ARTHRITIS
I29
&
RHEUMATISM OFFICIAL J O U R N A L OF T H E AMERICAN RHEUMATISM ASSOCIATION SECTION OF T H E A R T H R I T I S FOUNDATION
HISTOCOMPATIBILITYANTIGENS IN SYSTEMIC LUPUS ERYTHEMATOSUS MARC A. GOLDBERG, FRANK C. ARNETT, WILMA B. BIAS, and LAWRENCE E. SHULMAN Histocompatibility (HL-A) antigens were determined in 120 patients with systemic lupus erythematosus (SLE) and 120 matched controls. Increased frequencies of HL-AI and HL-A8 were found. HL-A1 was more strongly associated with SLE in black patients (71 patients), whereas HL-A8 was mote impressively associated with SLE in white patients (49 patients). In addition HL-AI appeared more frequently in those with early onset of disease in both races; and HL-AI, HL-AS, and the HLA1,8 phenotype seemed to be associated with severe SLE (renal and central nervous system involvement) in white patients. These data support the proposal that there are genetic influences in the pathogenesis and expression of SLE. Several reports have indicated an association between certain HL-A antigens and systemic lupus erFrom the Connective Tissue Division, Department of Medicine, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205. Presented in part at the American Rheumatism Association Meetings, June 7-8, 1973, Los Angeles, California. S u p p o r t e d by U S P H S G r a d u a t e T r a i n i n g G r a n t s 5TOlAM05033-17 and 18, and by Study of Human Genetics Grant SPOIGM19489. Marc A. Goldberg, M.D.: Fellow in Medicine, The Johns Hopkins University School of Medicine, presently a t the Watson US Army Hospital, Fort Dix, New Jersey; Frank C. Arnett, M.D.: Assistant Professor of Medicine, The Johns Hopkins University School of Medicine; Wilma B. Bias, Ph.D.: Associate Professor of Medicine, The Johns Hopkins University School of Medicine; Lawrence E. Shulman, M.D., Ph.D.: Associate Professor of Medicine, The Johns Hopkins University School of Medicine. Address reprint requests t o Frank C. Arnett, M.D., The Good Samaritan Hospital, 5601 Loch Raven Boulevard, Baltimore, Maryland 21239. Submitted for publication April 16, 1975; accepted July 18, 1975.
Arthritis and Rheumatism, Vol. 19, No. 2 (March-April 1976)
ythematosus (SLE); however the actual HL-A specificities reported have varied widely. Waters et a1 (1) initially described an excess of W 15 (Thorsby-LND) in predominantly white persons of northern European backgrounds whose disease began before 25 years of age. Grumet et at (2) confirmed this finding and also reported an association of HL-A8 with SLE in a series of 40 predominantly white patients. Studying black SLE patients, Bitter et a1 (3) described an increase of HL-A7, HL-A5, and W5, whereas Stastny (4) found an association with W19 and HL-AS. Nies et a1 ( 5 ) studied 122 patients with SLE from various ethnic backgrounds and noted only increased HL-A5 among black patients. These variations may be attributable to differences in the size of the samples tested or to population variables including age, race, and geographic differences. Similarly the selection and number of controls along with nonuniformity of tissue-typing antisera and techniques may also be important. The present report describes a study of the HL-A specificities in 120 patients with SLE and 120 matched controls. Frequencies of HL-A1 and HL-A8 were increased in SLE patients. HL-A1 was more strongly associated with SLE in black patients, whereas HL-A8 was present more frequently in white patients. Additionally HL-A1 seemed to be associated with early onset of disease, and HL-Al, HL-A8, and the HL-Al, 8 phenotype appeared to be associated with more serious SLE (renal and central nervous system involvement).
MATERIALS AND METHODS One-hundred-and-twenty unrelated patients with SLE (49 whites, 71 blacks) were selected randomly from the in-
GOLDBERG ET AL
130
patient and outpatient facilities of The Johns Hopkins and Good Samaritan Hospitals. The diagnosis of SLE was based on both clinical and serologic findings. An equal number of normal unrelated controls (hospital employees, students, and normal relatives of potential allograft recipients) were matched with the SLE patients for age (within 5 years), sex, and race. HL-A typing was performed using a standard microlymphocytotoxicity test designed to detect 21 different HL-A specificities (6). Statistical analysis of these data was performed using the chi square test, and Yates’ correction was applied for small numbers when appropriate. The P value obtained was then additionally corrected by multiplying by the number of antigens tested for (7,8). Both the original and corrected P values are presented.
RESULTS In the 120 patients with SLE the frequency of the first segregant series antigen, HL-A1, was increased (22%) as compared to the 120 controls (9%) (P = 0.004, corrected P = 0.08) (Table 1). Similarly the second segregant series antigen, HL-A8, occurred in 19% of patients and 8% of controls (P = 0.007, corrected P = 0.14). The incidence of the HL-A1,8 phenotype in SLE patients was 12% compared t o 5% of controls (P = 0.03, corrected P = 0.30). Table 1. Frequency of HL-A Antigens in SLE Patients and Controls HL-A Antigen
SLE (120 patients)
First segregant series I L
3 9 10
11
II W19 W28
8 20
Second segregant series 5 7 8 12 13 w5 WIO W14 w15 W16 W17 w22 W27
*P
=
P
=
22%* 38 17 23
0.004,corrected P = 0.08. 0.007, corrected P = 0.14.
10
10%
30 19t 18 8 25 5 8 2 2 17 3 3
Controls (120patients) 9%* 43 23 22 14 8 14 4
16% 29 8t 23 3 23
Analysis of the series by race (Table 2) showed the 71 black patients with SLE to have an increased frequency of HL-A1 (14%) as compared to controls (0%) (P = O.OOO5, corrected P = 0.01). HL-A1 was slightly increased in the 49 white patients (33% compared to 22% in controls), but this difference was not statistically significant. HL-AS was found in 35% of white SLE patients and in 12% of controls (P = 0.004, corrected P = 0.08) but was not increased in the black patients. The HL-A1,8 phenotype showed a trend in white SLE patients (22%) over controls (12%) but was not increased in black patients. In order to determine if either HL-A1 or HL-A8 was related to certain specific attributes of SLE, patients were segregated according to the major manifestations of their disease and evaluated for the presence of these antigens. Specific features examined included nephritis, central nervous system (CNS) involvement, age a t onset of disease, serum complement depression, and antinuclear antibody titer levels. For the 17 white patients with unequivocal renal disease, the incidence of HL-A1 was 53% compared to 22% for the 32 white patients without apparent renal involvement (P = 0.014). Likewise the frequency of HLA8 was increased at 53% compared to 25% in white patients without this complication (P = 0.025). The HLA 1,8 phenotype occurred in 29% of white SLE patients with nephritis compared to 19% of white patients without nephritis, a nonsignificant difference. There were 15 white patients with, and 34 without, CNS involvement. Of those patients with CNS disease 40% had HL-A 1, whereas the incidence of this antigen in those without CNS disease was 29%. The frequency of HL-A8 was 47% in those with CNS disease, compared to 29% in patients who did not have CNS manifestations. These differences were not statistically significant. However the HL-A1,8 phenotype seen in 40% of whites with CNS disease and in only 15% of those without it was significantly increased (P= 0.025). Table 2. Frequency o i HL-A Antigens in SLE Patients and Controls by Racial Group Whites
I 6 5 I 15 3 4
Blacks
Antigen
SLE (49 )
Controls (49)
SLE (71)
Controls (71)
HL-AI HL-A8 HL-AI, 8
33% 35 22
22% 12. 12
14% 8 4
OW
*P tP
= =
0.004, corrected P = 0.08. 0.0005, corrected P = 0.01.
6 0
HL-A ANTIGENS IN SLE
Black patients analyzed in a similar manner failed to show an increased incidence of these or other antigens when examined for the presence or absence of renal or CNS disease. Analysis of data for age at onset of disease showed HL-A1 present in 46% of the white patients in whom SLE was documented by age 25 years or earlier. This is an increased incidence when compared to the HL-A1 frequency of 20% for white patients whose disease began after 25 years of age (P = 0.027). The incidence in black patients with disease onset by age 25 who demonstrated HL-A1 was 29%, whereas the incidence in those with onset after age 25 was 6% (P = 0.012). The occurrence of HL-A8 and the HL-A1,8 phenotype did not appear related to age of onset in either race. Analysis of the HL-A data with respect to decreased serum complement or to the titer of antinuclear antibodies failed to show any significant associations in either race.
DISCUSSION The present study suggests that there may be significant associations of HL-Al, HL-AS, and the HLA 1,8 phenotype.with certain attributes of systemic lupus erythematosus. Variability does exist between races, with HL-A1 being found in more black SLE patients and HL-A8 appearing more frequently in white SLE patients. For both racial groups there appears to be an association of HL-A1 with early onset of disease. Moreover for white patients it is suggested that an association exists between HL-Al, HL-A8, and the HL-A1,8 phenotype and a more severe form of disease, as manifested by renal and CNS involvement. This series supports the association of HL-A8 with SLE, initially described by Grumet et (11 (2) predominantly in white patients. W15 (LND), reported to be increased in the studies by Waters (1) and by Grumet (2), could not be evaluated adequately in the present study because of technical problems with antisera for this particular antigen. In light of the present results, it is of interest that Grumet (2) also demonstrated a high incidence of HLA1 (41%)in white patients, and that Waters (1) reported high frequencies of HL-A1 (29%), HL-A8(38%), and the HL-A1,8 phenotype (25%). All are comparable to the present data for white patients. Moreover, when only Waters' younger patients are considered, these frequencies are even higher, a fact suggesting that these antigens may be associated with an earlier and more severe form of SLE, as data from the present study may indicate. In
131
the light of these findings it is important to recall that the prognosis of SLE may be worse in children than in adults (9). Of great interest are the findings of Mackay and Morris (10) that HL-A1 and HL-A8 were significantly increased in a series of patients with chronic active hepatitis, a disease with clinical and serologic similarities to SLE. Likewise Kissmeyer-Nielsenet a1 (1 1 ) and Mickey et a1 (12) have described an increased prevalence of these antigens in patients with severe renal transplant rejection. One or both of these antigens have also been implicated in a number of other conditions associated with disordered immunity (13-17), and such findings support the proposal that HL-A 1 and HL-A8 may function as a haplotype associated with an immunopathic response to various antigens (10). The need for continued caution in interpreting data from these studies, including this one, is important. One must avoid overestimating the significance of an increased HL-A antigen because testing for large numbers of antigens promotes the probability that an excess antigen frequency will occur on the basis of chance alone. Several methods have been proposed for circumventing this problem (2,7,8). The most widely advocated maneuver entails multiplying the P value obtained from an increased antigen frequency by the number of antigens sought. When this method was applied to the present data, the association of HL-A1 with SLE in black patients remained statistically significant. The increased frequency of HL-A8 in white SLE patients did not retain statistical significance (P = 0.08); however the trend in this series and those of others (1,2) suggests that this association may be real. This proposal is further supported by the segregation of this antigen according to those attributes connoting disease severity. Therefore further evaluation of HL-A antigens in still larger numbers of SLE patients and with attention to specific disease features appears warranted. The importance of controls matched for age, sex, and race must be stressed. Racial differences in HL-A antigen frequencies have been well documented (1 8). The present study and that of Waters suggest that age may be an important factor in associating HL-A antigens with SLE. Moreover, in view of the predominance of SLE in women, appropriate attention to sex should be exercised when choosing controls. Therefore these data emphasize the need for further studies, especially in families in which several members have SLE. They do suggest, however, a possible role for histocompatibility antigens in the pathogenesis of this disorder.
GOLDBERG ET AL
132
ACKNOWLEDGMENTS The authors wish to thank Gary A. Chase, Ph.D. and Helen Abbey, Sc.D. for help in the statistical analysis of the data.
REFERENCES
9. 10.
11.
1. Waters H, Konrad P, Walford R: The distribution of HL-
A histocompatibility factors and genes in patients with systemic lupus erythematosus. Tissue Antigens 1:68-73, 1971 2. Grumet FC, Coukell A, Bodmer JG, et al: Histocompatibility (HL-A) antigens associated with systemic lupus erythematosus. A possible genetic predisposition to disease. N Engl J Med 285193-196, 1971 3. Bitter T, Mottironi WD, Terasaki PI: HL-A antigens associated with lupus erythematosus. N Engl J Med 285:435, 1972 4. Stastny P: The distribution of HL-A antigens in black patients with systemic lupus erythematosus (SLE). Arthritis Rheum 15:455, 1972 5 . Nies KM, Brown JC, Dubois EL, et al: Histocompatibility (HL-A) antigens and lymphocytotoxic antibodies in systemic lupus erythematosus (SLE). Arthritis Rheum 17:397-402, 1974 6. Amos DB, Bashir H, Boyle W, et al: A simple micro cytotoxicity test. Transplantation 7:220-222, 1969 7. Bach FH: Disease and the HL-A histocompatibility system. Ann Intern Med 75:962-964, 1971 8. McDevitt HO, Bodmer WF: Histocompatibility antigens,
12.
13.
14.
15. 16.
17. 18.
immune responsiveness and susceptibility to disease. Am J Med 52:l-8, 1972 Meislin AG, Rothfield N: Systemic lupus erythematosus in childhood. Pediatrics 42:37-49, 1968 Mackay IR, Morris PJ: Association of autoimmune active chronic hepatitis with HL-A l,8. Lancet 2:793-795, 1972 Kissmeyer-Nielsen F, Svejgaard A, Fjeldborg 0, et al: Scandiatransplant: preliminary report of a kidney exchange program. Transplant Proc 3: 1019-1029, 1971 Mickey M R , Kreisler M, Albert ED, et al: Analysis of HL-A incompatibility in human renal transplants. Tissue Antigens 157-67, 1971 Thorsby E, Engeset A, Lie SO: HL-A antigens and susceptibility to diseases. A study of patients with acute lymphoblastic leukaemia, Hodgkin’s disease and childhood asthma. Tissue Antigens 1:147-152, 1971 Falchuk Z M , Rogentine G N , Strober W: Predominance of histocompatibility antigen HL-A8 in patients with gluten-sensitive enteropathy. J Clin Invest 51:1602-1605, 1972 Falk J, Osoba D: HL-A antigens and survival in Hodgkin’s disease. Lancet 2: 1 1 18-1 120, 197 1 Stokes PL, Asquith P, Holmes G T K , et al: Histocompatibility antigens associated with adult coeliac disease. Lancet 2:162-164, 1972 White AG, Baretson RSt-G, DaCosta JAG, et al: HL-A and disordered immunity. Lancet 1:108, 1973 Albert ED, Mickey MR, McNicholas AC, et al: Seven new HL-A specificities and their distribution in three races, Histocompatibility Testing. Edited by PI Terasaki. Copenhagen, Munksgaard, 1970, pp 221-230
I29
&
RHEUMATISM OFFICIAL J O U R N A L OF T H E AMERICAN RHEUMATISM ASSOCIATION SECTION OF T H E A R T H R I T I S FOUNDATION
HISTOCOMPATIBILITYANTIGENS IN SYSTEMIC LUPUS ERYTHEMATOSUS MARC A. GOLDBERG, FRANK C. ARNETT, WILMA B. BIAS, and LAWRENCE E. SHULMAN Histocompatibility (HL-A) antigens were determined in 120 patients with systemic lupus erythematosus (SLE) and 120 matched controls. Increased frequencies of HL-AI and HL-A8 were found. HL-A1 was more strongly associated with SLE in black patients (71 patients), whereas HL-A8 was mote impressively associated with SLE in white patients (49 patients). In addition HL-AI appeared more frequently in those with early onset of disease in both races; and HL-AI, HL-AS, and the HLA1,8 phenotype seemed to be associated with severe SLE (renal and central nervous system involvement) in white patients. These data support the proposal that there are genetic influences in the pathogenesis and expression of SLE. Several reports have indicated an association between certain HL-A antigens and systemic lupus erFrom the Connective Tissue Division, Department of Medicine, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205. Presented in part at the American Rheumatism Association Meetings, June 7-8, 1973, Los Angeles, California. S u p p o r t e d by U S P H S G r a d u a t e T r a i n i n g G r a n t s 5TOlAM05033-17 and 18, and by Study of Human Genetics Grant SPOIGM19489. Marc A. Goldberg, M.D.: Fellow in Medicine, The Johns Hopkins University School of Medicine, presently a t the Watson US Army Hospital, Fort Dix, New Jersey; Frank C. Arnett, M.D.: Assistant Professor of Medicine, The Johns Hopkins University School of Medicine; Wilma B. Bias, Ph.D.: Associate Professor of Medicine, The Johns Hopkins University School of Medicine; Lawrence E. Shulman, M.D., Ph.D.: Associate Professor of Medicine, The Johns Hopkins University School of Medicine. Address reprint requests t o Frank C. Arnett, M.D., The Good Samaritan Hospital, 5601 Loch Raven Boulevard, Baltimore, Maryland 21239. Submitted for publication April 16, 1975; accepted July 18, 1975.
Arthritis and Rheumatism, Vol. 19, No. 2 (March-April 1976)
ythematosus (SLE); however the actual HL-A specificities reported have varied widely. Waters et a1 (1) initially described an excess of W 15 (Thorsby-LND) in predominantly white persons of northern European backgrounds whose disease began before 25 years of age. Grumet et at (2) confirmed this finding and also reported an association of HL-A8 with SLE in a series of 40 predominantly white patients. Studying black SLE patients, Bitter et a1 (3) described an increase of HL-A7, HL-A5, and W5, whereas Stastny (4) found an association with W19 and HL-AS. Nies et a1 ( 5 ) studied 122 patients with SLE from various ethnic backgrounds and noted only increased HL-A5 among black patients. These variations may be attributable to differences in the size of the samples tested or to population variables including age, race, and geographic differences. Similarly the selection and number of controls along with nonuniformity of tissue-typing antisera and techniques may also be important. The present report describes a study of the HL-A specificities in 120 patients with SLE and 120 matched controls. Frequencies of HL-A1 and HL-A8 were increased in SLE patients. HL-A1 was more strongly associated with SLE in black patients, whereas HL-A8 was present more frequently in white patients. Additionally HL-A1 seemed to be associated with early onset of disease, and HL-Al, HL-A8, and the HL-Al, 8 phenotype appeared to be associated with more serious SLE (renal and central nervous system involvement).
MATERIALS AND METHODS One-hundred-and-twenty unrelated patients with SLE (49 whites, 71 blacks) were selected randomly from the in-
GOLDBERG ET AL
130
patient and outpatient facilities of The Johns Hopkins and Good Samaritan Hospitals. The diagnosis of SLE was based on both clinical and serologic findings. An equal number of normal unrelated controls (hospital employees, students, and normal relatives of potential allograft recipients) were matched with the SLE patients for age (within 5 years), sex, and race. HL-A typing was performed using a standard microlymphocytotoxicity test designed to detect 21 different HL-A specificities (6). Statistical analysis of these data was performed using the chi square test, and Yates’ correction was applied for small numbers when appropriate. The P value obtained was then additionally corrected by multiplying by the number of antigens tested for (7,8). Both the original and corrected P values are presented.
RESULTS In the 120 patients with SLE the frequency of the first segregant series antigen, HL-A1, was increased (22%) as compared to the 120 controls (9%) (P = 0.004, corrected P = 0.08) (Table 1). Similarly the second segregant series antigen, HL-A8, occurred in 19% of patients and 8% of controls (P = 0.007, corrected P = 0.14). The incidence of the HL-A1,8 phenotype in SLE patients was 12% compared t o 5% of controls (P = 0.03, corrected P = 0.30). Table 1. Frequency of HL-A Antigens in SLE Patients and Controls HL-A Antigen
SLE (120 patients)
First segregant series I L
3 9 10
11
II W19 W28
8 20
Second segregant series 5 7 8 12 13 w5 WIO W14 w15 W16 W17 w22 W27
*P
=
P
=
22%* 38 17 23
0.004,corrected P = 0.08. 0.007, corrected P = 0.14.
10
10%
30 19t 18 8 25 5 8 2 2 17 3 3
Controls (120patients) 9%* 43 23 22 14 8 14 4
16% 29 8t 23 3 23
Analysis of the series by race (Table 2) showed the 71 black patients with SLE to have an increased frequency of HL-A1 (14%) as compared to controls (0%) (P = O.OOO5, corrected P = 0.01). HL-A1 was slightly increased in the 49 white patients (33% compared to 22% in controls), but this difference was not statistically significant. HL-AS was found in 35% of white SLE patients and in 12% of controls (P = 0.004, corrected P = 0.08) but was not increased in the black patients. The HL-A1,8 phenotype showed a trend in white SLE patients (22%) over controls (12%) but was not increased in black patients. In order to determine if either HL-A1 or HL-A8 was related to certain specific attributes of SLE, patients were segregated according to the major manifestations of their disease and evaluated for the presence of these antigens. Specific features examined included nephritis, central nervous system (CNS) involvement, age a t onset of disease, serum complement depression, and antinuclear antibody titer levels. For the 17 white patients with unequivocal renal disease, the incidence of HL-A1 was 53% compared to 22% for the 32 white patients without apparent renal involvement (P = 0.014). Likewise the frequency of HLA8 was increased at 53% compared to 25% in white patients without this complication (P = 0.025). The HLA 1,8 phenotype occurred in 29% of white SLE patients with nephritis compared to 19% of white patients without nephritis, a nonsignificant difference. There were 15 white patients with, and 34 without, CNS involvement. Of those patients with CNS disease 40% had HL-A 1, whereas the incidence of this antigen in those without CNS disease was 29%. The frequency of HL-A8 was 47% in those with CNS disease, compared to 29% in patients who did not have CNS manifestations. These differences were not statistically significant. However the HL-A1,8 phenotype seen in 40% of whites with CNS disease and in only 15% of those without it was significantly increased (P= 0.025). Table 2. Frequency o i HL-A Antigens in SLE Patients and Controls by Racial Group Whites
I 6 5 I 15 3 4
Blacks
Antigen
SLE (49 )
Controls (49)
SLE (71)
Controls (71)
HL-AI HL-A8 HL-AI, 8
33% 35 22
22% 12. 12
14% 8 4
OW
*P tP
= =
0.004, corrected P = 0.08. 0.0005, corrected P = 0.01.
6 0
HL-A ANTIGENS IN SLE
Black patients analyzed in a similar manner failed to show an increased incidence of these or other antigens when examined for the presence or absence of renal or CNS disease. Analysis of data for age at onset of disease showed HL-A1 present in 46% of the white patients in whom SLE was documented by age 25 years or earlier. This is an increased incidence when compared to the HL-A1 frequency of 20% for white patients whose disease began after 25 years of age (P = 0.027). The incidence in black patients with disease onset by age 25 who demonstrated HL-A1 was 29%, whereas the incidence in those with onset after age 25 was 6% (P = 0.012). The occurrence of HL-A8 and the HL-A1,8 phenotype did not appear related to age of onset in either race. Analysis of the HL-A data with respect to decreased serum complement or to the titer of antinuclear antibodies failed to show any significant associations in either race.
DISCUSSION The present study suggests that there may be significant associations of HL-Al, HL-AS, and the HLA 1,8 phenotype.with certain attributes of systemic lupus erythematosus. Variability does exist between races, with HL-A1 being found in more black SLE patients and HL-A8 appearing more frequently in white SLE patients. For both racial groups there appears to be an association of HL-A1 with early onset of disease. Moreover for white patients it is suggested that an association exists between HL-Al, HL-A8, and the HL-A1,8 phenotype and a more severe form of disease, as manifested by renal and CNS involvement. This series supports the association of HL-A8 with SLE, initially described by Grumet et (11 (2) predominantly in white patients. W15 (LND), reported to be increased in the studies by Waters (1) and by Grumet (2), could not be evaluated adequately in the present study because of technical problems with antisera for this particular antigen. In light of the present results, it is of interest that Grumet (2) also demonstrated a high incidence of HLA1 (41%)in white patients, and that Waters (1) reported high frequencies of HL-A1 (29%), HL-A8(38%), and the HL-A1,8 phenotype (25%). All are comparable to the present data for white patients. Moreover, when only Waters' younger patients are considered, these frequencies are even higher, a fact suggesting that these antigens may be associated with an earlier and more severe form of SLE, as data from the present study may indicate. In
131
the light of these findings it is important to recall that the prognosis of SLE may be worse in children than in adults (9). Of great interest are the findings of Mackay and Morris (10) that HL-A1 and HL-A8 were significantly increased in a series of patients with chronic active hepatitis, a disease with clinical and serologic similarities to SLE. Likewise Kissmeyer-Nielsenet a1 (1 1 ) and Mickey et a1 (12) have described an increased prevalence of these antigens in patients with severe renal transplant rejection. One or both of these antigens have also been implicated in a number of other conditions associated with disordered immunity (13-17), and such findings support the proposal that HL-A 1 and HL-A8 may function as a haplotype associated with an immunopathic response to various antigens (10). The need for continued caution in interpreting data from these studies, including this one, is important. One must avoid overestimating the significance of an increased HL-A antigen because testing for large numbers of antigens promotes the probability that an excess antigen frequency will occur on the basis of chance alone. Several methods have been proposed for circumventing this problem (2,7,8). The most widely advocated maneuver entails multiplying the P value obtained from an increased antigen frequency by the number of antigens sought. When this method was applied to the present data, the association of HL-A1 with SLE in black patients remained statistically significant. The increased frequency of HL-A8 in white SLE patients did not retain statistical significance (P = 0.08); however the trend in this series and those of others (1,2) suggests that this association may be real. This proposal is further supported by the segregation of this antigen according to those attributes connoting disease severity. Therefore further evaluation of HL-A antigens in still larger numbers of SLE patients and with attention to specific disease features appears warranted. The importance of controls matched for age, sex, and race must be stressed. Racial differences in HL-A antigen frequencies have been well documented (1 8). The present study and that of Waters suggest that age may be an important factor in associating HL-A antigens with SLE. Moreover, in view of the predominance of SLE in women, appropriate attention to sex should be exercised when choosing controls. Therefore these data emphasize the need for further studies, especially in families in which several members have SLE. They do suggest, however, a possible role for histocompatibility antigens in the pathogenesis of this disorder.
GOLDBERG ET AL
132
ACKNOWLEDGMENTS The authors wish to thank Gary A. Chase, Ph.D. and Helen Abbey, Sc.D. for help in the statistical analysis of the data.
REFERENCES
9. 10.
11.
1. Waters H, Konrad P, Walford R: The distribution of HL-
A histocompatibility factors and genes in patients with systemic lupus erythematosus. Tissue Antigens 1:68-73, 1971 2. Grumet FC, Coukell A, Bodmer JG, et al: Histocompatibility (HL-A) antigens associated with systemic lupus erythematosus. A possible genetic predisposition to disease. N Engl J Med 285193-196, 1971 3. Bitter T, Mottironi WD, Terasaki PI: HL-A antigens associated with lupus erythematosus. N Engl J Med 285:435, 1972 4. Stastny P: The distribution of HL-A antigens in black patients with systemic lupus erythematosus (SLE). Arthritis Rheum 15:455, 1972 5 . Nies KM, Brown JC, Dubois EL, et al: Histocompatibility (HL-A) antigens and lymphocytotoxic antibodies in systemic lupus erythematosus (SLE). Arthritis Rheum 17:397-402, 1974 6. Amos DB, Bashir H, Boyle W, et al: A simple micro cytotoxicity test. Transplantation 7:220-222, 1969 7. Bach FH: Disease and the HL-A histocompatibility system. Ann Intern Med 75:962-964, 1971 8. McDevitt HO, Bodmer WF: Histocompatibility antigens,
12.
13.
14.
15. 16.
17. 18.
immune responsiveness and susceptibility to disease. Am J Med 52:l-8, 1972 Meislin AG, Rothfield N: Systemic lupus erythematosus in childhood. Pediatrics 42:37-49, 1968 Mackay IR, Morris PJ: Association of autoimmune active chronic hepatitis with HL-A l,8. Lancet 2:793-795, 1972 Kissmeyer-Nielsen F, Svejgaard A, Fjeldborg 0, et al: Scandiatransplant: preliminary report of a kidney exchange program. Transplant Proc 3: 1019-1029, 1971 Mickey M R , Kreisler M, Albert ED, et al: Analysis of HL-A incompatibility in human renal transplants. Tissue Antigens 157-67, 1971 Thorsby E, Engeset A, Lie SO: HL-A antigens and susceptibility to diseases. A study of patients with acute lymphoblastic leukaemia, Hodgkin’s disease and childhood asthma. Tissue Antigens 1:147-152, 1971 Falchuk Z M , Rogentine G N , Strober W: Predominance of histocompatibility antigen HL-A8 in patients with gluten-sensitive enteropathy. J Clin Invest 51:1602-1605, 1972 Falk J, Osoba D: HL-A antigens and survival in Hodgkin’s disease. Lancet 2: 1 1 18-1 120, 197 1 Stokes PL, Asquith P, Holmes G T K , et al: Histocompatibility antigens associated with adult coeliac disease. Lancet 2:162-164, 1972 White AG, Baretson RSt-G, DaCosta JAG, et al: HL-A and disordered immunity. Lancet 1:108, 1973 Albert ED, Mickey MR, McNicholas AC, et al: Seven new HL-A specificities and their distribution in three races, Histocompatibility Testing. Edited by PI Terasaki. Copenhagen, Munksgaard, 1970, pp 221-230
