Several studies have previously reported the association of inflammatory and matrix degrading genes with LDD [9, 10, 14], and findings of the present study confirms that some of these genes influence the radiographic degeneration of the lumbar disc. The novelty of the present study is that it also highlights the influence of these variants on pain and disability at long term in patients treated with spinal fusion or cognitive intervention and exercises for CLBP.
This study observed associations of IL18RAP polymorphism rs1420100 and MMP3 polymorphism rs72520913 with severe degeneration and for rs1420100 also with more than one degenerated disc. Interestingly two other polymorphisms of IL18RAP, rs917997 and rs1420106, showed association with improvement in disability. Furthermore the MMP3 polymorphism, rs72520913, was observed to be associated with reduction in pain.
The associations between IL18RAP polymorphism rs1420100 with severity of degeneration and the number of degenerated discs are in line with results from a previous study by Videman et al. They were the first to identify an association of this SNP of IL18RAP gene with LDD . In their study of 588 monozygotic and dizygotic male twins from the Finish population, they reported a significant association with disc desiccation at both upper and lower lumbar levels, which is considered to be an earlier stage of degenerative process. They did not find an association with disc height narrowing or bulging. In contrast, we only assessed the lower lumbar segments (L4-L5 and L5-S1). These are usually more prone to severe degeneration and our association was observed in patients who had degeneration and chronic LBP and were considered for lumbar fusion. Although our findings strengthen the evidence in support of previously reported association of IL18RAP gene variant with radiographic lumbar disc degeneration, we could not rule out that these findings could be by chance. The small sample size and no correction for multiple testing, does not allow drawing firm conclusions.
IL18RAP is a subunit of IL18 receptor along with IL18R1  and is essential for IL18 signal transduction and ligand binding affinity. IL18R mediated activation of T cells and natural killer cells results in secretion of interferon gamma (IFN-γ), which induce macrophages to secrete cytokines TNF-α and IL-1, which are eventually involved in increased matrix degradation, by production of cytokines and proteases. These have been found to be secreted by the cells of degenerated and herniated discs [41–43]. This sheds light on a putative connection between inflammation and degeneration and a possible pathway for back pain.
Interestingly in the present study, two other tested IL18RAP polymorphisms, rs917997 and rs1420106, did not display association with severity of degeneration, but with reduction in disability. The reduction was less pronounced among T/T genotypes at rs917997 compared to C/C and C/T indicating that carrying even one copy of T allele was associated with reduced improvement in disability. Similar lower reduction appeared to be attributed to A allele for rs1420106 where less improvement was observed with A/A compared to the G/G and A/G genotypes. The alleles at these polymorphisms rs1420106-A and rs917997-T were in strong LD and both revealed similar association with improvement in disability. On the other hand, the alleles rs1420100-A and rs917997-T or rs1420106-A, were never carried on the same haplotype, but were in complete negative LD (D' = - 1). Therefore we speculate that different genetic variants at the same gene, i.e. IL18RAP, affect both the severity of degeneration and improvement in outcome. The polymorphisms rs1420106 and rs917997 were found to be associated with disc desiccation in the study by Videman et al, but no association of these polymorphisms was observed with severity of degeneration in the present study. The SNP rs917997 has previously been reported to be associated with differential expression of IL18RAP gene in coeliac disease patients .
We observed that a haplotype T-C-C-A-T spanning the IL1A, IL18R1 and IL18RAP genes located on chromosome 2, revealed association with improvements in pain and disability. IL-18 is a member of the IL-1 family of cytokines. These are related in terms of structure, receptor family and signal transduction pathways  and both are located on chromosome 2. Based on these similarities, we considered the IL1A rs2071375 along with IL18 receptor polymorphisms when testing the association of haplotypes.
The association of MMP3 SNP rs72520913 with severe degeneration is in agreement with previously reported associations with disc degeneration . This SNP has been reported to regulate the MMP3 gene expression where the 5A allele possesses twice the promoter activity compared to the 6A allele . A study involving elderly patients found the 5A/5A genotype to be associated with presence of osteophytes, narrowing of discs and end plate sclerosis . Another study evaluated the interaction between this polymorphism and whole body vibration (gene-environment interaction) and revealed susceptibility between the 5A allele and LDD . Together with IL1A it has been reported to be associated with type II Modic changes . In vitro studies have also highlighted the role of MMP3 in intervertebral disc degeneration through matrix degradation . Our finding supports the role of variant(s) in promoter region of MMP3 gene in LDD, and based on the fact that such polymorphisms may enhance the MMP3 expression, we speculate that they may ultimately disturb the balance of ECM of the lumbar disc. Contrary to this, the association of the same polymorphism with improvement in pain after treatment, is difficult to understand, but in line with other studies that have reported a low correlation between clinical and radiographic outcome [12, 32].
The inflammatory genes tested in the present study are interestingly also involved in the modulation of pain. They regulate the production of cytokines that may contribute to LBP by inducing growth of nerve and vascular elements [47, 48]. IL18 induced synthesis of tumor necrosis factor-alpha (TNFα) has also been proposed to play a role in discogenic pain . MMP3 and IL1A have been associated with neuropathic pain and sensitization of nocireceptors [18, 24].
Our finding of a higher yield of DNA extracted from saliva compared to blood was similar to Mitsouras et al report that DNA yield from saliva was superior to blood and buccal swabs . In accordance with our finding that saliva DNA was of high quality, a study had previously reported the saliva DNA extracted using similar kits and methods as ours, to be of high quality . Other studies have reported an equal or lower DNA yield from saliva compared to the blood [28, 52].
The main strengths of the present study include long term follow up of patients, a high genotype success rate alongside homogeneity of the sample population (all subjects were of homogenous Norwegian origin) and testing of polymorphisms from genes that have previously been reported to be involved in both degeneration and pain modulation.
Limitations were small sample size, no testing for multiple correction, dichotomy of images of degeneration into moderate and severe rather than using a continuous classification, inclusion of patients with only 1-2 levels involved rather than multi level degeneration and testing by three genetic models on the same data. Furthermore, we have only tested one or few variants from each of these inflammatory genes, thus we could have missed the involvement of genetic variation at each tested gene.
Despite these limitations, our study gives valuable replication of previous findings, support the importance of variation at inflammatory and matrix degrading genes in the degeneration process and LBP and provide hypothesis for future studies. Replication of these results in larger study populations is important for strengthening conclusions from the present study.