ARTHRITIS & RHEUMATISM Volume 38 Number 3, March 1995, pp 381-383 0 1995, American College of Rheumatology 38 1 OCCURRENCE OF ANKYLOSING SPONDYLITIS IN A NATIONWIDE SERIES OF TWINS PENTTI JARVINEN Objective. To obtain information on the rate of concordance for ankylosing spondylitis (AS) in a population-based series of twins. Methods. AS cases were identified by record linkage of the population-based Finnish Twin Cohort and the nationwide registry for fully reimbursed medications. A clinical examination was performed to establish concordance for AS. Results. There were 6 monozygotic (MZ) pairs and 20 dizygotic (DZ) pairs with at least 1 member affected by AS. Three MZ pairs and 3 DZ pairs were concordant for the disease. All affected subjects were HLA-B27 positive. The pairwise concordance rate was 50% in MZ twins and 20% in HLA-B27 positive DZ twins (95% confidence intervals 11.8-88.2% and 4.3%48.1 %, respectively). Conclusion. These results indicate that AS disease expression is largely, but not entirely, genetically based, with a gene or genes other than B27 probably playing a role. Both genetic and environmental influences play a role in the etiopathogenesis of ankvlosing spondylitis (AS) (1). This disease shows a striking association with a genetic marker, HLA-B27. Family studies have shown that -20% of the first-degree relatives f AS probands also have this disease. Among ranuomly selected B27 positive individuals, the risk is substantially smaller. These observations suggest either the influence of additional genes or the role of a shared environment. Indeed, the presence of HLA-Bw60 has been reported to increase susceptibility to AS in B27 positive patients (2). Supported by the Rheumatism Research Foundation, Finland. Pentti Jarvinen, MD: Rheumatism Foundation Hospital, Heinola, Finland. Address reprint requests to Pentti Jarvinen, MD, Kiljava Hospital, FIN45250 Kiljava, Finland. Submitted for publication March 21, 1994; accepted in revised form September 10, 1994. There have been 110 previously reported population-based studies of AS in twins. The published data, consisting of case reports and small case series, are too sparse to allow estimation of concordance rates for AS in monozygotic (MZ) and dizygotic (DZ) twins (3). The only definitive conclusion to be drawn from these reports is that there are MZ twins who are discordant for AS (4).The present study was undertaken to estimate rates of twin concordance for AS, using data from the Finnish Twin Cohort. PATIENTS AND METHODS Twenty-seven individual twins who had been diagnosed by their attending physicians as having AS, representing 26 twin pairs (6 MZ pairs anti 20 DZ pairs; both members of 1 DZ pair were identified independently of each other), were identified by record linkage of the population-based Finnish Twin Cohort (consisting of 4,137 MZ twins and 9,162 same-sexed DZ twins) and the nationwide registry for fully reimbursed medications (5). Twin zygosity was determined by examining the response of both members of the twin pair to questions on the similarity of appearance and confusions by strangers during childhood (6) and was verified in apparently MZ pairs by determining the genetic markers included in paternity examinations routinely performed in Finland (7). After approval of the study by the institutional ethics committee, clinical examinations were performed. AS was diagnosed according to the New York criteria for definite AS (8). In some instances twins were visited at home. Chest expansion was measured at the level of the fourth intercostal space and was recorded as abnormal if <2.5 cm. Anterior spinal flexion was measured by Schober’s test. Lateral spinal flexion and extension werle evaluated subjectively and recorded as abnormal if judged to be <75% of normal. Ophthalmologic sequelae of anterior uveitis were recorded. Functional outcome was graded 1-4 according to the Steinbrocker criteria (9). In the MZ pair with both AS and rheumatoid arthritis (RA) (see below), the complete HLA type was documented by antibtody-dependent cytotoxicity testing (lo), initiated for the present study. In others, only HLA-B27 was tested in the present study. The prevalence of HLA-B27 in the Finnish population is 14%. After informed consent was obtained, radiographs of the sacroiliac joints were taken if current ones were not available. However, JARVINEN 382 Table 1. Demographic and dislease characteristics of the probands with ankylosing spondylitis No. maleslno. females Mean (range) onset, years Mean (range) current age, years Mean radiologic grade of sacroiliitis, 0-4 scale Extraarticular manifestations, %I Monozygotic twins (n = 6) Dizygotic twins (n = 21)* 511 24.7 (15-38) 45.8 (37-57) t 3.3 I912 21.3 (1 1-33) 49.5 (36-68)$ 3.4 50 52 * Both members of I dizygotic pair were identified by record linkage independently of each other. t One index twin from a monozygotic pair had died at the age of 42. $ One index twin from a dizygcitic pair had died at the age of 68. 8 Anterior uveitis and aortitis. asymptomatic B27 negative twin sibs of B27 positive DZ probands were not requeste:d to undergo radiography. The radiographs were read by a radiologist who was unaware of the clinical data and were sclored according to the New York radiographic criteria (grader;0-4) (8). Existing spinal radiographs were reviewed, or new ones were obtained when clinically indicated. Two male probands, representing 1 MZ and 1 DZ pair, had died. Analysis of their hospital records and spinal radiographs showed that the MZ twin proband had definite AS, and the DZ twin probanid had classic AS with a bamboo spine. One asymptomatic and clinically unaffected B27 positive DZ twin refused to cooperate with respect to the sacroiliac radiography. RESULTS Pertinent data on the probands are shown in Table 1. All probands had AS according to the New York criteria. The were no significant differences between the MZ twin and DZ twin probands with regard to age at onset of AS, disease duration, severity of sacroiliitis, or frequency of extraarticular manifestations. The mean functional score was 2.3. In contrast, a normal functional score and a somewhat shorter duration of AS were noted in the AS cases identified as a result of the present followup study. For probands, the mean disease duration was 27 years. Twin concordance rates are shown in Table 2. In summary, both members of 3 MZ pairs and 3 DZ pairs had AS. Two MZ pairs were discordant for all features of AS. In addition, in 1 MZ pair and 4 DZ pairs, the less severely alffected member, who was not classified as having AS, did have grade 2 radiologic sacroiliitis. All affected subjects were HLA-B27 positive. In contrast, all 5 individual twins who were B27 negative were unaffected members of DZ pairs. Two MZ twin probands and 1 DZ twin proband had aortic insufficiency. Three DZ twin probands had had total hip replacements. Four individuals, representing 3 twin pairs (1 MZ), were diagnosed as having both AS and RA. DISCUSSION There are a number of issues that should be considered with regard to the study design of the current investigation. Because the clinical expression of AS varies widely and there is great variation in the need for treatment, most cases remain undiagnosed. A clinic-based twin series would obviously favor inclusion of severe cases as probands. In contrast, if the probands are recruited by advertising, MZ twins and twins concordant for the disease are preferentially recruited. (MZ twins are known to be more interested in participating in studies than are DZ twins.) To minimize these biases, AS cases in the present study were sought by record linkage from two sources of data: the population-based Finnish Twin Cohort and the nationwide registry for fully reimbursed medications. Those identified were used as probands. Both members of the available twin pairs were then examined, and concordance rates were determined based on fulfillment of the New York criteria for definite AS (8). In this study, all features of AS segregated with HLA-B27. The 20% pairwise concordance rate in B27 positive DZ twins is well in accordance with earlier observations made in family studies (11). The MZ concordance rate indicates the maximum level of Table 2. Pairwise concordance rates for 6 monozygotic (MZ) and 20 dizygotic (DZ) twin pairs with at least 1 member affected by ankylosing spondylitis* MZ pairs Ankylosing spondylitis Radiologic sacroiliitis (grade 2) Anterior uveitis DZ pairs (1 1.8-88.2) 20t (4.3-48.1) 66.7 (22.3-95.7) 33.3 (4.3-77.7) (23-77) 13.3t (1.7-40.5) 50 SO$ * Values for pairwise concordance rates are percentages. Values in parentheses are 95% confidence intervals; exact values were used (ref. 15). t Concordance was based on 15 B27 positive DZ twin pairs. The pairwise concordance rate for all 20 DZ twin pairs was 15% (95% confidence interval 3.2-37.9). $ Concordance was based on 14 B27 positive DZ twin pairs. AS CONCORDANCE IN TWINS 383 possible genetic contribution for AS. Since the pairwise concordance rate was greater in MZ twins than in B27 positive DZ twins, a role for additional genetic factors, such as HLA-Bw60, is possible, although the influence of a more similar environment in the MZ pairs cannot be excluded. The pairwise concordance rate for radiologic sacroiliitis was greater than that for definite AS. In this series, 3 MZ twin pairs were discordant for definite AS. All 3 of these pairs were well over 50 years old and will therefore probably remain discordant. Although not shown in the case of idiopathic AS, it is postulated that differences in environmental exposure between members of MZ pairs explain the discordance, despite the presence of known genetic markers of the disease (12). Interestingly, both probands of the totally discordant MZ sets had had infections that had not occurred in their unaffected twins. The findings of the present series are compatible with those of the Arthritis and Rheumatism Council (UK) Twin Survey, which was designed to avoid selection. In that study, only 1 of the 2 MZ and none of the DZ twin pairs was concordant for clinical and radiologic spondylitis (3). In a case series providing evidence for the importance of environmental factors, Eastmond and Woodrow (4) described 3 MZ pairs in which at least I member had AS and found discordance in 2 of them, and a difference in disease severity in the third. AS and RA can occur concomitantly in the same individual (13). In the present series, secondary seropositive RA occurred in as many as 4 subjects with AS. The MZ pair was completely discordant for both diseases despite the presence of HLA-B27, B40 (typing for the Bw60,61 split of B40 was not performed), and DR4. The results of this first systematic population- based study of twins with AS are compatible with the conclusion that genetic factcirs play a dominant role in the disease expression. The occurrence of clearly discordant MZ pairs indicates that environmental or stochastic events are also relevant. Close observation of MZ twin pairs who are discordant for AS may provide valuable clues on .the environmental conditions that favor disease expression (14). REFERE:NCES 1. Ivanyi P: Immunogenetics of the spondyloarthropathies. Curr Opin Rheumatol 5:436445, 1993 2. Robinson WP, van der Linden SM, Khan MA, Rentch H-U, Cats A, Russell A, Thomson G’: HLA-Bw60 increases susceptibility to ankylosing spondylitis in HLA-B27+ patients. Arthritis Rheum 32:1135-1141, 1989 3. Lawrence JS: Rheumatism in Populations. London, Heineman, 1977 4. Eastmond CJ, Woodrow JC: Discordance for ankylosing spondylitis in monozygotic twins. Ann Rheum Dis 36:360-364, 1977 5. Aho K, Koskenvuo M, Tuominen J, Kaprio J: Occurrence of rheumatoid arthritis in a nationwide series of twins. J Rheumato1 13:899-902, 1986 6. Sarna S, Kaprio J, Sistonen P, Koskenvuo M: Diagnosis of twin zygosity by mailed questionnaire. Hum Hered 28:241-254, 1978 7. Helminen P, Ehnholm C, Lokki M-L, Jeffreys A, Peltonen L: Application of DNA “fingerprinting” to paternity determinations. Lancet 1574-576, 1988 8. Bennett PH, Burch TA: Population Studies of the Rheumatic Diseases. Amsterdam, Excerpta Medica, 1968 9. Steinbrocker 0, Traeger CH, 13atterman RC: Therapeutic criteria in rheumatoid arthritis. JAMA 140:659-662, 1949 10. Amos D, Badir H, Boyle W, McQueen M, Tiilikainen A: A simple microcytotoxicity test. Transplantation 7:220-223, 1969 1 1 . Moller P: Genetics of ankylosirig spondylitis, psoriatic arthritis and Reiter’s syndrome. Clin Exp Rheumatol 5 (Suppl 1):3540, 1987 12. Gran JT, Husby G: The epidemiiology of ankylosing spondylitis. Semin Arthritis Rheum 22:319-334, 1993 13. MacGregor AJ, Fox H, Ollier WER, Snaith ML, Silman AJ: An identical twin pair discordant for rheumatoid arthritis and ankylosing spondylitis. Clin Exp Rheumatol 11:425-428, 1993 14. Vogel F, Motulsky AG: Humam Genetics: Problems and Approaches. Second edition. Berlin, Springer Verlag, 1986 15. Diem K, Lentner C: Scientific Tables. Seventh edition. Basel, Ciba-Geigy, 1970 Erratum In the list of reviewers for 1994 published in the December 1994 issue of Arthritis and Rheumatism, the name of Dr. Kaisa Granfors was inadvertently omitted. We regret the error.