Heritability of spinal pain and consequences of spinal painA comprehensive genetic epidemiologic analysis using a population-based sample of 15328 twins ages 2071 years.код для вставкиСкачать
Arthritis & Rheumatism (Arthritis Care & Research) Vol. 61, No. 10, October 15, 2009, pp 1343–1351 DOI 10.1002/art.24607 © 2009, American College of Rheumatology SPECIAL ARTICLE: EPIDEMIOLOGY OF THE RHEUMATIC DISEASES Heritability of Spinal Pain and Consequences of Spinal Pain: A Comprehensive Genetic Epidemiologic Analysis Using a Population-Based Sample of 15,328 Twins Ages 20 –71 Years JAN HARTVIGSEN,1 JAN NIELSEN,2 KIRSTEN OHM KYVIK,2 RENE FEJER,3 WERNER VACH,2 IVAN IACHINE,2 AND CHARLOTTE LEBOEUF-YDE4 Objective. To assess the relative contribution of genetic and environmental factors to different deﬁnitions of spinal pain and consequences of spinal pain. Methods. The Danish Twin Registry contains detailed survey information on spinal pain and its consequences in twins ages 20 –71 years. A classic genetic epidemiologic analysis was performed in order to establish heritability for a number of phenotypes, including location of pain, radiation of pain in the extremities or chest, pain duration, and combinations of pain in >1 spinal area. Consequences included reduced physical activity, sick leave, care seeking, change of work, and disability pension. The analysis included a biometric analysis based on the effect of shared genetic and common environmental factors. Furthermore, a bivariate twin model was ﬁtted to identify genetic and environmental correlations. Results. Altogether, data on 15,328 twin individuals (44% monozygotic and 56% dizygotic) from complete twin pairs were included. Genetic susceptibility explained ⬃38% of lumbar pain, 32% of thoracic pain, and 39% of neck pain. For patterns of pain, estimates were 7% for lumbar/thoracic, 24% for lumbar/cervical, 0% for thoracic/cervical, and 35% for pain in all 3 areas. Moderate to high genetic correlations indicated a common genetic basis for many spinal pain syndromes. In general, heritability was higher for women, and only a minor age effect was seen. Conclusion. Heritability estimates for pain in different spinal regions are quite similar and there is a moderate to high genetic correlation between the phenotypes. This may indicate a common genetic basis for a high proportion of spinal pain. INTRODUCTION Spinal pain and its consequences have developed into a modern epidemic. Unfortunately, effects of efforts to preData collection was supported by the Danish Foundation for Chiropractic Research and Postgraduate Education. 1 Jan Hartvigsen, DC, PhD: Clinical Locomotion Science, University of Southern Denmark, Campusvej 55, 5230 Odense M, and Nordic Institute of Chiropractic and Clinical Biomechanics, Clinical Locomotion Science, Forskerparken 10A, 5230 Odense M, Denmark; 2Jan Nielsen, MSc, Kirsten Ohm Kyvik, MD, PhD, Werner Vach, PhD, Ivan Iachine, PhD: University of Southern Denmark, JB Winsløwsvej 9A, 5000 Odense C, Denmark; 3Rene Fejer, DC, PhD: Clinical Locomotion Science, University of Southern Denmark, Campusvej 55, 5230 Odense M, Denmark; 4Charlotte Leboeuf-Yde, DC, PhD: Back Research Center, Clinical Locomotion Science, Lindevej 5, 5750 Ringe, and University of Southern Denmark, Odense, Denmark. Address correspondence to Jan Hartvigsen, DC, PhD, Institute of Sports Science and Clinical Biomechanics, University of Southern Denmark, Campusvej 55, DK-5230 Odense M, Denmark. E-mail: firstname.lastname@example.org. Submitted for publication January 14, 2009; accepted in revised form April 3, 2009. vent and treat the various manifestations have proven to be of modest beneﬁt, at least in a public health perspective (1–3). This is probably because the underlying mechanisms behind spinal pain and its consequences are poorly understood, making it difﬁcult to tailor speciﬁc interventions on a rational basis. Recently, however, 2 promising directions have emerged. First and foremost, it is becoming increasingly clear from population studies that the various forms of spinal pain, i.e., low back pain, mid-back pain, and/or neck pain, are probably not separate and site-speciﬁc entities, but rather in many cases part of a more general musculoskeletal pain syndrome (4 – 6) and maybe even an indicator of poor general health (7,8). Of course, this line of thinking has implications for both researchers and clinicians, because pain reported in a localized spinal area may be both a result of local tissue injury and an expression of a general musculoskeletal frailty syndrome (9). Second, there is now convincing evidence that genetic susceptibility plays a signiﬁcant role in many aspects of spinal pain and pathology, including reporting of low back pain at all ages (10 –13), reporting of neck pain (14), cer1343 1344 vical and lumbar disc degeneration observed on magnetic resonance imaging (MRI) (12,15), and lumbar osteophyte formation (16). Results of a recent study using a small but population-based twin sample indicated that up to 25% of genetic inﬂuences on back pain history were due to the same genetic inﬂuences that affected disc height narrowing, the degenerative ﬁnding most associated with back symptoms in the study sample (12). Furthermore, there is now convincing evidence that disc degeneration may be primarily caused by genetic inﬂuences and not, as once believed, primarily by aging and wear and tear (17). This may in part explain why interventions aimed at modifying environmental factors such as reduced workload have not resulted in less back pain in the population. Interestingly, there is also emerging evidence that the documented association between back and neck pain and depression (18 – 20) may be primarily due to common genetic effects between the two (21). Therefore, it is tempting to speculate that the pattern of co-occurrence of musculoskeletal and other symptoms observed in population studies may to a signiﬁcant extent be genetically based. One approach to this question would be to study patterns of heritability and genetic correlation between different phenotypes of spinal pain in the same population. Such research requires very large samples of twins, and the Danish Twin Registry, the world’s largest and oldest twin registry with information on more than 120 birth cohorts of twins (22), is well suited for such a purpose. In 2002, an extensive survey of all Danish twins born between 1931 and 1982 including information on pain in the 3 spinal areas (neck, mid back, and low back) and consequences of such pain was performed. Details of the prevalence and patterns of ﬁndings in an epidemiologic perspective have been reported elsewhere (4). We subsequently performed a comprehensive genetic epidemiologic analysis in order to answer the following research questions: 1) What is the relative genetic and environmental contribution to different phenotypes, i.e., different deﬁnitions of spinal pain and consequences of spinal pain?, 2) Is there a common pattern in the heritability estimates for spinal pain and its consequences between the spinal regions?, and 3) Can we identify correlations between the genetic and environmental factors inﬂuencing the region-speciﬁc phenotypes? PATIENTS AND METHODS Participants. The Danish Twin Registry covers birth cohorts from 1870 to 2001 (22). It is among the largest and most comprehensive twin registries in the world, and is considered representative of the Danish general population (23). In April 2002, all twins born between 1931 and 1982 and registered in the Danish Twin Registry who had previously agreed to participate in research received a 20-page questionnaire booklet containing questions on a wide variety of health-, work-, and lifestyle-related factors. The information letter stated the purpose of the projects as focusing on the twins’ health in general. The questionnaire was followed by one reminder, which is the number of reminders allowed by the Danish Scientiﬁc Ethical Com- Hartvigsen et al mittees. The study was approved by the Regional Scientiﬁc Ethics Committee and the Danish Data Protection Agency. Zygosity of the twin pairs was determined using answers to questions regarding resemblance between twins in a pair. This method has been shown to yield correct classiﬁcation in 97% of patients within this cohort (23). Variables. Information regarding spinal pain (neck pain, mid-back pain, and low back pain) was collected using the Standardized Nordic Questionnaire. This questionnaire has been validated for both test–retest reliability and validity (24 –26). Phenotypes of pain in the different spinal regions (low back pain, mid-back pain, and neck pain) included pain in the past year and the number of days with pain in the past year (later categorized into 1–30 days and ⬎30 days out of the past year). In addition, we studied those reporting different pain patterns, i.e., low back pain and mid-back pain but no neck pain, low back pain and neck pain but no mid-back pain, mid-back pain and neck pain but no low back pain, and low back pain, mid-back pain, and neck pain, as well as the 3 patterns of occurrence of pain in one speciﬁc region alone, in order to see if heritability for speciﬁc regional pain differed from pain in more than 1 region. Finally, pain reported as radiating from the lumbar, thoracic, and cervical areas was included, described as pain in the leg(s), chest, and arm(s), respectively. All variables relate to symptoms in the past year. Phenotypes of consequences of spinal pain during the past year included 1) reduced physical activity in the past year, 2) sick leave during the past year, 3) sought care during the past year, 4) changed work or tasks at work during the past year, and 5) disability pension (including being under consideration for disability pension and already having a disability pension). It is relevant to note that there are no ﬁnancial barriers to access the Danish health care system and that sick leave and compensation are available regardless of the cause of the disease. The work-related variables were only applied to actively working individuals. Additional variables were sex and age groups (20 –35, 36 –50, and 51–71 years). Statistical analysis. The analysis was stratiﬁed by sex, therefore limiting the analysis to same-sex twin pairs. Participants were then divided into 3 age groups (20 –35, 36 –50, and 51–71 years) (14). Accordingly for each variable, heritability estimates for the different phenotypes were obtained for men and women and each of the age groups separately. Heritability estimates were obtained using a classic biometric analysis based on the assumption that the total phenotypic variance in a trait can be partitioned into components attributable to genetic and environmental effects (27). In the univariate variance component model, one assumes that the phenotypic variance can be partitioned into an additive genetic component (A), a dominant genetic component (D), a shared environmental component (C), and a nonshared environmental component (E) (27). In all previous analyses of back and neck pain data from the Comprehensive Genetic Epidemiologic Analysis of Spinal Pain Danish Twin Registry using this and other cohorts (3,11,14,28), the AE biometric submodel showed the best ﬁt to the data, and therefore we chose to ﬁt only that model. This ensured comparability of estimates inside the analysis and with other analyses. To understand the genetic and environmental contribution to the association between the 3 region-speciﬁc phenotypes, we further ﬁtted a bivariate heritability twin model (29). From this model, we obtained estimates of the correlation between genetic and environmental factors inﬂuencing the single phenotypes. A high genetic/environmental correlation indicates that the phenotype is inﬂuenced by the same genetic/environmental factors. We also report bivariate heritabilities expressing to what degree the overall association between the 2 phenotypes can be attributed to correlations of genetic or environmental factors. The Mx software package was used for all biometric modeling (30). RESULTS Of 46,818 questionnaires sent out, 34,902 (74.5%) were returned immediately or after the ﬁrst reminder and regarded as valid. In this analysis, only twin pairs where both twin individuals had provided valid answers were included, and because analyses were performed separately for men and women, opposite-sex dizygotic pairs were excluded, as were twins without zygosity determination, triplets, and quadruplets. This resulted in a ﬁnal sample of 15,328 twin individuals (6,412 men and 8,916 women) comprising 7,664 complete twin pairs, of which 3,402 were monozygotic and 4,262 were dizygotic. Further details regarding the sample and prevalence of the various phenotypes have been reported in detail elsewhere (4). The prevalences of phenotypes included in this analysis for monozygotic and dizygotic twins are shown in Table 1. No statistical differences beyond chance were found between the 2 zygosities. Heritability estimates based in the univariate AE model estimating the overall genetic and environmental effect on the different phenotypes are shown in Figures 1, 2, and 3. For spinal pain in the 3 spinal regions, genetic susceptibility explained 38% (95% conﬁdence interval [95% CI] 33– 42) of low back pain, 32% (95% CI 26 –39) of mid-back pain, and 39% (95% CI 34 – 43) of neck pain (Figure 1). If this was deﬁned as pain in one region only, i.e., low back pain, mid-back pain, or neck pain alone with no pain reported in the other regions, the estimates were 21% (95% CI 14 –27) for low back pain only, 15% (95% CI 0 – 42) for mid-back pain only, and 18% (95% CI 9 –28) for neck pain only. For radiating pain in the leg, chest, and arms, genetic factors explained 31% (95% CI 26 –37), 40% (95% CI 30 –50), and 35% (95% CI 29 – 41), respectively. For practically all of the phenotypes, the heritability estimates for women were higher than for men, although this was not statistically signiﬁcant. When these phenotypes were broken down into age strata (Figure 1), no clear age effect was seen for low back pain or sciatica; for mid-back pain and pain radiating into the chest this was also the case, although low prevalences in the age strata resulted in 1345 wide conﬁdence intervals. For neck pain and mid-back pain and for pain radiating into the upper extremity or chest, there was a tendency of lower heritability with age for men and stable or slightly increasing estimates for women. For the different pain patterns (Figure 2), the overall heritability estimates were 7% (95% CI 7–32) for low back pain and mid-back pain but no neck pain, 24% (95% CI 16 –31) for low back pain and neck pain but no mid-back pain, 0% (95% CI 0 –34) for mid-back pain and neck pain but no low back pain, and 35% (95% CI 27– 43) for low back pain, mid-back pain, and neck pain. In spite of wide conﬁdence intervals, there was a clear tendency for higher heritability in women for all of the phenotypes (Figure 2). Furthermore, women showed increasing heritability with age, whereas the opposite was the case for men. For consequences of spinal pain, the results are shown in Figure 3. Overall, heritability estimates for consequences are quite similar between the spinal regions. The highest heritability was found for the most severe consequence, early pension. There are only minor differences between men and women, although estimates for early pension were consistently higher for men. Genetic, environmental, and phenotypic correlations and bivariate heritability estimates are shown in Table 2. For both men and women, genetic correlations between low back pain, mid-back pain, and neck pain were moderate to high regardless of age, and the bivariate heritability estimates were roughly equal to the univariate estimates, indicating that genetic factors are as important in explaining the pain pattern phenotypes as they are in explaining the region-speciﬁc phenotypes. DISCUSSION In a comprehensive genetic epidemiologic analysis using survey data from a large population-based twin cohort, we have shown that spinal pain and its consequences are to a signiﬁcant extent determined by genetic factors for both men and women throughout adulthood. Furthermore, the size of the genetic contribution to spinal pain is remarkably similar between the 3 spinal regions. We even found strong genetic correlations between all 3 regions for all of the age groups, indicating that to a large extent, the same genes are responsible for the pain regardless of the location. The fact that the bivariate heritability estimates were of the same magnitude as the univariate estimates further supports the conclusion that genetic factors are important in explaining and understanding the interplay between pain in the various regions of the spine. We found that genetic inﬂuences on pain patterns involving 2 regions were consistently lower than for pain involving any speciﬁc region alone. This indicates that genetic predisposition to pain in more than one spinal region is probably not very common. In a recently published PhD thesis, Nyman concluded quite the opposite, namely that heritability estimates nearly doubled (from ⬃30% to 60%) when low back pain and neck pain were considered together instead of separately (31). Although a population-based sample of twins was used in this Swed- 1346 Hartvigsen et al Table 1. Prevalence of included phenotypes in 15,328 Danish twins ages 20 –71 years* Age 20–35 years Phenotype Men, no. Women, no. Total, no. LBP in the past 12 months Men Any pain 1–30 days ⬎30 days Sciatica Women Any pain 1–30 days ⬎30 days Sciatica MBP in the past 12 months Men Any pain 1–30 days ⬎30 days Sciatica Women Any pain 1–30 days ⬎30 days Sciatica NP in the past 12 months Men Any pain 1–30 days ⬎30 days Sciatica Women Any pain 1–30 days ⬎30 days Sciatica Pain in ⬎1 area Men LBP ⫹ MBP LBP ⫹ NP MBP ⫹ NP LBP ⫹ MBP ⫹ NP Women LBP ⫹ MBP LBP ⫹ NP MBP ⫹ NP LBP ⫹ MBP ⫹ NP Consequences of LBP Men Reduced physical activity Sought care Change work duty Sick leave Early pension Women Reduced physical activity Sought care Change work duty Sick leave Early pension MZ DZ 1,453 2,092 3,545 1,587 1,914 3,501 Age 36–50 years P† MZ DZ 1,298 1,550 2,848 2,037 2,299 4,336 Age 51–71 years P† MZ DZ 1,220 1,380 2,600 2,461 2,562 5,023 P† 36.0 29.1 7.5 10.2 39.0 30.9 8.5 13.3 0.108 0.285 0.311 0.011 49.3 38.2 11.6 24.0 45.6 35.7 10.5 23.6 0.048 0.162 0.363 0.766 38.8 28.6 10.5 21.3 36.8 24.6 11.6 22.2 0.258 0.014 0.379 0.547 41.7 30.9 11.1 15.3 44.0 32.1 12.5 16.0 0.175 0.462 0.191 0.594 50.3 35.9 15.3 27.2 51.2 35.2 17.1 26.9 0.637 0.670 0.166 0.816 41.4 23.6 16.1 27.5 40.8 22.5 16.3 26.5 0.770 0.468 0.888 0.565 14.9 12.4 2.7 4.4 15.3 12.3 3.5 4.6 0.785 0.939 0.246 0.806 12.1 9.8 3.7 4.5 10.5 9.1 2.2 3.4 0.159 0.477 0.024 0.113 6.6 4.6 2.3 3.7 7.0 4.1 2.7 4.7 0.683 0.528 0.492 0.172 15.8 11.6 4.6 4.4 15.7 11.4 4.8 4.5 0.994 0.866 0.737 0.816 17.8 13.6 5.6 7.3 17.9 13.1 6.5 6.8 0.926 0.653 0.296 0.529 10.4 5.3 6.2 5.7 12.1 6.1 6.0 7.1 0.126 0.388 0.839 0.129 23.1 18.6 4.8 6.9 25.6 20.6 5.9 7.4 0.126 0.194 0.229 0.614 31.3 24.3 9.1 12.8 27.9 22.4 7.4 11.9 0.039 0.246 0.112 0.451 21.0 14.9 8.1 11.9 22.4 14.9 8.4 12.2 0.350 0.959 0.825 0.837 35.2 24.5 11.4 12.6 37.1 25.6 12.4 13.5 0.242 0.434 0.362 0.414 47.1 32.3 16.3 25.2 46.4 30.9 17.7 25.0 0.713 0.387 0.316 0.903 33.9 17.4 16.0 23.4 34.0 17.7 15.6 22.4 0.955 0.875 0.788 0.545 2.7 7.4 1.7 6.3 2.7 8.7 1.4 6.9 0.966 0.189 0.650 0.520 1.8 11.8 0.8 7.7 1.8 10.6 0.8 5.4 0.992 0.277 0.840 0.008 1.1 7.5 0.6 2.9 1.0 7.4 0.2 3.5 0.625 0.945 0.123 0.286 1.8 11.0 1.4 9.9 1.9 13.1 1.2 9.6 0.787 0.047 0.518 0.699 1.3 15.3 1.2 13.0 1.9 15.1 1.3 11.8 0.164 0.871 0.829 0.316 0.7 8.4 0.6 5.9 0.7 7.7 0.8 6.8 0.854 0.441 0.480 0.256 12.1 12.2 5.1 8.6 0.4 12.0 13.6 5.8 11.1 0.7 0.905 0.263 0.454 0.066 0.307 21.2 18.5 8.7 12.4 1.5 18.9 17.6 8.1 12.4 1.3 0.120 0.503 0.537 0.955 0.746 17.2 14.4 6.2 7.7 2.4 17.5 15.0 7.6 8.6 3.9 0.839 0.669 0.122 0.436 0.027 11.4 13.1 6.3 6.8 0.6 12.6 14.5 6.2 8.4 0.9 0.283 0.202 0.820 0.143 0.218 18.7 20.6 9.0 9.4 1.9 19.6 20.6 10.7 11.1 2.8 0.514 0.999 0.101 0.135 0.086 18.3 19.8 8.3 11.4 4.9 19.1 0.536 20.0 0.881 8.0 0.683 10.0 0.325 5.1 0.779 (continued) Comprehensive Genetic Epidemiologic Analysis of Spinal Pain 1347 Table 1. (Cont’d) Age 20–35 years Phenotype Consequences of MBP Men Reduced physical activity Sought care Change work duty Sick leave Early pension Women Reduced physical activity Sought care Change work duty Sick leave Early pension Consequences of NP Men Reduced physical activity Sought care Change work duty Sick leave Early pension Women Reduced physical activity Sought care Change work duty Sick leave Early pension Age 36–50 years Age 51–71 years MZ DZ P† MZ DZ P† MZ DZ P† 3.9 4.6 1.7 3.9 0.1 5.0 4.9 1.9 3.2 0.3 0.182 0.699 0.620 0.402 0.322 4.1 4.3 1.7 2.2 0.2 2.7 4.4 1.3 1.7 0.2 0.025 0.871 0.402 0.344 0.931 2.2 2.9 0.5 1.3 0.8 3.0 3.1 1.2 1.3 1.0 0.187 0.665 0.042 0.894 0.591 3.2 5.5 1.9 1.6 0.2 3.9 6.1 2.3 2.5 0.5 0.218 0.437 0.394 0.125 0.222 4.5 7.9 2.9 2.2 0.6 4.4 7.3 2.7 2.2 1.0 0.934 0.507 0.799 0.954 0.298 4.5 5.6 2.0 2.7 1.4 5.2 6.0 2.2 3.2 1.7 0.411 0.670 0.644 0.497 0.529 5.5 8.0 1.6 4.2 0.1 5.3 8.1 2.3 4.5 0.3 0.858 0.888 0.187 0.792 0.485 8.4 10.5 3.5 3.8 0.7 6.7 9.8 2.6 3.8 0.5 0.067 0.538 0.163 0.953 0.582 5.3 7.6 1.9 3.3 1.1 7.6 9.3 2.5 2.6 1.7 0.009 0.089 0.287 0.348 0.158 6.9 12.7 3.3 5.0 0.4 8.7 14.2 3.5 7.0 0.7 0.036 0.206 0.721 0.055 0.210 14.3 20.7 6.6 8.3 1.3 12.8 21.5 6.1 7.1 1.8 0.187 0.597 0.571 0.243 0.269 12.8 16.2 5.2 5.0 3.8 11.7 17.3 5.4 6.4 3.5 0.307 0.402 0.801 0.211 0.595 * Prevalences are the percentage for all categories for monozygotic (MZ) and dizygotic (DZ) twins separately unless otherwise indicated. LBP ⫽ low back pain; MBP ⫽ mid-back pain; NP ⫽ neck pain. † Test for equality. Figure 1. Relative inﬂuences of genetic and environmental components in liability to low back pain, mid-back pain, and neck pain during the past year for Danish men and women ages 20 –71 years. 1348 Hartvigsen et al Figure 2. Relative inﬂuences of genetic and environmental components in liability to pain in multiple spinal areas during the past year for Danish men and women ages 20 –71 years. LBP ⫽ low back pain; MBP ⫽ mid-back pain; NP ⫽ neck pain. ish study, participation rates were lower and the deﬁnitions of the phenotypes were different from ours. It is unclear whether these factors explain this apparent paradoxical ﬁnding, but it emphasizes that heritability estimates are likely to differ for different deﬁnitions of spinal pain. When considering consequences of spinal pain, the more severe and costly consequences such as early pension demonstrate the highest heritability. However, early pension is not solely related to structural pathology or even to pain, but reﬂects a complex phenotype in the border area between physiology, psychology, and sociology, all 3 of which have shown strong genetic inﬂuences (29,32). In other words, new and even more comprehensive analyses including phenotype deﬁnitions comprising extensive spinal pain information (such as in this study) combined with clinical information (12,33) and psychosocial information (21) are needed to unravel the genetic basis for complex biopsychosocial phenomena like spinal pain. In general, regardless of subgroup, heritability estimates for spinal pain in women were higher than for men, although most conﬁdence intervals were overlapping. There have been conﬂicting conclusions from previous twin studies regarding sex differences in heritability; however, overlapping conﬁdence intervals between the sex-speciﬁc estimates and different deﬁnitions of back and neck pain, i.e., different phenotypes, have hampered interpretability (11–14,28). Our results raise the question of whether this consistent difference is due to a sex-speciﬁc genetic inﬂuence. Fejer et al examined this heterogeneity in relation to lifetime prevalence of neck pain using so-called sex-limitation models, and found that sex differences in the heritability of neck pain were not due to any sex-speciﬁc genetic inﬂuence (34). Rather, they (also) proposed that neck pain (and probably spinal pain in general) is a com- Figure 3. Relative inﬂuences of genetic and environmental components in liability to consequences of spinal pain during the past year for Danish men and women ages 20 –71 years. Comprehensive Genetic Epidemiologic Analysis of Spinal Pain 1349 Table 2. Genetic, environmental, and phenotypic correlations and bivariate heritability between pain in different spinal areas during the past year for 15,328 Danish twins ages 20 –71 years* Phenotype LBP ⫹ MBP Men, years of age 20–35 36–50 51–71 Women, years of age 20–35 36–50 51–71 LBP ⫹ NP Men, years of age 20–35 36–50 51–71 Women, years of age 20–35 36–50 51–71 MBP ⫹ NP Men, years of age 20–35 36–50 51–71 Women, years of age 20–35 36–50 51–71 Genetic correlation (95% CI) Environmental correlation (95% CI) Phenotypic correlation (95% CI) Bivariate heritability (95% CI) 0.54 (0.25–0.85) 1.00 (0.41–1.00) 1.00 (0.30–1.00) 0.41 (0.26–0.55) 0.37 (0.24–0.50) 0.49 (0.34–0.64) 0.45 (0.39–0.51) 0.46 (0.40–0.52) 0.54 (0.48–0.60) 0.39 (0.16–0.62) 0.36 (0.11–0.59) 0.30 (0.05–0.55) 0.74 (0.57–0.93) 0.79 (0.55–1.00) 0.66 (0.45–0.89) 0.50 (0.38–0.61) 0.52 (0.39–0.64) 0.54 (0.39–0.68) 0.59 (0.55–0.63) 0.61 (0.56–0.65) 0.59 (0.54–0.64) 0.47 (0.34–0.60) 0.43 (0.28–0.58) 0.45 (0.28–0.62) 0.62 (0.40–0.84) 0.86 (0.51–1.00) 0.54 (0.14–0.99) 0.47 (0.34–0.59) 0.35 (0.23–0.47) 0.60 (0.48–0.70) 0.52 (0.47–0.57) 0.47 (0.42–0.51) 0.57 (0.53–0.61) 0.42 (0.24–0.59) 0.44 (0.24–0.63) 0.24 (0.04–0.42) 0.75 (0.61–0.89) 0.64 (0.47–0.80) 0.74 (0.58–0.90) 0.46 (0.36–0.55) 0.49 (0.39–0.59) 0.53 (0.42–0.63) 0.57 (0.53–0.61) 0.55 (0.51–0.58) 0.61 (0.57–0.65) 0.51 (0.39–0.63) 0.45 (0.37–0.56) 0.47 (0.34–0.60) 0.69 (0.43–0.98) 0.76 (0.00–1.00) 1.00 (0.34–1.00) 0.52 (0.37–0.65) 0.54 (0.42–0.67) 0.53 (0.41–0.65) 0.58 (0.52–0.63) 0.54 (0.49–0.60) 0.60 (0.54–0.66) 0.41 (0.21–0.59) 0.11 (0.00–0.34) 0.26 (0.02–0.50) 0.73 (0.56–0.91) 0.72 (0.49–1.00) 0.84 (0.65–1.00) 0.54 (0.42–0.64) 0.57 (0.44–0.69) 0.58 (0.43–0.71) 0.61 (0.57–0.65) 0.62 (0.57–0.66) 0.68 (0.64–0.73) 0.46 (0.33–0.59) 0.39 (0.24–0.54) 0.50 (0.35–0.64) * 95% CI ⫽ 95% conﬁdence interval; LBP ⫽ low back pain; MBP ⫽ mid-back pain; NP ⫽ neck pain. plex phenomenon with great psychological and social inﬂuences, which in turn has been shown to inﬂuence genetic liability to neck pain (33,34). We found a negligible overall effect of age on the various heritability estimates. Previously, conﬂicting results regarding the inﬂuence of age have been shown, and using a different cohort of the Danish Twin Registry (The Longitudinal Study of Aging Danish Twins) and different deﬁnitions of back and neck pain, we found a strong environmental (and thus small genetic) inﬂuence in back and neck pain in persons age ⬎70 years (13,28). Therefore, an association between age and heritability estimates is not apparent in comparisons between studies (12), although Hestbaek et al found that common environment was most important in relation to back pain in very young patients (11). Given that physical and psychosocial exposures suspected of being risk factors for spine pain vary widely over a lifetime, this lack of association may indicate a substantial interaction between genes and environmental exposures. Therefore, the value of estimating heritability from twins is in establishing the likely success of speciﬁc strategies to detect the action of individual genes under the inﬂuence of a range of environmental factors (33). Several twin studies have dealt with the relationship between the heritability of both cervical and lumbar disc degeneration and back and neck pain (15,17,33). Collectively, an important genetic inﬂuence on degeneration has been found, explaining as much as 75% of the underlying disposition to disc degeneration (15), and Battié et al recently concluded that a substantial minority (23%) of the genetic inﬂuence on pain was due to the same genetic inﬂuences affecting disc degeneration (12). In other words, disc degeneration is one pathway through which genes may inﬂuence back pain, and rather than aging and wear and tear, genetic inﬂuences appear to be responsible for the larger part of disc degeneration (17). We did not have access to MRI and our results are based entirely on survey data. Therefore, we cannot investigate possible pathways between imaging and pain. Nevertheless, as attractive as imaging and hard data may seem, these are not necessarily the best predictors of prognosis in patients with back pain (35), whereas patient reporting has been found to be of substantial value (36,37). We therefore believe that our phenotypes are good, clinically relevant, and easy to relate to, for example, primary care clinicians. It is important, however, to remember that heritability estimates are dependent on both phenotype time and place. Therefore, the estimates presented here cannot be directly applied to other deﬁnitions of spinal pain, to future generations, or to persons in other cultures (38). Heritability estimates based on genetic epidemiologic analyses are vulnerable to a number of factors and should be interpreted with caution. Our estimates are most likely too conservative because nondifferential misclassiﬁcation, 1350 Hartvigsen et al which is likely to happen in large questionnaire-based surveys, will tend to lower the heritability estimates (39). Furthermore, dissimilar twin pairs, which are most often dizygotic, are less likely to participate in twin studies, resulting in increased dizygotic correlations and therefore again conservative heritability estimates (40). On the other hand, if disease concordant pairs were more likely to participate than disease discordant pairs, heritability would be biased upward (39). However, this is less of a problem in large population-based samples such as this, but more likely to bias studies based on volunteers responding to media campaigns or studies based on twins who are seeking care for a particular condition. The major strength of the current analysis is the large and representative population-based sample (22). However, the results cannot be applied directly to the clinical setting, and a comprehensive analysis involving several subgroups with varying prevalence many times results in thin strata and consequently imprecise estimates, even in conditions as common as back and neck pain. Furthermore, we lack analyses including information on genetic and environmental factors in spinal pain, comorbidities, clinical tests, psychosocial factors, and risk factors from life history. Including such information would allow us to paint a comprehensive picture of the interplay between genetic and environmental factors and to assess the genetic correlations among all of these factors. Such an approach would have the potential to correlate candidate genes to clusters of relevant variables showing genetic correlations. AUTHOR CONTRIBUTIONS All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the ﬁnal version to be published. Dr. Hartvigsen had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study conception and design. Hartvigsen, Nielsen, Kyvik, Fejer, Vach, Iachine, Leboeuf-Yde. Acquisition of data. Hartvigsen, Kyvik, Fejer, Iachine, LeboeufYde. Analysis and interpretation of data. Hartvigsen, Nielsen, Kyvik, Fejer, Vach, Iachine, Leboeuf-Yde. ACKNOWLEDGMENT We gratefully acknowledge Professor Tom Bendix, DrMedSci, for reading and commenting on manuscript drafts. REFERENCES 1. Deyo R. Treatments for back pain: can we get past trivial effects? [letter]. Ann Intern Med 2004;141:957– 8. 2. Croft P. The question is not “have you got it”? But “how much have you got”? [letter]. Pain 2008;141:6 –7. 3. Haldeman S, Carroll L, Cassidy JD, Schubert J, Nygren A. The Bone and Joint Decade 2000-2010 Task Force on Neck Pain and Its Associated Disorders: executive summary. Spine 2008; 33:S5–7. 4. Leboeuf-Yde C, Nielsen J, Kyvik KO, Fejer R, Hartvigsen J. Pain in the lumbar, thoracic or cervical regions: do age and gender matter? A population-based study of 34,902 Danish twins 20-71 years of age. BMC Musculoskelet Disord 2009;10: 39. 5. Croft P, Dunn KM, von Korff M. Chronic pain syndromes: you can’t have one without another [letter]. Pain 2007;131:237– 8. 6. Aggarwal VR, McBeth J, Zakrzewska JM, Lunt M, Macfarlane GJ. The epidemiology of chronic pain syndromes that are frequently unexplained: do they have common associated factors? Int J Epidemiol 2006;35:468 –76. 7. Hartvigsen J, Frederiksen H, Christensen K. Back and neck pain exhibit many common features in old age: a populationbased study of 4,468 Danish twins aged 70-102. Spine 2003; 29:576 – 80. 8. Hestbaek L, Leboeuf-Yde C, Kyvik KO. Is comorbidity in adolescence a predictor of adult low back pain? A prospective study of a young population. BMC Musculoskelet Disord 2006;7:29. 9. Natvig B, Bruusgaard D, Eriksen W. Localized low back pain and low back pain as part of widespread musculoskeletal pain: two different disorders? J Rehabil Med 2001;33:21–5. 10. El-Metwally A, Mikkelsson M, Stahl M, Macfarlane GJ, Jones GT, Pulkkinen L, et al. Genetic and environmental inﬂuences on non-speciﬁc low back pain in children: a twin study. Eur Spine J 2008;17:502– 8. 11. Hestbaek L, Iachine IA, Leboeuf-Yde C, Kyvik KO, Manniche C. Heredity of low back pain in a young population: a classical twin study. Twin Res 2004;7:16 –26. 12. Battie MC, Videman T, Levalathi E, Gill K, Kaprio J. Heritability of low back pain and the role of disc degeneration. Pain 2007;131:272– 80. 13. Hartvigsen J, Christensen K, Frederiksen H, Pedersen HC. Genetic and environmental contributions to back pain in old age: a study of 2,108 Danish twins aged 70 and older. Spine 2004;29:897–902. 14. Fejer R, Hartvigsen J, Kyvik KO. Heritability of neck pain: a population-based study of 33,794 Danish twins. Rheumatology (Oxford) 2005;45:589 –94. 15. Sambrook PN, MacGregor AJ, Spector TD. Genetic inﬂuences on cervical and lumbar disc degeneration: a magnetic resonance imaging study in twins. Arthritis Rheum 1999;42:366 – 72. 16. Sakai Y, Matsuyama Y, Hasegawa Y, Yoshihara H, Nakamura H, Katayama Y, et al. Association of gene polymorphisms with intervertebral disc degeneration and vertebral osteophyte formation. Spine 2007;32:1279 – 86. 17. Battie MC, Videman T, Kaprio J, Gibbons LE, Gill K, Manninen H, et al. The Twin Spine Study: contributions to a changing view of disc degeneration. Spine J 2009;9:47–59. 18. Leino P, Magni G. Depressive and distress symptoms as predictors of low back pain, neck-shoulder pain, and other musculoskeletal morbidity: a 10-year follow-up of metal industry employees. Pain 1993;53:89 –91. 19. Caroll LJ, Cassidy JD, Cote P. Depression as a risk factor for onset of an episode of troublesome neck and low back pain. Pain 2004;107:134 –9. 20. Hartvigsen J, Frederiksen H, Christensen K. Physical and mental function and incident LBP in seniors: a populationbased two year prospective study of 1,387 Danish twins aged 70-100 years. Spine 2005;31:1628 –32. 21. Reichborn-Kjennerud T, Stoltenberg C, Tambs K, Roysamb E, Kringlen E, Torgersen S, et al. Back-neck pain and symptoms of anxiety and depression: a population-based twin study. Physiol Med 2002;32:1009 –20. 22. Skytthe A, Kyvik K, Holm NV, Vaupel JW, Christensen K. The Danish Twin Registry: 127 birth cohorts of twins. Twin Res 2002;5:352–7. 23. Kyvik KO, Green A, Beck-Nielsen H. The new Danish Twin Register: establishment and analysis of twinning rates. Int J Epidemiol 1995;24:589 –96. 24. Kuorinka I, Jonsson B, Kilbom A. Standardized Nordic Questionnaire for the analysis of musculoskeletal symptoms. Appl Ergon 1987;18:233–7. 25. Palmer K, Smith G, Kellingray S, Cooper C. Repeatability and validity of an upper limb and neck discomfort questionnaire: the utility of the standardized Nordic questionnaire. Occup Med (Lond) 1999;49:171–5. 26. Franzblau A, Salerno DF, Armstrong TJ, Werner RA. Test- Comprehensive Genetic Epidemiologic Analysis of Spinal Pain 27. 28. 29. 30. 31. 32. 33. retest reliability of an upper extremity discomfort questionnaire in an industrial population. Scand J Work Environ Health 1997;23:299 –307. Neale MC, Cardon LR. Methodology for genetic studies of twins and families. Dordrecht: Kluwer Academic; 1992. Hartvigsen J, Petersen HC, Frederiksen H, Christensen K. Small effect of genetic factors on neck pain in old age: a study of 2,108 Danish twins 70 years of age and older [published erratum appears in Spine 2005;30:710]. Spine 2005;30:206 – 8. Plomin R, DeFries JC. Multivariate behavioral genetics and development: twin studies. Prog Clin Biol Res 1981;69:25–33. Neale MC, Boker SM, Xie G, Maes H. Mx: statistical modelling. Richmond (VA): Virginia Commonwealth University; 2002. Nyman T. Low back and neck shoulder pain: work and heritability [dissertation]. Stockholm (Sweden): Stockholm University; 2008. Viding E, Jones AP, Frick PJ, Mofﬁtt TE, Plomin R. Heritability of antisocial behaviour at 9: do callous-unemotional traits matter? Dev Sci 2008;11:17–22. MacGregor A, Andrew T, Sambrook PN, Spector TD. Structural, psychological, and genetic inﬂuences on low back and neck pain: a study of adult female twins. Arthritis Rheum 2004;51:160 –7. 1351 34. Fejer R, Hartvigsen J, Kyvik KO. Sex differences in heritability of neck pain. Twin Res Hum Genet 2005;9:198 –204. 35. Carragee E, Alamin T, Cheng I, Franklin T, van den Haak E, Hurwitz E. Are ﬁrst-time episodes of serious LBP associated with new MRI ﬁndings? Spine J 2006;6:624 –35. 36. Axen I, Jones JJ, Rosenbaum A, Lovgren PW, Halasz L, Larsen K, et al. The Nordic Back Pain Subpopulation Program: validation and improvement of a predictive model for treatment outcome in patients with low back pain receiving chiropractic treatment. J Manipulative Physiol Ther 2005;28:381–5. 37. Mallen CD, Peat G, Thomas E, Dunn KM, Croft PR. Prognostic factors for musculoskeletal pain in primary care: a systematic review. Br J Gen Pract 2007;57:655– 61. 38. Sneider H. Path analysis of age-related disease traits. In: Spector TD, Sneider H, MacGregor AJ, editors. Advances in twin and sib-pair analysis. 1st ed. London: Greenwich Medical Media; 2000. p. 120 –9. 39. Strachan DP. Inﬂuence of design on the outcome of twin studies. In: Spector TD, Sneider H, MacGregor AJ, editors. Advances in twin and sib-pair analysis. 1st ed. London: Greenwich Medical Media; 2000. p. 12–21. 40. Lykken DT, McGue M, Tellegen A. A recruitment bias in twin research: the rule of two-thirds reconsidered. Behav Genetic 1987;17:343– 62.