Future cooperation in the field of rheumatoid arthritis and related diseases*

MEMORANDA Future cooperation in the field of rheumatoid arthritis and related diseases * This Memorandum discusses two areas of research into rheumatic disease in which it is considered international cooperation would be valuable. These are: (a) standardization and cooperative evaluation of immunological tests used in diagnosis, clinico-pathological assessment, and epidemiological work; and (b) standardization and cooperative assessment of criteria for application to disease classification, evaluation of therapeutic agents, epidemiological studies, and public health programmes. During the last few decades evidence has accumu- lated to indicate that rheumatoid arthritis and cer- tain allied diseases are closely associated with im- munological abnormalities. Rheumatoid arthritis was once regarded as an infectious disease, and in recent years advances in microbiology and immunol- ogy have again brought the concept of infection to the forefront. It has been increasingly suggested that rheumatoid arthritis and allied diseases arise from the development of immunopathological responses, following established or latent infection with a pos- sibly wide range of microorganisms, which lead to pathology of the locomotor, vascular, and lymphoid systems. An extension of this view is that the different clinical syndromes within this group of disorders arise from immunoregulatory defects, the expression of which could be influenced by both genetic and environmental factors. These concepts are likely to have important implications for work directed towards the etiology, epidemiology, and health care aspects of this group of diseases. At intervals since 1953 the World Health Organi- zation has convened scientific and study groups through the activities of its Geneva headquarters (1-3 and unpublished report to WHO, 1967) and its Regional Office for Europe in Copenhagen.a A Scientific Group on the Diffuse Connective Tissue Diseases was held in Geneva during June-July 1966 to consider criteria for the classification of systemic lupus, polyarteritis, systemic sclerosis, and dermato- * This memorandum was drafted by the signatories listed on page 606, on the occasion of a symposium held at the Kennedy Institute of Rheumatology, London, England in March, 1974. Requests for reprints should be addressed to: Division of Non-Communicable Diseases, World Health Organization, 1211 Geneva 27, Switzerland. a Unpublished reports to WHO Regional Office for Europe, Copenhagen, 1964, 1967 and 1972. r - P ,- C) myositis. This Group recommended that validation of the proposed criteria be implemented by the establishment of collaboration between investigators in different centres. An international reference centre was subsequently established in Paris for the collec- tion of data by means of a protocol-questionnaire for the validation of diagnostic criteria in the diffuse connective tissue diseases. Since then, the question- naire has been agreed by 26 rheumatological socie- ties. In September 1966, the WHO Regional Office for Europe convened a working group in Copen- hagen to discuss epidemiological studies in chronic rheumatoid arthritis. The Copenhagen group recom- mended the establishment of a project for epidemiol- ogical research and for the validation of diagnostic criteria for rheumatoid arthritis that would relate to the etiology of arthritis syndromes and to the public health aspect of this group of diseases. This project led to a working group on chronic rheumatoid arthritis, which met at Disley, England in 1970 to evaluate developments and to consider possible future work. Examples of subjects not discussed in this Memo- randum are: the biochemistry of connective tissue; purine metabolism; the surgery of osteoarthritis; and heritable disorders of connective tissue. However, in view of the particular interests of signatories in the field of immunology and infection in rheumatoid arthritis and allied diseases, it was felt that discussion of this particular field would be of additional value to national and international bodies concerned with the promotion of investigative rheumatology. This Memorandum therefore discusses two prin- cipal areas of potential cooperation in the field of rheumatoid arthritis and diffuse connective tissue diseases: (a) standardization and cooperative eval- uation of immunological tests used in diagnosis, a/- A ' , V 3305 A 597 - BULL. WORLD HEALTH ORGAN., Vol. 51, 1974 MEMORANDA clinicopathological assessment, and epidemiological work; and (b) standardization and cooperative assessment of criteria for application to disease classification, evaluation of therapeutic agents, epi- demiological studies, and public health programmes. POTENTIAL VALUE OF IMMUNOLOGICAL TESTS IN SURVEY WORK If microbial infection plays a part in determining the prevalence or severity of rheumatic disease, it might be anticipated that further studies of anti- microbial and autoimmune responses might yield information of etiological significance. It should be considered, therefore, whether environmental leads might emerge from the introduction of immunol- ogical tests into survey work on rheumatoid arthritis and the systemic connective tissue disorders. Epidemiological background In rheumatoid arthritis, point-prevalence epi- demiological studies in many different countries have so far yielded only a little information of apparent etiological significance (4-8). In the USSR, marked differences have been found in the severity (but not in the prevalence) of rheumatoid arthritis in different climatic areas; these are mote likely to be related to environmental factors than to genetic influences. In one study there appeared to be less rheumatoid arthritis among persons living at high altitudes. The mild nature of rheumatoid arthritis among Afri- cans (9) and the possibility of protection from the disease by chronic malarial infection or other envi- ronmental factors is worthy of further investigation. There appear to be marked differences in the preva- lence of systemic lupus in the UK from that in the USA and the USSR; this suggests that both environ- mental and genetic factors are worthy of continued study. Twins could be studied to identify both environmental and genetic factors (10-13); the study of monozygotic twins concordant or discordant for rheumatoid arthritis would allow genetic influences to be equalled so that environmental hypotheses could be tested. In rheumatoid arthritis new longi- tudinal population studies similar to those carried out on heart disease in Framingham, USA (14-16) should probably not be started on a large scale, although existing studies should be continued. In relation to isolated populations that are free from rheumatoid arthritis, it may be possible to identify the presence or absence of important environmental factors. In practice, however, " proving " the absence of rheumatoid arthritis in a population is probably not feasible. In view of the present epidemiological uncertain- ties, considerable difficulty may be encountered in establishing a major international project incorporat- ing immunological studies to test hypotheses of the etiology of rheumatoid arthritis and the systemic connective tissue diseases. However, smaller-scale studies from individual centres should be encour- aged, particularly where internationally acceptable clinical and laboratory criteria are used (see Stan- dardization of test reagents and laboratory test proce- dures, p. 600). The introduction of standardized immunological tests into survey work should be discussed, and a special conference could be held to define both clinical and laboratory criteria to be used in such epidemiological studies. Tests of humoral immune status Further studies of autoantibodies (particularly antiglobulin and antinuclear antibodies) are to be preferred to studies of antimicrobial antibodies, since autoantibody studies have been used more widely and have yielded much information of rheumato- logical value. It would be of interest to obtain parallel information on infective and rheumatic disease in different countries against which the significance of differences in antimicrobial and autoimmune status could be interpreted. For example, since rheumatoid arthritis is a mild disease in Africans, further work on antiglobulin factors in relation to patterns of microbial infection and rheumatic disease in different countries might be fruitful. Likewise, since systemic lupus is less prevalant in the UK than in the USA and USSR, a study of antinuclear factors in age- and sex-matched groups in different countries might dis- close geographical variations with regard to these antibodies. The following tests of humoral immune status are recommended: Detection of antibodies. These should include tests for rheumatoid factors, antinuclear antibodies, other autoantibodies and possibly antimicrobial antibodies if such can be more closely related to rheumatic disease. Tests for " hidden " rheumatoid factors may be important but need further elaboration tech- nically. Detection of immune complexes. Several tests for the detection of immune complexes are available. Such tests are potentially of great value in survey work, but further standardization of the tests are 598 COOPERATION IN RHEUMATIC DISEASE RESEARCH needed. The use of radiolabelled Clq or rheumatoid factor will probably be most important for future survey work. Techniques of histological immunology. Detection of humoral immune responses in the tissues should be used to a much larger extent in survey work, through either biopsies or surgically removed tissues. To date attention has mainly been focussed on circulating antibodies, and the local inflammatory response has not been so well characterized. Recent studies indicate that there is closer clinical correla- tion with local immune responses than with those apparent in the general circulation. Complement investigations. The complement sys- tem is closely associated with tissue damage and inflammation induced by humoral immune re- sponses. Not only the components of the classical pathway (initiated through activation of Clq) but also the properdin system (alternate pathway) should be investigated. Humoral responses to defined antigens. Since the problem of immunodeficiency has been raised in relation to rheumatoid arthritis, and is well estab- lished in the case of X-linked hypogammaglobulinae- mia, studies including the testing of antibody re- sponse to defined antigens should be encouraged. Potential value of lymphocyte tests During the last few years, a considerably increased understanding of immune reactions in rheumatoid arthritis and systemic lupus has focused attention on cell-mediated immunity as an important factor in these diseases. Firstly, there is evidence that in both adult and juvenile rheumatoid arthritis and in sys- temic lupus there is depression of cellular immune reactivity to common microbial antigens. Secondly, lymphocytes from rheumatoid patients cause migra- tion inhibition of leucocytes when stimulated with aggregated IgG. Thirdly, cytotoxicity of rheumatoid synovial fluid lymphocytes has also been demon- strated. Fourthly, striking deviations from the normal ratios between various lymphocyte cell popu- lations in blood have been found among synovial fluid lymphocytes. Some of the difficulties in determining lymphocyte function and in quantitating lymphocyte cell popula- tions must be acknowledged. However, some of these tests appear to measure parameters that may be intimately associated with rheumatoid disease pro- cesses; they should therefore be included in im- munological survey work and in the monitoring of clinical trials evaluating immunosuppressive or im- munopotentiating agents. Immunogenetic studies are becoming increasingly important since they may distinguish genetically predisposed groups; these groups should be carefully surveyed for rheumatic disease with a view to prophylactic intervention. In the full assessment of immunodeficiency, attention should also be paid to tests of monocyte/macrophage and polymorph function. In assessing evidence related to the extent of immunodepression or immunopotentiation, the vari- ous lymphocyte test systems recommended can be classified as follows: Marker tests for human peripheral blood lympho- cytes. The present evidence indicates that most anti- body-forming cells (" B cells ") bear surface im- munoglobulin, and that immunofluorescence reac- tions provide useful marker tests for this subpopula- tion provided that appropriate reagents are used and that the immunoglobulin is shown to be truly syn- thesized by the enumerated cells. Likewise, most thymus-dependent cells (" T cells ") have receptors for sheep erythrocytes, and the rosette test provides a useful marker for the T-cell subpopulation. Errors in delineating B cells and T cells by these techniques may arise: (a) in lymphoproliferative disease accom- panied by monoclonal or heterogenetic B cell anti- body responses; and (b) where antilymphocyte auto- antibodies (or cell-adsorbed immune complexes) in- terfere with T cell identification. Techniques have recently been introduced by which the lymphocyte transformation test can be done using only 2-3 ml of whole blood, and further improvements in the standardization of methodology and test reagents are anticipated. Lymphokine production offers an additional ap- proach to assessing lymphocyte function when quan- titative assays of (for example) migration-inhibition factor (MIF) and mitogenic factor (MF) are devel- oped. Lymphokines are non-antibody products of activated lymphocytes that are generated within 1-3 days of antigenic stimulation. Recent evidence sug- gests that MF is generated by T cells and that MIF can be generated by both T cells and B cells in human peripheral blood. A simplification of the MIF test is provided by the leucocyte migration inhibition test (LMT). In this test, the ability of putative antigen (e.g., mycoplasma, DNA, aggregated IgG, or tuberculin) to inhibit the migration of buffy coat leucocytes is taken as an index of lymphocyte sensi- 599 MEMORANDA tization and lymphokine production, although the relative contributions of T cells and B cells is as yet unclear. The leucocyte migration test can be used in principle to identify selective immunodeficiency (as recently reported for myxoviruses in multiple sclero- sis). The future applicability of this operationally simple test to survey work in the rheumatic diseases is dependent on its standardization; a miniaturised LMT is now available that requires samples of only 5-10 ml of blood. Lymphoid cell cytotoxicity tests of two principal kinds have been studied at the research level, particu- larly in rheumatoid arthritis, systemic lupus, and polymyositis. These are: (i) direct (T-cell mediated) cytotoxicity of patients' lymphocytes for cultured "' target " cells bearing antigens to which the lympho- cytes are putatively sensitized; and (ii) antibody- mediated cytotoxicity in which lymphoid cells with Fc receptors are cytotoxic for antibody-coated " target " cells in culture. These tests are of growing interest in rheumatological research (e.g., for detect- ing new cellular antigens or for characterizing lym- phoid and other cells bearing Fc receptors), and when further elaborated they may be applicable to rheumatological survey work. Skin testing to detect delayed hypersensitivity in health and disease was discussed at a WHO meeting in 1973 (17), and recently several studies have re- vealed depression of delayed hypersensitivity to ubi- quitious microbial antigens in rheumatoid arthritis and systemic lupus. Although the delayed hyper- sensitivity reaction is a complex cellular and vascular response, its relative operational simplicity justifies its introduction into survey work that might correlate the prevalence of rheumatic and infectious disease. Immunogenetic studies. The discovery of correla- tions between the second-segregant series histo- compatibility antigen (HL-A) W27 and both ankylosing spondylitis and Reiter's syndrome has given great encouragement to survey work studying the frequency of HL-A specificities in rheumatic disease. Assuming that the technical problems of lymphocyte transport and the standardization of reagents and test procedures can be overcome, such survey work could also help in understanding the unusual variation of HL-A types in systemic lupus. Because the second-segregant series of histocom- patibility antigens in certain animals is linked to immune-response genes, associations between certain HL-A types and rheumatic disease would suggest that their pathogenesis was more directly linked to some aberrant, genetically-controlled, cellular im- mune response. However, recent experiments in- dicate that immune response genes are much more closely related to the mixed lymphocyte reaction (MLR) locus, and studies in this direction may further clarify the relationship between genetic traits and the rheumatic diseases. STANDARDIZATION OF TEST REAGENTS AND LABORATORY TEST PROCEDURES Selection of tests for standardization Although a large number of laboratory tests are used in rheumatology, some of which require further clarification, it is generally agreed that greater uni- formity and precision in individual tests is desirable. In the testing of biological substances it is customary to consider two ways in which greater uniformity and precision can be achieved: (a) the establishment and provision of reference standards; and (b) delineation of test procedures that provide acceptable re- producibility and specificity and that permit quality control. When the relative importance of different laboratory tests, particularly in survey work in the rheumatic diseases, are considered, those the most in need of standardization are as follows: (a) anti- globulin and antinuclear factors; (b) anti-DNA; (c) detection and quantitation of complement (C3) and of circulating immune complexes; (d) lympho- cyte function and surface marker tests; and (e) leuco- cyte antigen (HLA/MLC) tests. Serologic tests for antimicrobial antibodies are not proposed for stan- dardization in this context; however if research were to show closer correlation with rheumatic disease, their standardization would be needed. Reference preparations and test reagents Much serologic experience in various fields has shown that relating the potency of a serum under study in any one laboratory to the potency of a reference standard in that same laboratory is an effective way of defining the potency of that serum in terms that can be approximated in other laboratories using divergent or even similar test procedures. The principle is therefore to titrate unknown sera in parallel with a reference (" standard ") serum and to specify the results in units related to the designated activity of the standard serum. In view of the very great diversity of results obtained when pooled serum samples are distributed and tested, the follow- ing three tests are felt to merit the inclusion of reference reagents: 600 COOPERATION IN RHEUMATIC DISEASE RESEARCH 601 Rheumatoid factor. An international standard rheumatoid factor serum is now available (18). The limited demand for this preparation over the past several years suggests that its availability should be made more widely known. Antinuclear factor. A well-standardized pooled serum is available for this test (24). This serum gives a homogeneous nuclear stain when cryostat sections of unfixed rat liver are used as the substrate. DNA-binding antisera. No recognized standard serum is available but it is possible that the standard antinuclear factor serum (19) may function in this way. Dr E. V. Barnett, University of California, Los Angeles (UCLA) is investigating this particular point. These efforts at standardization by the provision of reference sera would improve the comparability of results in epidemiological studies and also in studies of the disease processes themselves. It must be pointed out, however, that the quantities of such reference preparations are not unlimited; it is there- fore suggested that standard preparations be used by national agencies with the capacity to produce com- parable secondary standards for local use. To ensure proper use of the primary standards, experts should take responsibility for the approval of applications. Reference sera and reagents would also be desir- able in many other tests of potential interest, but it is felt that these are not yet sufficiently advanced. For example, there are several tests available for the detection of IgG antiglobulins and for rheumatoid factors reactive with particular antigenic deter- minants (e.g., Fab). Secondly, a radiolabelled antigen reagent might be developed for the detection of anti- DNA antibodies. Thirdly, there are several tests for the presence of soluble immune complexes in serum (e.g., adsorption tests, rheumatoid factor inhibition, coprecipitation with Clq or monoclonal rheumatoid factor, competition for Fc receptor binding), and for their clinical application it would be desirable to provide reference preparations of antigen and anti- body components for the generation of " standard " complexes and to develop reference standards of precipitating agents such as Clq and antiglobulins. The standardization of immunofluorescence proce- dures is already being undertaken in other contexts, as is the development of reagents with which to delineate lymphocyte subpopulations and to study lymphocyte function. It is felt that the provision of cryopreserved lymphocytes and standard typing sera for immunogenetic studies is more the concern of transplantation research workers. Information about the availability of international standards, reference preparations, and reference re- agents may be obtained regularly from WHO, and certain of the biological substances have been tested with the collaboration of the National Institute for Biological Standards and Control, London, UK. The International Union of Immunological Societies (IUIS) and WHO also have a joint standardization committee that has encouraged the development of immunoglobulin, complement, and rheumatoid factor reference standards. It is recommended that efforts in these areas should be further developed. Test procedures and quality control Whilst only general guidelines for the performance of tests have been considered desirable, extremely wide variations between laboratories are often en- countered. An example stems from a coopbrative study by different laboratories in their estimation of rheumatoid factor activity in pools of human sera, where titres were found to differ by a thousand fold.a It is unlikely that differences in methodology alone could account for such results, and some form of " quality control " would therefore appear to be desirable at least in cooperative studies. A fairly recently introduced practice of quality control in the clinical National Health Service labo- ratories in the UK has also emphasized this problem of inter-laboratory variation. Participating labora- tories receive at regular intervals standard samples for analysis and the results are subsequently pub- lished. Strict anonymity is maintained. The results now cover the majority of the common estimations requested by clinical laboratories and once again emphasize the unacceptably wide range. Here again a proportion of the variation is attributable to differ- ences in method and technique. There can, however, be no doubt that the ability to measure in standard deviations the amount by which one's own results differ from the mean of all participating laboratories exerts a salutory effect. Laboratories that show a consistently wide deviation from the mean soon become identified and improvements in technique can also be readily monitored. The erythrocyte sedimentation rate (ESR) test has been replaced in many parts of the UK by tests of plasma viscosity, which are amenable to automation. a Bozsoky, S. (1962) The problem of standardization in rheumatoid arthritis serology. Unpublished document WHO/BS/574. MEMORANDA In addition to the provision of acceptable refer- ence reagents for inter-laboratory comparison, some kind of quality control should be instituted that would identify those laboratories in which results consistently fall outside the statistically acceptable range of variation. With adequate safeguards for anonymity, and adequate notice of increased devia- tion of results, it is likely that a reliable laboratory would readily accept a form of quality control. Where allocation of funds is concerned, quality control of the kind envisaged may also prove a useful guide to the reliable recipient. STANDARDIZATION OF CLINICOPATHOLOGICAL CRITERIA FOR COMPARISONS BETWEEN GROUPS Development of classification criteria In rheumatic fever the original and modified Duckett Jones criteria (3) have been highly successful and may be considered essential to general progress. Likewise the American Rheumatism Association (ARA) criteria for rheumatoid arthritis (20) have been very successful in describing clinical phenomena associated with the disease and in assuring a high degree of diagnostic certainty in drug trials. They have been less successful, however, in epidemio- logical surveys and have therefore been modified for specific epidemiological objectives (21). Neither set of criteria, whether designed for clinical or for epidemiological purposes, does more than establish the diagnosis; whilst providing needed baseline in- formation, they have been ineffective in assessing health care needs and the economic impact of the disease in terms of disability. It seems certain that the epidemiological criteria for ankylosing spondylitis (22) will require modifica- tion in the light of present immunogenetic informa- tion. The need for diagnostic criteria for gout is shown by the activity of an ARA committee, which is currently analysing large amounts of data on the diagnosis of gout. The ARA criteria for juvenile rheumatoid arthritis (23) may be too complex for ordinary use, and there is a need for further improve- ment in the classification criteria of the polyarthri- tides. Systemic connective tissue diseases. Although the report of the 1966 Geneva conference was not published, preliminary criteria for systemic lupus erythematosus (SLE) have been put out by the ARA based on a multicentre study (24). These preliminary ARA criteria have received some testing (25-27). Other unpublished criteria for SLE include the UCLA criteria (28). The various criteria have ap- parently not been applied to the study of systemic rheumatic disorders in childhood (29). A survey coordinated by the WHO International Reference Centre in Paris, which is still unfinished, compares in the first instance the symptoms and signs for SLE mentioned in the report of the 1966 Geneva confer- ence with those of 4 other diseases (polyarteritis nodosa, dermatomyositis, scleroderma, and rheuma- toid arthritis) in terms of specificity and sensitivity by means of the Youden index. Recommendations. Further clarification is needed on the relative importance of characters, and on the use to which different sets of criteria should be put (e.g., epidemiological, diagnostic, or therapeutic trials). Clinical, radiological, and immunological cri- teria should be supplemented by those based on the morphological description of biopsy specimens (see also Recommendations, p. 604). It is hoped that when further study groups are convened their discussions and reports will take advantage of these findings and recommendations and those of the 1966 meeting. Such a report should also refer to published and unpublished validations of sets of criteria, by means of frequency-analysis of signs and symptoms. Indica- tions of specificity and sensitivity should also be included where possible; these should include data supplied by the International Reference Centre as well as those from other studies. Criteria for functional disability and health care Additional methods of assessment and reporting could possibly be devised for rheumatoid arthritis for the purpose of health care planning, emphasizing the extent and location of disability and impairment. Thus on a regional basis it would be useful to know: the requirements for orthopaedic surgery in rheuma- toid arthritis; the prevalence of severe rheumatoid arthritis requiring much laboratory monitoring; the prevalence of deformity requiring the provision of appliances; and the extent of regional variations in morbidity patterns in rheumatoid arthritis. More detailed information on the type and extent of individual joint involvement might also be of value. Functional disability should be assessed in relation to both environment and occupation. (For example, a given lesion or disease will produce different disabilities in left-handed than in right-handed people, or in army personnel than in typists). 602 COOPERATION IN RHEUMATIC DISEASE RESEARCH Criteria for polymyalgia rheumatica Polymyalgia rheumatica is frequently diagnosed in rheumatic clinics, yet it has not been defined for international purposes. In one study the syndrome accounted for 4.5% of all rheumatic referrals and was diagnosed second in frequency to rheumatoid arthritis and equal in incidence to ankylosing spon- dylitis. Much interest in the polymyalgia rheumatica syndrome is related to its association with giant cell arteritis or temporal arteritis (30), and it is debatable whether the syndrome should be regarded as arising from single or from multiple disease entities. In- clusive criteria are: age over 60 years; high ESR; proximal limb stiffness; weight loss; mild anaemia; and response to steroids. Synovitis is not regarded as part of the syndrome. There is a great need for data on this condition. CRITERIA FOR THE EVALUATION OF THERAPEUTIC AGENTS The introduction and evaluation of new therapeu- tic agents or procedures runs along fairly well estab- lished lines. Drug studies have both preclinical (animal model; cell-culture) and clinical (therapeutic trial) phases. The purpose of animal data, if appro- priately obtained with well-defined experimental criteria, should be 3-fold: (i) to eliminate drugs with toxic effects; (ii) to indicate likely properties includ- ing anti-inflammatory, immunosuppressive and un- wanted effects; and (iii) to provide clues to the possible mode of action of a drug. Whereas preclin- ical drug screening is usually directed towards the suppression of experimental inflammation (adjuvant arthritis; carrageenin oedema; crystal synovitis), these tests detect preferentially drugs with effects on acute inflammation. There is a need for the wider introduction of other animal test systems (e.g., chronic allergic monoarthritis; granuloma forma- tion) and for the development of new experimental approaches that would detect the cellular and vascular mechanisms involved in such suppression. At present the lack of such knowledge may limit the application of preclinical studies to the design of therapeutic trials. Collaboration between the basic scientist (bio- chemist; pharmacologist; immunologist; patholog- ist) and the clinician in studies of potential antirheu- matic drugs may yield multidisciplinary criteria of greater precision relative both to the aims of treat- ment and to the mechanisms of drug action. A cooperating panel could explore these possibilities at an international level. Such cooperation would be demonstrably effective if it succeeded in pointing the way to the development of more effective antirheu- matic drugs with both specificity and selectivity for certain well-defined pathological processes. Therapeutic trials The well-known difficulties of therapeutic assess- ment in rheumatoid arthritis and allied diseases stem from several sources: (i) the limited effectiveness of available therapeutic agents differing pharmacolo- gically and in their degree of toxicity; (ii) difficulties in predicting the natural history and prognosis of rheumatoid arthritis in any one new patient; (iii) the lack of objective clinical criteria that would eliminate observer bias and placebo reactors; and (iv) the need for further information on the likely mechanism of action of potential drugs on rheumatoid inflamma- tion, and on delineating the onset and duration of their suppressive effects. In view of the increasing number of analgesic, anti-inflammatory, and anti- proliferative drugs appearing for preclinical and clinical study, it would seem an appropriate time to set up an international multidisciplinary panel to discuss cooperation in the development and pre- sentation of potential antirheumatic agents. Improvements are desirable in the actual standard of trial programmes (numbers of patients treated; random selection of cases; adequately controlled procedures), in the standardization of various clinical tests used (e.g., grip strength), and in the recording of data concerning observer differences. Reviews have recently appeared on problems relating to trial design, group selection, clinical measurement, and the interpretation of short-term and long-term effects (31, 32). Important difficulties arise when considering the detailed design and analysis of clinical trials in relation to five principal categories of antirheumatic agents (non-steroidal anti-inflammatory and anal- gesic agents; corticosteroids; antiproliferative drugs; other systemically administered drugs, e.g., gold and penicillamine; and topical administration). There is also a need for internationally accepted criteria for the evaluation of the surgical treatment of arthritic disorders. Orthopaedic surgery now plays an important part in the management of the arthritic patient and some means of standardization of results obtained by different groups is needed. Immunological evaluation In most therapeutic trials more attention is given and more value is ascribed to results of clinical 60)3 MEMORANDA assessment than to the results of laboratory tests. There may be several reasons for this. Firstly, the relevance of certain specific immunological tests to changes in clinical status in this field has been uncertain; and secondly, the failure of most single laboratory tests to predict clinical change has limited their assessment in depth (with the possible exception of complement levels and serial renal biopsies in systemic lupus). There is undoubtedly a need for much wider study of the behaviour of a panel of laboratory tests on a basis that can be achieved only by close collaboration between clinical and labora- tory workers. Operational considerations within in- dividual laboratory departments have generally limited the application of sufficiently inclusive panels of tests to therapeutic trials or even to the longitudi- nal assessment of individual patients. Renewed interest arises from considering whether antirheumatic therapy should be aimed at selectively suppressing or stimulating immune status. It is not yet clear whether the antirheumatic effects of anti- proliferative drugs can be ascribed to selective im- munodepression; nor has the potential of immuno- restorative therapy (e.g., with transfer factor) been settled, although results of pilot trials have been reported. Recent studies in systemic lupus suggest that a pattern of immunological abnormalities char- acteristic of a given patient can be established during the initial response to treatment, and that the order in which immunological tests revert towards normal- ity tends to be reversed in a subsequent exacerbation (33). This implies that, for individual patients and syndromes, a panel of selective laboratory tests repeated at regular intervals (during clinical remis- sion) may provide advance warning of impending relapse. For example, more information is needed on the predictive value of tests for circulating or tissue- bound immune complexes in both rheumatoid and lupus syndromes. Preclinical assessment of the anti- inflammatory action of antiproliferative drugs may yield closer parallels with clinical efficacy than arbi- trary experimental tests of immunosuppressive potency. In any therapeutic study of disorders involving the immune system, immunological testing should be carried out of both humoral and cellular immune systems, and attention should also be paid to tissue reactions. Selective immunodeficiency or heightened reactivity towards exogenous (microbial) agents would be worth investigation. Too many trials of immunosuppressive (antiproliferative) drugs have been performed without adequate immunological study and control, and it is recommended that such trials be accompanied by the maximum of attention to toxicity and follow-up. Immunopotentiation stu- dies should also be accompanied by monitoring of immune status. In view of the need to determine the predictive value of immunological tests in antirheumatic ther- apy, such information may more readily emerge from cooperative rather than from single-centre stu- dies. Although this presents problems, there is ample precedent for the concept of collaborative trials (34- 37). An international multidisciplinary group might agree upon a " division of labour " between labora- tory departments cooperating with clinical centres in attempts to establish laboratory criteria of predictive value. Registration of diseases Different centres may vary both in laboratory expertise and in the availability and organization of clinical material, yet both laboratory and clinical excellence are needed to obtain the maximum in- formation from therapeutic trials. Inter-centre co- operation at both national and international levels would benefit from the establishment of regional and national disease registers to provide information on the availability of patients in different categories for therapeutic assessment. The setting up of such regis- ters should be encouraged by national funding bodies. Disease registers would be of value to any multidisciplinary group concerned with the improve- ment of therapeutic criteria. Cost of therapeutic research Funding of research is inadequate, and in any therapeutic service a proportion of the expenditure should be devoted to further research in that field. Project research grants are often found to be inap- propriate for their particular purpose, and a greater use of grants to individuals or departments may be a better way of allocating funds. RECOMMENDATIONS The following recommended areas of further re- search reflected the special interests of the signatories in the relationships of microbial infection and im- munology to the inflammatory rheumatic diseases. (1) Further epidemiological studies on a larger scale would be of value in comparing the prevalence of different patterns of rheumatic and endemic in- fective disease in different geographical areas. Such 604 COOPERATION IN RHEUMATIC DISEASE RESEARCH comparative studies could be supplemented by pro- spective or retrospective enquiries concerning the relationship between epidemics of infectious disease and the incidence of rheumatic syndromes in both children and adults. (2) Smaller-scale survey work concerning the pre- valence and incidence of rheumatic diseases in com- munities should be encouraged where enquiry is directed towards particular hypotheses or questions and wheie internationally accepted clinical and labo- ratory criteria of rheumatic disease and activity are used. Examples are: (a) the determination of micro- bial infection as a cause of relapse in chronic rheumatic disease; (b) the prospective value of cer- tain laboratory test abnormalities in predicting re- lapse in rheumatoid arthritis and systemic lupus; and (c) the diagnostic importance of a given clinical sign, symptom, or laboratory test in the assessment of multisystem autoimmune disease. (3) In smaller community survey work, it would be desirable for some of the following test procedures to be introduced; (a) measurement of titre, class, subclass and, where appropriate, affinity or avidity, of autoantibodies to immunoglobulin determinants, to nuclear constituents, and to cytoplasmic tissue antigens; (b) the detection, measurement, and anal- ysis of immune complexes in serum, in inflammatory exudate, and (in sequestered form) in cells and tissues, by precipitation and extraction techniques; (c) determination of selected complement com- ponents and their state of activation in serum and inflammatory exudates; (d) delineation by surface marker tests of lymphocyte subpopulations in blood and inflammatory exudates; (e) delineation of lym- phocyte function by the cellular response in vitro to immunogenic or metabolic stimulators, or to modu- lators of lymphocyte activity; (f) the study of poly- morph/monocyte function in peripheral blood and inflammatory exudate by means of biochemical, surface-marker, and phagocytosis tests; and (g) the assessment of delayed hypersensitivity status to " re- call" antigens. Such studies would provide a basis for assessing the need for the standardization of particular test procedures. (4) In view of the association of the W-27 leuco- cyte antigen (HLA) type with ankylosing spondylitis and reactive arthritis states it would be of interest to extend HLA typing in different rheumatic syndromes by means of the techniques of mixed lymphocyte culture. Where there is evidence of leucocyte antigen restriction in a given syndrome, it would then be of great interest to study in parallel the responsiveness of such subjects to primary immunization and their immune status to " recall " antigens. The presence of these correlations would add weight to the evidence suggesting that expression of rheumatic disease may be genetically linked to abnormalities in the im- munological response to extrinsic antigens. (5) The standardization of immunological test procedures in widespread use, as well as those that are still in a developmental stage (e.g., cell-binding tests for immune complexes and cytotoxicity tests for lymphocyte function), is most important. Three aspects of this field are of particular interest: (a) the establishment of greater uniformity and of optimal conditions for the individual test procedures; (b) the development of working reagents of test components and of reaction products, which permit laboratory standardization for use in calibration; and (c) the need for cooperative work between different centres evaluating individual tests or reagents and for the introduction of quality control to assist the inter- pretation of results from cooperating centres. (6) When the value of a working standard in a laboratory test procedure is being assessed, an important criterion is that of reduced inter-test variation achieved by the inclusion of the standard reagent as part of the test design. The exchange of suitable working standards among cooperating centres would greatly facilitate the interpretation of immunological laboratory tests in individual pa- tients, in survey work, and in the monitoring of therapeutic trials. (7) Further studies are needed of the relative weighting of different clinicopathological criteria for the classification of polyarthritis and of multisystem rheumatic disease, as well as for survey work, for diagnosis, and for therapeutic monitoring. (8) The importance of expanding research into the clinical pharmacology of the rheumatic diseases arises from the increasing number of analgesic, anti- inflammatory, and antiproliferative drugs being offered for trial without detailed definition of their relative action upon particular pathological proces- ses in these diseases. For example, such detailed information would contribute to an understanding of the pathogenesis of those rheumatic syndromes ac- companied by immunological abnormalities, and could provide a scientific basis for assessing the relative potential of immunosupportive or immuno- suppressive therapy. 605 606 MEMORANDA (9) It is important for the integration of current scientific and clinical thinking into health services research that rheumatological centres should ex- change information in the different areas of in- vestigative interest. Useful examples would be: (a) the design of therapeutic trials; (b) information concerning drug toxicity; (c) the establishment of criteria for functional disability and for the evalua- tion of surgical treatment; (d) the evaluation of clinicopathological and immunological criteria in disease classification, survey work, and patient moni- toring; (e) the development of standard reagents to improve uniformity in laboratory test procedures; and (f) the establishment of registers that would indicate the impact on the community of rheumatic disease. (10) The future role of the World Health Organi- zation in the field of rheumatic diseases will be in facilitating cooperation between different centres, including the International League against Rheu- matism. In the fulfilment of these objectives it would be of great help if rheumatological centres were to inform WHO of their particular fields of interest and expertise. * * * M. A. Akhmeteli, Director, Division of Non-Communi- cable Diseases, World Health Organization, Geneva, Switzerland P. Barcelo, President, International League against Rheu- matism, Barcelona, Spain E. V. Barnett, Center for the Health Sciences, University of California, Los Angeles, USA W. W. Buchanan, Centre for Rheumatic Diseases, Glas- gow, Scotland E. G. L. Bywaters, Director, MRC Rheumatism Research Unit, Canadian Red Cross Memorial Hospital, Tap- low, Maidenhead, England J. Decker, Chief, Rheumatology Service, National Insti- tute of Arthritis, Metabolism and Digestive Diseases, National Institute of Health, Bethesda, MD, USA F. Delbarre, Directeur, Centre de Recherches sur les Maladies Osteoarticulaires, Hopital Cochin, Paris, France A. St J. Dixon, Royal National Hospital for Rheumatic Diseases, Bath, England D. C. Dumonde, Kennedy Institute of Rheumatology, Hammersmith, London, England L. E. Glynn, Director, Kennedy Institute of Rheumato- logy, Hammersmith, London, England V. Khatchatourov, Division of Non-Communicable Diseases, World Health Organization, Geneva, Switzer- land F. Maldyk, Director, Institute of Rheumatology, War- saw, Poland P. Matulis, Director, Institute of Experimental and Clinical Medicine, Vilnus, USSR V. A. Nassonova, Director, Institute of Rheumatology, Academy of Medical Sciences, Moscow, USSR J. B. Natvig, Institute of Immunology and Rheumato- logy, Rikshospitalet and Oslo Sanitetsforening Univer- sity Hospitals, Oslo, Norway I. M. Roitt, Department of Immunology, Middlesex Hospital Medical School, London, England J. T. Scott, Head, Clinical Research Division, Kennedy Institute of Rheumatology, Hammersmith, London, England L. E. Shulman, Johns Hopkins University School of Medicine, Baltimore, MD, USA T. Strasser, Cardiovascular Diseases, World Health Organization, Geneva, Switzerland G. Torrigiani, Immunology, World Health Organization, Geneva, Switzerland P. H. N. Wood, ARC Epidemiology Research Unit, Stopford Building (University of Manchester), Man- chester, England M. Ziff, University of Texas Southwestern Medical School, Dallas, TX, USA REFERENCES 1. WHO EXPERT COMMrTrEE ON RHEUMATIC DISEASES. WHO Technical Report Series, No. 78, 1954. 2. WHO EXPERT COMMITTEE ON RHEUMATIC DISEASES. Prevention of rheumatic fever. WHO Technical Re- port Series, No. 126, 1957. 3. WHO EXPERT COMMITrEE ON RHEUMATIc DISEASES. Prevention of rheumatic fever. WHO Technical Re- port Series, No. 342, 1966. 4. KELLGREN, J. H. The epidemiology of rheumatic diseases. Annals of the rheumatic diseases, 23: 109 (1964). 5. KELLGREN, J. H. Epidemiology of rheumatoid arthri- tis. Arthritis and rheumatism, 9: 658 (1966). 6. LAWRENCE, J. S. ET AL. Geographical studies of rheumatoid arthritis. Annals of the rheumatic dis- eases, 25: 425 (1966). COOPERATION IN RHEUMATIC DISEASE RESEARCH 607 7. WOLFE, A. M. The epidemiology of rheumatoid arthritis: a review. Part I. Surveys. Bulletin on rheumatic diseases, 19: 518 (1968). 8. MASI, A. T. The epidemiology of rheumatoid arthri- tis: a review. Part II. Incidence and diagnostic criteria. Bulletin on rheumatic diseases, 19: 524 (1968). 9. GREENWOOD, B. M. Polyarthritis in Western Nigeria. I. Rheumatoid arthritis. Annals of the rheumatic diseases, 28: 488 (1969). 10. ALLAN, G. Twin research: problems and prospects. In: Steinberg, H. G. & Bearn, H. G., ed. Progress in medical genetics. London, Heinemann, 1965, p. 242. 11. WORLD HEALTH ORGANIZATION. The use of twins in epidemiological studies. Acta geneticae medicae et gemellologiae, 15: 109 (1966). 12. O'BRIEN, W. M. Twin studies in rheumatic diseases. Arthritis and rheumatism, 11: 81 (1968). 13. BERNRSCHKE, K. & KM, C. K. Multiple pregnancy. New England journal of medicine, 2*8: 1276, 1329 (1973). 14. GORDON, T. ET AL. Some methodologic problems in the long-term study of cardiovascular disease: ob- servations on the Framingham study. Journal of chronic diseases, 10: 186 (1959). 15. DAWBER, T. R. ET AL. An approach to longitudinal studies in a community: the Framingham study. Annals of the New York Academy of Sciences, 107: 539 (1963). 16. FREDMAN, G. D. An evaluation of follow-up methods in the Framingham heart study. American journal ofpublic health, 57: 1015 (1967). 17. WORLD HEALTH ORGANIZATION. Cell-mediated im- munity and resistance to infection. WHO Technical Report Series, No. 519, 1973. 18. ANDERSON, S. G. ET AL. International reference preparation of rheumatoid arthritis serum and a preparation of rabbit antibody to sheep red blood cells (amboceptor). Bulletin of the World Health Organization, 42: 311 (1970). 19. ANDERSON, S. G. ET AL. Anti-nuclear factor serum (homogeneous): an international collaborative study of the proposed research standard 66/233. Annals of the New York Academy of Sciences, 177: 337 (1971). 20. RopEs, M. W. ET AL. The 1958 revision of diagnostic criteria for rheumatoid arthritis. Bulletin on rheu- matic diseases, 9: 175 (1958). 21. BENNETr, P. H. & WOOD, P. H. N. (EDITORS). Population studies of the rheumatic diseases. New York, Exerpta Medica, 1968. 22. BENNETr, P. H. & BURCH, T. A. New York sym- posium on population studies in the rheumatic dis- eases: new diagnostic criteria. Bulletin on rheumatic diseases, 17: 453 (1967). 23. FINK, C. W. ET AL. Criteria for the classification of juvenile rheumatoid arthritis. Bulletin on rheumatic diseases, 23: 712 (1972). 24. COHEN, A. S. ET AL. Preliminary criteria for the classification of systemic lupus arythematosus. Bul- letin on rheumatic diseases, 21: 643 (1971). 25. LIE, T. H. & ROTHFIELD, N. An evaluation of the preliminary criteria for the diagnosis of systemic lupus erythematosus. Arthritis and rheumatism, 15: 532 (1972). 26. DAVIS, P. ET AL. Criteria for the classification of SLE. British medical journal, 3: 88 (1973). 27. FRIES, J. F. & SIEGEL, R. C. Testing the preliminary criteria for the classification of SLE. Annals of the rheumatic diseases, 32: 171 (1973). 28. BARNETT, E. V. (EDrroR). Proceedings of the first workshop on the immunopathology of nucleic acids, Los Angeles, March, 1974. Los Angeles, University College Press, 1974. 29. HANSON, V. & KORNREICH, H. Systemic rheumatic disorders (" collagen disease ") in childhood: lupus erythematosus, anaphylactoid purpura, dermato- myositis and scleroderma (parts I and II). Bulletin on rheumatic diseases, 17: 435, 441 (1967). 30. PLOTZ, C. M. & SPIERA, H. Polymyalgia rheumatica. Bulletin on rheumatic diseases, 20: 578 (1969). 31. DEODHAR, S. D. ET AL. Measurement of clinical response to anti-inflammatory drug therapy in rheu- matoid arthritis. Quarterly journal of medicine, 42: 387 (1973). 32. LEE, P. ET AL. The evaluation of anti-rheumatic drugs. Current medical research and opinion, 1: 427 (1973). 33. EPSTEIN, W. In: Barnett, E. V. ed. Proceedings of the first worshop on the immunopathology of nucleic acids, Los Angeles, March, 1974. Los Angeles, University College Press, 1974. 34. MEDICAL RESEARCH CouNCIL/NuFFIELD FOUNDA- ION. Report on a comparison of predisolone with aspirin or other analgesics in the treatment of rheumatoid arthritis. Annals of the rheumatic dis- eases, 18: 173 (1959). 35. EMPIRE RHEUMATISM COUNCIL. Report of a multi- centre controlled trial on gold therapy in rheumatoid arthritis. Annals of the rheumatic diseases, 19: 95 (1960). 36. AMERICAN RHEUMATIM ASSoCATION Co-OPERATNG CLINCS. A seven-day variability study of 499 patients with peripheral rheumatoid arthritis. Arthritis and rheumatism, 8: 302 (1965). 37. AMERICAN RHEUMATISM ASSOCIATION CO-OPERATING CLimcs. A three-month trial of Indomethacin in rheumatoid arthritis with special reference to analysis and inference. Clinical pharmacology and therapeu- tics, 8: 11 (1967).