The Reliabilities of Radiographic Measurements of Cervical Sagittal Alignment in Ankylosing Spondylitis.

Journal of Spinal Disorders and Techniques Publish Ahead of Print DOI:10.1097/BSD.0000000000000105

The reliabilities of radiographic measurements of cervical sagittal alignment in ankylosing spondylitis

Jong Ki Shin · Jung Sub Lee · Sung Shik Kang · Jae Myung Lee · Bu Hyun Youn

J. K. Shin , M.D. : Orthopaedic Surgeon, Assistant professor J. S. Lee, M.D., Ph.D.: Orthopaedic Surgeon, Professor (corresponding author) Medical Research Institute, Department of Orthopaedic Surgery, Pusan National University School of Medicine, Busan, Republic of Korea S. S. Kang, M.D. : Orthopaedic Surgeon, Assistant professor Medical Research Institute, Department of Orthopaedic Surgery, Pusan National University Yangsan Hospital, Yangsan, Republic of Korea J. M. Lee, Ph.D.: Professor Department of Naval Architecture and Ocean Engineering, Pusan National University, Busan, Republic of Korea -1 Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.

B. H. Youn, Ph.D.: Associate Professor Department of Biologic Sciences, College of Natural Sciences, Pusan National University, Busan, Republic of Korea

Corresponding author; Jung Sub Lee, MD, PhD Medical Research Institute, Department of Orthopaedic Surgery, Pusan National University School of Medicine, 1-10 Ami-Dong, Seo-Gu, Busan, Republic of Korea, 602-739 Tel; 82-51-240-7248, Fax; 82-51-247-8395, E-mail; [email protected]

Conflicts of Interest: None

Source of Funding: This work was supported by the Convergence Research Grant funded by the Pusan National University(PNU, Convergence Research Grant)(PNU-2013-1312-0001).

-2 Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.

Study Design: Prospective study. Objective: To test the inter- and intra-observer reliabilities of five specific measures of global cervical sagittal alignment in patients with ankylosing spondylitis (AS), and to suggest a better means of measuring cervical sagittal alignment. Summary of Background Data: The intraobserver and interobserver reliabilities of several different methods of measuring cervical lordosis have been reported. However, they have not been studied till yet in patients with AS. Methods: Inter- and intra-observer reliabilities of five specific measures of cervical lordosis were evaluated in patients with AS. Eighty patients with AS were allocated to a non-ankylosis or an ankylosis group, and the reliabilities of the Cobb C1-C7, Cobb C2-C7, centroid, posterior tangent, and Ishihara index methods were evaluated. Results: The intra- and interclass correlation coefficients (ICCs) of all five methods were generally high. For the 80 study subjects, ICCs were ≥0.84 (excellent) for all five radiographic methods. However, comparison of the ICCs, 95% confidence intervals (CI) and mean absolute differences (MAD) between groups with varying degrees of ankylosis showed that the reliability of lordosis measurements decreased as the severity of ankylosis increased. The five methods consistently demonstrated higher ICCs for both inter- and intra-observer comparisons in the nonankylosis group. However, in the ankylosis group, the Cobb C1-C7 method demonstrated high ICCs for both inter- and intra-observer comparisons whereas the other 4 methods had high ICCs only for intra-observer comparisons. The intraobserver MADs were similar for the five methods (2.4-3.9), but the interobserver MADs of measurement methods in both groups showed low measurement reliability except for the Cobb C1-C7 methods. -3 Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.

Conclusion: This study provides a reliability analysis of different cervical lordosis measurement methods in AS, and shows that the Cobb C1-C7 method provides a reliable means for measuring cervical lordosis in AS. Key words: ankylosing spondylitis, cervical lordosis, radiographic measure


Introduction Relationships between sagittal alignment and cervical spine disorders have received considerable attention, as the configurations of sagittal spinal curves can influence cervical postsurgical outcomes and alignment affects load distributions on intervertebral discs. Therefore, appropriate measurements of the cervical curvature are important for clinical decision making. For this reason, several methods have been developed to measure cervical sagittal alignment,1-7 and the interobserver and intraobserver reliabilities of several methods of quantifying cervical lordosis have been reported.6,7 These devised methods include the Cobb, centroid, posterior tangent, and Ishihara index methods. Harrison et al6 compared the Cobb and the posterior tangent methods in cervical lordosis and demonstrated higher reliability for the centroid method. Ohara et al7 reported higher reliability for the Cobb C1-C7 method. However, no reports have been issued regarding cervical alignment measurements made under specific conditions, such as ankylosing spondylitis (AS). Recognition of the importance of reliability analysis for spine disease has led to several studies and radiological analyses of AS before and after surgery have been demonstrated in several studies.8,9 However, to date the reliabilities of lordosis angle measurements in AS have not been reported. In particular, sagittal alignment of the cervical spine is an important value in AS, the role of which in deformity progression, symptoms, and outcomes needs to be determined. AS is a chronic, inflammatory rheumatic disease characterized by inflammatory back pain and/or neck pain due to sacroiliitis and spondylitis, and the formation of syndesmophytes leading to ankylosis.10 In addition, AS is considered the most common and most typical form of


spondyloarthropathy. Arthritic changes and osteoporosis in a vertebral body could interfere accurate measurements of cervical sagittal alignment, and therefore, the availability of appropriate methods for measuring cervical sagittal alignment is important for clinical decision making in AS. The aim of this study was to test the inter- and intra-observer reliabilities of five specific measures of global cervical sagittal alignment in patients with AS, and to suggest a better means of measuring cervical sagittal alignment.

Patients and Methods Eighty consecutive AS patients, 10 women and 70 men of average age 43.4 ± 12.3 years, were enrolled. AS patients were recruited from the patients attending an outpatient clinic. Demographic and disease characteristics were documented. Patients that met the current New York criteria11 were considered eligible to participate in this study if they had received medical treatment for at least 1 year. The exclusion criteria applied were an age of >60 years, or a concomitant neurological or psychiatric disease. Patients with a history of a spinal fracture, spinal disc herniation, spinal surgery, or another orthopaedic condition of the spine or lower extremities (as prosthesis) were also excluded. In addition, 16 patients with unclear lower margin of C7 in the cervical radiographs were excluded. The 80 study subjects underwent plain cervical lateral radiopraphy. Radiographs were taken by one technician using a standard technique and the same x-ray machine in the standing position. The distance between the radiographic tube and the film was 150cm and the film was centered on C4. According to the ankylosis status of the curve in the radiographs, the radiographs were divided into two groups: non-ankylosis (no osseous bridge in the cervical -6 Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.

spine) and ankylosis (cervical spines were partially or completely ossified). The radiographs were composed of 40 non-ankylosis and 40 ankylosis patients. A delayed and repeated measurement design was used to evaluate the reliability of the three examiners. All 80 radiographs were measured twice by three examiners with 2 weeks delay between first and second measurements. All measurement were carried out using a computerbased digital radiogram on a picture achieving computer system (PACS Expertise, Marosis, South Korea), which allowed easy and accurate determinations. A total of 400 measurements were taken for each of the five different radiographic methods. The five methods included the Cobb C1-C7, Cobb C2-C7, centroid, posterior tangent C2-C7, and Ishihara index methods.

Cobb C1-C7 method and C2-C7 method (Fig. 1) The angle between the inferior margin of C1 and the superior endplate of C7, and the angle between the inferior endplate of C2 and the superior endplate of C7 are measured.

Centroid method (Fig. 2) The intersections of four vertebral body corner are connected diagonally as wertebral centroids (C3, C6, C7). The intersection of the perpendicular lines drawn from the proximal and distal line made the cervical lordosis. The proximal line connects the bisection of the inferior endplate of the C2 and C3 centroid, and the distal line the C6 and C7 centroids.


Posterior tangent C2-C7 (Fig. 3) The posterior tangent method uses the superior-posterior and inferior-posterior body corners.

Ishihara index method (Fig. 4) Posterior lines are drawn through the posterior inferior corners of C2 and C7. The 4 distances, a3, a4, a5, and a6, between the posterior line and the posterior inferior corners of C3-C6, respectively, were measured. The Ishihara index was then calculated using the following formula: Ishihara index = (a3+a4+a5+a6) / posterior line length x 100

Statistical analysis was performed using SPSS 13 software for Windows (SPSS, Chicago, IL, USA) and MedCalc software (MedCalc, Mariakerke, Belgium). Data were expressed by mean ± standard deviation. Means, standard deviations, interclass and intraclass correlation coefficients (ICCs), and 95% confidence intervals (CI) were calculated. ICCs were considered by Shrout and Fleiss’s classifications (poor, 0.75).12,13 The mean absolute differences (MAD) of observer measurements were used in the error analysis. All statistical calculations for reliability results were performed based on the assumptions that measurement is crossed with the examiner and patient, and that the examiner was a random factor rather than a fixed factor.


Results Table 1 lists the overall means and standard deviations of the outcome measures for the two readings of the 80 radiographs by the three examiners. ICCs, 95% CIs and MADs are provided for the two Cobb methods and the centroid, posterior tangent, and Ishihara’s method. Table 2 provides the data for each group of radiographs classified by degree of curve ankylosis. In general, ICCs of all five methods were high. In the 80 study subjects, ICCs were all ≥0.84 (excellent) for the five radiographic methods (Table 1). However, a comparison of ICCs and 95% CIs between groups with varying degrees of ankylosis showed that the reliability of lordosis measurements decreased as the severity of ankylosis increased (Table 2). The five methods consistently demonstrated high ICCs for inter- and intra-observer comparisons in the nonankylosis group. However, in the ankylosis group, the Cobb C1-C7 method produced high ICCs for both inter- and intra-observer comparisons, whereas the other 4 methods had high ICCs only for intra-observer comparisons. Regarding error analysis, in all study subjects, intraobserver MADs for the five methods and interobserver MAD for the Cobb C1-C7 method were similar (2.9-3.8), whereas the other 4 methods showed higher interobserver MAD. A comparison of MADs between the ankylosis and non-ankylosis groups showed that the reliability of lordosis measurements decreased as the severity of ankylosis increased. Intraobserver MADs were similar for the five methods (2.4-3.9), but interobserver MADs of the Cobb C2-C7, centroid, posterior tangent and Ishihara index methods showed less measurement reliability than Cobb C1-C7 method in the ankylosis and non-ankylosis groups.


Discussion Intra- and interobserver reliabilities of several different measurements of cervical lordosis in the normal population have been reported.6,7 Ohara et al7 compared the Cobb C1-C7, Cobb C2-C7, centroid, sagittal tangent, and Ishihara index methods with respect to measures of global cervical lordosis in a normal population. They reported that the inter- and intra-observer reliabilities of all global angles were high. Harrison et al6 compared the Cobb C1-C7, Cobb C2-C7 and posterior tangent methods in measurement of global and segmental cervical lordosis, and reported intraand inter-observer reliabilities of measuring all cervical lordotic angles were high (ICC>0.7). Furthermore, they concluded that the posterior tangent method is more reliable for measuring cervical lordosis than the Cobb method. However, reliability might be different under specific conditions such as AS. In this study, we assumed that ICCs of cervical lordosis measurement methods would differ for AS radiographs. Measurements of cervical lordosis are important for treatment of AS. These patients often show cervical hypolordosis and cervicothoracic kyphosis, which can cause significant disability due to loss of horizontal gaze, functional limitation, chin-on-chest deformity with swallowing difficulties, neck pain, and weakness due to stretching of the spinal cord over the apex, and they also increase the risk of fall-related injuries.14-19 Surgery aims to restore horizontal gaze, sagittal balance, improve function, diminish social disability and provide durable correction.14,16,17,20-23 Thus, cervical lordotic angle is an important factor in AS, and accurate measurements are needed, but measurements of cervical lordosis in AS are difficult due to spondylitis and the formation of syndesmophytes in vertebral bodies and endplates. In addition, osteopenic changes of the spine often blur the vertebral silhouette. For these reasons, we considered that the reliabilities of cervical lordosis measurements in AS may be different -10Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.

from that in the normal population. In this study, we analyzed the reliability of cervical lordosis measurement methods with the aim of suggesting a better method in AS patients. Measures of segmental cervical lordosis are also clinically important in AS surgery. However, it is difficult to draw multiple lines in a small deformed segment and global cervical lordosis is more valuable than segmental lordosis in AS patients because AS is a systemic disease which often involves whole spine. Therefore, in the present study, we did not compare the reliabilities of segmental lordosis measurement methods. We found that the ICCs of the 80 radiographs were all high (ICCs≥0.84). However, in the ankylosis group, intra- and interobserver reliabilities were lower than in the non-ankylosis group. In the ankylosis group, the intraobserver reliabilities of five methods showed consistently high ICCs (≥0.75). However, higher ICC for interobserver reliability (ICC=0.84) and lower MADs (≤3.6) were demonstrated only in the Cobb C1-C7 method. MAD has been useful for determining the most reliable method when the reliabilities of methods are similar. Harrison et al5 concluded that high ICCs and a low MAD mean high reliability, and proposed that MAD should be utilized in error analysis during reliability studies. Similarly, Lee et al24, who compared several methods for measuring lumbar lordosis in AS, found that the Cobb L1-L5 method had the lowest MAD, and concluded that the Cobb L1-L5 method wass better than the other methods, which had high ICCs and low MAD. In the present study, only the Cobb C1-C7 method showed higher reliability. Furthermore, the Cobb method is the technique most commonly used by clinicians because it provides a simple and quick measurement of cervical lordosis. Previous studies have reported that the Cobb method is sensitive to contour changes, such as wedge, biconcave and brush, which resulted

-11Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.

from osteoporosis.4,25 Harrison et al6 reported that the Cobb C1-C7 angle overestimated the cervical lordosis, whereas the Cobb C2-C7 angle underestimated the lordosis. On the other hand, Ohara et al7 reported that the Cobb C1-C7 method showed better reliability than the other methods. For the above reasons, several authors have used other measurement methods, such as the posterior vertebral body tangent, centroid and Ishihara index methods. However, it might be different in AS patients. Although regardless of degenerative change, the posterior vertebral body margins of C2 and C7 are quite visible in the normal population, it is difficult to draw the posterior body margin precisely for AS patient because of spondylitis and formation of the syndesmophytes. In the present study, the intra- and interobserver reliabilities for the global C1-C7 Cobb angle was high with ICCs≥0.84 in all groups, and Cobb C1-C7 method showed better reliability than the Cobb C2-C7 and posterior tangent method. This result is in line with the fact that the C2 vertebra is often affected by degenerative changes and its contour is usually invisible in arthritic patients.25 In addition, the reliability of the Ishihara method was less that of the Cobb methods. The Ishihara index value is calculated using 4 numerators and 1 denominator, which are not angles. Furthermore, when there is no lordosis, the Ishihara index value is extremely small because the 4 numerators are small and the denominator is large. Actually, of the 5 methods examined, when alignment was straight, Ishihara index values approached zero and varied very little, which indicates that the Ishihara method is not very reliable when the alignment is straight. Some limitations of this study require consideration. In particular, only AS patients with radiographs in which C1-C7 were clearly visible were enrolled: those in which C7 was not clearly seen were excluded (16 patients).


The present study is the first to undertake reliability analysis on different global cervical lordosis measurement methods in AS. The findings of this study show that the Cobb C1-C7 method is reliable for measuring the global cervical lordosis in AS. Clinicians may find these results useful when interpreting cervical sagittal alignment values obtained using different measurement methods.

Disclosures None


References 1.

Chen YL. Vertebral centroid measurement of lumbar lordosis compared with the Cobb technique. Spine 1999;24:1786-1790.

2.

Chernukha KV, Daffner RH, Reigel DH. Lumbar lordosis measurement: A new method versus Cobb technique. Spine 1998;23:74-80.

3.

Cobb J. Outline for the study of scoliosis. Instructional course lectures. American Academy of Orthopaedic Surgeons, Ann Arbor, MI, 1948; 261-275.

4.

Genant HK, Li J, Wu CY, et al. Vertebral fractures in osteoporosis: a new method for clinical assessment. J Clin Densitom 2000;3:281-290.

5.

Harrison DE, Harrison DD, Cailliet R, et al. Radiographic analysis of lumbar lordosis: centroid, Cobb, TRALL, and Harrison posterior tangent methods. Spine 2001;26:E235E242

6.

Harrison DE, Harrison DD, Cailliet R, et al. Cobb method or Harrison posterior tangent method: which to choose for lateral cervical radiographic analysis. Spine 2000;25:20722078.

7.

Ohara A, Miyamoto K, Naganawa T, et al. Reliabilities of and correlations among five standard methods of assessing the sagittal alignment of the cervical spine. Spine 2006;31:2585-2591.

8.

Kim KT, Lee SH, Suk KS, et al. Spinal pseudarthrosis in advanced ankylosing spondylitis with sagittal plane deformity: clinical characteristics and outcome analysis. Spine 2007;32:1641-1647.

9.

Suk KS, Kim KT, Lee SH, et al. Significance of chin-brow vertical angle in correction of kyphotic deformity of ankylosing spondylitis patients. Spine 2003;28:2001-2005. -14-

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.

10. Zochling J, van der Heijde D, Burgos-Vargas R, et al. ASAS/EULAR recommendations for the management of ankylosing spondylitis. Ann Rheum Dis 2006;65:442-452. 11. van der Linden S, Valkenburg HA, Cats A. Evaluation of diagnostic criteria for ankylosing spondylitis. A proposal for modification of the New York criteria. Arthritis Rheum 1984;27:361-368. 12. Dunn G. Design and analysis of reliability studies. Stat Methods Med Res 1992;1:123-157 13. Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing reter reliability. Psychol Bull 1979;86:420-428. 14. Belanger TA, Milam RA, Roh JS, et al. Cervicothoracic extension osteotomy for chin-onchest deformity in ankylosing spondylitis. J Bone Joint Surg 2005;87-A:1732–1738. 15. Chavanne A, Pettigrew DB, Holtz JR, et al. Spinal cord intramedullary pressure in cervical kyphotic deformity. Spine 2011;36:1619–1626. 16. Etame AB, Thoi KD, Wang AC, et al. Surgical management of symptomatic cervical or cervicothoracic kyphosis due to ankylosing spondylitis. Spine 2008;33:E559–E564 17. Hoh DJ, Khoueir P, Wang MY. Management of cervical deformity in ankylosing spondylitis. Neurosurg Focus 2008;24:E9. 18. Mehdian SMH, Arun R. A safe controlled instrumented reduction technique for cervical osteotomy in ankylosing spondylitis. Spine 2011;36:715–720 19. Mehdian SMH, Freeman BJC, Licina P. Cervical osteotomy for ankylosing spondylitis: an innovative variation on an existing technique. Eur Spine J 1999;8:505–509 20. Chin KR, Ahn J. Controlled cervical extension osteotomy for ankylosing spondylitis utilizing the Jackson operating table. Spine 2007;32:1926–1929. 21. Khoueir P, Hoh DJ, Wang MY. Use of hinged rods for controlled osteoclastic correction


of a fixed cervical kyphotic deformity in ankylosing spondylitis. J Neurosurg Spine 2008;8:579–583. 22. McMaster MJ. Osteotomy of the cervical spine in ankylosing spondylitis. J Bone Joint Surg 1997;79-B:197–203. 23. Mummaneni PV, Mummaneni VP, Haid RW, et al. Cervical osteotomy for the correction of chin-on-chest deformity in ankylosing spondylitis. Neurosurg Focus 2003;14:e9. 24. Lee JS, Goh TS, Park SH, et al. Radiographic measurement reliability of lumbar lordosis in ankylosing spondylitis. Eur Spine J 2013;22:813-818. 25. Polly DW, Kilkelly FX, McHale KA, et al. Measurement of lumbar lordosis: Evaluation of intraobserver, interobserver, and technique variability. Spine 1996;21:1530-1536.


Legend Fig. 1 The Cobb C1-C7 and C2-C7 methods Fig. 2 The centroid method Fig. 3 The posterior tangent C2-C7 method Fig. 4 The Ishihara index method


Tables Table 1. Mean, standard deviations and ICC values of each measurement method Methods

Mean and SD Intraobserver reliability ICC 95% CI

MAD

Interobserver reliability ICC 95% CI

MAD

Cobb C1-C7

43.0 ± 11.3

0.96 0.95-0.97 2.9

0.92 0.90-0.94 3.4

Cobb C2-C7

14.2 ± 10.7

0.95 0.94-0.96 3.8

0.88 0.94-0.90 5.0

Centroid

14.1 ± 8.2

0.93 0.91-0.95 3.1

0.84 0.79-0.88 4.9

Posterior tangent 20.3 ± 10.5

0.96 0.95-0.97 3.1

0.87 0.84-0.90 5.2

Ishihara

0.96 0.95-0.97 3.0

0.91 0.88-0.93 4.7

15.1 ± 9.5


Table 2. Mean, standard deviations and ICC values of each measurement method for each group divided by the severity of ankylosis Methods

Mean and SD Intraobserver reliability ICC 95% CI

MAD

Interobserver reliability ICC 95% CI

MAD

Non-ankylosis Cobb C1-C71

42.9 ± 14.9

0.97 0.96-0.98 2.4

0.95 0.93-0.97 3.2

Cobb C2-C7

15.7 ± 12.8

0.97 0.95-0.98 3.6

0.95 0.92-0.96 4.3

Centroid

15.4 ± 11.2

0.96 0.94-0.97 3.0

0.95 0.93-0.97 4.7


0.98 0.97-0.99 2.6

0.96 0.95-0.97 4.5

Ishihara

15.7 ± 12.6

0.97 0.96-0.98 2.6

0.95 0.93-0.97 4.2

Cobb C1-C71

43.1 ± 6.4

0.91 0.87-0.94 3.4

0.84 0.77-0.89 3.6

Cobb C2-C7

14.4 ± 7.3

0.92 0.88-0.94 3.9

0.69 0.56-0.78 5.6

Centroid

12.6 ± 5.1

0.75 0.64-0.82 3.1

0.42 0.29-0.60 5.0


0.89 0.85-0.93 3.6

0.74 0.62-0.81 5.8

Ishihara

0.88 0.82-0.91 3.5

0.55 0.36-0.69 5.2

Ankylosis

13.9 ± 4.6


Figure (TIF or EPS Only!!! Resolution of at least 300 dpi!) Click here to download high resolution image








Comparison of Sagittal Spinopelvic Alignment in Patients With Ankylosing Spondylitis and Thoracolumbar Fracture.

Analysis of sagittal balance of ankylosing spondylitis using spinopelvic parameters.

Cervical spine fracture in advanced ankylosing spondylitis.

Predictive factors of radiographic progression in ankylosing spondylitis.

Cervical spine fracture associated with ankylosing spondylitis.

Cervical fracture in ankylosing spondylitis: value of multidimensional imaging.

Sagittal alignment of cervical spine in adult idiopathic scoliosis.

Management of cervical spine injuries in patients with ankylosing spondylitis.

Correlations of Cervical Sagittal Alignment before and after Occipitocervical Fusion.

Cervical spine injury in patients with ankylosing spondylitis.

The impact of standing regional cervical sagittal alignment on outcomes in posterior cervical fusion surgery.

Cervical spine surgery in ankylosing spondylitis: Review and current concept.

Comparison of radiographic abnormalities of the sacroiliac joint in degenerative disease and ankylosing spondylitis.

Is It Possible to Evaluate the Parameters of Cervical Sagittal Alignment on Cervical CT scan?

The Nonradiographic Axial Spondyloarthritis, the Radiographic Axial Spondyloarthritis, and Ankylosing Spondylitis: The Tangled Skein of Rheumatology.

Anterior Cervical Discectomy and Fusion Alters Whole-Spine Sagittal Alignment.

Body mass index and estrogen predict radiographic progression in the spine in ankylosing spondylitis.

The heart in ankylosing spondylitis.

Treatment of ankylosing spondylitis.

Treatment of ankylosing spondylitis.

Aspects of ankylosing spondylitis.

Ankylosing spondylitis.

Alclofenac in ankylosing spondylitis.

Ankylosing spondylitis in women.