Pmwn. mdwrd Drff Vol. 6. No. 2. pp. 283-285.1985 Prmted in Great Britain 0191-8869 85 S3.00 co.00 Pergamon Press Ltd A possible association between HLA B-27 and vulnerability to schizophrenia WAGNER F. GATTAZ,’ MICHAEL SEITZ’ and HELMUT BECKMANN’ ’ Cenrral Ittsritule of Menial Health, 6800 Mmnheim I-J.5, F.R.G. ’ Rheumaric Diseases Ambulatory C:nit, Unrl~ersit~ q/ Heidelberg, 6900 Heidelberg, F. R.G. (Received IO April 1984) Summary-The frequency of HLA B-27 has been found to be increased in schizophrenic patients, especially in those with a positive family history of schizophrenia. In the present study, psychiatrically healthy arthropathic patients bearing HLA B-27 were found to have significantly higher P scores as compared to arthropathic patients without this antigen. Psychoticism could be seen as a trait of the personality at high-risk for schizophrenia. Thus, the present finding suggests an association between HLA B-27 and the genetic vulnerability to some forms of schizophrenia. INTRODUCTION The human leucocyte antigens (HLA) are individual traits genetically determined by a locus on chromosome 6. Several diseases with a known or suspected hereditary element or with a possible immunological basis have been shown to correlate positively with specific HLAs (Dick, 1978). Similarly. associations between the HLA system and schizophrenia have been reported [for a review see Gattaz and Beckmann (1981)] but results from different studies have frequently been contradictory. Several factors could account for the discrepancies of the published findings: besides the possibility of no association at all between the HLA system and the disease (findings by chance), the study of psychopathologically heterogeneous patient groups and the investigation of ethnically heterogeneous populations should be considered, as the frequencies of HLA show significant differences among different racial groups (Bodmer and Bodmer, 1978). Indeed, consideration of these factors has yielded considerable agreement between different studies (McGuRin, Farmer and Yonace, 1981; Gattaz. Beckmann and Mendlewicz, 1981a). Against this background, a notable similarity was found in the results from three studies carried out under identical methodological procedures in a population of central European Caucasoid probands: Gattaz, Ewald and Beckmann (1980) found a significantly increased incidence of HLA B-27 in 100 schizophrenic patients as compared to healthy controls. Subsequently. Mendlewicz, Verbanck, Linkowski and Govaerts (1981) and Riisler, Bellaire, Giannitsis, Jarovici and Gress (1982) obtained similar results in 64 and 107 schizophrenics. respectively, selected according to the same diagnostic criteria (Feighner, Robins, Guze, Woodruff, Winokur and Munoz. 1972). As no significant differences arose from the comparisons of the HLA frequencies in controls and patients from the three studies. data could be pooled. The analysis of the total patient group (N = 271) against controls (n = 1185) showed a highly significant increase of HLA B-27 in schizophrenics (14%) as compared to controls (6%; P < 0.0001). This difference remained significant after the correction of the P-values for the number of antigens studied (P x 26). In this population, the relative risk for schizophrenia in individuals with B-27 is 2.2 times higher than in individuals without this antigen. These findings suggest that B-27 could be associated with greater vulnerability to schizophrenia. This is particularly of interest in view of the observations that: B-27 is strongly associated with some forms of arthropathies (i.e. ankylosing spondylitis) but not with others (rheumatoid arthritis) (Sachs and Breweron, 1978); and schizophrenia and arthropathies seem to occur very seldom in the same individual (iisterberg, 1978). If there is an association between B-27 and schizophrenia. psychiatrically healthy individuals bearing this antigen might be expected to have some personality characteristics which would in turn be correlated with the ‘high-risk’ for schizophrenia (Garmezy and Streitman, 1974). To investigate this question, we carried out the present study in a group of arthropathic patients who had no past history of psychiatric illness. MATERIALS AND METHODS Fifty-two arthropathic patients (29 B-27 positive and 23 B-27 negative) of the Rheumatic Diseases Ambulatory Unit (University of Heidelberg) were contacted by mail. Through a letter explaining the nature of the study, probands were asked to answer an enclosed (code-numbered) Eysenck Personality Questionnaire (EPQ). The EPQ contains 90 items which are grouped in four scores: Psychoticism (P), Extraversion (E), Seuroticism (N) and Lie(L) scores (Eysenck and Eysenck, 1975). Thirty-three probands answered the questionnaire, 17 (59’,,) B-27 positive and 16 (70%) B-27 negative. Data from both groups are summarized on Table I. The activity of the disease (arthropathy) and the degree of functional impairment were evaluated in each patient by means of a 4-point scale (O-3) based on the blood sedimentation rates, clinical manifestations and on the Steinbrocker classification (for arthritis patients only). Data were evaluated by means of non-parametric tests (Mann-Whitney U-test and Spearman correlation coefficient, both two-tailed) as well as x2 tests when indicated. RESULTS B-27 positive probands showed significantly higher P scores than B-27 negative (P < 0.01). No significant differences concerning E, N and L scores from the EPQ were found between the groups (Table 2). ‘83