Objective. To estimate the genetic influence on the occurrence of radiologic osteoarthritis (ROA) in the knees, hips, and hands and disc degeneration of the spine in the general population. Methods. A random sample of 1,583 individuals was drawn to estimate the prevalence of ROA and disc degeneration in the general population. Of 118 pro-bands with multiple affected joint sites who were derived from this sample, we were able to recruit 257 siblings. The variance of ROA and disc degeneration within sibling pairs was compared with the variance between sibling pairs. Heritability estimates for ROA in the knees, hips, and hands and for disc degeneration of the spine were calculated. OA was defined according to radiologic criteria, using the Kellgren/Lawrence grading system. Results. We observed that hand ROA and disc degeneration of the spine were statistically significantly more frequent in siblings than in the random sample, whereas the prevalence of knee and of hip ROA was similar and lower, respectively. Heritability estimates for hand ROA and disc degeneration were statistically significant, P 0.56 (95% confidence interval [95% CI] 0.34-0.76) and P 0.75 (95% CI 0.30-1.00), respectively. For knee and hip ROA, no evidence of a genetic effect in the general population was found. Finally, the heritability estimate for a score that summed the number of joints affected in the knees, hips, hands, and spine was 0.78 (95% CI 0.52-0.98). All heritability estimates were adjusted for age, sex, body mass index, and bone mineral density. Conclusion. The present study shows that in the general population, there is a strong genetic effect for hand ROA and disc degeneration of the spine. The findings on the total number of joints affected at multiple sites suggest genetic susceptibility to generalized OA. A genetic effect on osteoarthritis (OA) was initially recognized in 1941 by Stecher, who showed that Heberden's nodes of the fingers were more common in sisters of affected subjects than in the general population (1). In 1963, Kellgren et al (2), in a study of subjects derived from an outpatient clinic, reported that first-degree relatives of subjects with generalized radiologic OA (ROA) were twice as likely to have ROA than was expected in the general population. In 1996, Spector et al (3) measured ROA in the hands and knees of female twins and suggested that genetic factors might explain up to 65% of the variability in ROA of the hand and knee. This familial aggregation of hand and knee ROA was confirmed recently in 2 population-based studies (4,5). Furthermore, Felson et al (4), in a segregation analysis, found evidence of effects of a major recessive gene with a residual multifactorial component. With the exception of the study by Kellgren et al (2), previous studies concerning the role of genetic factors in the occurrence of ROA were limited to OA of the hands and knees (3-5). The contribution of genetic factors to other common forms of OA, including hip

Correspondence: Dorrett Boomsma, Department of Bi ol ogi cal Psychol ogy, Vri je Uni versi tei t, A msterdam, The Netherl ands.

Thi s study r epor ts on zygosi ty deter mi nati on i n tw i ns of chi l dhood age. Par ents r esponded to questionnaire items dealing with twin similarity in physical characteristics and frequency of mi stak i ng one tw i n for another by par ents, r el ati ves and str anger s. The accur acy of zygosi ty di agnosi s w as eval uated acr oss tw i ns aged 6, 8, and 10 and acr oss par ents. I n addi ti on, i t w as exami ned w hether the use of mul ti pl e r ater s and the use of l ongi tudi nal data l ead to an i mpr ovement of zygosi ty assi gnment. Compl ete data on zygosi ty questi ons and on geneti c mar k er s or bl ood pr ofi l es w er e avai l abl e for 618 tw i n pai r s at the age of 6 year s. The method used w as pr edi cti ve di scr i mi nant anal yses. A gr eement betw een zygosi ty assi gned by the r epl i es to the questi ons and zygosi ty deter mi ned by DNA mar k er s/bl ood typi ng w as ar ound 93%. The accur acy of assi gnment r emai ned constant acr oss age and par ents. A nal yses of data pr ovi ded by both par ents and col l ected over mul ti pl e ages di d not r esul t i n better pr edi cti on of zygosi ty. Detai l s on the di scr i mi nant functi on ar e pr ovi ded. Twin Research (2000) 3, 134-141. Keyw or ds: tw i n zygosi ty, chi l dhood, questi onnai re, revi ew, di scri mi nant anal ysi s I ntr oducti onIn 1927, Si emens 1 suggested that the di agnosi s of zygosi ty i n tw i ns can take pl ace by eval uati ng the degree of resembl ance on geneti cal l y determi ned trai ts. Devel opment of thi s method resul ted i n the frequent use of questi onnai res, often i ncl udi ng those cri teri a ori gi nal l y proposed by Si emens, for exampl e.2 Several studi es have show n that the establ i shment of zygosi ty based on mai l ed questi onnai res i s of consi derabl e accuracy, w i th around 95% correctl y cl assi fi ed compared w i th bl ood or DNA typi ng. Studi es on the di agnosi s of zygosi ty by mai l ed questi onnai res are summari sed i n A ppendi x 1.

management of CPSP as difficult, and there was a tension between their desire to help patients and the concern that keeping patients 'in the system' of secondary care might not always be helpful. Identifying ways to improve services, participants described specialist services, specifically multidisciplinary pain clinics and pain management programmes, as having a vital and specific role in helping patients with CPSP. Conclusion: Despite healthcare professionals' commitment to helping patients with CPSP, coherent referral pathways and services to manage CPSP were described as lacking, resulting in an unsatisfactory approach. By corollary, better services would provide a joined up pathway that would assist patients and healthcare professionals involved in their care. This small-scale study complements ongoing work that highlights the diversity of services for CPSP after knee replacement across the UK, in which some examples of clear pathways are apparent. Future research is needed to assess the impact of clear referral pathways and integrated service provision for patients with CPSP after total knee replacement.