Prevalence of hyperplastic articular pillars in the cervical spine and relationship with cervical lordosis

Background: Cervical lordosis is often used as an indicator for a number of clinical conditions ranging from traumatic to degenerative. Previous research has indicated that a number of factors may change the lordosis. However, the link between hyperplastic articular pillars and cervical lordosis measurements has never been studied. Objective: To investigate the reliability of determining articular pillar hyperplasia, to determine its prevalence, and to compare lordosis measurements between persons with and without hyperplasia of the cervical articular pillars. Methods: Twelve normal neutral lateral cervical radiographs were chosen for the reliability study. Two chiropractors independently evaluated the articular pillars of C3 through C6 by drawing lines along the planes of the superior and inferior articular surfaces of each pillar. Each pillar was categorized as "normal" or "hyperplastic" depending on the convergence or divergence of these lines. One examiner repeated this procedure after a 1-month interval. Percent agreement and kappa statistics were calculated for interexaminer and intraexaminer agreement. Forty-eight normal neutral lateral cervical radiographs with a horizontal Chamberlain's line were evaluated for the presence or absence of pillar hyperplasia. Two measurement techniques were used to assess the cervical lordosis: the method of Jochumsen and the "angle of the cervical lordosis." The unpaired t test was used to compare the angle of cervical lordosis measurements between the two groups. The Mann-Whitney U test was used to analyze data obtained with the method of Jochumsen. The prevalence of pillar hyperplasia was calculated on 100 normal lateral cervical radiographs. Results: Interexaminer and intraexaminer reliability of determining pillar hyperplasia was fair to substantial (kappa = 0.4 to 0.61; 75% to 92%). Forty-six percent of the patients demonstrated pillar hyperplasia of at least one cervical level. Patients with pillar hyperplasia had a mean cervical curve of 31.52 degrees, whereas the curve in patients without pillar hyperplasia was 44.76 degrees (P =.0001). The method of Jochumsen also detected a significant difference in the curve measurements between the two patient groups (P =.0127). Conclusion: Articular pillar hyperplasia is common and significantly reduces the cervical lordosis measurements. Accepted normal values for lordosis evaluation need to be reassessed and articular pillar configuration considered when treatment plans include attempts to restore a "normal" cervical lordosis. Previous studies attempting to link lordosis measurements with pathologic conditions or symptoms may need to be reevaluated.