Report of a three-year study on the systemic and articular indexes in rheumatoid arthritis.код для вставкиСкачать
Report of a Three-Year Study on the Systemic and Articular Indexes in Rheumatoid Arthritis Theoretic and Clinical Considerations By JOHN The author’s methods of systemic and articular indexes have been an important advance in the field of evaluating disease activity and response to therapy in patients with rheumatoid arthritis. The present paper summarizes his experience with these methods and his present views on this controversial subject. LANSBURY Per su methodologia del indices systemic e articular, le autor ha promovite importantemente le evalutation del activitate marbide e del responsa therapeutic in patientes con arthritis rheumatoide. Le presente articulo summarisa su experientias in le us0 de su methodos e su vistas currente con respecto a iste thema controverse. I N THIS PAPER we propose first to discuss some of the fundamental principles which should be taken into account in devising any scheme of evaluation of the status of patients suffering from rheumatoid arthritis and to report on a three-year study of the Systemic and Articular Indexes in order to see how far they go toward solving the problem. The ultimate aim of drug therapy in rheumatoid arthritis is to control the disease process at the most fundamental level of its pathogenesis. The need for accurate methods of assessment is obvious and will probably become more pressing with the expected introduction of new antirheumatic agents. When the etiology of the disease is better understood, we may be in a position to devise laboratory tests which will tell us not only how well, but at what level, the activity of the disease is being controlled. But, until then, we shall have to continue to rely on indirect, clinical assessment of the inflammatory manifestations of the disease and of the slowly progressive structural changes which accompany them. Evaluation of therapeutic responses to drug therapy of rheumatoid arthritis may be carried out in two quite different but equally important ways. These are: 1.) the detection of an antirheumatic effect by statistical analysis of responses to a given drug in a large group of cases observed for a short time in a multi-clinic situation, and 2.) serial numerical quantitations of rheumatoid disease in individual patients over a period of years. When such observations are graphed, one has a picture of the course of the disease as it is affected by treatment and other factors which permits the physician to judge the effect of drugs accurately and to plan his treatment accordingly. Both of these approaches require an accurate, quantitative method of evaluation. SOME BASICCONSIDERATIONS An acceptable method for evaluating a case of rheumatoid arthritis should meet the following requirements: It should be capable of being performed by ~ From the Arthritis Unit of the Department of Medicine, Temple University School of Medicine, Philadelphia, Penna. 505 506 JOHN LANSBURY any doctor; it should require only a few minutes to carry out; it should require no special apparatus or complicated laboratory procedures; it should be applicable to the great majority of cases; its results should be capable of being expressed as a numerical value; it should be shown to reflect accurately the general state of rheumatoid disease; and the degree of inter-observer error should be insignificant. In table 1 are listed most, if not all, the clinical manifestations of rheumatoid arthritis which might conceivably be used in forming an opinion as to the status of a rheumatoid patient. It will be noted that some of these reflect the activity of the disease while others reflect the structural changes caused by or associated with this activity; some are closely related to the joints themselves, while others are more general or “systemic” in nature. In devising a practical scheme for evaluation, it at once becomes apparent that a21 these manifestations cannot be taken into account. It is likewise clear that no single item can, of itself, be regarded as a reliable guide to all the others. Thus one must select a group of crucially important items and then find a way of summing up or averaging the information which they furnish. It is mandatory that only those items which can be accurately defined, measured and referred to a standard base line should be taken into consideration. Poorly defined clinical impressions should be avoided at all costs. This rule applies equally to both articular and constitutional findings. As clear a distinction as is possible should be maintained between those items which reflect the inflammatory activity of the disease and those which reflect permanent structural changes. (In the case of restricted joint motion and general functional disability this distinction may be hard to make because it is often difficult to know in advance whether such restrictions are due to TABLE1 . 4 h s e m n b l e Items in Rheumatoid Arthtjtis A. QUANTITATNE ITEMS FOUND IN ALL GRADES OF HHEUMATOIU ACTIVITY: Stiffness, fatigue, pain, muscle weakness, rapid sedimentation rate, total amount of joint inflammation and joint dysfunction. B. SIMILAR ITEMS WHICH ARE NOT ALWAYS PRESENT IN ACTIVE DISEASE: Weight-loss, fever, tachycardia, anemia, leukocytosis, leukopenia, increased globulins, decreased serum albumin, low B.M.R., positive C-reactive protein and other laboratory tests. C. QUALITATIVE ITEMS OF DUBIOUS VALUE IN ASSESSING DISEASE ACTIVITY: Depression, anorexia, achlorhydria, hydremia, paresthesias, cyanosis, blanching, clamminess, muscle twitching and spasm, serologic tests such as: sheep-cell agglutination, latex and bentonite fixation, etc. D. ITEMS INDICATING ANATOMIC CHANGES CAUSED BY I)ISEASE ACTIVITY: Muscle-atrophy, tenosynovitis, tendon rupture, joint contractures, fibroiis ankylosis, osteoporosis, cartilage erosion, bone absorption, bony ankylosis, rheumatic nodules. vasculitis, pneumonitis, valvulitis, nephritis, neuritis, lymphadenopathy, splenomegaly, iridocyclitis, recurrent conjunctivitis, amyloidosis. SYSTEMIC AND ARTICULAR INDEXES: &YEAR STUDY 507 current disease activity or to deformities resulting from previous disease activity.) Almost every one of the items listed in table 1 is open to some sort of objection; but this must be accepted as a brute fact, for, unfortunately, there are no other, more perfect items which will do the job for us. In the case of items listed under “A”, our quantitations of pain, stiffness and fatigue are, we think, no less accurate than are clinical impressions of gradations in joint inflammation and, until proof to the contrary is forthcoming, we shall continue to hold this view. We believe that it matters little whether a given item is classified as “subjective” or “objective.” The important thing is the accuracy with which it can be observed. Theoretically, the greater the number of items taken into consideration, the more closely would the result approach an “absolute” measure of disease activity, but, on the other hand, an attempt to observe a large number of items may defeat its own purpose because of sheer complexity and unwieldiness. In a recent attempt to use 18 items in the clinical assessment of a drug, Howell and Vaughn’ found that only those listed under Category A in table 1were useful and reliable. For general use we therefore recommend this small sample of the many possible parameters. They have the great advantage of having been thoroughly studied and tested, and, as will be seen, have been shown to function well both in the long-term study of individual cases and in multi-clinic drug trials. Before the data which any acceptable group of observable items may furnish can become usable, a way must be found for putting them together in a standardized manner. The following three methods suggest themselves: 1. A matching system in which cases are allocated to plausible but arbitrary categories, as in the present official A.R.A. scheme. 2. A point system, using a scale of, say, 10 crucial items, the grade being determined by adding up the number of positive items. (Functional status and anatomic stage might be thus graded, but we feel sure that disease activity cannot be so graded). 3. An index system in which all information is quantitated by reference to an objectively determined numerical standard. (We believe that diseaseactivity can best be determined by this method.) In a multi-clinic drug testing program with the limited aim of detecting a qualitative antirheumatic effect in a statistically significant number of cases, one might be content to record each item as merely showing a plus, minus or zero change. In this situation special items from categories B and C might be included (also special measurements, such as ring sizes: depending on the effect which one might be looking for, and there might well be no need for the methods of summation (which will be presented later), since it would be advantageous to deal with each item separately. But in following individual patients over long periods of time the situation is different. Here 508 JOHN LANSBURY one must have a quick method based on the clinical features of the disease which uses only the simplest kind of calculation for summation. The Indexes fulfil both these requirements. Whatever scheme of summation is used, whether “matching,” “points,” or “indexes,” the information which it furnishes will naturally fall into each of two major categories: 1.) “Grade of Inflammatory Activity,” and 2 . ) “Anatomic Stage of the Disease.” But in what follows we shall be concerned mainly with the ability of the Indexes to accurately reflect gradations of rheumatoid activity. The possible applications of the Articular Index to measurements of joint function and anatomical stage have been dealt with elsewhere:’,’ and will not be considered here. Neither shall we discuss the role of the items under Category D (table I), in assessing the anatomic stage of rheumatoid disease, except to say that we think that extra-articular changes should also be taken into account. Amyloidosis or renal disease are, we think, at least as important as are ankylosis or restricted joint motion. THE ROLE OF THE INDEXESIN EVALUATION The Systemic Index For a detailed description of the Systemic Index, the reader is referred to previous article^.^.^ In essence, it consists of a quick ( 2 to 5 minute) objective quantitation of the five cardinal manifestations of rheumatoid activity: stiff ness, pain, fatigue, muscle weakness and increased sedimentation rate. These are present in some degree in all phases of active disease and are quantitated, respectively by: duration after rising, aspirin need, time of onset after rising, grip strength and either Cutler or Westergren sedimentation methods. Their values are summated by referring to table 2. The index is especially useful in following the trend of the disease and its responses to therapy in individud patients. Among those who have applied the method there is general agreement that the items can be accurately elicited. Dif6culties at first encountered with measuring grip strength have been eliminated by using a standardized sphygmomanometer cuff and by being sure that the valve at the top of the mercury column is open. A recent study by Mikkelsen showed a high degree of inter-observer reliability for this item.6 Identification of the time of onset of fatigue is the most difficult item to be sure of, but, so far, we have been unable to find a more objective method of qnantitating this highly important item. The other items have presented no serioia problem. The proof of the accuracy of this index (or of any other method of evaluation) is, of course, of the greatest importance. It is also very difficult because there is no direct way by which the accuracy of m y method can be tested. Testing one method against another is valid only if neither is subject to the same kind of error. One’s natural impulse is to match information derived from a given method against one’s clinical impression. But this is rather illogical because one is then using as a criterion the very thing which the method is designed to avoid! Despite this dilemma, our three year’s experience convinces us that clinical impressions, the patient’s own evaluation and com- 509 SYSTEMIC AND ARTICULAR INDEXES: 3-YEAR STUDY parison with other methods, can and do, when taken together, give convincing proof of the accuracy of the Systemic Index. The supporting evidence for this contention will now be briefly presented. 1.) We do not remember having observed a change of 20 or more points in the index which went contrary to our own or the patient’s impression of improvement or worsening. 2.) Since a full clinical remission is clearly identifiable, it is possible to trace the serial regression of activity, as measured by the Systemic Index, towards this point and to compare such trends with those of other cases pursuing a static course. Numerous, impressive examples of this steady type of decline have been published.3 3.) The Index registers major exacerbations, such as would be expected, following surgical operations, acute infections and severe emotional stress. Examples of such temporary exacerbations are given in figure 1. Lesser exacerbations were also found to be significant in a series of 15 cases taken in alphabetic order. Twenty consecutive rises of 20 points or more, each of which had occurred in the space of one month or less, were studied in retrospect. A definite exacerbating factor had been recorded in 18 (90 per cent) of these, and only two were r STUOY OF 4 MAJOR EXACER8ATlONS AS REVEALED BY THE SYSTEMIC B ARTICULAR tNOEXES- a-* SYSTEMIC INDEX o------ ARTICULAR INDEX I d FIG. 1.-These four graphs show how both the Systemic and Articular Indexes registered major exacerbations in disease activity in situations where such an exacerbation could be predicted. The exacerbating factors, from left to right were: unaccustomed severe physical activity while exposed to cold; escape from suppressive action of gold salts which were withdrawn five months previously because of toxicity; the last two exacerbations followed about one month after a hysterectomy in each instance. In all graphs the parallelism of the two Indexes is obvious, each thus validating the other. 510 JOHN LANSBURY unexplained. The following factors were noted in order of frequency: emotional upsets (4),infection (4),steroid reduction ( 4 ) , stopping gold and chloroquin ( 2 ) , and one case of each of: chilling, surgery, severe menorrhagia and escape from steroid control. Thus, almost all the rises occurred in situations where a rise would be expected-a fact which further suggests that the Index correctly registered moderate flares of disease activity, even when we take into account the known weakness of retrospective studies. 4.) The Indexes also registered response and relapse associated with a known antiDATA FRDH OCWELE-ELIND MULrl-CLINIC S Z W Y IMCX X 3 I I FIG 2.-The above graph indicates the result of a study which was carried out in 1955 to test the ability of the Indexes to register the effect of a drug of known potency in a multi-clinic trial. Observations were made on 21 patients by 15 doctors working in eight arthritis clinics in seven states. The standard double-blind technic was used except that the patients were not randomized. Patients were selected in inatching pairs as far as possible, and drug A administered to one, and drug B to the other. After one week, the tablets were switched without the knowledge of the patients, and a final observation was made at the end of another week. The test drug was prednisone given in dosage of 5 mg. 3 times daily, and the control drug was an identical-appearing tablet composed of an inert substance. The two groups of patients are categorized as follows: 8 cases on sequence B-A, with an average age of 46 yrs., duration of disease 8.7 years, anatomical stage 3.8 and functional class 3.8 by the A.R.A. classification. Three were males and five females. The group of 13 cases receiving drugs in the A-B sequence was not as closely comparable as could be wished. The average age was 52.1 years, duration of disease 7.3 yrs., anatomical stage 2.2 and functional class 2.1. Six were males and seven females. It is obvious from the above graph that drug B (prednisone) was registered by the Indexes as showing a significant lowering of activity. The more striking results in the B-A group as compared with the A-B group is difficult to explain. In general, the interobserver accuracy of the Systemic Index seems to he superior to that of the graded Articular Index. ( N o change was noted in the simple spread of the disease in either g r o u p t h i s was probably because of the short time interval). Of 29 patients who stated that they felt better, 25 showed a lowering of the Systemic Index and only four an increase, the average decrease being minus 33 points. Of eight patients claiming to be worse, six showed a rise in the Systemic Index and two a fall. The average was plus 29 points. Of 22 patients stating that they felt no change, the average change was plus 0.31 points. Thus, the Systemic Index also agreed, on the average, with the patients' own evaluations. 511 SYSTEMIC AND ARTICULAR INDEXES: 3-YEAR STUDY inflammatory agent ( prednisone ) administered in a “doubIe-blind” situation, as illustrated by the graphed data in figure 2, and also accurately reflected the patient’s impressions of change for better or worse. (These data were provided by the Committee for Evaluation of New Therapeutic Agents of the A.R.A.). 5 . ) Similar responses have been registered by controlled studies in groups of cases receiving gold salts, camoquin and chloroquin; a negative response was also found in a double-blind study using manganese as the test drug. The bar graphs in figure 3 summarize the results of these various studies, Group -Responses lo Six Drugs Meosured by ?he Indexes GOLD MEDVISONE CAMOQUIN ARALEN 102 r”] .....:.:. TEST8 CONTROLS :::::.:.: TEST BUTAZOL IDWE MANGANESE I03 CONTROL FIG.3.-In the above bar-graphs, the initial values for both test and control groups of patients have been brought to a value of 100 and subsequent changes calculated as percentages. Since the various drugs were used in various doses and over varying periods of time, caution is advised in interpreting their relative therapeutic merits. 1. Sodium-Auro-Thio sulfate, 25 mg. per week for (average) 10 months in 12 consecutioe cases. Personally treated by the author. No controls. 2. Prednisone, 15 mg. per day for one week in 21 patients. Data from a double-blind, multi-clinic study carried out by the Committee for the Evaluation of New Therapeutic Agents of the A.R.A., 1957. 3. Camoquin, 200 mg. a day for eight months. Part of a double-blind study in progress involving 14 patients, Temple University Hospital. 4. Arden (chloroquin) 500 mg. a day for 10 weeks; part of a double-blind study involving 21 patients. Reported by Cohen and Calkins’ at the 9th International Congress on Rheumatic Diseases, Toronto, 1957. 5. Phenylbutazone (Butazolidine) 360 mg. per day (average) dose in 20 patients for five months. Part of a comparative study of cortisone, aspirin and phenylbutazone in which a modification of the Articular Index, rather than the Systemic Index, was used. Smyth and Clark, 1957.8 6. Manganese (glycerophosphate) 300 mg. a day for two months involving 18 patients (double-blind study) by Bepler and Rogers, 1957: 512 JOHN LANSBURY and suggest that the Systemic Index functions well in collecting data on therapeutic responses by multiple observers. 6.) In figure 4 successive relapses are recorded following serial reductions in prednisone dosage, thus showing the ability of the Systemic Index to reflect flares following decrements of as little as 1.25 milligrams of this drug. The above 6 examples, taken together, provide impressive evidence that the Systemic Index correctly reflects changes in disease activity. But even stronger evidence for its accuracy can now be furnished by a study of the parallelism between the Systemic and Articular Indexes. Since these are derived respectively from entirely different types of observations they are not subject to the same kind of error. Thus, each acts as a valid check on the other. This parallelism will be illustrated and discussed after a brief description of the Articular Index. The Articular Index In order to estimate the total amount of joint inflammation, a simple count of involved joints is not enough. Joint size must be taken into account, for as figure 5 shows, the amount of inflammation that can be contributed by a knee is many times (25 times to be exact) greater than the amount which could possibly be contributed by a knuckle. The Articular Index is simply A CASE OF RH€UMATOlD ARTHRITIS UND€R TREATMENT Flores Following Sieroid Reduction: A - 8 IOmg., E-F l.25mg., I-J l.25mg. - L -Ml.ZSmg. I20 -sYsr..mINDEX 60 E 0 -I -1-1-1- 2 I- -1-1 I- 6 I2 19 I6 18 MONTHS lMPROEA4E"TS.' 8-C hospiiolizothn. 8 - I gofd iherapx 0- P Chloroquin iheropy 9 6 FIGURE 4 SYSTEMIC AND ARTICXJLAR INDEXES: $YEAR STUDY 513 the sum of the articulating areas of all involved joints and takes even less time to calculate than does simply writing down their names. As to accuracy, our figures for normal joint sizelo cannot, we think, be seriously questioned. However, summations of total articular inflammation can only be as good as are our methods of joint evaluation. Such evaluations should therefore be based on simple and absolutely clearly defined end-points. At present little is known about inter-observer accuracy in grading joint inflammation, nor is there likely to be general agreement until we use standardized end-points which are at least as objective as are those which now constitute the Systemic Index. Smyth and Clarks have reported valuable explorations in this direction, but in our work we have so far been content to merely identify the presence or absence of joint inflammation without attempting to grade its severity. We have used as our end-points tenderness and pain on passive motion. Thus, we disregard subjective arthralgias FIG.5 . T h e s e x-rays of a knuckle and a knee illustrate the great differences in joint size. Obviously a knuckle would contribute far less than would a knee to the totality of joint inflammation. Therefore, joint size should be taken into account in estimating the total “amount” of joint inflammation. Since the great majority of involved joints manifest an “average” degree of inflammation, the total amount of joint involvement, as determined by the Articular Index, is a measure of rheumatoid activity even without taking into account gradations in the severity of inflammation in individual joints. 514 JOHN LANSBURY in normal-appearing joints, also residual thickening or contractures in joints which are free of pain or tenderness. As rheumatoid arthritis evolves, the pattern of spread to involve successive new joints is well known. Such a spread can be recorded by the Articular Index. Less well known is the fact (which our studies have demonstrated) that as rheumatoid arthritis improves, “ o l d joints successively become free of involvement, the signs of inflammation falling below the level of clinical detection. Such a recession can also be recorded by the Articular Index, which thus becomes a measure of the fluctuations in total joint inflammation and, in fact, an alternate method for evaluating rheumatoid activity. This is illustrated in figure 6. By using these simple end-points we avoid the difficulties inherent in trying to grade joint inflammation. Whether it would also be worthwhile to give only a half value for joints where the activity is doubtful and an added half value in the case of severely inflamed joints manifesting rest pain, or severe fluid accumulation and heat must be determined by further study. A pilot study in our clinic indicated that the great majority of affected joints fall into the category of “average” inflammation. PARALLELISM OF SYSTEMIC AND ARTICULARINDEXES In comparing the Systemic and Articular Indexes, we are really comparing total body inflammatory response with total articular inflammation. Theoretically, there need be little correlation between the two, since rheumatoid disease is so variable. The following two extreme cases illustrate this fact. Case 1 is that of a 38 year old extxutive whose Systemic Index has varied betwen 50 and 60 for three years, but whose Articular Index ( X 3 ) has slowly risen from 64 to 153 over the same period of time. These figures clearly express what obviously took place-a major increase in joint involvenient with little or no change in general nianifestations. This type of rase is, however, unique in our records. Case 2. A young woman giving a history of previous multiple, severe joint involvemmt now shows objective signs of arthritis in only one knuckle. Nevertheless, she has persistent stiffness, diffuse (fibrositic) pain, fatigue, weak grips and an increased E.S.R. Her Articular Index ( X 3 ) was only 1.2 but her Systemic. Index was 58. This case is also unique in our records. These glaring discrepancies need not alarm us. They are no greater than those which occur between simultaneous records of temperature and pulse rate in certain special clinical situations, and they do not alter the fact that, in general, there is a fairly constant relation between the two. It is therefore not surprising that we found a significant correlation between 80 random pairs of Systemic and Articular Indexes in a retrospective study.ll Since then we have made simultaneous records of the two Indexes at monthly intervals in 20 private patients. Of these, 15 (75 per cent) show significantly parallel trends. Of the remaining 5, two are quoted above and in the other three the disparity could be attributed respectively to variable steroid administration or to paucity of joint involvement but not to faulty registration. These parallel trends are illustrated in figure 7, which clearly indicates that the long-term trend of the disease is well reflected by both Indexes. In figures 8 SYSTEMIC AND ARTICULAR INDEXES: %YEAR STUDY 515 FIG. 8.-Progression and regression of arthritis registered by the Articular Index. To understand the significance of the above graphs it is important to realize that they register only the “spread” of the disease and that they do not take into account any gradations in joint inflammation. Case 1 shows a continuing spread involving new joints as the disease progresses, a phenomenon with which we are all familiar. But in case 2 the opposite is taking place. Here the patient is going into remission and previously involved joints are successively becoming free of any detectable sign of active inflammation. Thus the Articular Index quantitatively registers both the progression and regression of joint inflammation in rheumatoid arthritis and so provides an alternate method for assessing the activity of the disease. and 9 the parallelism in 12 cases is so obvious as to constitute a crucial proof that both Indexes independently register the same quantitative information. This concludes a survey of evidence for the accuracy of the Indexes. It seems clear that they reflect the course of rheumatoid activity with acceptable exactitude, and we have no hesitation in saying that the biometric information which they furnish compares favorably with the reliability of other, standard quantitations, such as basal metabolic rates, oral temperatures, blood counts and many blood chemical determinations in relation to other diseases. Also, they are quick and easy to apply, although unfortunately the tedious explanations of their rationale cannot help giving the opposite impression. Tables and instructions for using them are given on pages 519 and 520. IMPROVING THE INDEXES Although the Indexes as they now stand fulfil the criteria laid down in the first part of this paper, a simplification of calculations would be desirable, 516 JOHN LANSBURY COMPARISON O f SYSTEMIC AND ARTICULAR INKXES AT ONE YEAR INTERVAL 160 srsrmic INDEX ARTICULAR INDEX s r s r E u i c INDEX ~ ARTICULAR INDEX 14 \ , \ ‘. 100 60 4.--. , 0 L FIG. 7.-From left to right, the first two columns show improvement a s registered by both the Systemic and Articular Indexes in sev(w cases over a period of one year. Thesc cases were selected according to the following criteria: none had had steroid therapy, and all showed a 50 per cent or better improvement as registered by the Systemic Index. The last two columns show the course of five cases in which no improvement was apparent. The criteria for their selection was othrrwise the smw. In both grniips of cases hoth Indexes gave essentially the saine informtition, etic.11 thus validating the cther. provided it does not introduce a significant degree of error. In tables 2 and 3 we present what we believe to be the maximum possible simplification with only a minimal error. In the case of the table for the Systemic Index the percentages are given at 1/5th their previous value, thus making it unnecessary to divide their sum by five and at the same time making addition easier by substituting smaller numbers. The error introduced is not more than plus or minus 3 points for the full scale of numbers. We have also changed the scale for fatigue from an 8 to a 12 hour span and have merged the slightly different values for male and female grip strengths into a set of single values. In the case of the Articular Index the streamlined values given in table 3 introduce an error of plus or minus 2.5 per cent. They also cause some “skewing” which can be corrected by multiplying by a factor of 1.2. For routine clinical use one need not make this correction. Wallace and Ragan,I2 in a comprehensive review of the problem of evaluation, have tentatively suggested a radical simplification of the Indexes in which they combine both Systemic and Articular findings. Their method of calculation consists of categories which are added up on a point system. Unfortunately, in thus passing from a continuous to an interrupted scale, an SYSTEMIC AND ARTICULAR INDEXES: 3-yEAR STUDY 517 r 'T I/40 T I 4 "i" "p 9" 4f 20 I 0 I T 4" 40 I 20 I 0 FIGS.8 (top) AND 9 . T h e s e 12 cases, most of which were observed for about one year, show a striking parallelism between the Systemic and Articular Indexes. All are on the same scale. The Systemic Index is registered as , the Articular as 0- - -0- - -0- - - . About 75 per cent of our cases show this degree of parallelism, and obviously, each Index is rcflecting essentially the same trend of rheumatoid activity. .--.-- -. I 518 JOHN LANSBURY inadmissibly large mathematical error (plus or minus 30 points) is introduced. This was clearly shown when our data were recalculated according to their method. Their method has the fuither disadvantage of substituting an undefined clinical impression for fatigue, weakness and spread in place of the much more accurate and objective measurements which we have made available for these three items. In the last three years we have collected enough data on the Indexes to make possible, for the first time, a study of the effect of the various combinations, substitutions and omissions of the six component items. We give below a brief summary of some exploratory studies in this direction which are intended merely to serve as examples: Combination.-Incorporation of the Articular Index (as a measure of total joint inflammation) as a sixth item in the Systemic Index has the theoretic advantage of providing a broader base for estimating “total rheumatoid activity.” But in applying this change to 84 sequential pairs of observations in 10 long-range cases, we found only minimal (average 4 points) differences, and these hardly ever altered the trend of the Systemic Index itself. In the few cases showing large deviations between the two Indexes, we found that a simple average of the two, while theoretically measuring “total rheumatoid activity” more accurately, actually concealed information as to whether the rheumatoid process was preponderantly articular or systemic. So, for the present, we are continuing to chart the Indexes separately on our own records. Substitution.-In cases where any one of the five items constituting the Systemic Index cannot be elicited, the values for the Articular Index as given in table 3 (XO.24 to bring them to scale) may be substituted. For example, a substitution of the Articular Index for fatigue was made in 74 consecutive observations in eight patients observed for long periods of time. The maximum discrepancy was 12 points, and the average for all of the variations was 6 points difference. The trends were unchanged except for minor discrepancies which were found in only 12 isolated, scattered instances. Omission.-We have also experimented with a limited Index composed of the following four items: duration of morning stiffness, grip strength, sedimentation rate and Articular Index. In a series of 129 sequential observations in 13 unselected cases we compared this limited, 4 point Index with the Systemic Index, with the following results: Although there were considerable variations in amplitude (averaging 10 points) of the curves as compared with the Systemic Index, the trend between any two consecutive observations was divergent in only 18 instances. The general impression given by looking at the graphed data was that the over-all trends were seldom sufficiently divergent to be clinically misleading although they were considerably greater than in the other type of substitution just noted. Although future study may show that the above or similar variations are to be preferred over the present In:!exes, this will not alter the basic principles of itemization, quantitation and summation on which they are based. We believe that these steps have a fundamental and enduring place in any scheme for evaluating rheumatoid activity. Itemization consists of selecting items which reflect gradations of activity at all levels; quantitation consists of ob- SYSTEMIC AND ARTICULAR INDEXES: TABLE2.-For MORNING STIFFNESS Min % 10 1 20 30 45 2 3 4 HB 1.0 6 1.5 9 2.0 1 1 2.5 14 3.0 17 3.5 20 4.0 23 4.5 26 5.0 29 5.5 31 6.0 34 6.5 37 7.0 40 7.5 43 8.0 46 3-YM 519 STUDY Calclrlating the Systemic Index FATIGUE AFTER ASPIRIN / DAY WEAKNESS (GRIPS) SED RATE NESTERGREN Hrs. % Tabs % rnm. % mm. % 0 0 1 2 2 5 3 7 4 10 5 12 6 15 7 17 8 20 9 22 10 25 1 1 27 12 30 13 32 14 35 15 37 16 40 17 42 18 45 19 47 20 50 21 52 22 55 23 57 24 60 25 62 260 0 250 1 240 2 230 3 220 4 210 6 200 7 190 8 180 9 170 1 1 160 12 150 13 12.0 11.0 10.0 9.0 8.0 7.0 6.0 5.0 4.5 4.0 3.5 3.0 2.5 2.0 0 2 4 7 10 13 16 18 20 21 23 24 26 27 1.5 28 1.0 30 0.5 31 0 32 140 15 130 120 110 100 90 80 70 60 50 40 30 20 10 0 16 17 19 20 21 22 23 24 25 26 27 28 29 30 - 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 145 0 2 3 5 7 8 10 12 13 15 17 18 20 22 23 25 27 28 30 32 33 35 37 38 40 42 43 45 To calculate articular index, simply add values for all affected joints. (See also detailed also detailed instructions on following page.) TABLE 3.-For Calculating the Articular Index 7 Knuckles, Toes, Jaw & Acrorniocavicular Bunion Sternoclavicular Wrist Joint Carpals, Chopart's and Subastragalar Ankle G Tarsal Area Elbow and Shoulder Knee and Hip - 12 25 JOHN LANSBURY Directions for Using Tables of Systemic and Articular Indexes Systemic Index 1. Vurution of morning sti%ness.-Note the tinie at which patient usually rises. Ask if he is stiff. If so, ask by what time the stiffness wears off or when he “limbers up.” Estimate the dwition of his stiffness and write down the percentage value for this as indicated in the table. 2. Tittie of onset of futigue.-Using the same type of questioning, ascertain the number of lioiirs after rising, before he gets tired or exhausted, and write down the percentage rqiiivdent from the table. (If patient is tired on waking due to poor sleep, do not count this as fatigue at zero hours, but find out when true fatigue comes on later in the clay. If he takes a scheduled afternoon nap, ask him to observe his fatigue on a day (such as Sunday) when the nap is omitted. Be sure not to confuse the need to “get off one’s feet becaiise of foot pain” with fatigue.) 3. Asliirin need.-Note the average daily number of aspirin tablets (or other salicylate), writc down its percentage valne from the table. (If patient does not take aspirin have him try a prcscribed dose of 8 to 12 tablets a day for a few days but record only the amount takrn thereafter on an “ad lib” basis. Be sure it is being taken for pain and not from liahit, and not a s a prescribed close or as a sedative on retiring.) 1. Grip strength.-An ordinary sphygmomanometer, re&tering to 300 mm. of Hg, is preferred. The valve at the top of the mercury column must be open. The rubber part of the cuff is folded exactlv twice, and the cotton flap rolled not too tightly around it. The system is inflated to 20 mni. of Hg. With the mercury column in full view and with his arm unsnpported, the patient sqiieezes the cuff as hard as possible. Three tries are allowed for c*:ichliiuitl, and the average of the highest reading for each hand is noted. Its percentage value (from the table) is recortlcd. 5. Scdittietitutivn rate.-Either the Westergren or Cutler technic may be used. ( I n the of the Cutler method, record the maximum fall in any one 5 minute period. Tables for this can be supplied.) Record the percentnge value of the reading. ( Wintrobe readings x c not acccptable. ) USI- To coniplette the Systemic Index, simply add up the percentage values for the five items. (The patient may keep a diary of the first three items, and their average values for 1 week prec-(ding the office visit may be used. ) hticulur Index For estimating the amount of active joint inflammation, one esamines all peripheral joints for tenderness and/or pain on forced, passive motion. When doing this, have the table i n front of you and simply write down the numerical value for each “positive” joint. Add these up. This is the Articular Index. (\Vhen one is not sure whether the joint is active or not, a half value may be given. When a joint has rest pain, or marked heat and fluid accumulation, it may be given an extra half value. The great majority of joints are in the “average” range. When fleeting pain is said to be present without objective evidence of arthritis (thickening, etc.) this Jors not count. Old, painless deformities likewise do not count.) In following cases, both Indexes should be recorded on ordinary graph paper a t nioiithly intervals. This permits one to see at a glance the trend of the disease and responses t8 1 tliwipy. SYSTEMIC AND ARTICULAR INDEXES: %YEAR STUDY 521 jective numerical measurements of the intensity of these, and summation consists of a standardized method for combining the information thus obtained and expressing it as an average, numerical value. CONCLUDING REMAHKS On the basis of a three-year study of the problem of evaluation of rhciimatoid arthritis we conclude that both the Systemic and Articular Indeses in their present form, taken either separately or together, have, in our hands, provided a highly satisfactory measure of the activity of the disease. We also have reason to believe that other observers have found them to be functionally satisfactory. However, we would like to stress that all our observations have been made on ambulatory, private patients and that almost none of them were receiving corticosteroid therapy. They must therefore be regarded as a selected group of patients, We therefore do not claim that our results automatically apply to groups of patients which are severely crippled, bedridden or who have been receiving corticosteroids for many years; but neither do we know that our methods will not apply in this situation. Before the Indexes can be recommended for general use, clearly defined end-points for judging joint inflammation must be agreed on and a comprehensive study of inter-observer error in the case of both Systemic and Articular Indexes is mandatory. This should preferably be carried out by a committee of independent observers from widely separated clinics. Finally, we should like to emphasize that none of the controversial questions concerning the Indexes can be settled on the basis of preconceived opinions or official proclamations, but rather, by further diligent study. SUMMARY Drug evaluation in rheumatoid arthritis can be carried out either by detecting an “anti-rheumatic” effect in a large group of patients in a multi-clinic project, or by long-range observations of the course of the disease in individual patients by the same doctor. Accurate methods of evaluation are necessary in both situations. In any scheme for evaluation of rheumatoid arthritis, only parameters which can be accurately defined and measured should be used, and the information which they furnish must be capable of being numerically summated. A survey of possible parameters reveals six out of about 30 which seem suitable for evaluation of the “activity” of the disease, and about 20 which seem suitable for evaluating the anatomic changes resulting from this activity. Methods of summation may be classified as: 1.) matching according to arbitrary categories, 2.) addition on a point system of positive observable items, and 3.) averaging of quantitated values for a selected number of observable items. The role of the Systemic and Articular Indexes both in multi-clinic drug trials and in following individual patients over a period of three yenrs is illustrated and discussed. (These Indexes, which have been described elscwhere, are based on quantitations of: stiffness, fatigue, pain, muscle weakness, total joint inflammation and sedimentation rate). A strong case is made out 522 JOHN LANSBURY for the accuracy of these new methods of evaluation on the basis of the following facts: 1.) The Indexes register changes in rheumatoid activity in those situations where they would be expected and agree with the general. over-all clinical estimate of changes in activity. 2.) The Systemic and Articular Indexes show a significant degree of parallelism, which, because each is based on entirely different data, is strong proof that they both register rheumatoid activity with acceptable accuracy. We conclude that the Indexes function well and accurately in the assessment of rheumatoid activity, and we recommend them for further trial. ACKNOWLEDGMENTS We wish to thank the members of the Committee fm the Evaluation of New Therapeutic Agents of the American Rheumatism Association for their cooperation and constructive comments. Thanks are also due to the following doctors who took part in the doubleblind prednisone study: Dr. A. A. Asadi, Dr. J, J. Calabro, Dr. G. Clark, Dr. F. Forestier, Dr. H. R. Goehrs, Dr. J. Hollander, Dr. J. G. Mayne, Dr. W. Mikkelsen, Dr. R. Pike, Dr. N. Radiou, Dr. C. Short, Dr. E. J. Southwood, Dr. S. F. Strain, Jr., and Dr. P. Vaughn. W e gratefully acknowledge the work of Drs. H. N. Baier, C. H. Bepler, S. McCracken and F. B. Rogers in making a small, pilot study of the interrelationships of the various manifestations of joint inflammation. We wish to especially acknowledge statistical help and advice on the part of Dr. Donald Mainland in planning the prednisone study, and the help of Miss Irene Pelczar with calculations. We also thank the Schering Corporation for supplying Meticorten with matching placebos directly to the clinics undertaking the double-blind prednisone study, and Parke, Davis and Company for supplying specially prepared 50 mg. camoquin pills with matching placebos. REFERENCES 1. Howell, D.S. and Vaughn, P.: Personal antirheumatic agent. Arthritis and communication. Rheumatism 1 :297, 1958. 2. Hart, D. F. and Clark, C. J. M.: 8. Smyth, C. J. and Clark, G. M.: PhenylMeasurement of digital swelling in butazone in rheumatoid arthritis. J. rheumatoid arthritis. Lancet I :775, Chron. Dis. 5734, 1957. 1951. 9. Bepler, C. R. and Rogers, F. B.: A 3. Lansbury, J.: A numerical method for double-blind study using manganese evaluating the status of rheumatoid against a placebo in rheumatoid ararthritis. Ann. Rheumat. Dis., 17:101, thritis. Am. J. Med. Sc. 234:459, 1957. 1958. 1 0 . Lansbury, J. and Haut, D. D.: 4. Area 4. - and Mueller, E. E.: 7. A numerical of joint surfaces as an index to total method for summing up total dejoint inflammation and deformity. Am. formity. Am. J. Med. Sc. 235154, J. Med. Sc. 232~150,1958. 1958. 11. and Free, S. M., Jr.: 6. Correlation 5. -: 5. A method for summation of the of systemic and joint findings. Am. systemic indices of rheumatoid acJ. Med. Sc. 233375, 1957. tivity. Am. J. Med. Sc. 222300, 1956. 8. Mikkelsen, W. M.: Personal conimiinicaIS. Wallace, S. L. and Ragan, C.: The tion. problem of therapeutic evaluation in 7. Cohen, A. S. and Calkins, E.: A conrheumatoid arthritis. Arthritis and trolled study of chloroquin as an Rheumatism I :20, 1958. -___ John Lambuy, M.D., Professor of Clinical Medicine and Chief of Arthritis Clinic, Temple Unitiersify Medical Center, Philadelphia, Penna.