Supplement I . Vol. 8, Aust. N.Z. J . Med. (197S), pp. 114 115
Measurement in Rheumatoid Arthritis Peter Lee
From the Rheurnatology Unit. Middlemore Hospital, Otahu
Despite the growing stature of research both at cellular and subcellular levels in rheumatoid arthritis (RA), the current methods for assessing disease activity are essentially indirect and crude measures of the underlying inflammatory process. At present, there is no single ideal measurement and despite the innumerable methods which have been put forward, there is a lack of international agreement as to which indices should be employed. There are also basic problems related to the investigators ignorance of the natural history of the disease, and the influence of apparently unrelated factors. An ideal measurement in RA should be sensitive, simple, clinically applicable and economical. The present generation of non-steroidal anti-inflammatory drugs are not particularly potent and as a consequence, the resulting clinical changes are small. For this reason it is essential to select the most sensitive indices available when clinical trials are being designed (Table 1). Despite the craving in the past for objective measurements such as the erythrocyte sedimentation rate, joint circumference, and various tests of physical function, these have been found to be the least sensitive. Radioisotope scanning using technetium (99" TC) and thermography have both been used as measures of joint inflammation and reflect the underlying increase in synovial vascularity. When employed in the evaluation of singlejoints, there is good correlation between the technetium count and other clinical parameters, but changes in a single joint d o not necessarily reflect general disease activity. In an attempt to overcome this problem, a technetium index was calculated by the summation of counts from ten separate joint areas in patients with RA and significant differences were found between these patients and normal subjects.' Although there was a Correspondence'
Dr Peter Lee, Specialist Rheumatologist. Middlemore Hospital, Otahuhu, Auckland, 6, New Zealand
Auckland, New Zealand
TABLE 1 The most sensitive measurements in R A 1. Pain index or measure of pain relief.
2. Articular index of joint tenderness or count of active joints. 3. Patients' own drug preference or estimation of drug efectiveness.
significant correlation between the technetium index and other clinical parameters such as the pain score and articular index, it was found to be a relatively insensitive measure. After ten days treatment with prednisone the technetium index showed a mean percentage change of only 11.4 compared with 3 1.4",, change in the articular index. These sophisticated techniques are therefore both insensitive and have the additional disadvantage of requiring expensive equipment. It is often not realised that results of various clinical measurements can be strongly influenced by factors unrelated to the underlying disease. An ideal measurement should be reproducible and have a low observer error. The measurement of digital joint circumference has been one of the favourite measurements for over two decades, and is still used extensively. However, the procedure is associated with a small but significant intra-observer error and an even greater inter-observer error.' For this reason, the same observer and same instrument should be employed in making serial measurements during the course of a clinical trial. Likewise, considerable intra- and inter-observer errors were observed in the measurement of grip ~ t r e n g t h . Although ~ the intra-observer error might seem small, it may be sufficient to influence the outcome and should be considered in assessing the results of drug trials. Hospital inpatients are often the subject of clinical trials but the magnitude of the clinical changes which can be induced by hospitalisation alone is not often realised. In a study of the effects of inpatient treatment on RA, it was
JlJNE
1978
MEASUREMENT IN RHEUMATOII> ARTHRITIS
115
~
found that significant improvements in pain severity and duration of morning stiEness had occurred by the end of the first week of hospitalisation alone with a significant reduction in articular index score by the end of the third week.' If this phenomenon is not appreciated such improvement may well be attributed to the therapeutic regime under test. The basis of therapeutic trials in RA has been the double-blind-cross-over method employing multiple indices. Although the soundness of such techniques are not in question, they are both time consuming and expensive. With the increase in demand for clinical trials in RA, the feasibility of simpler and more economical methods should be explored. In Glasgow, a simple and rapid, single-blind, non-crossover technique has been developed which utilises entirely subjective parameter^.^ A pain chart is employed on which the patient can record his pain severity on a day to day basis and at the end of the trial. the patient records a satisfaction score which is his own estimate of drug efficacy. In addition, the number of days withdrawn from the trial are noted. In the analysis, initial differences between treatment groups can be
corrected, enabling results from separately conducted trials to be compared. In a study of ten anti-rheumatic drugs in 684 patients with RA, results were obtained which were comparable to those previously found utilising more complex and detailed methods. There is an urgent need for an internationally acceptable assessment programme in which the individual measurements are standardised. Unfortunately, at the present time, opinions about this differ widely. but if data from individual centres are to be correlated in a meaningful way, such a programme is essential.
indicrr In the a,\eciment US di\ca\s actt\ity in rheumatoid arthut!,, J Rkt~uniuir,i. 1, 432 2 Wbss. J . D O W h i l . W W , Dlrr. W. C and L t t . P (1973): Evaluation 0 1 digital joint circumlerenw mra\uremenl, in rheumatoid arthncrr. . S r u d J
Rheum 2, 127.
3 LLL. P . B A X I L RA,. . Diix. W C and Wrss. J (19741 An a s b e s m c n l of grip 5trcngth ineacurenirnt in rheumatoid arthritis. Scond J Rheum 3, 17 4 L I I . P.. Ktzhtuy. A C , ALI)IRION. J A and B i I i H A N A X . W W 11974) T h e lherapeutic benefiis of hosprlal m p a t i r n t ireatmcnl in rheumatoid arthriti5. Quuri J .Wed 43. 205 5 L E I . P . ANlllRSIIN. J A . M i l l r n . J . Wluu. J and B ~ c H A \ A \ .W W . 11976): Evaiuaiwn of analgesic action and ethcscy 01 anti-rhrumalic druga Study 01 tcn drugr in 683 patient\ wLth rheumatoid arthriti\. .I Rhcu,,uiio/ 3. 2x3
Discussion Room : Could you comment on acute phase reactants other than the ESR and their usefulness in following patients with rheumatoid arthritis'! Lee: I have not had personal experience with C reactive protein but McConkey' has found this acute phase reactant to be a useful assessment. We have not found the ESR particularly helpful in short term clinical trials but it might be of more value in long term drug assessments. Burry: In your simplified questionnaire assessment you rely on the premise that pain correlates with disease activity and that drug efficacy can be measured purely in terms of pain relief.
Personally, I am not of that opinion but you may have some facts to support your view. Lee: Yes, I take your point. Since pain is the patient's main complaint in rheumatoid arthritis, it should be one of the essential. parameters to follow. We base our overall assessment on three parameters: the pain response; the patient's own assessment of the drugs effectiveness, and the number of days withdrawn from the study. In recent trials we have also incorporated the severity of morning stiffness. Reference I
, k f l ' C l ) \ K l Y . B . CROCh\Oh. R A and < ' R W K \ O h . A P (19721 T h r a\\~ssilnenlof rheumatoid arthrilis A rtudy b a d on mcabur~mcnt\01 $erum acute-phase reactant\. Quorr J Wed 41. II 5
Measurement in Rheumatoid Arthritis Peter Lee
From the Rheurnatology Unit. Middlemore Hospital, Otahu
Despite the growing stature of research both at cellular and subcellular levels in rheumatoid arthritis (RA), the current methods for assessing disease activity are essentially indirect and crude measures of the underlying inflammatory process. At present, there is no single ideal measurement and despite the innumerable methods which have been put forward, there is a lack of international agreement as to which indices should be employed. There are also basic problems related to the investigators ignorance of the natural history of the disease, and the influence of apparently unrelated factors. An ideal measurement in RA should be sensitive, simple, clinically applicable and economical. The present generation of non-steroidal anti-inflammatory drugs are not particularly potent and as a consequence, the resulting clinical changes are small. For this reason it is essential to select the most sensitive indices available when clinical trials are being designed (Table 1). Despite the craving in the past for objective measurements such as the erythrocyte sedimentation rate, joint circumference, and various tests of physical function, these have been found to be the least sensitive. Radioisotope scanning using technetium (99" TC) and thermography have both been used as measures of joint inflammation and reflect the underlying increase in synovial vascularity. When employed in the evaluation of singlejoints, there is good correlation between the technetium count and other clinical parameters, but changes in a single joint d o not necessarily reflect general disease activity. In an attempt to overcome this problem, a technetium index was calculated by the summation of counts from ten separate joint areas in patients with RA and significant differences were found between these patients and normal subjects.' Although there was a Correspondence'
Dr Peter Lee, Specialist Rheumatologist. Middlemore Hospital, Otahuhu, Auckland, 6, New Zealand
Auckland, New Zealand
TABLE 1 The most sensitive measurements in R A 1. Pain index or measure of pain relief.
2. Articular index of joint tenderness or count of active joints. 3. Patients' own drug preference or estimation of drug efectiveness.
significant correlation between the technetium index and other clinical parameters such as the pain score and articular index, it was found to be a relatively insensitive measure. After ten days treatment with prednisone the technetium index showed a mean percentage change of only 11.4 compared with 3 1.4",, change in the articular index. These sophisticated techniques are therefore both insensitive and have the additional disadvantage of requiring expensive equipment. It is often not realised that results of various clinical measurements can be strongly influenced by factors unrelated to the underlying disease. An ideal measurement should be reproducible and have a low observer error. The measurement of digital joint circumference has been one of the favourite measurements for over two decades, and is still used extensively. However, the procedure is associated with a small but significant intra-observer error and an even greater inter-observer error.' For this reason, the same observer and same instrument should be employed in making serial measurements during the course of a clinical trial. Likewise, considerable intra- and inter-observer errors were observed in the measurement of grip ~ t r e n g t h . Although ~ the intra-observer error might seem small, it may be sufficient to influence the outcome and should be considered in assessing the results of drug trials. Hospital inpatients are often the subject of clinical trials but the magnitude of the clinical changes which can be induced by hospitalisation alone is not often realised. In a study of the effects of inpatient treatment on RA, it was
JlJNE
1978
MEASUREMENT IN RHEUMATOII> ARTHRITIS
115
~
found that significant improvements in pain severity and duration of morning stiEness had occurred by the end of the first week of hospitalisation alone with a significant reduction in articular index score by the end of the third week.' If this phenomenon is not appreciated such improvement may well be attributed to the therapeutic regime under test. The basis of therapeutic trials in RA has been the double-blind-cross-over method employing multiple indices. Although the soundness of such techniques are not in question, they are both time consuming and expensive. With the increase in demand for clinical trials in RA, the feasibility of simpler and more economical methods should be explored. In Glasgow, a simple and rapid, single-blind, non-crossover technique has been developed which utilises entirely subjective parameter^.^ A pain chart is employed on which the patient can record his pain severity on a day to day basis and at the end of the trial. the patient records a satisfaction score which is his own estimate of drug efficacy. In addition, the number of days withdrawn from the trial are noted. In the analysis, initial differences between treatment groups can be
corrected, enabling results from separately conducted trials to be compared. In a study of ten anti-rheumatic drugs in 684 patients with RA, results were obtained which were comparable to those previously found utilising more complex and detailed methods. There is an urgent need for an internationally acceptable assessment programme in which the individual measurements are standardised. Unfortunately, at the present time, opinions about this differ widely. but if data from individual centres are to be correlated in a meaningful way, such a programme is essential.
indicrr In the a,\eciment US di\ca\s actt\ity in rheumatoid arthut!,, J Rkt~uniuir,i. 1, 432 2 Wbss. J . D O W h i l . W W , Dlrr. W. C and L t t . P (1973): Evaluation 0 1 digital joint circumlerenw mra\uremenl, in rheumatoid arthncrr. . S r u d J
Rheum 2, 127.
3 LLL. P . B A X I L RA,. . Diix. W C and Wrss. J (19741 An a s b e s m c n l of grip 5trcngth ineacurenirnt in rheumatoid arthritis. Scond J Rheum 3, 17 4 L I I . P.. Ktzhtuy. A C , ALI)IRION. J A and B i I i H A N A X . W W 11974) T h e lherapeutic benefiis of hosprlal m p a t i r n t ireatmcnl in rheumatoid arthriti5. Quuri J .Wed 43. 205 5 L E I . P . ANlllRSIIN. J A . M i l l r n . J . Wluu. J and B ~ c H A \ A \ .W W . 11976): Evaiuaiwn of analgesic action and ethcscy 01 anti-rhrumalic druga Study 01 tcn drugr in 683 patient\ wLth rheumatoid arthriti\. .I Rhcu,,uiio/ 3. 2x3
Discussion Room : Could you comment on acute phase reactants other than the ESR and their usefulness in following patients with rheumatoid arthritis'! Lee: I have not had personal experience with C reactive protein but McConkey' has found this acute phase reactant to be a useful assessment. We have not found the ESR particularly helpful in short term clinical trials but it might be of more value in long term drug assessments. Burry: In your simplified questionnaire assessment you rely on the premise that pain correlates with disease activity and that drug efficacy can be measured purely in terms of pain relief.
Personally, I am not of that opinion but you may have some facts to support your view. Lee: Yes, I take your point. Since pain is the patient's main complaint in rheumatoid arthritis, it should be one of the essential. parameters to follow. We base our overall assessment on three parameters: the pain response; the patient's own assessment of the drugs effectiveness, and the number of days withdrawn from the study. In recent trials we have also incorporated the severity of morning stiffness. Reference I
, k f l ' C l ) \ K l Y . B . CROCh\Oh. R A and < ' R W K \ O h . A P (19721 T h r a\\~ssilnenlof rheumatoid arthrilis A rtudy b a d on mcabur~mcnt\01 $erum acute-phase reactant\. Quorr J Wed 41. II 5
