This study attempted to validate an association between rs6746030 in the SCN9A gene and reporting pain at multiple body sites in four independent cohorts. No evidence of association with CWP was observed, conflicting with a previous report that the minor (A) allele was associated with increased odds of reporting pain in multiple sites
The findings of this study and the Valdes et al. study are equivocal; this may be due to a number of factors. Firstly, the criteria used to classify cases with widespread pain and controls differed between studies. The majority (76.1%) of subjects in the Valdes et al. study were recruited as part of case–control study of OA with the remainder recruited from two population based studies. Subjects were classified as having multiple regional pain using different definitions in each sub-study, however, the authors reported no between-study heterogeneity. Secondly, the prevalence of WP/CWP reported here, 8-18% was higher than in Valdes et al. (2011), 4-9%. It is possible that the case phenotype in the present study was more heterogeneous and/or includes subjects with less severe or disabling pain and rs6746030 may associate with severe/disabling widespread pain. Thirdly, in the present study, controls were subjects reporting no pain to eliminate the possibility of a false negative association occurring due to the presence of regional pain in the control group. However, in the Valdes et al. study control subjects may or may not have had pain suggesting that the association was specifically with multi-site pain rather than pain per se. Finally, the original observation was adjusted for age, sex and body mass index. Individuals with widespread pain have increased rates of other painful disorders such as irritable bowel syndrome and chronic oro-facial pain and the results from the Valdes et al. (2011) study may reflect a higher “pain-load” that was not detected in this study due to phenotype heterogeneity.
The limitations of this study were that the pain assessment and definition differed between cohorts. A joint pain assessment, which did not include chronicity, was used for the Framingham study. Consequently, cases in the Framingham cohort may represent a different group to the other cohorts; however, the majority of chronic pain reported is at joint sites and widespread pain is usually chronic. In addition, the controls in the Framingham cohort may have had some pain at non-joint sites and may not represent a truly pain-free group. The cohorts also differed in age, gender, geographic location and individual cohorts lacked power to detect an association. No data on ethnicity was available for the EPIFUND cohort, however, the majority are thought to be white British
 and only information on self-reported ancestry was available in EMAS.
CWP will resolve in some individuals and persists in others. Genetic factors may play a more important role in persistent and/or disabling CWP; although, this has yet to be demonstrated. If so, comparing individuals with persistent and/or disabling CWP to pain-free controls may be a more powerful approach to identifying susceptibility loci for chronic pain. Given that pain was assessed at a single time-point in the majority of the cohorts utilised in this study and associated disability data was not available, we were unable to test this in the current study; however, this approach could be fruitful if adopted in future studies.
Our findings also conflict with those of Reimann et al. (2010), who observed that the A allele of rs6746030 was associated with reporting higher pain scores in subjects with a variety of clinical pains and increased pain sensitivity in healthy individuals
. This may suggest that the SNP is associated with the severity of pain reported within a pain state rather than the presence of widespread pain, as tested here. However, Valdes et al. (2011) did not observe an association between rs6746030 and pain scores in knee OA in over 1000 subjects
The lack of association reported here, despite the large sample size (1071 cases and 3214 controls), suggests that this SNP is not a susceptibility marker for CWP. Other SNPs within the gene or in regulatory regions for SCN9A may be important in CWP susceptibility and may influence the effect of rs6746030 but have yet to be investigated. Environmental factors, not controlled for in the present study could also potentially mask a true effect.
In conclusion, we find no evidence of association for rs6746030 in SCN9A, previously shown to play a role in pain perception, with CWP in a meta-analysis of four population-based cohorts. However, SCN9A remains a strong candidate gene for chronic pain and is worthy of investigation in large-scale genetic association studies. Harmonisation of pain phenotypes to allow cohorts to be combined is required to facilitate such studies and to elucidate the genetic risk factors for chronic pain.