.0 6.5 .. .. 6.0 6.5 1.0 7.6 8.0 8.6 9.0 9.5 10.0 10.6 11.0 11.5 12.0 12.5 13.0 13.5 14.0 14.5 16.0 15.6 16.0 WesterSed.Rate 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 160 160 170 180 190 200 210 220 230 240 260 260 210 280 290 300 10 15 20 26 30 35 40 46 60 66 60 0 8 17 26 as 110 116 120 126 130 136 140 146 160 42 60 68 67 76 83 92 100 108 117 126 la8 142 160 158 167 176 183 192 200 208 217 226 233 .. .. 66 70 75 80 86 90 96 100 106 .. .. The percentile equivalenta are given in the nearest whole numbers and the scales are arranged b~ taking the average value for each item as 100 with 0 as normal. The Westergren ESR scale b based on data furnished by Drs. Ziff, Duff and Galindez. The scale for hemoglobin valuea wan established by taking extreme values ( 0 and 200%) and arranging the scale to agree. (Reprinted from Iansbury, J. : Quantitation of the activity of rheumatoid arthritis. 5. A method for summation of the systemic indices of rheumatoid activity. Am. J. Med. Sc. 4S2:300, 1956.) subjective, the examiner's impression as to their presence and severity is used rather than an attempt to measure them. These evidences of rheumatoid activity are divided only into four degrees of severity-severe, moderate, mild, or not present-are given point values and are added to give an index of activity. These two methods, like others suggested for summating rheumatoid activity, are entirely empirical. Extensive experience is necessary before either or both can be shown to have intrinsic value and before they can be used in the characterization of patients with rheumatoid arthritis and in the evaluation of therapy. TECHNICS FOR CARRYING OUT THE STUDY The generally accepted methods8J'Jofor studying the effects of new drugs include: less than 30 minutes 30 minutes or more; less than 3 hours 3 hours or more MILD ( 1 ) MODE RATE (2) SEVERE (3) pronounced moderate !.light none Fatigue & Weakness more than 8 ASA tablets with poor relief 8 ASA tablets or more with good relief less than 8 ASA tablets with good relief none Joint Pain persistent joint swelling, redness, heat; many joints involved persistent joint swelling, redness, heat; few joints involved transient joint heat, redness and 'or swelling _ _ I _ Il(;Ilc Acute Joint Manifestations Add total points: 1 for each in the mild category, 2 for moderate, and 3 for severe. none Morning Stiffness NONE ( 0 ) Activity TAFILE4.-A Simplified Systemic Index over SO inin. hr. 31-50 mm.,hr. 1.5-30 nim. 'hr. nornial Sedimentation Rate (Westergren) below 9.5 Gni. 95-11 Gm. 11-12.5 Gni. nornial Hemoglobin THERhpEUTIC EVALUATION IN RHEUMATOID ARTHRITIS 25 1. Prospective rather than retrospective study. 2. Randomization of patient assignment to the various therapeutic groups, so that these groups may be comparable in all respects. In some studies, subgroups are set up prior to the allocation of patients to therapy to ensure the strict equivalence of all groups. 3. Double or triple blind testing to eliminate observer error of assessment. In such testing, neither the patient nor the observer is aware of the material used. The agent to be assessed, a placebo, and preferably a standard of reference (an agent with known therapeutic activity) should be used. All agents should be used in all patients, and the drug and placebo should be administered to the various patients in random order. 4. Statistical evaluation of results. Not all these standard technics are feasible in the study of drugs in rheumatoid arthritis. The long-term administration of placebos known to be ineffective in this disease cannot be justified ethically. This leaves only a comparison between the drug to be studied and a drug whose effectiveness is known. Another cause of difEculties in the assessment of drugs is the variability of the natural history of both untreated and conservatively treated rheumatoid arthritis. Approximately one-quarter of patients with this disease follow an intermittent course with spontaneous remissions alternating with periods of active arthritis." In two large s e r i e ~ ' ~ about , ~ ~ 50 per cent of conservatively treated patients with rheumatoid arthritis showed improvement on 9 to 11 year average follow-ups. Any drug studied, in order to be assessed properly, would need to be administered over a long period of time. It would be dif6cult to evaluate in the same patient the effects both of the drug under study and of the standard of reference. In addition, the variability of the disease makes it difficult to set up a therapeutic study with treatment groups that are comparable in all respects. To ensure equivalence, subgroups would have to be created to include such variables as age, sex, duration of disease prior to therapy, presence or absence of nodules, results of serologic testing, classifications of progression and inflammatory activity of the disease and functional capacity of the patient. If such subgroupings were not made prior to therapy, then the patients treated with the unknown agent and those treated with the standard of reference would have to be shown to be similar in these respects. A large number of patients with rheumatoid arthritis, necessarily from many clinics, would be required in order to obtain information of possible statistical significance. The subgroupings would need to be done at each center. Even then, the likelihood of multiple observer bias and error in a multicenter study is present. The question has been raised of the need for blind comparison of the agent whose therapeutic effecbveness is known and the one with unknown activity. Lasagna9 and Hilllo maintain that with rigid safeguards it is possible to obtain significant information from studies not using the double blind technic. In such studies there must be complete reliance on objective criteria in the measurement of results. Unless strictly objective standards are used, a thera- 26 STANLEY L. WALLACE AND CHARLES RAGAN peutic evaluation not using the double blind technic may suffer from marked observer error. In many studies in the past, no control group has been employed, but attempts have been made to compare the results of therapy in one group of patients with results from an untreated or differently treated group from the literature or from the earlier files of the same institution. This comparison is usually statistically invalid because comparability of the treated and supposed control groups cannot be established. METHODSOF MEASURING RESULTS The point in time at which appropriate measurements should be made must be determined before the results of therapy can be evaluated. Ideally, only end results should be considered, but in a disease as chronic and as capricious as rheumatoid arthritis, this is not practicable. The arbitrary selection of one, two or three year end points, however, may give too little consideration to the natural variability of the disease. Many ways of measuring the results of therapy in rheumatoid arthritis have been used. Changes in individual objective phenomena (sedimentation rate, hemoglobin, joint tenderness, joint range of motion, strength of grip, time to climb stairs or tie knots) and subjective phenomena (morning stiffness, pain, fatiguability ) have been measured. Sociologic changes as in employment status have been considered. The American Rheumatism AssociationGhas established a classification of response to therapy that unfortunately suffers from the fact that it considers simultaneously changes in inflammation and in permanent joint deformity. These need not parallel one another. A modification of this classification based on new data is now in process. Three separate major responses to therapy should be recorded; these are changes in the degree of progression of the rheumatoid arthritis (table l), in functional capacity (table 2 ) and in activity of inflammation (tables 3 and 4), if this last index proves reliable with further experience. It must be noted, however, that the measurement of these changes involves in large part the consideration of subjective rather than objective phenomena. This becomes an obstacle to the proper interpretation of results in a drug evaluation not done by the double blind technic. In addition to the changes in progression, functional capacity and activity of inflammation, note should be made of changes in the individual objective findings listed above. Changes in x-ray findings need to be considered separately. In order to remove the possibility of bias, the x-rays should be read by radiologists not acquainted with the findings in the patient or his therapy.10 The importance of serial joint x-rays in the evaluation of treatment, and the possibility of radiologic evidence of progression of rheumatoid arthritis in the face of clinical quiescence, has been emphasized.14 Data resulting from the study would then be analyzed statistically, and the significance, if any, determined of possible differences between the results THERAPEUTIC EVALUATION IN RHEUMATOID ARTHRITIS 27 of therapy in the groups treated with the known agent and with the unknown. CONCLUSIONS The difficulties in the controlled clinical evaluation of therapeutic agents in rheumatoid arthritis have been considered. It is not likely that an ideal, controlled study can be designed at this time. Nevertheless, valuable information may be obtained from studies embodying the following principles: ( 1.)complete comparability of therapeutic groups; (2.) blind comparison of the untested agent with a standard of reference; (3.) appropriate methods of measuring response to therapy; and (4.)statistical evaluation of results. If the comparison of the two drugs is not blind, then it is necessary that the methods of measuring results be completely objective. In essence, this approximates the procedures used by the Medical Research Council-Nuffield Foundation' and Empire Rheumatism Council2 studies in Great Britain in comparing the effects of aspirin and cortisone in rheumatoid arthritis. Even with a well set up study, the variability and chronicity of rheumatoid arthritis and the doubt as to the proper end pciint at which to make measurements make it uncertain that the resulting information would be valid. Whether the information obtained warrants the enormous expenditure in time, facilities, and personnel implicit in such a study is also questionable. There is, however, no easier way to obtain information which may be of value. REFERENCES 1. Joint Committee of the Medical Research Council and the Nuffield Foundation: A comparison of cortisone and aspirin in the treatment of early cases of rheumatoid arthritis. British Med. J. 1~1223,1954. 2. Empire Rheumatism Council: Multicentre controlled trial comparing cortisone acetate and acetylsalicylic acid in the long-term treatment of rheumatoid arthritis. Annals Rheum. Dis. 14~353,1955. 3. Cohen, A. S. and Calkins, E.: A controlled study of chloroquine as an antirheumatic agent, presented at the Ninth International Congress on Rheumatic Diseases, Toronto, Canada, June 26, 1957. 4. Freyberg, R. H.: The place of gold compounds in the treatment of rheumatoid arthritis. J. Chronic Dis. 5723, 1957. 5. Ropes, M. W.,Bennett, G. A., Cobb, S., Jacox, R. and Jessar, R. A.: Pro- posed diagnostic criteria for rheumatoid arthritis. J. Chronic Dis. 5330, 1957. 0. Steinbrocker, O., Traeger, C. H. and Batterman, R. C.: Therapeutic criteria in rheumatoid arthritis. J.A.M.A. 140~659,1949. 7. Lansbury, J.: Quantitation of the activity of rheumatoid arthritis. 5. A method for summation of the systemic indices of rheumatoid activity. Am. J. bled. Sc. 232~300,1956. 8. Beecher, H. K.: Experimental pharmacology and measurement of the subjective response. Science 116~167, 1952. 9. Lasagna, L.: The controlled clinical trial: Theory and practice. J. Chronic Dis. 1~353,1955. 10. Hill, A. B.: The clinical trial. New England J. Med. 247~113,1952. 11. Reynolds, W.A. and Short, C. L.: The clinical manifestations of rheumatoid 28 STANLEY arthritis. Med. Clinics N. A. March 1955, p. 365. 12. Ragan, C.: The general management of rheumatoid arthritis. J.A.M.A. 141: 124, 1949. 13. Short, C. L.and Bauer, W.: The course of rheumatoid arthritis in patients re- L. WALLACE AND CHARLES RAGAN ceiving simple medical and orthopeJ. Med. dic measures. New England 238:142, 1948. 14. Bollet, A. J. and Bunim, J. J.: The importance of serial x-rays in the evaluation of treatment of rheumatoid arthritis. Med. Clinics N. A. March, 1955, p. 439. Stanley L. W a h c e , M.D., Trainee, National Institute of Arthritis and Metabolic Diseases; Clinical Instructor, Department of Medicine, State University of New York College of Medicine at Brooklyn. Charles Ragan, M.D.,Associate Professor of Clinical Medicine, Columbia University College of Physicians and Surgeons; Associate Attending Physician, Presbyterian Hospital, New York, N.Y.