Diagnosis of rheumatism by means of universal criteria, obtained by the cybernetic method*

Diagnosis of rheumatism by means of universal criteria, obtained by the cybernetic method * L. I. BENEVOLENSKAJA, N. A. ANDREEV, M. E. KURMAEVA, E. A. NADEZDINA, M. M. BRZEZOVSKIJ, G. I. CYGANOVA, A. E. FELDMANIS, L. V. GARGASAS, P. V. MAL'CEV, & A. V. ORLOV-MOROZOV With a view to establishing standardized criteria for the diagnosis of rheumatism suitable for widespread use in polyclinics and in large-scale epidemiological investigations, 3000 adults and children with rheumatism or one of seven similar diseases were subjected to multiple screening in eight centres in the USSR using a specially prepared standard card and standard methods. The information obtained was processed by cybernetic methods. Following determination of the sensitivity, specificity, and informational value of 486 signs of rheumatism, 68 of the most indicative signs were selected as criteria of rheumatism. These were included in the final table and weighted according to their indicational value both individually and in groups as syndromes. Using the table it is possible by simple calculation, without using a computer, to arrive at an objective evaluation of the reliability of the clinical diagnosis of rheumatism in patients with one of a group of eight similar diseases. The final diagnostic table was tested on 562 patients selected epidemiologically orfrom polyclinics or clinics and was shown to result in a high degree ofdiagnostic precision (higher than 95 %). Extensive research on different aspects of the epi- demiology of rheumatism a is being carried out in various countries of the world. These studies cover the prevalence, incidence, mortality, etiological fac- tors, and social aspects of the disease. However, a comparison of the data obtained and evaluation * From the Laboratory for the Epidemiology of Collagen Diseases and Arthritides (Chief, Dr Med. Sci. L. I. Bene- volenskaja), Institute for Research on Rheumatism of the Academy of Medical Sciences of the USSR, Moscow, and the Division of Medical Cybernetics (Chief, Dr Med. Sci. N. A. Andreev) of the Central Research Laboratory of the Riga Medical Institute, Riga, Latvian SSR. a In the USSR the term "rheumatism " connotes acute rheumatic fever, chronic rheumatic heart disease and also a definite history of rheumatic fever but without subsequent formation of heart defects. Under the classification accepted in our country since 1956, active and inactive phases of the diseases have been defined. Acute rheumatic fever always corresponds to the active phase of the illness. The two other conditions can be found, either in a patient in the course of rheumatic fever, i.e. in the active phase, or in a patient in the inactive phase. of the results of the research are rendered somewhat difficult owing to the absence of standard criteria for diagnosis adequate for the epidemiological group involved. Diagnostic criteria for rheumatism were first formu- lated by the well known Soviet paediatrician, A. A. Kisel' in the 1930s and 1940s (1, 2). Later, the cri- teria drawn up by the American cardiologist, T. Jones (3), as modified by the American Rheu- matological Association (4) and supplemented by Academician A. I. Nesterov (5, 6) became widely used. However, the majority of clinicians are some- what dissatisfied with the existing criteria. The reasons for this are as follows. The Kisel'-Jones-Nesterov criteria are based on the diagnosis of rheumatic fever alone and are appli- cable mainly under clinical conditions. This makes it very difficult to assess a case under outpatient conditions, particularly in mass preventive screen- ing, where patients with rheumatic fever are rela- 3458 - 453 BULL. WQRLD HEALTH ORGAN., Vol. 53, 1976 L. I. BENEVOLENSKAJA ET AL. tively rare and patients with rheumatic heart defects or with some history of rheumatic fever without subsequent formation of heart defects are most common. The existing criteria were established for clinical forms of rheumatism which, in recent years, have become less frequent. The number of patients with acute onset of the disease and marked exudative manifestations has diminished, while blurred, atypi- cal, and latent active forms have shown a relative increase. Thus there has been a certain lack of cor- respondence between the changed clinical picture of rheumatism and the criteria for its diagnosis. More- over, the existing diagnostic criteria are based on a syndromic evaluation of the main manifestations of the disease, whereas distinct criteria for the diag- nosis of the syndromes themselves have not been drawn up. Finally, the different variants and modifications of the diagnostic criteria mentioned above (1-10) are largely empirical, being based mainly on deter- mination of the frequency with which the symptoms and syndromes are encountered. Their sensitivity and specificity have not been defined mathematically as was rightly pointed out by a WHO Expert Committee (11) in 1966. The use of mathematical methods and computers, of course, makes it possible to detect and select the most informative symptoms. Criteria for diagnosing rheumatism must ensure maximum reliability, be universal (i.e., must be adequate for any group of patients to be subjected to examination), and must also be widely used and standardized. MATERIALS AND METHODS From 1970 to 1973, 3000 patients with definite rheumatism or one of seven other similar diseases were examined by combined clinical, laboratory, instrumental, and graphic methods under the guid- ance of the Institute of Rheumatism of the Academy of Sciences of the USSR in eight different centres of the country with a view to drawing up standardized criteria. The seven other diseases were tonsillogenic cardiopathy (425.500), infectious/allergic polyar- thritis (714.9), rheumatoid arthritis, acute myocar- ditis (422), congenital heart defects, thyrotoxic heart disease, and neurocirculatory asthenia (305.3). From the outset particular attention was paid to ensuring the reliability of the primary medical information. This was achieved by: (1) use of a standardized system and methods for examining the patient; (2) use of data directly obtained from study of the patient and not of material from files; (3) the participation of highly qualified specialists in the work. In order to exclude any possible subjectivity in the interpretation of individual symptoms, a seminar for those participating directly in the work was held at the Institute of Rheumatism and, based on their experience, indications were drawn up for the meth- ods of completing examination cards. The patients were examined by a standard method using a standard card containing at the outset 486 signs of the diseases listed above, enumeration of the symptoms of rheumatism being as complete as possible. The diagnoses were made by a group of three experienced physicians. In addition, the opinions of specialists in functional diagnosis and radiology were taken into account. In our study, the diagnosis of rheumatism was based on the ARA criteria as modified in 1964 by Johns. Selective evalu- ation of the results of the laboratory findings was also carried out. To this end 30 sera were obtained from each centre for a second testing. The recruitment of research workers from various scientific centres and the collection of primary infor- mation on patients from different climatic and geo- graphical areas of the Soviet Union helped in estab- lishing standardized diagnostic criteria free from the influence of any single school or local peculiarities of the disease. The information obtained was processed on M 220-A and GE-400 computers. A standardized cybernetic diagnostic system drawn up by the Visnevskij Surgical Institute was used. An algorithm was employed based on the application of Bajesov's deciding principle. A matrix table of probability was established for the diagnosis of rheumatism. The sensitivity, specificity, and indicational value (12, 13) of each of the 486 signs were determined and 68 of the most informative symptoms were selected as criteria of rheumatism. RESULTS AND DISCUSSION The diagnostic criteria are given below. In view of the many and varied clinical manifestations of the disease and the absence of pathognomonic symp- toms for this disease, the original syndromic ap- proach to diagnosis has been maintained. 454 DIAGNOSIS OF RHEUMATISM BY MEANS OF UNIVERSAL CRITERIA Diagnostic criteria of rheumatism Vi No. Symptoms and syndromes info in 1. Onset of the disease 1. Age 7-15 years 2. 1 to 2 weeks after nasopharyngeal infection 3. Appearance of, or increased, breathlessness 4. Appearance of increased pulse rate 5. 6. 7. 8. ilue irmal bits 3.1 2.2 1.9 1.5 Appearance of, or increased, articular pains 1.0 Appearance of increased erythrocyte sedimenta- tion rate 2.0 Appearance of increased changes in ECG 2.8 Appearance of, or increased, fever 2.3 Threshold intervals for the evaluation of the syndrome in conventional units - 20. ion 21. (i) 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. ri< 8.3 - 88.3-11.5 11.6-16.8 0 conventional 1 conventional 3 conventional units unit units II. History ofjoint conditions 9. Arthralgia a 10. Pains in the large and medium joints equivalent to arthritis b 11. Pains in the large and medium joints lasting up to one week 12. Migratory pains in the joints c 13. Swelling of two or more joints 14. Swelling of large or medium joints 15. Swelling of joints lasting up to a week 16. Symmetrical nature of swelling IV. Heart defects Significant apical systolic murmur Long systolic murmur (more than 1/2 systole) Systolic murmur transmitted towards the axilla region Systolic murmur combined with first sound or high frequency (phonocardiogram) Apical mid-diastolic murmur Apical presystolic murmur Protodiastolic murmur in the 3rd and 4th inter- costal spaces to the left of the sternum Signs of enlargement or overloading of the left auricle (X-ray or ECG) Considerable enlargement of left ventricle in oblique projection (X-ray) Signs of enlargement or overloading of left ven- tricle (X-ray or ECG) Signs of enlargement or overloading of the right ventricle (X-ray or ECG) Slight opening snap of the mitral valve (phono- cardiogram) Q-1 sound over 0.07 s: for children over 0.05 s 4.1 2.3 3.3 2.5 4.4 6.1 5.3 3.5 3.5 1.3 3.0 4.7 2.9 si<7.0-0 7.0-10.4---2 10.5-46.9---5 1.9 33. 34. 2.4 35. 36. 3.1 37. 3.2 38. 3.2 3.1 39. 3.1 40. 3.1 ri<5.8 - 0 5.8-11.9-1i 12.0-23.1 -3 111. History of other indicative signs 17. Chorea 3.7 18. Positive effect of treatment by drugs against rheumatism 2.8 19. Diagnosis of rheumatism confirmed by reliable medical documentation expertly assessed by a doctor 3.0 | i<3.0-0 3.0-5.8---2 5.9-9.5----5 41. 42. 43. 44. 45. 46. 47. V. Carditis Tachycardia Muffled first sound at apex Third sound Changes in intensity and timbre of third sound Moderate apical systolic murmur Changes in intensity, timbre, and duration of systolic murmur Changes in intensity, timbre, and duration of diastolic murmur Systolic murmur in the 3rd and 4th intercostal spaces to the left of the sternum (more than 1/2 systole) Change in duration of PQ interval (ECG) T waves 1, II V4-6 reduced in amplitude (less than 2 mm, ECG) Flattening of T wave (ECG) Alteration in T-wave changes (ECG) Signs of moderate enlargement of left ventricle(X-ray) Change in heart dimensions (X-ray) Shallow pulsations (X-ray) zi<4.0-0 4.0-11.1-1 112.0-29.3-3 VI. Polyarthritis 48. Pains in the large or medium-sized joints equiva- lent to arthritis b 49. Pains in the large or medium-sized joints lasting up to a week 50. Pains in the large or medium-sized joints of migra- tory character c 455 0.6 0.7 1.9 28 0.6 3.8 3.1 1.5 2.8 1.2 1.7 2.7 0.6 3.9 2.0 a Arthralgia-attacks of pain in the regions of the joints lasting up to several hours. b Pains in the joints equivalent to arthritis-pains lasting not less than 24 hours and impeding movement. C Migratory pains in the joints-pains occurring in differentjoints over a period of 1-7 days. 2.2 2.8 2.8 L.-_ L. I. BENEVOLENSKAJA ET AL. 51. Swelling of the large or medium-sized joints 1.9 52. Symmetrical swelling of the joints 2.0 Xi<4.0 0 4.0-6.1 1 6.2-11.7-3 53. 54. 55. 56. VII. Chorea Hyperkinesia of the extremities Muscular hypotonia " Flaccid shoulders" symptom Positive finger-nose or heel-knee test zi<3.1 0 3.1-3.7-2 3.8-7.8-5 VIII. Skin changes 57. Annular erythema 2.0 58. Subcutaneous rheumatic nodules 1.7 .ri<1.0 0 1.0-2.0 2 2.1-3.9-3 IX. Symptoms diminishing the reliability of diagnosis as " rheumatism " 59. Heart murmur discovered at age under 1 year (-0.7) 60. Affection of fresh joints occurring after 1-2 months without appreciable reduction of the affection in other joints ( -0.8) 61. Three or more attacks of arthritis or carditis without the development of any heart defect (-1.5) 62. Occurrence or history of morning stiffness (-0.7) 63. Trembling fingers in Romberg's position (-0.6) 64. Enlargement of the thyroid gland (-0.9) I =i (-0.6) 0 (-0.7)-(-1.4)--1 (-1.5)-(-5.2)--3 X. Laboratory data 65. Erythrocyte sedimentation rate more than 30 mm/h 1.2 66. ASL-0 titre 625 units and over 2.0 67. Streptococcal antigens in the blood (titre 1 :40 and over) 0.6 68. Seromucoid >82 mg/100 ml (optical density > 0.21 ) 0.6 -Pi<2.0 0 2.0-4.4-1 Total of conventional Diagnosis units for all syndromes <I Diagnosis of rheumatism rejected 2-4 Rheumatism probable >5 Definite rheumatism The symptoms are combined in 10 groups or syndromes, of which I, II, IV and IX may be encountered in both the active and inactive phases of the disease, while syndromes V, VI, VII, VIII, and X indicate the active phase of the rheumatic process. In the criteria described an important place is given to a history of rheumatism, the main signs of which are equally as informative as the criteria indicative of current carditis and polyarthritis. The history is represented by three groups of symptoms reflecting the beginning of illness, with an indication of any chronological link with nasopharyngeal infection, arthritis suffered in the past, and other informative signs typical of rheumatism. Rheumatic defect of the heart is typified by 13 highly indicative symptoms. In particular, men- tion should be made of marked apical systolic murmur transmitted towards the axilla, apical pre- systolic and mid-diastolic murmurs, protodiastolic murmur in the 3rd and 4th intercostal spaces to the left of the sternum, slight opening snap of the mitral valve, and also signs of overloading of the left or right ventricles or of the left auricle. For rheumatic carditis the most informative symp- toms reflecting the dynamics of the process are changes in the sounds, murmurs, heart dimensions, and electrocardiographic signs. Naturally, the shorter the period over which the changes have occurred the more confidently can the doctor suspect carditis. Such symptoms as muffling of the heart sounds, apical systolic murmur, and tachycardia are not highly specific and their significance for differential diagnosis is therefore small. However, these symp- toms can draw the doctor's attention to the heart conditions at an early stage. Among the signs of cardiac involvement, the char- acter of the systolic murmur presents the greatest danger of subjective evaluation. We evaluated the quality of the systolic murmur on a five-point scale: very weak, weak, moderate, strong, very strong. For rheumatic affection of the heart, the most character- istic systolic murmur is of medium strength. This murmur, usually long with a blowing timbre, is audible immediately after placing the stethoscope and after exercise; it may change after a deep breath but it does not completely disappear. Both weak murmurs and very strong murmurs are doubtful criteria for carditis. The polyarthritis syndrome, which is detectable on objective examination by a combination of signs, is not essentially different from the similar grouping of symptoms in the history of joint conditions (Syn- 456 DIAGNOSIS OF RHEUMATISM BY MEANS OF UNIVERSAL CRITERIA drome IL), although fewer signs are included in the polyarthritis group as a doctor is better able to evaluate the pathological changes. The most frequent form of affection of the nervous system in rheumatism is chorea and this can be defined by four signs. In recent years, besides the typical manifestations of chorea, atypical variants of chorea minor have often been observed. These patients presented distal myoclonic hyperkinesia. Thus, even slight involuntary irregular movements of the extremities must be taken into account. Among the earliest manifestations of chorea are known to be changes in the mental and endocrino- vegetative sphere. The appearance of emotional insta- bility, disorders of the memory and attention, and heightened perspiration and sensitivity to cold are all symptoms that should alert the doctor. However, a diagnosis of chorea can be considered reliable only when there is development of typical hyper- kinesia accompanied by muscular hypotonia or impairment of movement coordination. Annular erythema and rheumatic nodule are ex- tremely specific signs of rheumatism. Their appear- ance with a background of carditis or polyarthritis confirms the rheumatic nature of these signs and, in combination with an acquired heart defect, the changes indicated provide evidence of the activation of the process. It is, however, essential to note that these signs, particularly the rheumatic nodules, are at present very seldom encountered and for this reason they are of very little practical importance. The laboratory indications are not very specific and have no independent significance. But, as supple- mentary criteria, an increase in the erythrocyte sedi- mentation rate, and particularly an increased anti- streptolysin-0 titre, are here of most diagnostic weight. For the first time so-called " negative" symptoms (group IX), which diminish the doctors' confidence in a diagnosis of rheumatism, have been introduced. These are the most specific symptoms of conditions that can be differentiated from rheumatism. In the table there is no breakdown into basic and supplementary criteria, although the diagnostic im- portance of the syndromes varies. Only two of the ten syndromes (acquired heart defect and chorea) are independent indicators of a diagnosis of rheu- matism. The others, including isolated carditis, and acute polyarthritis, provide reliable evidence of rheumatism only where they are definitely combined and pronounced. The diagnostic value of each sign is given in special mathematical units-bits-characterizing the quantity of information contributed by each sign towards the diagnosis. These bits are summated and are used to indicate the importance of the syndromes in conventional units. Diagnosis using the table was carried out in the following way. Patients were examined by a phy- sician for all the signs enumerated. Where a sign was present in a patient, the corresponding number of bits (i) was noted and these were added up for each syndrome (Zi). An evaluation was then made of this total in conventional units in accordance with the three threshold intervals worked out for each syndrome. Thus, it was possible even at this stage to determine which of the patient's syndromes were reliable and which only probable. The overall quantity of conventional units for all the syndromes was also calculated, the number of negative con- ventional units of the syndrome IX being subtracted from the overall total. The final total of conventional units was compared with the definitive threshold values shown at the end of the table. If this total was equal to 5 or more, we considered the diagnosis of rheumatism to be definite; if it was less than 2 such a diagnosis was rejected. Where intermediate figures were obtained the diagnosis of rheumatism was considered " probable" with different degrees of probability. The diagnostic aid using the above criteria was tested on 562 patients, adults and children, from epidemiological, polyclinic, and clinic groups. The overall number of patients with rheumatism was 214; patients with other diseases numbered 348. The epidemiological group (162 subjects) included people suspected of having rheumatism following prelimin- ary selection by mass screening of the population. The polyclinic group (258 subjects) consisted of a continuous series of patients who had visited the rheumatologist at a polyclinic for consultation. Among the patients in the clinic group, the cases that were most complicated to diagnose were spe- cially selected. Thus, out of 133 persons in the clinic group, 67 had primary rheumatism, 13 rheumatism with a latent and prolonged, mild course, 24 were suffering from incipient forms of rheumatoid arthri- tis, and 29 were classified as cases of acute myo- carditis. The results of diagnosis according to the criteria tabulated above were compared with the clinical diagnosis as confirmed by the computer. Such clini- 457 L. I. BENEVOLENSKAJA ET AL. cal diagnosis was considered to be well substantiated, since in earlier work with another series of pa- tients (14) computer diagnosis based on a matrix table ensured precise diagnosis in 92% of cases. In the group of patients with rheumatism the diag- nosis arrived at using the table of criteria agreed completely in 95.4% of cases and if those diagnosed as " probable " rheumatism were taken into account the agreement was 100%. There was no appreciable difference in the correctness of diagnosis in children and in adults. False positive diagnosis of rheuma- tism occurred in 6.3 % of cases, mainly in those diagnosed as " probable " rheumatism. Using the table, " probable" rheumatism was diagnosed mainly among the following groups: chil- dren with primary acute rheumatic fever in the absence of a typical history and extra-cardiac mani- festations; patients with acute myocarditis, infec- tious/allergic polyarthritis, and incipent forms of rheumatoid arthritis, i.e., in the most difficult cases to diagnose and verify. In order to establish a definitive diagnosis in these patients, a detailed exam- ination and observation over a period is essential and this has been reflected in our criteria. Thus, we have shown that the diagnostic criteria for rheumatism worked out according to cybernetic methods can be used for both children and adults in epidemiological investigations, in polyclinics, and in clinical situations. The application of standardized diagnostic criteria for rheumatism will thus make it possible for the doctor to evaluate the patients' symp- toms quantitatively, and will facilitate greater pre- cision and standardization of diagnosis, which is an indispensable condition for mass screening, case- finding, and follow-up and for the organization of curative and preventive care. RItSUME * CRITiRES UNIVERSELS POUR LE DIAGNOSTIC DU RHUMATISME iTABLIS PAR DES MflTHODES CYBERNtTIQUES * Les criteres de diagnostic elabores par Kisel', Jones et Nesterov, largement utilises dans la pratique m6dicale et les etudes scientifiques, ne sont congus que pour le diagnostic du rhumatisme articulaire aigu. II en resulte de grandes difficultes quand il s'agit de poser un diagnostic dans des conditions correspondant a celles d'une poly- clinique et, en particulier, lors des d6pistages preventifs de masse, etant donne que la majorite des malades presentent soit des cardiopathies rhumatismales, soit des antecedents de rhumatisme proprement dit, mais n'ayant pas entraines de sequelle cardiaque. D'autre part, une tendance du rhumatisme, observ6e au cours des dernires annees, a evoluer d'une facon plus benigne milite aussi pour une revision des criteres existants. Une objectivite maximale du diagnostic suppose une evaluation precise et faite dans les memes termes de chaque symptome et de chaque syndrome de la maladie. Une telle evaluation est possible si on definit la valeur d'indication des sympt6mes et des syndromes. Afin d'etablir des criteres normalises pour le diagnostic du rhumatisme, largement applicables dans la pratique clinique et lors des enqu8tes epidemiologiques de masse, on a proce&d dans huit centres en URSS a un examen multiple de 3000 adultes et enfants souffrant soit de rhumatisme, soit de l'une de sept autres maladies ana- logues, selon des proc6d6s standard comprenant l'utilisa- tion d'une carte speciale. Les donn6es obtenues ont ete traitees par la methode cybern6tique. On a choisi comme criteres de diagnostic 68 signes pr6sentant la plus grande valeur d'indication, apres avoir defini la sensibilite, la specificit6 et la valeur d'indication de 486 signes de rhumatisme. Les 68 signes choisis ont ete inscrits dans un tableau final et pond6r6s suivant leur valeur d'indication, aussi bien en tant que sympt8mes individuels que, groupes, en tant que syndromes. Ce tableau a permis d'evaluer objectivement la fiabilite du diagnostic clinique de rhumatisme pose dans le cas de malades atteints d'une maladie faisant partie d'un groupe de huit maladies similaires, et cela par un calcul simple, sans recours a l'ordinateur. Le tableau de criteres de diagnostic a 6t6 mis a l'6preuve sur 562 malades souffrant de rhumatisme (214 sujets) ou d'autres maladies (348 sujets) choisis lors d'enqu8tes epidemiologiques ou bien dans les polycliniques ou parmi divers groupes de malades. Les resultats obtenus grace au tableau ont et6 compares avec les diagnostics cliniques pos6s par un groupe d'experts et confirm6s par ordinateur Dans le cas du groupe de malades atteints de rhumatisme, le diagnostic a coIncid6 dans 95,4% des cas et, si l'on tient compte des diagnostics de rhumatisme . probable *, dans 100% des cas. Des diagnostics de rhumatisme faussement positifs chez les malades ne souffrant pas de cette maladie ont ete poses dans 6,3% des cas, la majorit6 de ces demiers etant diagnostiques comme des cas de rhumatisme # probable *. 458 DIAGNOSIS OF RHEUMATISM BY MEANS OF UNIVERSAL CRITERIA 459 REFERENCES 1. KISEL', A. A. In: Izbrannye trudy [Selected works], Moscow, 1960, pp. 40-49. 2. KISEL', A. A. In: Izbrannye trudy [Selected works], Moscow, 1960, pp. 203-206. 3. JONES, T. D. J. Am. med. Ass., 126: 481 (1944). 4. Circulation, 32: 664-668 (1965). 5. NESTEROV, A. I. Tezisy doklada. XIII Vsesojuznoy Nau6noy Konferencii Terapevtov, Moscow, 1963, pp. 116-119. 6. NESTEROV, A. I. Vopr. revm., 1: 3-12 (1967). 7. GRIFFITH, G. C. & SCHEINKOPF, G. A. Acta med. scand., Vol. 142, Suppl. 226, p. 429 (1952). 8. DAVIs, E. Lancet, 1: 1043 (1970). 9. DOLGONOLOVA, A. V. & KUZ'MINA, N. N. Pediatria, 4: 27-31 (1969). 10. KOLLE, G. Therapiewoche, 19: 240 (1969). 11. WHO Technical Report Series, 1966, No. 342. 12. BRIEzovsKu, M. M. ET AL. Ter. arhiv., 3: 62-64 (1974). 13. BYHOVSKU, M. L. & VYSNEvsKu, A. A. Kiberneti- ceskie sistemy v medicine [Cybernetic systems in medicine], Moscow, 1971, pp. 15-20. 14. BENEVOLENSKAYA, L. T. ET AL. In: Dixon, A. St. J- et al., ed., XIII International Congress of Rheumato. logy (Abstracts), Excerpta Medica, 1973, p. 93.