35 1
EPIDEMIOLOGIC AND IMMUNOGENETIC ASPECTS OF POLYMYALGIA RHEUMATICA AND GIANT CELL ARTERITIS IN NORTHERN ITALY CARLO SALVARANI, PIERLUIGI MACCHIONI, FABIO ZIZZI, WILMA MANTOVANI, FULVIA ROSSI, CONCETTA CASTRI, NICHOLAS CAPOZZOLI, ROBERTO BARICCHI, LUIGI BOIARDI, FRANC0 CHIARAVALLOTI, and ITALO PORTIOLI
We studied the epidemiology of giant cell arteritis (GCA) and polymyalgia rheumatica (PMR) in a Mediterranean population. Ninety-nine patients with PMR and/or GCA were identified over a 9-year period (19801988) in Reggio Emilia, Italy. The average annual incidence of PMR and GCA was 12.7/100,000 and 6.9/100,000, respectively, in a population aged 50 years or older. Frequencies of HLA antigens were determined in 49 patients with PMR and/or GCA who were followed by staff at our rheumatology unit during the 1980-1988 period. When compared with HLA findings in 242 healthy controls, DR4 was not found to be significantly associated with PMR (24% in PMR patients versus 14% in controls). Patients with GCA also showed an increased frequency of DR4 compared with controls (36% versus 14%), but this difference was also not statistically significant. The immunogenetic features of PMR and GCA and the relationship between the immunogenetic and epidemiologic patterns in different populations are discussed. From the Divisione di Medicina and Divisione di Oculistica, Reggio Emilia, Servizio di Medicina Interna, Unita Artrite Reumatoide, Bologna, Laboratorio di Patologia Clinica, Ospedale Malpighi, Bologna, Italy. Carlo Salvarani, MD: Arcispedale S. Maria Nuova; Pierluigi Macchioni, MD: Arcispedale s. Maria Nuova; Fdbio Zizzi, MD: Unita Artrite Reumatoide; Wilma Mantovani: Ospedale Malpighi; Fulvia Rossi, MD: Arcispedale S . Maria Nuova; Concetta Castri, MD: Arcispedale S. Maria Nuova; Nicholas Capozzoli, MD: Arcispedale S. Maria Nuova; Roberto Baricchi, MD: Arcispedale S . Maria Nuova; Luigi Boiardi, MD: Arcispedale S. Maria Nuova; Franco Chiaravalloti, MD: Arcispedale S. Maria Nuova; Italo Portioli, MD: Arcispedale S. Maria Nuova. Address reprint requests Carlo Salvarani, MD, Divisione di Medicina, Arcispedale S . Maria Nuova, V.le Umberto 1" N50,42100 Reggio Emilia, Italy. Submitted for publication February 23, 1990; accepted in revised form August 13, 1990. Arthritis and Rheumatism, Vol. 34, No. 3 (March 1991)
The pathogenesis of polymyalgia rheumatica (PMR) and giant cell (temporal) arteritis (GCA) is unknown. It has been suggested that genetic factors play a role in these diseases. There is evidence of familial aggregation (1,2). The highest incidence has been reported in northern Europe (Sweden and Denmark) (3-5) and in Minnesota (6-8), the population of the latter being composed primarily of people of northern European descent. A lower incidence has been observed in more southern regions (Israel, Italy, Tennessee, and Department Loire-Atlantique, France) (9-12). Some investigators have suggested that environmental factors, such as infectious agents, exposure to sunlight, and contact with birds, may also be involved in the pathogenesis of these diseases (13-15). It is likely that there are environmental or genetic factors, as yet unelucidated, that significantly influence the risk for these diseases in different populations. Recent studies have documented an association between the HLA-DR4 antigen and PMR, with or without GCA (16-18). In northern Italy, there is a low incidence of DR4 in the general population and in the population of patients with rheumatoid arthritis (RA) (19,20). To confirm the low incidence of PMR and GCA observed in a Mediterranean population, we have extended our previous epidemiologic study (10) to include the period 1980-1988. We also studied the frequency of HLA antigens in 49 PMR and/or GCA patients followed at our rheumatology unit during the period 1980-1988, to determine if genetic factors are important in explaining the racial distribution of these diseases.
352
SALVARANI ET AL
Table 1. Clinical groups and results of temporal artery biopsy in 99 patients with giant cell arteritis (GCA) and polymyalgia rheumatica (PMR)* All patients Clinical group
Men
Women
Total
GCA alone GCA + PMR PMR alone Total
10 (29) 5 (14) 20 (57) 35 (100)
13 (20) 15 (24) 36 (56) 64 (100)
23 (23) 20 (20) 56 (57) 99 (100)
Patients with histologic evidence of arteritis 11 (48) 9 (45)
* Values are the number (%).
PATIENTS AND METHODS Patients. Records of all patients who were diagnosed as having PMR or GCA at the Departments of Medicine, Ophthalmology, Neurology, or Rheumatology at Reggio Emilia Hospital between 1980 and 1988 were reviewed. We investigated only patients who resided in the Reggio Emilia metropolitan area (census tract number USL n.9) at the time of diagnosis. This area has a current population of -170,000. Annual information on the number of USL n.9 inhabitants and the age and sex distribution of the population was provided by the Statistics Department of Reggio Emilia province. During the study period (1980-1988), there were no significant variations in the total population or in the age and sex distribution. We believe all patients in whom a diagnosis of GCA was confirmed by biopsy were identified, since we examined the records of all patients in whom a temporal artery biopsy was performed at Reggio Emilia Hospital. Patients with PMR who had been treated at the Departments of Medicine and Rheumatology were identified from a review of the records of all patients with this diagnosis. Virtually all specialized medical care for the Reggio Emilia metropolitan area is provided at the Reggio Emilia Hospital. There are no private hospitals or rheumatologists in private practice in this area. The diagnosis of definite GCA was based on positive findings on temporal artery biopsy. However, we also considered patients with negative biopsy results, and a diagnosis of probable GCA was made if they had at least 4 of the following classic manifestations of GCA: abnormal temporal arteries (tender and swollen), loss of vision, jaw claudication, presence of PMR, and response to steroid therapy (7). The diagnosis of PMR was made if all the following were present: 1) persistent pain (for at least 1 month) involving 2 of the following areas: neck, shoulders, and pelvic girdle; 2) morning stiffness lasting more than 1 hour; 3) rapid response to prednisone (520 mgiday); and 4) absence of other diseases capable of causing the musculoskeletal symptoms (21). For both diagnoses, only patients over 50 years old and with an erythrocyte sedimentation rate >40 mm/hour were included. Ninety-nine patients with GCA or PMR were identified. The distribution of patients in each clinical group (GCA alone, PMR with GCA, and PMR alone) is shown in Table 1.
The diagnosis had been made by a family practitioner prior to the patient’s hospitalization in only 2 cases. Temporal artery biopsy was performed in 69 patients (a bilateral biopsy was performed in only 2 cases). Prior to 1986, all patients with signs of symptoms of either GCA or PMR had a temporal artery biopsy performed (56 patients). From 1986 on, biopsy was performed only in patients with signs or symptoms of arteritis (13 patients). The mean length of the temporal artery biopsy specimen was 0.5 cm, and multiple sections of each specimen were examined without prior knowledge of the clinical diagnosis. HLA typing. HLA typing was performed on 49 randomly selected patients with PMR and/or GCA. The healthy control group consisted of a pool of 242 unrelated blood donor volunteers from the same geographic area. Peripheral blood leukocytes were isolated by FicollHypaque density-gradient centrifugation (22). B lymphocytes were separated from T lymphocytes by an ox red blood cell rosetting technique, using monoclonal antibody H4 (One Lambda, Los Angeles, CA). The determination of class I (HLA-A, B , and C) and class I1 (HLA-DR and DQ) major histocompatibility complex antigens was performed using a modification of the NIH 2-stage complement microcytotoxicity test (23). Lymphocytes were typed using a panel of 180 antisera recognizing approximately 80 different specificities, from the Eighth and Ninth International Histocompatibility Workshops. Statistical analysis. Statistical analysis was performed using the Statistical Package for the Social Sciences (24). Chi-square with Yates’ correction and Fisher’s exact test were used to compare the frequencies of HLA antigens. P values were corrected for the number of antigens tested. Relative risks (RR) were calculated using the method of Woolf.
RESULTS Incidence of PMR, 1980-1988. A total of 76 patients with PMR (25 men and 51 women) were identified. The mean -t SD age at diagnosis was 69.7 -t 9 years (67.9 ? 10 for men, 70.6 +- 8 for women), with a range of 48-87 years. Age- and sex-specific incidence Table 2. Annual incidence of polymyalgia rheumatica in residents of Reggio Emilia, Italy, 1980-1988, by age and sex, per 100,000 population*
Men
Women
Total
Age group (years)
No.
Rate
No.
Rate
No.
Rate
0-49 50-59 60-69 70-79 280 Total 250
0 6 6 10 3 2s 25
0 6.3 7.2 17.8 17.7 3.4 9.7
1 6 15 25 4 51 50
0.2 5.7 14.0 29.9 10.1 6.4 14.9
1 12 21 35 7 76 75
0.1 6.0 11.0 25.0 12.0 4.9 12.7
* The total
population was 169,950 and the number of individuals 250 years old was 65,711.
PMR AND GCA IN NORTHERN ITALY
353
Table 3. Annual incidence of giant cell arteritis in residents of Reggio Emilia, Italy, 1980-1988, by age and sex, per 100,000 population*
Table 4. Frequency (9%) of HLA class I1 antigens in patients with polymyalgia rheumatica (PMR) with giant cell arteritis (GCA), in patients with PMR without GCA, and in controls
Women
Men
0-49 50-59 60-69 70-79 280 Total
250
Total
No.
Rate
No.
Rate
No.
Rate
Antigen
All PMR (n = 41)
0 2 6 6 1 15 15
0 2.1 7.2 10.7 5.9 2.0 5.8
2 5 8 10 3 28 26
0.4 4.8 7.5 11.8 7.6 3.5 7.8
2 7 14 16 4 43 41
0.2 3.5 7.3 11.4 7.1 2.8 6.9
DR I DR2 DR4 DRw53 DQ 1 DQ2 DQ3
22 19 24 46 46 27 51
PMR with GCA (n = 6)
PMR without GCA (n = 35)
17 17 33 83 34 67 50
23 20 23 40
48 20 51
Controls (n = 242) 16 22 14 43 51 40 47
* The total
population was 169,950 and the number of individuals 2 5 0 years old was 65,71 1.
rates of PMR are shown in Table 2. There was an increased incidence of PMR with increasing age, with a maximum in the eighth decade of life (17.8/100,000 for men and 29.9/100,000 for women). The incidence rate in the population 2.50 years old was approximately 1.5 times higher in women than in men. Incidence of GCA, 1980-1988. Forty-three patients (28 women and 15 men) met the diagnostic criteria for GCA. Twenty (46.5%) had a positive temporal artery biopsy result and 23 (53.3%) had a negative biopsy result but met the clinical criteria. The mean ? SD age at diagnosis was 67.6 ? 10 years (68.9 10 for men, 66.9 ? 11 for women), with a range of 46-85 years. The age-specific incidence rate in women increased from 4.8/100,000in the sixth decade of life to 11.8/100,000in the eighth decade (Table 3 ) . There was a corresponding increase in the incidence rates in men, with a peak level of 10.7/100,000in the eighth decade. There was a slightly greater incidence of GCA in women than in men in every age group. Temporal artery biopsy rate, 1980-1988. The mean rate of performance of temporal artery biopsy during the study period was 16.3/100,000 250 years old and 25.5/100,000 women 250 years old. The biopsy rate increased with age, with the maximum rate occurring during the eighth decade of life (37.3/100,000 men and 35.8/100,000 women). However, the rate decreased in the oldest patients (280 years old), with an incidence of 17.7/100,000 men and 22.7/100,000 women. HLA typing results. The distribution of HLA class I antigens in the patients did not differ significantly from that in the controls (data not shown). Tables 4 and 5 show the distribution of HLA-DR and HLA-DQ phenotypes in the 49 patients compared with the control group. The frequencies of class I1
*
antigens did not differ significantly in the patient subgroups compared with the controls. An increase of DR4 was found in all patient subgroups, particularly in the GCA patients (36% versus 14% in controls; RR = 3.3,95% confidence interval 1.0-10.5); however, none of these differences was statistically significant. The frequency of DRw53, which is in linkage disequilibrium with DR4, was not significantly increased in GCA patients compared with controls (78% versus 43%; RR = 4.9, 95% confidence interval 1.4-17.9).
DISCUSSION Our epidemiologic results show a lower incidence of PMR in a Mediterranean population than in the northern European and the Minnesota populations (3-8) (Table 6). An annual incidence of PMR of 12.7/ 100,000 population aged 2 5 0 years was observed in Reggio Emilia, Italy, compared with 53.7/100,000 in Olmstead County, Minnesota (6) and with 20.4/100,000 (recalculated to exclude patients with GCA alone) in Goteborg, Sweden (3,4). The annual rate of incidence of GCA found in the present study (Table 7) is similar to that reported in Table 5. Frequencies (%) of HLA class I1 antigens in patients with giant cell arteritis (GCA) with polymyalgia rheumatica (PMR), in patients with GCA without PMR, and in controls
Antigen
DR 1 DR2 DR4 DRw53 DQ 1 DQ2 DQ3
All GCA (n = 14)
GCA with PMR (n = 6)
GCA without PMR (n = 8)
7 14 36 78 21 50 64
17 17 33 83 34 67 50
0 12
37 75 12 37 75
Controls (n = 242) 16 22 14 43 51 40 47
SALVARANI ET AL
354
Table 6. Comparison of the incidence rates of polymyalgia rheumatica in various epidemiologic studies* Incidence/ I00,000/year Location, years Olmstead County, Minnesota, 1970-1979 Goteborg, Sweden, 1973-1975 Ribe County, Denmark, 1982-1985 Reggio Emilia, Italy, 1980-1988
All age groups
250
11.1
53.7
Age
6.7t
20.41
NR
68.3
4.9
12.7
* Refs. 3-8 and present study. NR = not reported. t Recalculated to exclude patients with giant cell arteritis alone
a French population in Department Loire-Atlantique (12), but is lower than that reported in the Olmsted County population and the Goteborg population (3,8). The lowest incidence of GCA has been observed in Israel (9) and in Shelby County, Tennessee (1 1) (especially among Sephardic Jews and blacks, respectively). Differences in the rates of performing of temporal artery biopsy may have resulted in a reduced propensity of physicians in Reggio Emilia to pursue a diagnosis of GCA, compared with physicians in Olmsted County. (The rates during 1980-1988 in our study were approximately one-third those recorded during 1980-1985 in Olmsted County during 1980-85 [8].) Conversely, the lower rate of performing temporal artery biopsy in the present study may be related to an actual lower incidence of GCA in the population of Reggio Emilia. Consistent with other studies (3,5,1I), we observed a reduction in the age-specific incidence of GCA in patients aged 80 years or older, whereas in Olmsted County, the rate continues to increase in the oldest patients (8). A declining incidence reported in this age group may be related to a reduced suspicion of the disease in the oldest patients, resulting in underdiagnosis. The relatively low rate of temporal artery biopsy in male and female patients aged 80 years or older supports this hypothesis. Our study included cases of “probable” GCA (negative biopsy result), based on criteria proposed by Huston et a1 (7). Of our 43 patients with clinical evidence of GCA, only 20 (46.5%) had a positive temporal artery biopsy result, in contrast with the Olmsted County study (8), in which the diagnosis of GCA was confirmed by biopsy in 94% of the patients
studied. These differing results could be explained by the differences in the mean length of the biopsy specimen (0.5 cm in our study versus 5.4 cm in the Olmsted County study) and by the practices in Olmsted County of performing a bilateral biopsy when findings in the first specimen are negative (25) (a second biopsy was performed in only 2 of our patients). However, none of our patients’ diagnoses were modified during the followup period (mean duration 64 months), making possible errors in the diagnosis unlikely. We cannot exclude the possibility that patients with a mild form of PMR or GCA were treated as outpatients, therefore escaping our attention. It is also possible that elderly patients with PMR may be misdiagnosed as having RA and treated only for RA by their family physicians (only 2 patients in our study had a correct diagnosis of PMR prior to admission). It is difficult to evaluate the impact of the low level of awareness of these diseases among general practitioners, but it is unlikely that this factor can fully account for the low incidence of PMR and GCA found in Reggio Emilia. Table 7. Comparison of the incidence rates of giant cell arteritis in various epidemiologic studies* Incidence/100,000/year, in individuals 2 5 0 years old
Location, years
Biopsyproven only
Including probable casest
NR NR
17.0 24.1
16.8
18.3$
NR
23.3
1.78 0.36
2.24 0.36
NR
9.4
NR NR
0.49 0.86
NR
6.9
~
Olmsted County, Minnesota, 1950-1 985 1980-1985 Goteborg, Sweden, 1973-1 975 Ribe County, Denmark, 1982-1985 Shelby County, Tennessee, 1971-1980 Whites Blacks Department Loire-Atlantique, France, 1970-1979 Israel 1960-1978 1975-1978 Reggio Emilia, Italy, 1980-1988
* Refs. 3-9, 11, 12, and present study. NR = not reported. t Probable clinical diagnosis in patients with a negative
biopsy result, or in patients in whom a biopsy was not performed. $ Recalculated to exclude patients with polymyalgia rheumatica only.
PMR AND GCA IN NORTHERN ITALY
355
It has been suggested that genetic factors may play a pathogenic role in both PMR and GCA. Richardson et a1 (17), in a metaanalysis of previous studies of HLA antigens in GCA, showed an association of HLA-DR4 with PMR, rather than with GCA alone. Cid et a1 (18) confirmed that, in Spanish patients with biops y-proven GCA, DR4 is mainly associated with PMR. The incidence of GCA and/or PMR has been shown to be low in blacks (26), paralleling the low frequency of HLA-DR4 observed in this population (27). Interestingly, there is a corresponding low incidence of PMR and GCA coupled with a low incidence of DR4 in the general population (range 12-14% [19,20]) in northern Italy. In the present study, in which we examined a group of patients with PMR and a second group with GCA, the frequency of DR4 was not significantly increased in either group compared with controls, and the highest incidence of this antigen was observed in the group of patients with GCA, a finding that is inconsistent with the results of other studies. In addition, the increased frequency of DRw53, which is in linkage disequilibrium with DR4, was not statistically significant in either PMR or GCA patients. Ferraccioli and Savi (19) reported a somewhat higher relative risk for DR1 (RR = 2.1) than for DR4 (RR = 1.3) in RA patients in northern Italy. In our PMR patients, a similarly weak and nonsignificant association with DR1 and DR4 was observed (RR = 1.4 and RR = 1.6, respectively). Conversely, we observed an increased frequency of DR4 and a reduction of DR1 (RR = 3.3 and RR = 0.7, respectively) among patients with GCA. However, this result is difficult to interpret due to the small number of GCA patients. Our results confirm a trend toward a decrease in the frequency of DR2 in PMR patients. In contrast to the results of Cid et a1 (18), we observed a reduction of the DQw2 antigen only in PMR patients without GCA, but this difference was again not statistically significant. Our study shows no significant increase in any of the HLA class I antigens, similar to the findings in other studies (17). The significant increase of HLA-DR4 found among PMR patients in other study populations ( 1 6 18) was thus not confirmed in our northern Italian population. A weak or nonexistent association with DR4 has also been reported in Italian patients with RA, as well as in RA patients of other Mediterranean countries (19,28-30). The evidence of a weak or absent association of DR4 with RA or PMR gives rise to the question whether disease susceptibility is linked to the
same epitope or combination of epitopes in all populations, or whether genetic heterogeneity is involved. Further studies using oligonucleotide probes are needed to define which epitopes are most closely related to PMR.
ACKNOWLEDGMENTS The authors thank Drs. Gene G. Hunder and David J. Ballard for their clinical review of the manuscript, Rita Farella for clinical assistance in the course of the study, and Jacqueline Costa for preparation of the manuscript.
REFERENCES 1. Granato JE, Abben RP, May WS: Familial association
of giant cell arteritis: a case report and brief review. Arch Intern Med 141:115-1 17, 1981 2. Kemp A: Monozygotic twins with temporal arteritis and ophthalmic arteritis. Acta Ophthalmol (Copenh) 55: 183190, 1977 3. Bengtsson B-A, Malmvall B-E: The epidemiology of giant cell arteritis including temporal arteritis and polymyalgia rheumatica. Arthritis Rheum 24:89%904, 1981 4. Bengtsson B-A, Malmvall B-E: Giant cell arteritis: incidence of giant cell arteritis. Acta Med Scand [Suppl] 658: 15-17, 1982 5 . Boesen P, Sgrensen SF: Giant cell arteritis, temporal arteritis, and polymyalgia rheumatica in a Danish County: a prospective investigation, 1982-1985. Arthritis Rheum 30:294-299, 1987 6. Chuang TY, Hunder GG, Ilstrup DM, Kurland LT: Polymyalgia rheumatica: a 10-year epidemiologic and clinical study. Ann Intern Med 97:672-680, 1982 7. Huston KA, Hunder GG, Lie JT, Kennedy RH, Elveback LR: Temporal arteritis: a 25-year epidemiologic, clinical and pathologic study. Ann Intern Med 88:162167, 1978 8. Machado EBV, Michet CJ, Ballard DJ, Hunder GG, Beard CM, Chu CP, O’Fallon M: Trends in incidence and clinical presentation of temporal arteritis in Olmsted County, Minnesota, 1950-1985. Arthritis Rheum 31:745749, 1988 9. Friedman G, Friedman B, Benbassat J: Epidemiology of temporal arteritis in Israel. Isr J Med Sci 18:241-244, 1982 10. Salvarani C, Macchioni PL, Tartoni PL, Rossi F, Baricchi R, Castri C, Chiaravalloti F, Portioli I: Polymyalgia rheumatica and giant cell arteritis: a 5-year epidemiologic and clinical study in Reggio Emilia, Italy. Clin Exp Rheumatol 5:205-215, 1987 11. Smith CA, Fidler WJ, Pinals RS: The epidemiology of giant cell arteritis: report of a ten-year study in Shelby County, Tennessee. Arthritis Rheum 26: 1214-1219, 1983 12. Barrier J, Pion P, Massari R, Peltier P, Rojouan J ,
SALVARANI ET AL
356
13.
14.
15. 16.
17.
18.
19.
20.
Grolleau JY: Approche epidemiologique de la maladie de Horton dans le department de Loire-Atlantique: 110 cas en 10 ans (1970-1979). Rev Med Interne 3:13-20, 1983 Elling H , Skinhoj P, Elling P: Hepatitis B virus and polymyalgia rheumatica: a search for HBsAg, HBsAb, HBcAb, HBeAg and HBeAb. Ann Rheum Dis 3951 1513, 1980 O’Brien JP: A concept of diffuse actinic arteritis: the role of actinic damage to elastin in ‘age change’ and arteritis of the temporal artery and in polymyalgia rheumatica. Br J Dermatol 98:l-13, 1978 Fessel WJ: Polymyalgia rheumatica, temporal arteritis and contact with birds. Lancet 11:124%1250, 1969 Armstrong RD, Panayi GS, Welsh KI: Polymyalgia rheumatica and rheumatoid arthritis: similarity of HLA antigen frequencies (letter). Arthritis Rheum 27: 14381439, 1984 Richardson JE, Gladman DD, Fam A, Keystone EC: HLA-DR4 in giant cell arteritis: association with polymyalgia rheumatica syndrome. Arthritis Rheum 30: 1293-1297, 1987 Cid M-C, Ercilla G, Vilaseca J, Sanmarti R, Villalta J, Ingelmo M, Urbano-Marquez A: Polymyalgia rheumatica: a syndrome associated with HLA-DR4 antigen. Arthritis Rheum 31 :678-682, 1988 Ferraccioli GF, Savi M: Association between DR antigens, rheumatoid arthritis with and without extraarticular features and SLE in northern Italy. J Rheumatol 15:51-53, 1988 Salvarani C, Macchioni PL, Zizzi F, Mantovani W, Rossi F, Baricchi R, Ghirelli L, Frizzier0 L, Portioli I:
21. 22.
23. 24. 25. 26.
27. 28.
29.
30.
Clinical subgroups and HLA antigens in Italian patients with psoriatic arthritis. Clin Exp Rheumatol 7:391-396, 1989 Healey LA: Long-term follow-up of polymyalgia rheumatica: evidence of synovitis. Semin Arthritis Rheum I3 :322-328, 1984 Bdyum A: Separation of leucocytes from blood and bone marrow. Scand J Clin Lab Invest [Suppl] 97:21-26, 1968 Terasaki PI, McClelland JD: Microdroplet assay of human serum cytotoxins. Nature 204:998-1000, 1964 SPSS” User’s Guide. Cary, NC, SAS Institute, 1986 Hall S, Lie JT, Kurland LT, Persellin S, O’Brien PC, Hunder GG: The therapeutic impact of temporal artery biopsy. Lancet II:1217-1222, 1983 Love DC, Rapkin J, Lesser GR, Shmookler BM, Kolsky MP, Jackson B, Barth WF: Temporal arteritis in blacks. Ann Intern Med 105:387-389, 1986 Alarcon GS, Koopman WJ, Acton RT, Barger BO: DR antigen distribution in blacks with rheumatoid arthritis. J Rheumatol 10579-583, 1983 Papasteriades CA, Kappou ID, Skopouli FN, Barla MN, Fostiropulos GA, Moutsopoulos HM: Lack of HLAantigen association in Greek rheumatoid arthritis patients. Rheumatol Int 5:201-203, 1985 Mercier P, Dallest R, Roux H: HLA-DR et polyarthrite rhumatoide de I’adulte: etude dans le Sud de la France. Rev Rhum Ma1 Osteoartic 47:465-470, 1980 Brautbar C, Naparstek Y, Yaron M, Amar A, Ehrenfeld M, Eliakim M, Bentwich Z, Cohen T, Korman SH, Albert ED: Immunogenetics of rheumatoid arthritis in Israel. Tissue Antigens 28:8-14, 1986
EPIDEMIOLOGIC AND IMMUNOGENETIC ASPECTS OF POLYMYALGIA RHEUMATICA AND GIANT CELL ARTERITIS IN NORTHERN ITALY CARLO SALVARANI, PIERLUIGI MACCHIONI, FABIO ZIZZI, WILMA MANTOVANI, FULVIA ROSSI, CONCETTA CASTRI, NICHOLAS CAPOZZOLI, ROBERTO BARICCHI, LUIGI BOIARDI, FRANC0 CHIARAVALLOTI, and ITALO PORTIOLI
We studied the epidemiology of giant cell arteritis (GCA) and polymyalgia rheumatica (PMR) in a Mediterranean population. Ninety-nine patients with PMR and/or GCA were identified over a 9-year period (19801988) in Reggio Emilia, Italy. The average annual incidence of PMR and GCA was 12.7/100,000 and 6.9/100,000, respectively, in a population aged 50 years or older. Frequencies of HLA antigens were determined in 49 patients with PMR and/or GCA who were followed by staff at our rheumatology unit during the 1980-1988 period. When compared with HLA findings in 242 healthy controls, DR4 was not found to be significantly associated with PMR (24% in PMR patients versus 14% in controls). Patients with GCA also showed an increased frequency of DR4 compared with controls (36% versus 14%), but this difference was also not statistically significant. The immunogenetic features of PMR and GCA and the relationship between the immunogenetic and epidemiologic patterns in different populations are discussed. From the Divisione di Medicina and Divisione di Oculistica, Reggio Emilia, Servizio di Medicina Interna, Unita Artrite Reumatoide, Bologna, Laboratorio di Patologia Clinica, Ospedale Malpighi, Bologna, Italy. Carlo Salvarani, MD: Arcispedale S. Maria Nuova; Pierluigi Macchioni, MD: Arcispedale s. Maria Nuova; Fdbio Zizzi, MD: Unita Artrite Reumatoide; Wilma Mantovani: Ospedale Malpighi; Fulvia Rossi, MD: Arcispedale S . Maria Nuova; Concetta Castri, MD: Arcispedale S. Maria Nuova; Nicholas Capozzoli, MD: Arcispedale S. Maria Nuova; Roberto Baricchi, MD: Arcispedale S . Maria Nuova; Luigi Boiardi, MD: Arcispedale S. Maria Nuova; Franco Chiaravalloti, MD: Arcispedale S. Maria Nuova; Italo Portioli, MD: Arcispedale S. Maria Nuova. Address reprint requests Carlo Salvarani, MD, Divisione di Medicina, Arcispedale S . Maria Nuova, V.le Umberto 1" N50,42100 Reggio Emilia, Italy. Submitted for publication February 23, 1990; accepted in revised form August 13, 1990. Arthritis and Rheumatism, Vol. 34, No. 3 (March 1991)
The pathogenesis of polymyalgia rheumatica (PMR) and giant cell (temporal) arteritis (GCA) is unknown. It has been suggested that genetic factors play a role in these diseases. There is evidence of familial aggregation (1,2). The highest incidence has been reported in northern Europe (Sweden and Denmark) (3-5) and in Minnesota (6-8), the population of the latter being composed primarily of people of northern European descent. A lower incidence has been observed in more southern regions (Israel, Italy, Tennessee, and Department Loire-Atlantique, France) (9-12). Some investigators have suggested that environmental factors, such as infectious agents, exposure to sunlight, and contact with birds, may also be involved in the pathogenesis of these diseases (13-15). It is likely that there are environmental or genetic factors, as yet unelucidated, that significantly influence the risk for these diseases in different populations. Recent studies have documented an association between the HLA-DR4 antigen and PMR, with or without GCA (16-18). In northern Italy, there is a low incidence of DR4 in the general population and in the population of patients with rheumatoid arthritis (RA) (19,20). To confirm the low incidence of PMR and GCA observed in a Mediterranean population, we have extended our previous epidemiologic study (10) to include the period 1980-1988. We also studied the frequency of HLA antigens in 49 PMR and/or GCA patients followed at our rheumatology unit during the period 1980-1988, to determine if genetic factors are important in explaining the racial distribution of these diseases.
352
SALVARANI ET AL
Table 1. Clinical groups and results of temporal artery biopsy in 99 patients with giant cell arteritis (GCA) and polymyalgia rheumatica (PMR)* All patients Clinical group
Men
Women
Total
GCA alone GCA + PMR PMR alone Total
10 (29) 5 (14) 20 (57) 35 (100)
13 (20) 15 (24) 36 (56) 64 (100)
23 (23) 20 (20) 56 (57) 99 (100)
Patients with histologic evidence of arteritis 11 (48) 9 (45)
* Values are the number (%).
PATIENTS AND METHODS Patients. Records of all patients who were diagnosed as having PMR or GCA at the Departments of Medicine, Ophthalmology, Neurology, or Rheumatology at Reggio Emilia Hospital between 1980 and 1988 were reviewed. We investigated only patients who resided in the Reggio Emilia metropolitan area (census tract number USL n.9) at the time of diagnosis. This area has a current population of -170,000. Annual information on the number of USL n.9 inhabitants and the age and sex distribution of the population was provided by the Statistics Department of Reggio Emilia province. During the study period (1980-1988), there were no significant variations in the total population or in the age and sex distribution. We believe all patients in whom a diagnosis of GCA was confirmed by biopsy were identified, since we examined the records of all patients in whom a temporal artery biopsy was performed at Reggio Emilia Hospital. Patients with PMR who had been treated at the Departments of Medicine and Rheumatology were identified from a review of the records of all patients with this diagnosis. Virtually all specialized medical care for the Reggio Emilia metropolitan area is provided at the Reggio Emilia Hospital. There are no private hospitals or rheumatologists in private practice in this area. The diagnosis of definite GCA was based on positive findings on temporal artery biopsy. However, we also considered patients with negative biopsy results, and a diagnosis of probable GCA was made if they had at least 4 of the following classic manifestations of GCA: abnormal temporal arteries (tender and swollen), loss of vision, jaw claudication, presence of PMR, and response to steroid therapy (7). The diagnosis of PMR was made if all the following were present: 1) persistent pain (for at least 1 month) involving 2 of the following areas: neck, shoulders, and pelvic girdle; 2) morning stiffness lasting more than 1 hour; 3) rapid response to prednisone (520 mgiday); and 4) absence of other diseases capable of causing the musculoskeletal symptoms (21). For both diagnoses, only patients over 50 years old and with an erythrocyte sedimentation rate >40 mm/hour were included. Ninety-nine patients with GCA or PMR were identified. The distribution of patients in each clinical group (GCA alone, PMR with GCA, and PMR alone) is shown in Table 1.
The diagnosis had been made by a family practitioner prior to the patient’s hospitalization in only 2 cases. Temporal artery biopsy was performed in 69 patients (a bilateral biopsy was performed in only 2 cases). Prior to 1986, all patients with signs of symptoms of either GCA or PMR had a temporal artery biopsy performed (56 patients). From 1986 on, biopsy was performed only in patients with signs or symptoms of arteritis (13 patients). The mean length of the temporal artery biopsy specimen was 0.5 cm, and multiple sections of each specimen were examined without prior knowledge of the clinical diagnosis. HLA typing. HLA typing was performed on 49 randomly selected patients with PMR and/or GCA. The healthy control group consisted of a pool of 242 unrelated blood donor volunteers from the same geographic area. Peripheral blood leukocytes were isolated by FicollHypaque density-gradient centrifugation (22). B lymphocytes were separated from T lymphocytes by an ox red blood cell rosetting technique, using monoclonal antibody H4 (One Lambda, Los Angeles, CA). The determination of class I (HLA-A, B , and C) and class I1 (HLA-DR and DQ) major histocompatibility complex antigens was performed using a modification of the NIH 2-stage complement microcytotoxicity test (23). Lymphocytes were typed using a panel of 180 antisera recognizing approximately 80 different specificities, from the Eighth and Ninth International Histocompatibility Workshops. Statistical analysis. Statistical analysis was performed using the Statistical Package for the Social Sciences (24). Chi-square with Yates’ correction and Fisher’s exact test were used to compare the frequencies of HLA antigens. P values were corrected for the number of antigens tested. Relative risks (RR) were calculated using the method of Woolf.
RESULTS Incidence of PMR, 1980-1988. A total of 76 patients with PMR (25 men and 51 women) were identified. The mean -t SD age at diagnosis was 69.7 -t 9 years (67.9 ? 10 for men, 70.6 +- 8 for women), with a range of 48-87 years. Age- and sex-specific incidence Table 2. Annual incidence of polymyalgia rheumatica in residents of Reggio Emilia, Italy, 1980-1988, by age and sex, per 100,000 population*
Men
Women
Total
Age group (years)
No.
Rate
No.
Rate
No.
Rate
0-49 50-59 60-69 70-79 280 Total 250
0 6 6 10 3 2s 25
0 6.3 7.2 17.8 17.7 3.4 9.7
1 6 15 25 4 51 50
0.2 5.7 14.0 29.9 10.1 6.4 14.9
1 12 21 35 7 76 75
0.1 6.0 11.0 25.0 12.0 4.9 12.7
* The total
population was 169,950 and the number of individuals 250 years old was 65,711.
PMR AND GCA IN NORTHERN ITALY
353
Table 3. Annual incidence of giant cell arteritis in residents of Reggio Emilia, Italy, 1980-1988, by age and sex, per 100,000 population*
Table 4. Frequency (9%) of HLA class I1 antigens in patients with polymyalgia rheumatica (PMR) with giant cell arteritis (GCA), in patients with PMR without GCA, and in controls
Women
Men
0-49 50-59 60-69 70-79 280 Total
250
Total
No.
Rate
No.
Rate
No.
Rate
Antigen
All PMR (n = 41)
0 2 6 6 1 15 15
0 2.1 7.2 10.7 5.9 2.0 5.8
2 5 8 10 3 28 26
0.4 4.8 7.5 11.8 7.6 3.5 7.8
2 7 14 16 4 43 41
0.2 3.5 7.3 11.4 7.1 2.8 6.9
DR I DR2 DR4 DRw53 DQ 1 DQ2 DQ3
22 19 24 46 46 27 51
PMR with GCA (n = 6)
PMR without GCA (n = 35)
17 17 33 83 34 67 50
23 20 23 40
48 20 51
Controls (n = 242) 16 22 14 43 51 40 47
* The total
population was 169,950 and the number of individuals 2 5 0 years old was 65,71 1.
rates of PMR are shown in Table 2. There was an increased incidence of PMR with increasing age, with a maximum in the eighth decade of life (17.8/100,000 for men and 29.9/100,000 for women). The incidence rate in the population 2.50 years old was approximately 1.5 times higher in women than in men. Incidence of GCA, 1980-1988. Forty-three patients (28 women and 15 men) met the diagnostic criteria for GCA. Twenty (46.5%) had a positive temporal artery biopsy result and 23 (53.3%) had a negative biopsy result but met the clinical criteria. The mean ? SD age at diagnosis was 67.6 ? 10 years (68.9 10 for men, 66.9 ? 11 for women), with a range of 46-85 years. The age-specific incidence rate in women increased from 4.8/100,000in the sixth decade of life to 11.8/100,000in the eighth decade (Table 3 ) . There was a corresponding increase in the incidence rates in men, with a peak level of 10.7/100,000in the eighth decade. There was a slightly greater incidence of GCA in women than in men in every age group. Temporal artery biopsy rate, 1980-1988. The mean rate of performance of temporal artery biopsy during the study period was 16.3/100,000 250 years old and 25.5/100,000 women 250 years old. The biopsy rate increased with age, with the maximum rate occurring during the eighth decade of life (37.3/100,000 men and 35.8/100,000 women). However, the rate decreased in the oldest patients (280 years old), with an incidence of 17.7/100,000 men and 22.7/100,000 women. HLA typing results. The distribution of HLA class I antigens in the patients did not differ significantly from that in the controls (data not shown). Tables 4 and 5 show the distribution of HLA-DR and HLA-DQ phenotypes in the 49 patients compared with the control group. The frequencies of class I1
*
antigens did not differ significantly in the patient subgroups compared with the controls. An increase of DR4 was found in all patient subgroups, particularly in the GCA patients (36% versus 14% in controls; RR = 3.3,95% confidence interval 1.0-10.5); however, none of these differences was statistically significant. The frequency of DRw53, which is in linkage disequilibrium with DR4, was not significantly increased in GCA patients compared with controls (78% versus 43%; RR = 4.9, 95% confidence interval 1.4-17.9).
DISCUSSION Our epidemiologic results show a lower incidence of PMR in a Mediterranean population than in the northern European and the Minnesota populations (3-8) (Table 6). An annual incidence of PMR of 12.7/ 100,000 population aged 2 5 0 years was observed in Reggio Emilia, Italy, compared with 53.7/100,000 in Olmstead County, Minnesota (6) and with 20.4/100,000 (recalculated to exclude patients with GCA alone) in Goteborg, Sweden (3,4). The annual rate of incidence of GCA found in the present study (Table 7) is similar to that reported in Table 5. Frequencies (%) of HLA class I1 antigens in patients with giant cell arteritis (GCA) with polymyalgia rheumatica (PMR), in patients with GCA without PMR, and in controls
Antigen
DR 1 DR2 DR4 DRw53 DQ 1 DQ2 DQ3
All GCA (n = 14)
GCA with PMR (n = 6)
GCA without PMR (n = 8)
7 14 36 78 21 50 64
17 17 33 83 34 67 50
0 12
37 75 12 37 75
Controls (n = 242) 16 22 14 43 51 40 47
SALVARANI ET AL
354
Table 6. Comparison of the incidence rates of polymyalgia rheumatica in various epidemiologic studies* Incidence/ I00,000/year Location, years Olmstead County, Minnesota, 1970-1979 Goteborg, Sweden, 1973-1975 Ribe County, Denmark, 1982-1985 Reggio Emilia, Italy, 1980-1988
All age groups
250
11.1
53.7
Age
6.7t
20.41
NR
68.3
4.9
12.7
* Refs. 3-8 and present study. NR = not reported. t Recalculated to exclude patients with giant cell arteritis alone
a French population in Department Loire-Atlantique (12), but is lower than that reported in the Olmsted County population and the Goteborg population (3,8). The lowest incidence of GCA has been observed in Israel (9) and in Shelby County, Tennessee (1 1) (especially among Sephardic Jews and blacks, respectively). Differences in the rates of performing of temporal artery biopsy may have resulted in a reduced propensity of physicians in Reggio Emilia to pursue a diagnosis of GCA, compared with physicians in Olmsted County. (The rates during 1980-1988 in our study were approximately one-third those recorded during 1980-1985 in Olmsted County during 1980-85 [8].) Conversely, the lower rate of performing temporal artery biopsy in the present study may be related to an actual lower incidence of GCA in the population of Reggio Emilia. Consistent with other studies (3,5,1I), we observed a reduction in the age-specific incidence of GCA in patients aged 80 years or older, whereas in Olmsted County, the rate continues to increase in the oldest patients (8). A declining incidence reported in this age group may be related to a reduced suspicion of the disease in the oldest patients, resulting in underdiagnosis. The relatively low rate of temporal artery biopsy in male and female patients aged 80 years or older supports this hypothesis. Our study included cases of “probable” GCA (negative biopsy result), based on criteria proposed by Huston et a1 (7). Of our 43 patients with clinical evidence of GCA, only 20 (46.5%) had a positive temporal artery biopsy result, in contrast with the Olmsted County study (8), in which the diagnosis of GCA was confirmed by biopsy in 94% of the patients
studied. These differing results could be explained by the differences in the mean length of the biopsy specimen (0.5 cm in our study versus 5.4 cm in the Olmsted County study) and by the practices in Olmsted County of performing a bilateral biopsy when findings in the first specimen are negative (25) (a second biopsy was performed in only 2 of our patients). However, none of our patients’ diagnoses were modified during the followup period (mean duration 64 months), making possible errors in the diagnosis unlikely. We cannot exclude the possibility that patients with a mild form of PMR or GCA were treated as outpatients, therefore escaping our attention. It is also possible that elderly patients with PMR may be misdiagnosed as having RA and treated only for RA by their family physicians (only 2 patients in our study had a correct diagnosis of PMR prior to admission). It is difficult to evaluate the impact of the low level of awareness of these diseases among general practitioners, but it is unlikely that this factor can fully account for the low incidence of PMR and GCA found in Reggio Emilia. Table 7. Comparison of the incidence rates of giant cell arteritis in various epidemiologic studies* Incidence/100,000/year, in individuals 2 5 0 years old
Location, years
Biopsyproven only
Including probable casest
NR NR
17.0 24.1
16.8
18.3$
NR
23.3
1.78 0.36
2.24 0.36
NR
9.4
NR NR
0.49 0.86
NR
6.9
~
Olmsted County, Minnesota, 1950-1 985 1980-1985 Goteborg, Sweden, 1973-1 975 Ribe County, Denmark, 1982-1985 Shelby County, Tennessee, 1971-1980 Whites Blacks Department Loire-Atlantique, France, 1970-1979 Israel 1960-1978 1975-1978 Reggio Emilia, Italy, 1980-1988
* Refs. 3-9, 11, 12, and present study. NR = not reported. t Probable clinical diagnosis in patients with a negative
biopsy result, or in patients in whom a biopsy was not performed. $ Recalculated to exclude patients with polymyalgia rheumatica only.
PMR AND GCA IN NORTHERN ITALY
355
It has been suggested that genetic factors may play a pathogenic role in both PMR and GCA. Richardson et a1 (17), in a metaanalysis of previous studies of HLA antigens in GCA, showed an association of HLA-DR4 with PMR, rather than with GCA alone. Cid et a1 (18) confirmed that, in Spanish patients with biops y-proven GCA, DR4 is mainly associated with PMR. The incidence of GCA and/or PMR has been shown to be low in blacks (26), paralleling the low frequency of HLA-DR4 observed in this population (27). Interestingly, there is a corresponding low incidence of PMR and GCA coupled with a low incidence of DR4 in the general population (range 12-14% [19,20]) in northern Italy. In the present study, in which we examined a group of patients with PMR and a second group with GCA, the frequency of DR4 was not significantly increased in either group compared with controls, and the highest incidence of this antigen was observed in the group of patients with GCA, a finding that is inconsistent with the results of other studies. In addition, the increased frequency of DRw53, which is in linkage disequilibrium with DR4, was not statistically significant in either PMR or GCA patients. Ferraccioli and Savi (19) reported a somewhat higher relative risk for DR1 (RR = 2.1) than for DR4 (RR = 1.3) in RA patients in northern Italy. In our PMR patients, a similarly weak and nonsignificant association with DR1 and DR4 was observed (RR = 1.4 and RR = 1.6, respectively). Conversely, we observed an increased frequency of DR4 and a reduction of DR1 (RR = 3.3 and RR = 0.7, respectively) among patients with GCA. However, this result is difficult to interpret due to the small number of GCA patients. Our results confirm a trend toward a decrease in the frequency of DR2 in PMR patients. In contrast to the results of Cid et a1 (18), we observed a reduction of the DQw2 antigen only in PMR patients without GCA, but this difference was again not statistically significant. Our study shows no significant increase in any of the HLA class I antigens, similar to the findings in other studies (17). The significant increase of HLA-DR4 found among PMR patients in other study populations ( 1 6 18) was thus not confirmed in our northern Italian population. A weak or nonexistent association with DR4 has also been reported in Italian patients with RA, as well as in RA patients of other Mediterranean countries (19,28-30). The evidence of a weak or absent association of DR4 with RA or PMR gives rise to the question whether disease susceptibility is linked to the
same epitope or combination of epitopes in all populations, or whether genetic heterogeneity is involved. Further studies using oligonucleotide probes are needed to define which epitopes are most closely related to PMR.
ACKNOWLEDGMENTS The authors thank Drs. Gene G. Hunder and David J. Ballard for their clinical review of the manuscript, Rita Farella for clinical assistance in the course of the study, and Jacqueline Costa for preparation of the manuscript.
REFERENCES 1. Granato JE, Abben RP, May WS: Familial association
of giant cell arteritis: a case report and brief review. Arch Intern Med 141:115-1 17, 1981 2. Kemp A: Monozygotic twins with temporal arteritis and ophthalmic arteritis. Acta Ophthalmol (Copenh) 55: 183190, 1977 3. Bengtsson B-A, Malmvall B-E: The epidemiology of giant cell arteritis including temporal arteritis and polymyalgia rheumatica. Arthritis Rheum 24:89%904, 1981 4. Bengtsson B-A, Malmvall B-E: Giant cell arteritis: incidence of giant cell arteritis. Acta Med Scand [Suppl] 658: 15-17, 1982 5 . Boesen P, Sgrensen SF: Giant cell arteritis, temporal arteritis, and polymyalgia rheumatica in a Danish County: a prospective investigation, 1982-1985. Arthritis Rheum 30:294-299, 1987 6. Chuang TY, Hunder GG, Ilstrup DM, Kurland LT: Polymyalgia rheumatica: a 10-year epidemiologic and clinical study. Ann Intern Med 97:672-680, 1982 7. Huston KA, Hunder GG, Lie JT, Kennedy RH, Elveback LR: Temporal arteritis: a 25-year epidemiologic, clinical and pathologic study. Ann Intern Med 88:162167, 1978 8. Machado EBV, Michet CJ, Ballard DJ, Hunder GG, Beard CM, Chu CP, O’Fallon M: Trends in incidence and clinical presentation of temporal arteritis in Olmsted County, Minnesota, 1950-1985. Arthritis Rheum 31:745749, 1988 9. Friedman G, Friedman B, Benbassat J: Epidemiology of temporal arteritis in Israel. Isr J Med Sci 18:241-244, 1982 10. Salvarani C, Macchioni PL, Tartoni PL, Rossi F, Baricchi R, Castri C, Chiaravalloti F, Portioli I: Polymyalgia rheumatica and giant cell arteritis: a 5-year epidemiologic and clinical study in Reggio Emilia, Italy. Clin Exp Rheumatol 5:205-215, 1987 11. Smith CA, Fidler WJ, Pinals RS: The epidemiology of giant cell arteritis: report of a ten-year study in Shelby County, Tennessee. Arthritis Rheum 26: 1214-1219, 1983 12. Barrier J, Pion P, Massari R, Peltier P, Rojouan J ,
SALVARANI ET AL
356
13.
14.
15. 16.
17.
18.
19.
20.
Grolleau JY: Approche epidemiologique de la maladie de Horton dans le department de Loire-Atlantique: 110 cas en 10 ans (1970-1979). Rev Med Interne 3:13-20, 1983 Elling H , Skinhoj P, Elling P: Hepatitis B virus and polymyalgia rheumatica: a search for HBsAg, HBsAb, HBcAb, HBeAg and HBeAb. Ann Rheum Dis 3951 1513, 1980 O’Brien JP: A concept of diffuse actinic arteritis: the role of actinic damage to elastin in ‘age change’ and arteritis of the temporal artery and in polymyalgia rheumatica. Br J Dermatol 98:l-13, 1978 Fessel WJ: Polymyalgia rheumatica, temporal arteritis and contact with birds. Lancet 11:124%1250, 1969 Armstrong RD, Panayi GS, Welsh KI: Polymyalgia rheumatica and rheumatoid arthritis: similarity of HLA antigen frequencies (letter). Arthritis Rheum 27: 14381439, 1984 Richardson JE, Gladman DD, Fam A, Keystone EC: HLA-DR4 in giant cell arteritis: association with polymyalgia rheumatica syndrome. Arthritis Rheum 30: 1293-1297, 1987 Cid M-C, Ercilla G, Vilaseca J, Sanmarti R, Villalta J, Ingelmo M, Urbano-Marquez A: Polymyalgia rheumatica: a syndrome associated with HLA-DR4 antigen. Arthritis Rheum 31 :678-682, 1988 Ferraccioli GF, Savi M: Association between DR antigens, rheumatoid arthritis with and without extraarticular features and SLE in northern Italy. J Rheumatol 15:51-53, 1988 Salvarani C, Macchioni PL, Zizzi F, Mantovani W, Rossi F, Baricchi R, Ghirelli L, Frizzier0 L, Portioli I:
21. 22.
23. 24. 25. 26.
27. 28.
29.
30.
Clinical subgroups and HLA antigens in Italian patients with psoriatic arthritis. Clin Exp Rheumatol 7:391-396, 1989 Healey LA: Long-term follow-up of polymyalgia rheumatica: evidence of synovitis. Semin Arthritis Rheum I3 :322-328, 1984 Bdyum A: Separation of leucocytes from blood and bone marrow. Scand J Clin Lab Invest [Suppl] 97:21-26, 1968 Terasaki PI, McClelland JD: Microdroplet assay of human serum cytotoxins. Nature 204:998-1000, 1964 SPSS” User’s Guide. Cary, NC, SAS Institute, 1986 Hall S, Lie JT, Kurland LT, Persellin S, O’Brien PC, Hunder GG: The therapeutic impact of temporal artery biopsy. Lancet II:1217-1222, 1983 Love DC, Rapkin J, Lesser GR, Shmookler BM, Kolsky MP, Jackson B, Barth WF: Temporal arteritis in blacks. Ann Intern Med 105:387-389, 1986 Alarcon GS, Koopman WJ, Acton RT, Barger BO: DR antigen distribution in blacks with rheumatoid arthritis. J Rheumatol 10579-583, 1983 Papasteriades CA, Kappou ID, Skopouli FN, Barla MN, Fostiropulos GA, Moutsopoulos HM: Lack of HLAantigen association in Greek rheumatoid arthritis patients. Rheumatol Int 5:201-203, 1985 Mercier P, Dallest R, Roux H: HLA-DR et polyarthrite rhumatoide de I’adulte: etude dans le Sud de la France. Rev Rhum Ma1 Osteoartic 47:465-470, 1980 Brautbar C, Naparstek Y, Yaron M, Amar A, Ehrenfeld M, Eliakim M, Bentwich Z, Cohen T, Korman SH, Albert ED: Immunogenetics of rheumatoid arthritis in Israel. Tissue Antigens 28:8-14, 1986
