Cervical spondylosis, as well as ossification of the posterior longitudinal ligament (OPLL), represent 2 of the most common causes of cervical spinal cord dysfunction internationally.1,2 Although spondylosis resulting in some degree of attendant spinal canal compromise is a ubiquitous finding with increased age, only a fraction of individuals develop spinal cord compression and symptomatic cervical spondylotic myelopathy (CSM).3,4 In contrast to age-related spondylosis, OPLL is an abnormal finding most prevalent in the Asian population, characterized by pathological ectopic bone formation in the posterior longitudinal ligament.5 OPLL has been shown to lead to myelopathy in approximately 17% of those affected.6 Although CSM and OPLL are distinct pathological entities, there is general consensus that both are multifactorial conditions representing a complex interaction of underlying genetic and environmental factors.
With respect to genetics, achieving an adept understanding of the heritable foundations of CSM and OPLL would not only aid in our understanding of how these conditions are passed between generations, but would also provide insight into the biological pathways that lead to their development and progression. Such investigation could lead to the identification of biomarkers predicting disease progression and/or surgical treatment response or, alternatively, to the development of novel therapies targeting specific pathological mechanisms.
We, therefore, sought to review the existing evidence critically relating to the genetics of CSM and OPLL. This was performed through the completion of a systematic review of the literature structured around 3 key questions (KQs):
- What is the evidence supporting a heritable predisposition for CSM and OPLL among relatives of affected individuals?
- What specific genetic polymorphisms have been associated with CSM and/or OPLL?
- What is the evidence supporting a genetic basis for predicting postoperative outcomes for patients with CSM and OPLL treated surgically?
MATERIAL AND METHODS
Electronic Literature Search
We conducted a systematic search in Medline (using PubMed) for literature published between 1980 and November 7, 2012. In addition, we searched 2 online association databases for potential publications: the Gene Association Database7 and Human Genome Epidemiology Network (HuGENet) Web site.8 The search results were limited to human studies published in the English language. Reference lists of key articles were also systematically checked to identify additional eligible articles. Articles concerned with ossification of the anterior longitudinal ligament or those primarily concerned with thoracic or lumbar OPLL were excluded. The full search strategy is included in the Supplemental Digital Content, available at http://links.lww.com/BRS/A828. For KQ 1, we included population-based family studies. For KQ 2, we included gene association studies evaluating mutations, single nucleotide polymorphisms (SNPs), and haplotypes. Gene linkage studies, studies conducted in children, or studies looking at myelopathy secondary to trauma, infection, or tumor were excluded. For KQ 3, studies were excluded if they were about patients treated without surgery (Table 1). Articles were screened for inclusion by 2 individuals who resolved disagreements through discussion.
From the included articles addressing the heritability of CSM or OPLL, the following data were extracted: study design, inclusion criteria, patient demographics, results of various methods for testing heritability, and segregation or risk ratios. From the included articles addressing gene association, the following data were extracted: study design, patient demographics, gene, SNPs and/or haplotype and alleles under investigation, and corresponding odds ratios (OR) and/or frequencies. In addition, outcomes after surgery were recorded for KQ 3.
Study Quality and Overall Strength of Body of Literature
For KQ 1 assessing the evidence for heritability from family studies, we assigned a class of evidence rating to each article independently by 2 reviewers (J.R.D., E.D.B.), using criteria set by The Journal of Bone & Joint Surgery9 for prognostic studies and modified to delineate criteria associated with methodological quality and risk of bias on the basis of recommendations made by the Agency for Healthcare Research and Quality (AHRQ).10,11 The appraisal system used in this article accounts for features of methodological quality and important sources of bias by combining epidemiologic principles with characteristics of study design to determine the class of evidence and is consistent with those used in previous focus issues.12
After individual article evaluation, the strength of the overall body of evidence with respect to each outcome was determined for KQ 1 on the basis of precepts outlined by the Grading of Recommendation Assessment, Development and Evaluation (GRADE) Working Group13,14 and recommendations made by the AHRQ.10,11 Qualitative analysis is performed considering AHRQ-required and additional domains.15 The initial strength of the overall body of evidence was considered high if the majority of the studies were class I or II and low if the majority of the studies were class III or IV. Criteria for downgrading the evidence 1 or 2 levels included (1) inconsistency of results, (2) indirectness of evidence, (3) imprecision of the effect estimates (e.g., wide confidence intervals [CIs]), or (4) non-a priori statement of subgroup analyses. Alternatively, the body of evidence could be upgraded 1 or 2 levels on the basis of the following factors: (1) large magnitude of effect or (2) dose-response gradient.
For KQs 2 and 3 (gene association studies), we assessed risk of bias in each article against criteria outlined in a 2009 article by Attia et al16 for genetic association studies. Studies were appraised by the following 5 standards.1 The disease phenotype should be properly defined and accurately recorded by one blind to the genetic information. For CSM, we looked for a description of clinical signs and symptoms of myelopathy confirmed by imaging evidence of spondylotic cord compression. For OPLL, we sought a description along the lines of “radiological presence of heterotopic bone formation in the posterior longitudinal ligament” in the methods or if OPLL was defined in the introduction and the methods indicated that experts using radiographical information performed diagnoses.2 Authors must have included similar disease and control groups, especially in the categories of age, sex, weight, ancestry, or metabolic conditions. We looked to see whether 2 or more of these demographic categories were controlled either by excluding those patients in the study design or by adjusting results using multivariate analyses.3 To assess measurement accuracy and potential measurement bias, we looked to see whether the authors indicated that bias in measuring genetic information was minimized by describing uniform sample collection for disease and control groups, genotyping method used, implementing quality checks, establishing rules to indicate valid genotyping results, or acknowledging missing data.4 We sought statements that the authors observed Hardy-Weinberg equilibrium in the population.5 Finally, we looked for statistical correction for multiple comparisons if 3 or more SNPs or haplotypes were examined.
We assessed the overall strength of evidence for KQs 2 and 3 using the following 3 domains: (1) the amount of evidence, (2) replication, and (3) protection from bias as suggested by The HuGENet Working Group in the Venice Interim Guidelines.17 The HuGENet scheme assigns an ordered score (A, B, or C) to each domain, with A representing the highest credibility and C representing the lowest credibility of that domain. The amount of evidence domain was graded A if the total sample size was 1000 or more, B in the case of 100 to 1000, and C in the case of less than 100. The level of replication domain was graded A when there was extensive replication including at least 1 well-conducted meta-analysis with little between-study inconsistency; B in well-conducted meta-analysis with some methodological limitations or moderate between-study inconsistency; and C when there was no association, no independent replication, failed replication, scattered studies, flawed meta-analysis, or large inconsistency. The domain of protection from bias was graded A if bias could affect the magnitude but probably not the presence of the association, B when there was no obvious bias that affected the presence of the association but there was considerable missing information on the generation of evidence, and C when there was considerable potential for or demonstrable bias that could affect the presence or absence of the association.
The scores from each domain in this system are combined and an algorithm produces an overall evidence rating of strong, moderate, or weak. Strong evidence rating is given if all 3 domains receive an A score (i.e., AAA). Moderate evidence is assigned if the ratings are a combination of A and B scores (e.g., ABA, BBA, etc.). Finally, weak evidence is designated if any of the domains receives a C score (e.g., AAC, CBB, etc.). To provide consistency among the KQs for this study and among the other publications for this focus issue, we modified the HuGENet guidelines by substituting the GRADE terminology (“high” instead of “strong,” “low” instead of “weak”) and added an “insufficient” evidence category if each of the 3 domains received a score of C (CCC) or if there was only a single study without replication and the sample size was less than 1000.
An overall strength of “high” means that we are very confident that the true effect lies close to that of the estimated effect. A “moderate” rating means that we are moderately confident in the effect estimate; the true effect is likely to be close to the estimated effect, but there is a possibility that it is substantially different. An overall strength of “low” means that our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate. Finally, a rating of “insufficient” means that we have very little confidence in the effect estimate: The true effect is likely to be substantially different than the estimated effect. In addition, this rating may be used if there is no evidence or it is not possible to estimate an effect.
For studies addressing the heritability of CSM or OPLL, means and information on variation (e.g., standard deviation) for continuous variables were abstracted from the report as available. Specifically, segregation ratios, defined as the proportion of offspring that can be expected to be of a particular genotype or phenotype, were summarized and the observed ratios (i.e., those among various degrees of relatives of participants with CSM or OPLL) were compared with the expected ratios (i.e., those of the general population). Relative risks (RRs) and 95% CIs were also abstracted when included in the article. Risks and prevalence were reported as the number of participants with OPLL or CSM divided by the total number of participants at risk.
For studies addressing the association of SNPs with the risk of the development of CSM or OPLL, the specific allele substitutions responsible for the SNP were abstracted and corresponding ORs and 95% CIs were included when available in the article.
For studies addressing the association of haplotypes with the risk of the development of CSM or OPLL, the frequency of haplotype expression within a specific gene was reported as the percent of cases and controls with that specific haplotype and, when available, corresponding ORs and 95% CIs were abstracted.
Clinical Recommendations and Consensus Statements
Clinical recommendations or consensus statements were made through a modified Delphi approach by applying the GRADE/AHRQ criteria that impart a deliberate separation between the strength of the evidence (i.e., high, moderate, low, or insufficient) and the strength of the recommendation. When appropriate, recommendations or statements “for” or “against” were given “strong” or “weak” designations on the basis of the quality of the evidence, the balance of benefits/harms, and values and patient preferences. In some instances, costs may have been considered. A more thorough description of this process can be found in the focus issue Methods article.
We identified a total of 23 studies from our search strategy that met the inclusion criteria: 3 studies addressing KQ 1, 19 studies addressing KQ 2, and 1 study addressing KQ 3 (Figure 1). For KQ 1, our initial search produced 16 possible studies. We excluded 10 after title and abstract review. Among the 6 articles retrieved for full text review, 3 were excluded for the following reasons: sample size less than 10, population with spondylosis but no myelopathy, and a genetic linkage study. For KQ 2, our initial search yielded 70 possible citations. We excluded 42 after title and abstract review, primarily because they did not investigate SNPs or mutations as genetic predictors. Of the 28 articles that underwent full text review, 9 were excluded, the majority of which looked at restriction fragment length polymorphisms or protein expression or were about patients with disease of the lumbar spine. For KQ 3, our initial search produced 28 possible articles, of which all but 1 were excluded after abstract review. The remaining article underwent full text review and was included.
What Is the Evidence Supporting a Heritable Predisposition for CSM and OPLL Among Relatives of Affected Individuals?
Three studies provided evidence to support an inherited predisposition to the development of CSM (1 study)18 and OPLL (2 studies).19,20 Patel et al18 in 2012 performed a case-control family study using the Utah Population Database, a computerized genealogy of the Utah founding pioneers and their descendants, and found that, compared with the general population, first-degree relatives of patients with CSM had a 5-time greater risk (RR = 5.21; 95% CI, 2.07–13.1; P = 0.0009) of developing CSM and third-degree relatives were at twice the risk (RR = 1.95; 95% CI, 1.04–3.7; P = 0.04) (Table 2). Furthermore, the Genealogical Index of Familiality revealed a significant excess of relatedness between cases when compared with controls (P < 0.001), supporting the hypothesis of excess familial clustering among all cases with CSM.
Two case-control family studies assessed the prevalence of OPLL in first-degree relatives (Table 3). Tanikawa et al19 compared the prevalence of OPLL in siblings and children of cases with OPLL with that of the general population of Tokyo and found it to be significantly higher: 27.7% versus 3.9% (RR = 7.1; 95% CI, 4.3–11.7; P < 0.00001). Terayama20 reported the prevalence of OPLL in siblings, children, and parents of probands (19.1%) and compared them with an estimate of the prevalence in the general Japanese population (3.7%). The RR of first-degree relatives having OPLL was 5.2; 95% CI, 3.7–7.3; P < 0.00001.20
What Specific Genetic Polymorphisms Have Been Associated With CSM and/or OPLL?
Nineteen studies investigated the association of specific genetic variations (SNPs or haplotypes) with OPLL (16 studies) and CSM (3 studies). All of the included studies were case control in design, comparing the frequency of a specific finding in case as compared with control subjects. Three are judged as class of evidence I, 4 as level II, 9 as level III, and 1 as level IV. In addition, all studies used the candidate gene approach in which genetic variation was studied within prespecified genes hypothesized to play a key role in disease pathobiology on the basis of laboratory research or previously completed human linkage studies.
OSSIFICATION OF THE POSTERIOR LONGITUDINAL LIGAMENT
Single Nucleotide Polymorphisms
Across 15 studies, a total of 120 specific SNPs have been examined in relation to 11 genes, with 40 of these having shown statistically significant associations with OPLL (Table 4). These findings are presented below according to the candidate gene considered.
Collagen-encoding genes were chosen as candidates in 5 studies, with 3 investigating the COL6A1 gene (encoding the α1 chain of type VI collagen) and 2 investigating COL11A2 (encoding chain 2 of type XI collagen). In relation to COL6A1, 22 of the 34 SNPs investigated were significantly associated with OPLL, with 1 having demonstrated significance in 2 studies. To elaborate, in the study by Tanaka et al,33 a genome-wide linkage study in affected sibling pairs was first performed to identify loci related to OPLL. Based on the linkage results, which strongly implicated the COL6A1 gene region on chromosome 21, an association study was undertaken comparing COL6A1-related SNPs between cases and controls. Of the 32 SNPs evaluated, 21 were statistically associated with OPLL (P < 0.05), with 1 (T to C substitution at intron 32(−29)) found to be particularly significant (OR = 1.82, P = 0.000003). This same pleomorphism was also evaluated and found to be significantly more prevalent in OPLL cases than in controls in a follow-up confirmatory study (OR = 1.89, P = 0.004). In relation to the COL11A2 gene, 5 of the 20 SNPs evaluated across the 2 studies were shown to be statistically associated with OPLL (P < 0.05). Of note, one of these, a T to A substitution at intron 6 (−4) was seen at a significantly higher frequency in patients with OPLL than in controls in both studies. This pleomorphism, in addition to the one described previously in relation to the COL6A1 gene, are the only 2 SNPs that have shown a significant association with OPLL in more than 1 study.
The transforming growth factor β1 gene, known to encode transforming growth factor β1, a cytokine protein of importance for bone formation and resorption, was chosen as the candidate gene in 2 Japanese studies. Of 2 SNPs considered, 1 was shown to be significantly more prevalent in OPLL cases even after adjustment for relevant confounders.
The gene-encoding nucleotide pyrophosphatase, a membrane-bound glycoprotein known to be of importance in the inhibition of bone calcification and mineralization, was the candidate gene investigated in 3 Japanese studies. Two of the 14 nucleotide pyrophosphatase gene–related SNPs evaluated were significantly associated with OPLL in individual studies (P < 0.05). Of these 2 SNPs, the base pair deletion at intron 20 position 11 found to be most strongly associated with OPLL in the first of these 3 studies was not found to be significantly associated with this condition in the 2 subsequent studies.
Bone morphogenic protein (BMP)-encoding genes were candidates in 4 studies, with BMP-2 gene (encoding BMP-2) investigated in 3 studies and BMP-4 gene (encoding BMP-4) investigated in 1 study. Both of the encoded proteins belong to the transforming growth factor β family of signaling proteins and have potent established osteoinductive effects. Overall, 2 of the 5 BMP-2 related polymorphisms and 6 of the 19 BMP-4 related polymorphisms were significantly associated with OPLL (P < 0.05). None of the significant associations have been replicated in subsequent studies.
In addition to those presented previously, genes encoding runt-related transcription factor 2 (RUNX2), leptin receptor, vitamin D receptor (VDR), retinoic X receptor β (RXRβ), and interleukin 15 receptor (IL15RA) have all been considered as candidates in individual studies. Each of the encoded proteins is of known importance in bone biology, mineralization, and/or energy metabolism. Although a total of 26 SNPs have been considered in relation to these genes (11 for RUNX2, 5 for leptin receptor, 5 for VDR, 3 for RXRB, and 2 for ILI5RA), only 2 of these—1 related to ILI5RA and 1 related to RXRB—were significantly associated with OPLL. Neither of these 2 significant associations has been corroborated by subsequent study.
In addition to considering individual SNPs, associations between OPLL and specific SNP combinations, called haplotypes, have also been investigated. Across 8 studies, a total of 32 haplotypes have been examined in relation to COL11A2 gene (3 studies, 12 haplotypes considered), COL6A1 (2 studies, 8 haplotypes), RUNX2, BMP-4, and IL15RA (each 1 study and 4 haplotypes, respectively) (Table 5). Regarding the collagen-encoding genes, 7 of the 12 COL11A2 haplotypes and 4 of the 8 COL6A1 haplotypes were significantly associated with OPLL. Of note, in the 2001 study by Maeda et al36 evaluating the COL11A2 gene, although the overall analysis demonstrated a significant association between 2 specific haplotypes and OPLL, after sex stratification the same haplotypes showed an even greater association with OPLL in males but no significant association in females. Of haplotypes considered in relation to the remaining genes, 2 of the 4 BMP-4 haplotypes were significantly associated with OPLL; however, none of the individual haplotypes considered in relation to ILI5RA or RUNX2 were shown to be significant.
CERVICAL SPONDYLOTIC MYELOPATHY
Across 3 studies, a total of 4 SNPs and 5 gene alleles (SNPs underlying these alleles not reported) have been evaluated in relation to 3 genes, with 4 of these shown to have a significant association with CSM (Table 6).
VDR is the only gene to have been studied in association with both CSM and OPLL. Although no VDR-related SNPs have shown significant associations with OPLL, 2 of 4 SNPs evaluated in a single study were observed in a significantly higher proportion of patients with clinically and magnetic resonance imaging–confirmed CSM than in matched control patients.
Two additional studies have, respectively, evaluated the association of specific apolipoprotein E (ApoE) and collagen IX gene alleles with CSM. ApoE is a serum protein involved in lipid transportation and has been suggested to play an important role in repair and regeneration of the central nervous system after injury. In a single study, the ApoE4 allele, which encodes the ApoE4 protein isoform, was observed in a significantly higher proportion of patients with spondylosis-related cervical cord compression and clinical evidence of myelopathy (35%) than in patients with cord compression and no evidence of myelopathy (13%). The authors concluded that the E4 polymorphism was associated with an increased susceptibility to myelopathy in the face of chronic spinal cord compression. Regarding collagen IX–related alleles, a 2012 study by Wang et al38 found that an SNP in the COL9A2 gene, resulting in a Gln326Trp substitution in the α2 chain of collagen IX (Trp2 allele), was observed in a higher proportion of patients with CSM (19.8%) than in controls without cord compression or clinical evidence of myelopathy (6.2%) (P < 0.05). This study suggests that the Trp2 allele represents an inherited predisposing cause of cervical spine spondylosis and cord compression. Of note, none of the significant findings related to VDR, ApoE4, or Trp2 in CSM have been validated in separate studies.
What Is the Evidence Supporting a Genetic Basis for Predicting Postoperative Outcomes for Patients With CSM and OPLL Treated Surgically?
As regards the identification of a genetic predictor of surgical outcome, only a single CSM-related study was included.40 In this analysis, the presence of the same ApoE allele discussed previously (ApoE4) was associated with reduced postoperative functional recovery among patients with CSM treated with surgical decompression. Of 60 patients considered, 30 of 39 (76.9%) patients without the ApoE4 allele demonstrated improvement, defined as a 1 grade modified Japanese Orthopaedic Association (mJOA) scale improvement at a mean of 18.8 months postsurgery, as compared with 5 of 21 (23.8%) patients carrying this allele. In the multivariable analysis, after adjusting for preoperative mJOA score, the ApoE4 allele was associated with an 8.6 times greater odds of no functional improvement postoperatively (95% CI: 5.1–20.6).
There is low evidence that the Intron 32(−29) SNP (T replacing C) on the COL6A1 gene and the Intron 6(−4) SNP (T replacing A) on the COL11A2 gene are associated with an approximate 2-fold increased risk of developing OPLL. No other SNPs or haplotypes were identified in more than 1 study to be associated with OPLL. There is insufficient evidence from single studies to determine whether VDR-related SNPs or ApoE or collagen IX alleles are associated with CSM. There is insufficient evidence from a single study to determine whether the ApoE allele is associated with postoperative mJOA improvement (Table 7).
The goal of this study was to review critically the existing evidence relating to the heritability and genetics of CSM and OPLL through the completion of a systematic review of the literature. The results of the review were used arrive at the series of summary statements.
Regarding OPLL, the quantity of published material varied substantially depending on the KQ considered. For question 1, although 2 studies supported a positive inherited predisposition for OPLL, the design and risk of bias within these studies led to an overall strength of evidence rating of low. Although we support the concept that a combination of environmental and heritable factors likely lead to the development of this condition, the current evidence confirming the heritability of OPLL is limited. For question 2, despite the reasonably large volume of pertinent material, we found there to be substantial heterogeneity between studies with respect to specific gene-related polymorphisms and haplotypes considered. Where multiple studies considered the same single pleomorphism, only in the case of 2 SNPs (COL6A1/Intron 32(–29) and COL11A2/Intron 6(–4)) was an effect replicated across 2 studies. However, even in the presence of this replication, moderate sample sizes and imbalances of potential confounding variables between comparison groups in the relevant studies limit the reliability of the effect estimates, thus precluding a strength of evidence rating higher than low. Interestingly, all of the OPLL studies included in this review investigated genetic associations with the radiographical presence of OPLL, with none evaluating genetic associations with OPLL in combination with myelopathy. Because only a fraction of OPLL cases are known to progress to myelopathy, future genetic association studies focusing on this clinically relevant subpopulation would be of interest.
With respect to CSM, our ability to address all 3 KQs was limited by the low number of published studies pertinent to this topic. The 1 case-control family study relevant to question 1 was thought to provide low evidence supporting an inherited predisposition for CSM. For questions 2 and 3, the paucity of available evidence relating to associated polymorphisms and genetic predictors of surgical outcome led to an insufficient strength grading. It should be acknowledged that unlike OPLL, which is primarily a radiographical diagnosis, diagnosis of CSM requires consideration of both radiological and clinical elements. Furthermore, because spondylotic changes are common to virtually all individuals with increased age, it becomes a challenge to define suitable “case” and “control” groups for a genetic association study in the context of CSM. The CSM-related case-control studies included in this review took slightly different approaches to overcoming this challenge. In their study of ApoE polymorphisms, Setzer et al,39 compared “cases,” with magnetic resonance imaging evidence of spondylotic cervical cord compression coupled with symptoms and signs of myelopathy, with “controls,” with spondylotic cord compression but no clinical evidence of myelopathy. As a result, in this study, only genetic differences related to the intrinsic healing and regenerative and/or tolerance properties of the spinal cord would be identified. In contrast, in the 2 studies by Wang et al,37,38 the frequency of polymorphisms was compared between “cases,” with radiographical cord compression and clinical symptoms and signs of myelopathy, and “controls,” with a normal imaging and clinical picture. As a consequence, for these studies, it is less clear whether observed genetic differences are related to the progression of spondylosis and bony changes or to the healing and/or tolerance properties of the spinal cord itself. Future studies exploring the genetics of CSM should specify comparative groups that will permit focused insights into how genetic variations impact specific aspects of disease pathophysiology.
For both OPLL and CSM, all of the studies included in relation to question 2 used the candidate gene approach to identify SNPs or haplotypes potentially associated with disease occurrence. The disadvantage of this approach is that it forces investigators to make a priori assumptions about the importance of a specific gene in relation to disease pathophysiology on the basis of either previous animal work or on extrapolations from other disease processes. In processes such as OPLL and CSM, in which disease biology remains incompletely understood, such assumptions may not correctly identify the most responsible genes and common polymorphisms. The alternative to the candidate gene approach is the genome-wide association study that can be used to investigate associations between polymorphisms and disease across the entire genome without making assumptions about the importance of a particular gene in that disease process. The main disadvantage of such genome-wide association studies is that they require a large number of cases and control subjects (typically several thousands) and as a result are very resource intensive. However, to gain a true understanding of the common genetic variations underlying CSM and OPLL, such studies will be required in the future.
Finally, all of the OPLL studies considered in this review emanated from case and control samples collected in East Asian populations. Although OPLL is known to be most common among individuals of Asian ancestry, with a prevalence rate of 1.9% to 4.3%, this condition is nonetheless frequently observed in the Caucasian population, with prevalence rates estimated up to 1.7%.41,42 Investigating for evidence of heritability and genetic associations related to OPLL in Caucasian and other non-Asian populations might allow for insights into how the biology of the disease differs within ancestral subgroups. Furthermore, such study might allow for an increased understanding of the factors underlying increased prevalence of OPLLs in Asian as compared with non-Asian populations.
Although we have focused this review on the genetics of OPLL and CSM, it is important to note that a variety of environmental and clinical factors have also been associated with these conditions. In the case of OPLL, the presence of non–insulin-dependent diabetes, elevated BMI, sleep deprivation, and a variety of nutritional deficiencies have all been implicated.43–45 Given that both of these conditions are presumed to be the result of both genetic and environmental cues, the ultimate goal would be to develop combinatorial models, reflecting this multifactorial etiology, allowing for prediction of disease development and progression.
First, one class of study not considered in this review was genetic linkage studies. In such studies, the genomes of family members affected with a disease are compared with one another to evaluate whether specific genetic loci are shared more frequently than would be expected on the basis of chance alone. These studies were not included because they were thought to be less relevant to the specified KQs and also were thought difficult to assess from a level of evidence perspective. Although inclusion of these studies may have contributed to KQ 1, we feel that it would not have dramatically altered the evidentiary landscape and would have diminished the clarity of conclusions. Second, as acknowledged previously, in relation to KQ 2 the definition of what constituted a CSM “case” differed between studies. In this situation, we chose to tolerate case description heterogeneity in order to maximize the number of studies included. Third, we did not consider any animal CSM or OPLL studies that have attempted to probe the genetic and biological origins of these conditions. Although several such studies have provided very important mechanistic insights into the pathobiology of these conditions,46 it was thought that the results of such studies could not be reasonably integrated and collated with the results of human-based genetic studies. Fourth, because our main intent was to explore the genetics and heritability of the most common causes of cervical myelopathy internationally, studies concerning ossification of the anterior longitudinal ligament and OPLL involving the thoracic and lumbar spine were excluded. Nonetheless, the genetics of these related conditions are of importance and will form the basis of future review articles. Finally, for practical purposes our report included only English language studies. In general, studies that evaluate the effect of excluding trials published in languages other than English in systematic reviews report little effect on summary treatment effects and do not seem to bias results.47–49
Although existing family association studies support the principle of an inherited predisposition to CMS and OPLL, the strength of evidence supporting these findings is of low. In multiple studies, 2 separate SNPs (COL6A1/Intron 32(−29) and COL11A2/Intron 6(−4)) have been observed at higher frequencies in OPLL cases than in controls and may be associated with its development; however, there is insufficient evidence to support the association between CSM and any genetic polymorphism at present time. There is also insufficient evidence to support a genetic predictor of surgical outcome in relation to CSM or OPLL.
Statement 1: Existing family studies provide support for the principle of an inherited predisposition to CSM and OPLL
Statement 2: Two SNPs related to the collagen 6A1 gene (COL6A1/Intron 32(–29)) and the collagen 11A2 gene (COL11A2/Intron 6(–4)) have been associated with OPLL in multiple studies and may be associated with its development.
Statement 3: No statement can be made from the literature regarding the association of specific SNPs or haplotypes with CSM.
Statement 4: No statement can be made from the literature regarding genetic predictors of surgical outcome in the context of OPLL or CSM.
- Existing family studies provide support for the principle of an inherited predisposition to CSM and OPLL.
- Two SNPs related to the collagen 6A1 gene (COL6A1/Intron 32(−29)) and the collagen 11A2 gene (COL11A2/Intron 6(−4)) have been associated with OPLL in multiple studies and may be associated with its development.
- No recommendation can be made from the literature regarding the association of specific SNPs or haplotypes with CSM.
- No recommendation can be made from the literature regarding genetic predictors of surgical outcome in the context of OPLL or CSM.
The authors are indebted to Nancy Holmes and Ms. Chi Lam for their administrative assistance. The authors thank Katie Moran for her help in interpreting and abstracting genetic data and helping with development of critical appraisal criteria and appraisal of articles.
Author contributions are as follows: J.W.: study concept, data analysis and interpretation, manuscript preparation, and manuscript revision; M.G.F.: study concept, data interpretation, and manuscript revision; A.A.P.: data interpretation and manuscript revision; E.D.B.: data analysis and interpretation, manuscript preparation, and manuscript revision; J.R.D.: study concept, data analysis and interpretation, manuscript preparation, and manuscript revision; and D.S.B.: data interpretation and manuscript revision.
Supplemental digital content is available for this article. Direct URL citation appearing in the printed text is provided in the HTML and PDF version of this article on the journal's web site (www.spinejournal.com).
1. Kalsi-Ryan S, Karadimas SK, Fehlings MG. Cervical spondylotic myelopathy: the clinical phenomenon and the current pathobiology of an increasingly prevalent and devastating disorder. Neuroscientist 2013;19:409–21.
2. Inamasu J, Guiot BH, Sachs DC. Ossification of the posterior longitudinal ligament: an update on its biology, epidemiology, and natural history. Neurosurgery 2006;58:1027–39; discussion 39.
3. Gore DR, Sepic SB, Gardner GM. Roentgenographic findings of the cervical spine in asymptomatic people. Spine (Phila Pa 1976) 1986;11:521–4.
4. Bednarik J, Kadanka Z, Dusek L, et al. Presymptomatic spondylotic cervical myelopathy: an updated predictive model. Eur Spine J 2008;17:421–31.
5. Sakou T, Matsunaga S, Koga H. Recent progress in the study of pathogenesis of ossification of the posterior longitudinal ligament. J Orthop Sci 2000;5:310–5.
6. Matsunaga S, Sakou T, Taketomi E, et al. Clinical course of patients with ossification of the posterior longitudinal ligament: a minimum 10-year cohort study. J Neurosurg 2004;100(suppl):245–8.
7. Yu W, Clyne M, Khoury MJ, et al. Phenopedia and Genopedia: disease-centered and gene-centered views of the evolving knowledge of human genetic associations. Bioinformatics 2010;26:145–6.
8. Yu W, Gwinn M, Clyne M, et al. A navigator for human genome epidemiology. Nat Genet 2008;40:124–5.
9. Wright JG, Swiontkowski MF, Heckman JD. Introducing levels of evidence to the journal. J Bone Joint Surg Am 2003;85-A:1–3.
10. West S, King V, Carey TS, et al. Systems to Rate the Strength of Scientific Evidence. Rockville, MD: Agency for Healthcare Research and Quality; 2002. Evidence Report/Technology Assessment No. 47 (Prepared by the Research Triangle Institute-University of North Carolina Evidence-based Practice Center, Contract No. 290-97-0011).
11. Agency for Healthcare Research and Quality. Methods Guide for Effectiveness and Comparative Effectiveness Reviews. AHRQ Publication No. 10(12)-EHC063-EF. Rockville, MD: Agency for Healthcare Research and Quality. Published April 2012 Available at: www.effectivehealthcare.ahrq.gov
. Accessed September 13, 2013.
12. Norvell DC, Dettori JR, Skelly AC, et al. Methodology for the systematic reviews on an adjacent segment pathology. Spine (Phila Pa 1976) 2012;37(suppl):S10–7.
13. Atkins D, Best D, Briss PA, et al. Grading quality of evidence and strength of recommendations. BMJ 2004;328:1490.
14. Balshem H, Helfand M, Schunemann HJ, et al. GRADE guidelines: 3. Rating the quality of evidence. J Clin Epidemiol 2011;64:401–6.
15. Owens DK, Lohr KN, Atkins D, et al. AHRQ series paper 5: grading the strength of a body of evidence when comparing medical interventions—Agency for Healthcare Research and Quality and the effective health-care program. J Clin Epidemiol 2010;63:513–23.
16. Attia J, Ioannidis JP, Thakkinstian A, et al. How to use an article about genetic association: B: are the results of the study valid? JAMA 2009;301:191–7.
17. Ioannidis JP, Boffetta P, Little J, et al. Assessment of cumulative evidence on genetic associations: interim guidelines. Int J Epidemiol 2008;37:120–32.
18. Patel AA, Spiker WR, Daubs M, et al. Evidence of an inherited predisposition for cervical spondylotic myelopathy. Spine (Phila Pa 1976) 2012;37:26–9.
19. Tanikawa E, Furuya K, Nakajima H. Genetic study on ossification of posterior longitudinal ligament. Bull Tokyo Med Dent Univ 1986;33:117–28.
20. Terayama K. Genetic studies on ossification of the posterior longitudinal ligament of the spine. Spine (Phila Pa 1976) 1989;14:1184–91.
21. Kamiya M, Harada A, Mizuno M, et al. Association between a polymorphism of the transforming growth factor-beta1 gene and genetic susceptibility to ossification of the posterior longitudinal ligament in Japanese patients. Spine (Phila Pa 1976) 2001;26:1264–6; discussion 1266–7.
22. Kawaguchi Y, Furushima K, Sugimori K, et al. Association between polymorphism of the transforming growth factor-beta1 gene with the radiologic characteristic and ossification of the posterior longitudinal ligament. Spine (Phila Pa 1976) 2003;28:1424–6.
23. Kim DH, Jeong YS, Chon J, et al. Association between interleukin 15 receptor, alpha (IL15RA) polymorphism and Korean patients with ossification of the posterior longitudinal ligament. Cytokine 2011;55:343–6.
24. Koga H, Sakou T, Taketomi E, et al. Genetic mapping of ossification of the posterior longitudinal ligament of the spine. Am J Hum Genet 1998;62:1460–7.
25. Kong Q, Ma X, Li F, et al. COL6A1 polymorphisms associated with ossification of the ligamentum flavum and ossification of the posterior longitudinal ligament. Spine (Phila Pa 1976) 2007;32:2834–8.
26. Koshizuka Y, Kawaguchi H, Ogata N, et al. Nucleotide pyrophosphatase gene polymorphism associated with ossification of the posterior longitudinal ligament of the spine. J Bone Miner Res 2002;17:138–44.
27. Liu Y, Zhao Y, Chen Y, et al. RUNX2 polymorphisms associated with OPLL and OLF in the Han population. Clin Orthop Relat Res 2010;468:3333–41.
28. Maeda S, Ishidou Y, Koga H, et al. Functional impact of human collagen alpha2(XI) gene polymorphism in pathogenesis of ossification of the posterior longitudinal ligament of the spine. J Bone Miner Res 2001;16:948–57.
29. Nakamura I, Ikegawa S, Okawa A, et al. Association of the human NPPS gene with ossification of the posterior longitudinal ligament of the spine (OPLL). Hum Genet 1999;104:492–7.
30. Numasawa T, Koga H, Ueyama K, et al. Human retinoic X receptor beta: complete genomic sequence and mutation search for ossification of posterior longitudinal ligament of the spine. J Bone Miner Res 1999;14:500–8.
31. Ren Y, Feng J, Liu ZZ, et al. A new haplotype
in BMP4 implicated in ossification of the posterior longitudinal ligament (OPLL) in a Chinese population. J Orthop Res 2012;30:748–56.
32. Tahara M, Aiba A, Yamazaki M, et al. The extent of ossification of posterior longitudinal ligament of the spine associated with nucleotide pyrophosphatase gene and leptin receptor gene polymorphisms. Spine (Phila Pa 1976) 2005;30:877–80; discussion 881.
33. Tanaka T, Ikari K, Furushima K, et al. Genomewide linkage and linkage disequilibrium analyses identify COL6A1, on chromosome 21, as the locus for ossification of the posterior longitudinal ligament of the spine. Am J Hum Genet 2003;73:812–22.
34. Wang H, Liu D, Yang Z, et al. Association of bone morphogenetic protein-2 gene polymorphisms with susceptibility to ossification of the posterior longitudinal ligament of the spine and its severity in Chinese patients. Eur Spine J 2008;17:956–64.
35. Wang H, Yang ZH, Liu DM, et al. Association between two polymorphisms of the bone morpho-genetic protein-2 gene with genetic susceptibility to ossification of the posterior longitudinal ligament of the cervical spine and its severity. Chin Med J (Engl) 2008;121:1806–10.
36. Maeda S, Koga H, Matsunaga S, et al. Gender-specific haplotype
association of collagen alpha2 (XI) gene in ossification of the posterior longitudinal ligament of the spine. J Hum Genet 2001;46:1–4.
37. Wang ZC, Chen XS, Wang da W, et al. The genetic association of vitamin D receptor polymorphisms and cervical spondylotic myelopathy in Chinese subjects. Clin Chim Acta 2010;411:794–7.
38. Wang ZC, Shi JG, Chen XS, et al. The role of smoking status and collagen IX polymorphisms in the susceptibility to cervical spondylotic myelopathy. Genet Mol Res 2012;11:1238–44.
39. Setzer M, Hermann E, Seifert V, et al. Apolipoprotein E gene polymorphism and the risk of cervical myelopathy in patients with chronic spinal cord compression. Spine (Phila Pa 1976) 2008;33:497–502.
40. Setzer M, Vrionis FD, Hermann EJ, et al. Effect of apolipoprotein E genotype on the outcome after anterior cervical decompression and fusion in patients with cervical spondylotic myelopathy. J Neurosurg Spine 2009;11:659–66.
41. Firooznia H, Benjamin VM, Pinto RS, et al. Calcification and ossification of posterior longitudinal ligament of spine: its role in secondary narrowing of spinal canal and cord compression. N Y State J Med 1982;82:1193–8.
42. Matsunaga S, Sakou T. OPLL: disease entity, incidence, literature search and prognosis. In: Yonenobu K, Nakamura K, Toyama Y, eds. OPLL: Ossification of the Posterior Longitudinal Ligament. New York, NY: Springer; 2006:11–7.
43. Okamoto K, Kobashi G, Washio M, et al. Dietary habits and risk of ossification of the posterior longitudinal ligaments of the spine (OPLL); findings from a case-control study in Japan. J Bone Miner Metab 2004;22:612–7.
44. Washio M, Kobashi G, Okamoto K, et al. Sleeping habit and other life styles in the prime of life and risk for ossification of the posterior longitudinal ligament of the spine (OPLL): a case-control study in Japan. J Epidemiol 2004;14:168–73.
45. Kobashi G, Washio M, Okamoto K, et al. High body mass index after age 20 and diabetes mellitus are independent risk factors for ossification of the posterior longitudinal ligament of the spine in Japanese subjects: a case-control study in multiple hospitals. Spine (Phila Pa 1976) 2004;29:1006–10.
46. Okawa A, Nakamura I, Goto S, et al. Mutation in NPPS in a mouse model of ossification of the posterior longitudinal ligament of the spine. Nat Genet 1998;19:271–3.
47. Juni P, Holenstein F, Sterne J, et al. Direction and impact of language bias in meta-analyses of controlled trials: empirical study. Int J Epidemiol 2002;31:115–23.
48. Moher D, Pham B, Lawson ML, et al. The inclusion of reports of randomised trials published in languages other than English in systematic reviews. Health Technol Assess 2003;7:1–90.
49. Morrison A, Polisena J, Husereau D, et al. The effect of English-language restriction on systematic review-based meta-analyses: a systematic review of empirical studies. Int J Technol Assess Health Care 2012;28:138–44.