Rheumatol Int (2000) 19:213±217

Ó Springer-Verlag 2000

ORIGINAL ARTICLE

M. A. Cimmino á C. Salvarani á P. Macchioni C. Montecucco á V. Fossaluzza á M. T. Mascia L. Punzi á C. Davoli á D. Filippini á R. Numo

Extra-articular manifestations in 587 Italian patients with rheumatoid arthritis Received: 21 February 2000 / Accepted: 30 March 2000

Abstract The aim of the study was to evaluate the frequency of extra-articular manifestations (EAMs) of rheumatoid arthritis (RA) in a series of patients from nine Italian rheumatology clinics. A total of 587 patients underwent direct questioning, complete physical evaluation, and review of medical records and laboratory data. The relationships between EAMs and the eosinophilic count, IgM rheumatoid factor (RF), and antinuclear antibodies (ANA) were studied. EAMs were present in 240/587 (40.9%) patients. The most common features were sicca syndrome (17.5%) and rheumatoid nodules (16.7%). EAMs were signi®cantly more frequent in male patients (OR ˆ 1.68), patients with ANA positivity (OR ˆ 2.82), high anatomical class M. A. Cimmino (&) Clinica Reumatologica, Dipartimento di Medicina Interna e SpecialitaÁ Mediche, UniversitaÁ degli Studi di Genova, Viale Benedetto XV 6, 16132 Genova, Italy e-mail: [email protected] Tel.: +39-010-3538905; Fax: +39-010-3538638 C. Salvarani á P. Macchioni Divisione di Reumatologia, Arcispedale S. Maria Nuova, Reggio Emilia, Italy C. Montecucco Servizio di Reumatologia, IRCCS S. Matteo, Pavia, Italy V. Fossaluzza  (deceased) Divisione di Medicina, Ospedale S. Maria della Misericordia, Udine, Italy M. T. Mascia Cattedra di Reumatologia, UniversitaÁ di Modena, Modena, Italy L. Punzi Cattedra di Reumatologia, UniversitaÁ di Padova, Padova, Italy C. Davoli Divisione di Medicina, Azienda Istituti Ospitalieri di Cremona, Cremona, Italy D. Filippini Divisione di Medicina, Ospedale Niguarda, Milan, Italy R. Numo Centro di Medicina Sociale di Reumatologia, Bari, Italy

(OR ˆ 2.3), and rheumatoid factor seropositivity (OR ˆ 2.22). EAMs were more common in patients from southern Italy than in those from northern Italy (P < 0.001). EAMs seem to be rarer in Italy than in the Anglo-Saxon populations of northern Europe and the USA. Di€erences in prevalence of EAMs can exist even within the same country. Key words Rheumatoid arthritis á Extra-articular manifestations á Epidemiology á SjoÈgren's syndrome á Rheumatoid nodules á Italy

Introduction Extra-articular manifestations (EAMs) of rheumatoid arthritis (RA) correlate with articular severity and with functional impairment [1]. They are associated with high concentrations of IgM rheumatoid factor (RF) and constitute one of the risk factors for early death in patients with RA [2]. Reports from the Mediterranean basin suggest that RA in this region has a milder course than RA in the Caucasian population of the USA and the UK [3, 4, 5, 6, 7]. Drosos et al. compared RA patients from Greece and the UK and found that articular involvement was more severe and EAMs more frequent in the latter group [6]. Greek patients, however, had an increased prevalence of xerostomia and/or xerophthalmia in comparison with the British ones. Salvarani et al. [7] performed a retrospective chart review on 141 Italian patients with RA and reported low frequency of rheumatoid nodules (16%) and vasculitis (2.1%). Possible explanations for these ethnic di€erences include genetical diversity and environmental factors. RA has a weak or insigni®cant association with human leukocyte antigen (HLA)-DR4 in southern Europeans [3, 4, 5, 7, 8], Africans [9], and African Americans [10]. RA in the southern European and African populations is characterized by a low prevalence and mild course. In particular, a recent Italian survey has shown a

214

prevalence of 0.33% [11]. Environmental factors, such as di€erences in climate, food, and prevailing infections, have been hypothesized for these ethnic di€erences [12, 13, 14]. This work is concerned with the frequency and severity of EAMs in di€erent geographical areas of Italy. EAMs were assessed by clinical examination, review of the patients' chart and laboratory data, and administration of a questionnaire in nine Italian clinics. The investigation was undertaken with a view to also gaining some insights into genetical and environmental factors that may be at work in RA.

Patients and methods A total of 587 consecutive, unselected patients with RA from nine Italian rheumatology clinics were studied. Each center contributed all patients seen in a ®xed 3-month period from 1 April to 30 June 1991. The number of patients enrolled by a single center ranged from a minimum of 50 to a maximum of 120. Eight centers were located in northern Italy (Milan, Pavia, Genoa, Cremona, Reggio Emilia, Modena, Padua, and Udine) and one in southern Italy (Bari). Four centers were university based and ®ve were hospitalbased tertiary-care referral centers. Centers were chosen because of similar referral patterns of the patients. All patients ful®lled the ACR revised criteria for RA [15]. Patients were considered seropositive if they had a Rose-Waaler test score greater than or equal to 1:64 or a latex test score greater than 1:40 on at least two occasions in the last 2 years. RF was considered highly positive if its concentration exceeded an arbitrary value of 1:640 on latex test. All seronegative patients had normal sacroiliac joints on radiographic examination and a negative family history of psoriasis and spondylitis. There were 451/587 (76.8%) women and 136/587 (23.2%) men in our series. The mean age was 58.9 ‹ 12.7 years (range 17±84 years) at the time of the study and 49.2 ‹ 14.4 years at the onset of RA. The mean duration of RA was 115.6 ‹ 102.9 months. To determine whether the patients were of northern or southern stock, the place of birth of their parents was considered. Of our patients, 484 (82.5%) were of northern stock, 84 (14.3%) of southern stock, and the remaining 19 (3.2%) originated from central Italy. To compare the frequency of EAMs among these subgroups, we matched 76 patients born in southern Italy with 76 patients born in northern Italy for sex, age (mean age 54.5 ‹ 10.9 vs 54 ‹ 10.8 years), and disease duration (109.2 ‹ 90.5 vs 108.5 ‹ 88.6 months). EAMs were identi®ed by direct questioning, complete physical evaluation including articular examination, and review of medical records. Disease activity was assessed by counting the number of swollen and tender joints; disease severity was assessed by evaluating the anatomical class [16] on a radiogram taken less than 6 months before clinical examination. Steinbrocker's functional class was also recorded [16]. The patients were considered to have EAMs when at least one of the following clinical features was reported or directly observed by the clinician during the disease course: 1. Subcutaneous nodules 2. Pulmonary disease including pleuritis, interstitial pneumonitis, and ®brosis con®rmed by a chest radiogram and by lung function tests 3. Echocardiographic evidence of pericardial e€usion 4. Secondary SjoÈgren's syndrome (sSS), diagnosed on the clinical evidence of dry eyes and dry mouth and con®rmed by (a) the objective demonstration of xerophthalmia (positive Schirmer's test), or (b) keratoconjunctivitis sicca (positive slit lamp examination after ¯uorescent staining), or on (c) the presence of salivary gland involvement evaluated by lip biopsy and/or salivary scintigraphy

5. Felty's syndrome, de®ned as seropositive RA with splenomegaly and less than 2.0 ´ 109/l granulocytes 6. Cutaneous vasculitis, including distal arteritis, cutaneous ulcerations, or palpable purpura 7. Non-compressive peripheral neuropathy con®rmed by neurophysiologic studies 8. Amyloid deposits con®rmed by biopsy Neutrophil, eosinophil, and platelet counts were derived from the means of three recent determinations. ANA were determined by indirect immuno¯uorescence using rat liver as substrate; the test was considered positive at titers greater than or equal to 1:80. Statistical analysis was carried out using the Statistical Package for Social Sciences software (SPSS Inc., Chicago, Ill., USA). Student's t-test and v2 test with Yates' correction were used. All the tests were two-tailed.

Results The distribution of patients according to anatomical and functional classes is reported in Fig. 1. The mean number of joints a€ected was 6.3 ‹ 5.3 (range 0±31). A family history of RA was reported by 70 (11.9%) patients. Of the patients, 92.9% with EAMs and 93.1% without EAMs were on disease-modifying drugs. Similar ®gures were found when the individual EAMs were considered. Of our patients, 240 (40.9%) showed at least one EAM. The types of EAMs are reported in Table 1. The most common features were sicca syndrome (17.5%) and rheumatoid nodules (16.7%); the least common were Felty's syndrome (0.7%) and amyloidosis (0.7%). Clinical and demographic data of patients with EAMs and patients without EAMs are compared in Table 2. Patients with EAMs had a higher relative male prevalence and a longer disease duration. RF and ANA positivity was signi®cantly more frequent in patients with EAMs than in those without EAMs. Likewise, in the former group, the functional and anatomical classes were worse, with a higher number of joints involved by RA, and higher counts of peripheral eosinophils. RF was present in 403 (68.7%) patients; its concentration was high in 69 (11.8%) patients. ANA were

Fig. 1 Distribution of the 587 patients with rheumatoid arthritis according to their anatomical (white column) or functional (gray column) classes. Numbers in the columns indicate the percent values

215

positive in 134 (22.8%) patients. The pattern was homogeneous in 61 patients, speckled in 38, nucleolar in 20, peripheral in 9 patients, and indistinct in the remaining 10. The risk of developing EAMs was higher in men than in women, with an odds ratio (OR) of 1.68 [95% con®dence interval (CI) 1.14±2.48]. Other relevant risk factors for EAMs were RF seropositivity (OR ˆ 2.22, 95% CI 1.52±2.23), ANA positivity (OR ˆ 2.82, 95% CI 1.90±4.20), and anatomical class III±IV vs I±II (OR ˆ 2.30, 95% CI 1.64±3.22). In view of the higher proportion of males and longer disease duration in patients with EAMs, these variables were adjusted to determine whether the other risk factors were independently di€erent. Only anatomical class was in¯uenced by sex and disease duration. The mean interval between onset of RA and appearance of each type of EAM is reported in Table 3. Pulmonary disease, in particular pleuritis, was an early manifestation. Pleuritis occurred after a mean interval of 28 months after the onset of RA. In contrast, cutaneous vasculitis and amyloidosis were late manifestations appearing 122 and 148 months, respectively, after the onset of the disease. The demographic and clinical characteristics between patients from northern and southern Italy were compared (Table 4). The main di€erence was an earlier onset of RA in southern Italy. In this area, a family history of RA was less frequent. When patients from southern

Italy were compared with a subgroup of patients from northern Italy matched for sex, age, and disease duration, there was a striking preponderance of EAMs in the patients from southern Italy (Table 5). However, amongst individual EAMs, only splenomegaly was signi®cantly more frequent in southern patients. The risk of presenting EAMs was increased (OR ˆ 3.9, 95% CI 1.97±7.63) in patients from southern stock in comparison with those from northern stock.

Discussion Classical data on the frequency of EAMs in RA patients were collected in the Caucasian population of North America and Britain [1, 17, 18]. More recent studies have been performed in di€erent geographical areas and have shown that patients from Greece [6], Nigeria [19], Malaysia [20], China [21], and Oman [22] present signi®cantly fewer EAMs than those from Britain. The only exception is sSS, which appears to be more prevalent in Greeks and Malaysians than in British patients. We found that EAMs in Italian RA patients show an intermediate prevalence between the British and Greek populations. The overall frequency of EAMs in Italy was 40.9%, a value lower than that of 76% in Canada [17] and of 66.4% in the UK [6]. In particular, rheumatoid nodules

Table 1 Frequency of extra-articular manifestations in 587 Italian patients with rheumatoid arthritis

Table 3 Mean time interval between onset of rheumatoid arthritis and appearance of each extra-articular manifestation

Manifestation

Patients (%)

n

n

Sicca syndrome Rheumatoid nodules Pulmonary disease Raynaud's phenomenon Cutaneous vasculitis Peripheral neuropathy Pericarditis Lymphadenopathy Splenomegaly Felty's syndrome Amyloidosis

17.5 16.7 6.3 6.3 3.7 2.2 1.8 1.7 1.5 0.7 0.7

103 98 37 37 22 13 11 10 9 4 4

Extra-articular manifestation

Mean interval (months)

Range (months)

Pulmonary disease Pleuritis Interstitial ®brosis Peripheral neuropathy Pericarditis Raynaud's phenomenon Sicca syndrome Rheumatoid nodules Cutaneous vasculitis Amyloidosis

37 15 25 10 4 31 99 89 20 4

44.7 27.7 49.2 58.6 67.5 79.5 83.1 104.9 121.9 147.7

0±168 1±20 0±168 4±240 1±204 1±378 1±444 1±312 4±340 27±336

Table 2 Comparison of the demographic and clinical data of rheumatoid arthritis patients with and without extra-articular manifestations (EAMs)

Number of patients Mean age ‹ SD (years) Age at onset ‹ SD (years) Mean duration of RA ‹ SD (months) Females/males (%) Seropositive/negative (%) ANA positive/negative (%) Eosinophils count ‹ SD (number 109/l) Anatomical class I, II/III, IV (%) Functional class I, II/III, IV (%) Mean number of involved joints ‹ SD Family history of RA (positive/negative %) Disease modifying treatment (yes/no %) *P < 0.05; **P < 0.01; ***P < 0.001

With EAMs

Without EAMs

240 59.7 ‹ 11.3 48.3 ‹ 13.7 134.9 ‹ 102 71/29 78.3/21.7 34.2/65.8 0.16 ‹ 0.27 40.8/59.2 57.1/42.9 6.8+5 14/86 92.9/7.1

347 58.3 ‹ 13.5 49.9+14.9 102.3 ‹ 101.5*** 81/19** 62/38*** 15.5/84.5*** 0.12 ‹ 19* 61.4/38.6*** 75.8/24.2*** 5.9 ‹ 5.5* 11/89 93.1/6.9

216 Table 4 Comparison of the demographic and clinical features in Italian patient from northern or southern stock

Mean age ‹ SD (years) Mean age at onset of RA ‹ SD (years) Mean duration of RA ‹ SD (months) Women/men (%) Family history of RA (positive/negative %) RF positive/negative (%) Mean number of involved joints ‹ SD Anatomical class I, II/III, IV (%) Functional class I, II/III, IV (%)

Northern Italy (n = 484)

Southern Italy (n = 84)

60 ‹ 12.4 50.4 ‹ 14.1 115.4 ‹ 103.3 77.3/22.7 13.2/86.8 67.6/32.4 6.2 ‹ 5.4 52.5/47.5 69.2/30.8

52.7 ‹ 12.6** 42.9 ‹ 14.4** 116.3 ‹ 98.3 73.8/26.2 3.6/96.4* 78.6/21.4 6.7 ‹ 4.5 58.4/41.6 60.7/39.3

*P < 0.05; **P < 0.001

Table 5 Frequency of extra-articular manifestations (EAMs) in Italian patients with rheumatoid arthritis from northern or southern stock (%) Extra-articular manifestation

Northern Italy (n = 76)

Southern Italy (n = 76)

Sicca syndrome Rheumatoid nodules Pulmonary disease Raynaud's phenomenon Cutaneous vasculitis Peripheral neuropathy Pericarditis Lymphadenopathy Splenomegaly Felty's syndrome Amyloidosis All EAMs

13 9 5 3 1 1 1 3 0 0 0 23

22 14 6 8 5 1 3 3 4 0 1 48

(17.3) (12) (6.7) (4) (1.3) (1.3) (1.3) (4) (0) (0) (0) (30.7)

(28.9) (18.4) (7.9) (10.5) (6.6) (1.3) (4) (4) (5.3)* (0) (1.3) (63.2)**

*P < 0.05; **P < 0.001

were present in 16.7% of patients in Italy, 53% in Canada [17], and 51.4% in the UK [6]. The prevalence of rheumatoid nodules was as low as 1% in Nigeria [19] and 5.6% in Greece [6]. All the EAMs, except sSS, were rarer in Italy than in Britain or Canada [6, 17]. These data con®rm the results of Salvarani et al. [7], who reviewed a small group of Italian RA patients. Salvarani et al. and most of the above-mentioned authors recruited not only outpatients but also inpatients. This fact introduces a selection bias because an abnormally high number of severe RA cases are recruited. Since severely invalidating RA is relatively rare, we feel that an unselected population of outpatients should re¯ect more closely the appearance of the disease in the general population. Comparisons between groups of RA patients derived from di€erent clinics are a€ected by the statistical bias common to multicenter studies; the most important is that patients are examined by di€erent physicians. Standardization of the procedures was achieved by common discussion of the clinical protocols and of laboratory work-up. In addition, the imbalance in the patients' demographic characteristics required a matching for age, sex, and disease duration. It is important to note that the frequency of EAMs in the complete group of patients from northern origin was similar to that

observed in the matched subgroup. We believe that despite these possible biases our results re¯ect real di€erences in EAMs frequency in our country. Secondary SS di€ered from the other EAMs because it did not show male predominance, nor an association with RF or with the indexes of disease severity. Secondary SS was strongly associated with ANA positivity (data not shown). These characteristics suggest that SS associated with RA is similar to the primary form of SS. Populations from the Mediterranean basin, such as Greeks and Italians, have a high incidence of SS [6, 7]. In this work, we reported that Italian RA patients born in the southern part of the country had a higher frequency of sSS than those of northern stock. The prevalence of sSS was 17.5%; a value higher than that of 11% observed in another Italian survey [7]. This di€erence may be due to the fact that only patients from northern Italy were considered in that survey. A possible explanation for di€erences between northern and southern Italians is that the coastal area of southern Italy was colonized by the Greeks between the eighth and ®fth centuries B.C. As a result, the populations from southern Italy and Greece share several common genetical traits that are associated with various diseases besides RA. An additional explanation is the higher prevalence of HLA-DR4 in the general population from southern Italy than that from northern Italy [23]. Although it is not clear whether RA is associated with HLA-DR4 also in southern Europe, a recent study [9] indicates that RA-associated alleles, identical to those of British patients, are signi®cantly associated with the disease in Greek patients. This view has not been con®rmed in HLA studies of Italian RA patients [5]. However, it is interesting to note that HLA studies of RA in Italy were performed only in patients from the North. A further characteristic of patients from southern Italy was a younger age at onset of RA and a lower degree of familiar aggregation of the disease than those from the North. A young age of RA onset has been repeatedly reported also in West Africa [19]. These differences may re¯ect di€erences in the age distribution of these populations. In 1990, the percentage of population in the age range between 0 and 14 years was 13.4% in northern Italy and 20.5% in southern Italy. Conversely, subjects aged more than 65 years accounted for 17% of

217

the population in northern Italy and for 12.4% in southern Italy [24]. The present study con®rmed, on a large population basis, the higher number of eosinophils in patients with EAMs than in patients without EAMs. RA patients with severe peripheral eosinophilia have been reported to have more functional impairment and more EAMs than controls [25]. The role of eosinophils in RA, in particular when it is accompanied by EAMs, may be related to the high levels of circulating immune complexes in these patients and to a defect in cell-mediated response [26]. Further studies on the relationship between the eosinophilic count and RA features could clarify whether it may be used as a marker of disease severity. Our data indicate that RA in Italy has a milder course and fewer EAMs than in Britain and North America. This ®nding is probably part of a more generalized geographical trend in which RA shows a decreasing severity from northern to southern latitudes. However, this same geographical trend was not found within Italy, where patients from the South had signi®cantly more EAMs than those from the North. The reasons for the milder course of RA in Italy are not known, although several hypotheses could be considered [27]. It has been postulated that bacterial and parasitic infections during childhood may determine a protective e€ect against RA and other autoimmune disease by inducing tolerance to the postulated infective trigger of RA [14]. Other environmental factors could include nutrition and climate. For instance, the intake of unsaturated fatty acids such as olive oil is higher in the Mediterranean area than in northern Europe. Recent data point to the possible anti-in¯ammatory activity of olive oil [13]. A warm climate and especially exposure to sunlight may have an e€ect on the immune system [12], which may be relevant to the pathogenesis of RA. Further studies comparing RA in di€erent countries or in di€erent ethnic groups and evaluating risk factors for severe disease could be useful for clarifying unknown aspects of the etiology of RA. Acknowledgments Prof. Alberto Piazza, Dipartimento di Genetica, Biologia e Chimica Medica, UniversitaÁ di Torino, helped with insightful comments. Dr. Carlo M. Pesce, Cattedra di Istituzioni di Anatomia Patologica, UniversitaÁ di Genova, reviewed the manuscript.

References 1. Hurd ER (1979) Extraarticular manifestations of rheumatoid arthritis. Semin Arthritis Rheum 8: 151±176 2. Leigh JP, Fries JF (1991) Mortality predictors among 263 patients with rheumatoid arthritis. J Rheumatol 18: 1307±1312 3. Jajic Z, Jajic I (1988) Antigens of HLA-DR locus in rheumatoid arthritis. Br J Rheumatol 27 [Suppl 2]: 139±140 4. Mercier P, Dallest R, Roux H (1980) HLA-DR et polyarthrite rhumatoide de l'adulte. Etude dans le Sud de la France. Rev Rhum Mal Osteoartic 47: 465±470 5. Salvarani C, Macchioni PL, Mantovani W, Bragliani M, Collina E, Cremonesi T, Battistel B, Boiardi L (1998) HLADRB1 alleles associated with rheumatoid arthritis in northern

6. 7.

8.

9. 10. 11. 12. 13.

14. 15.

16. 17. 18. 19. 20. 21. 22. 23.

24. 25. 26. 27.

Italy: correlation with disease severity. Br J Rheumatol 37: 165±169 Drosos AA, Lanchbury JS, Panay GS, Moutsopoulos HM (1992) Rheumatoid arthritis in Greek and British patients. Arthritis Rheum 35: 745±748 Salvarani C, Macchioni P, Mantovani W, Rossi F, Veneziani M, Boiardi L, Lodi L, Portioli I (1992) Extraarticular manifestations of rheumatoid arthritis and HLA antigens in Northern Italy. J Rheumatol 19: 242±246 Boki KA, Panay GS, Vaughan RW, Drosos AA, Moutsopoulos HM, Lanchbury JS (1992) HLA class II sequence polymorphisms and susceptibility to rheumatoid arthritis in Greeks: the HLADRB shared-epitope hypothesis accounts for the disease in only a minority of Greek patients. Arthritis Rheum 35: 749±755 Mody GM, Hammond MG, Naidoo PD (1989) HLA associations with rheumatoid arthritis in African blacks. J Rheumatol 16: 1326±1328 AlarcËon GS, Koopman WJ, Acton RT, Barger BO (1983) DR4 antigen distribution in blacks with rheumatoid arthritis. J Rheumatol 10: 579±583 Cimmino MA, Parisi M, Moggiana G, Mela GS, Accardo S (1998) Prevalence of rheumatoid arthritis in Italy: the Chiavari study. Ann Rheum Dis 57: 315±318 Hersey P, Bradley M, Hasic E, Haran G, Edwards A, McCarthy WH (1983) Immunological e€ects of solarium exposure. Lancet 2: 545±548 Brzesky M, Madhok R, Capell HA (1991) Evening primrose oil in patients with rheumatoid arthritis and side e€ects of nonsteroidal anti-in¯ammatory drugs. Br J Rheumatol 30: 370±372 Adebajo AO (1991) Is rheumatoid arthritis an infectious disease? BMJ 303: 706 Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, Healey LA, Kaplan SR, Liang MIFI, Luthra HS, Medsger TA Jr, Mitchell DM, Neustadt DH, Pinals RS, Schaller JG, Sharp JT, Wilder RL, Hunder GG (1988) The American Rheumatism Association 1987 revised criteria for the classi®cation of rheumatoid arthritis. Arthritis Rheum 31: 315±324 Steinbroker O, Traeger CH, Batterman RC (1949) Therapeutic criteria in rheumatoid arthritis. JAMA 140: 659 Gordon DA, Stein JL, Broder I (1973) The extra-articular features of rheumatoid arthritis. A systemic analysis of 127 cases. Am J Med 54: 445±452 Dudley Hart F (1969) Rheumatoid arthritis: extraarticular manifestations. BMJ 3: 131±136 Adebajo AO, Reid DM (1991) The pattern of rheumatoid arthritis in West Africa and comparison with a cohort of British patients. QJM 292: 633±640 Veerapen K, Mangat G, Watt I, Dieppe P (1993) The expression of rheumatoid arthritis in Malaysian and British patients: a comparative study. Br J Rheumatol 32: 541±545 Cohen MG, Li EK, Ng PY, Chan KL (1993) Extra-articular manifestations are uncommon in Southern Chinese with rheumatoid arthritis. Br J Rheumatol 32: 541±545 Pountain G (1991) The prevalence of rheumatoid arthritis in the sultanate of Oman. Br J Rheumatol 30: 24±28 Piazza A, Olivetti E, Gri€o RM, Rendine S, Amoroso A, Barbanti M, Caruso C, Conighi C, Conte R, Favoino B, Galliano L, Luciani G, Magri D, Martinetti M, Mattiuz PL, Menicucci A, Misefari V, Moro L, Purpura M, Savi M (1989) The distribution of HLA antigens in Italy. Gene Geogr 3: 141±164 (1990) Popolazione e movintento anagra®co dei comuni: anno 1988. ISTAT, Istituto Centrale di Statistica Panush RS, Franco AE, Schur PH (1971) Rheumatoid arthritis associated with eosinophilia. Ann Intern Med 75: 199±205 Davis P (1981) Eosinophils and the connective tissue diseases. In: Carson Dick W (ed) Immunological aspects of rheumatology. Elsevier, North Holland, pp 171±182 Cimmino MA (1999) Geographical di€erences in the prevalence and severity of rheumatoid arthritis in Europe. CPD Rheumatol 1: 28±34