Reliability and development of a new classification of lumbosacral spondylolisthesis
- Jean-Marc Mac-Thiong†1, 2, 3Email author,
- Hubert Labelle†1, 3,
- Stefan Parent†1, 2, 3,
- Michael Timothy Hresko†4,
- Vedat Deviren†5,
- Mark Weidenbaum†6 and
- members of the Spinal Deformity Study Group
© Mac-Thiong et al; licensee BioMed Central Ltd. 2008
Received: 14 October 2008
Accepted: 10 December 2008
Published: 10 December 2008
A classification of lumbosacral spondylolisthesis has been proposed recently. This classification describes eight distinct types of spondylolisthesis based on the slip grade, the degree of dysplasia, and the sagittal sacro-pelvic balance. The objectives of this study are to assess the reliability of this classification and to propose a new and refined classification.
Standing posteroanterior and lateral radiographs of the spine and pelvis of 40 subjects (22 low-grade, 18 high-grade) with lumbosacral spondylolisthesis were reviewed twice by six spine surgeons. Each radiograph was classified based on the slip grade, the degree of dysplasia, and the sagittal sacro-pelvic balance. No measurements from the radiographs were allowed. Intra- and inter-observer reliability was assessed using kappa coefficients. A refined classification is proposed based on the reliability study.
All eight types of spondylolisthesis described in the original classification were identified. Overall intra- and inter-observer agreement was respectively 76.7% (kappa: 0.72) and 57.0% (kappa: 0.49). The specific intra-observer agreement was 97.1% (kappa: 0.94), 85.0% (kappa: 0.69) and 88.8% (kappa: 0.85) with respect to the slip grade, the degree of dysplasia, and the sacro-pelvic balance, respectively. The specific inter-observer agreement was 95.2% (kappa: 0.90), 72.2% (kappa: 0.43) and 77.2% (kappa: 0.69) with respect to the slip grade, the degree of dysplasia, and the sacro-pelvic balance, respectively.
This study confirmed that surgeons can classify radiographic findings into all eight types of spondylolisthesis. The intra-observer reliability was substantial, while the inter-observer reliability was moderate mainly due to the difficulty in distinguishing between low- and high-dysplasia. A refined classification excluding the assessment of dysplasia, while incorporating the assessment of the slip grade, sacro-pelvic balance and global spino-pelvic balance is proposed, and now includes five types of lumbosacral spondylolisthesis.
Recently, Mac-Thiong and Labelle  proposed a new classification of spondylolisthesis specifically intended to guide the evaluation and treatment of lumbosacral spondylolisthesis. This classification clarifies the concept of low- and high-dysplasia introduced by Marchetti and Bartolozzi  and incorporates recent knowledge in the study of sagittal sacro-pelvic balance and morphology. Eight distinct types of spondylolisthesis are described in this classification based on the following: 1) slip grade (low- vs. high-grade), 2) degree of dysplasia (low- vs. high-dysplastic), and 3) sagittal sacro-pelvic balance. The classification organizes groups and subgroups in an ascending order of severity to develop an associated intuitive surgical algorithm of increasingly complex surgery as the severity of the spondylolisthesis increases.
The radiological files of all patients with lumbosacral developmental spondylolisthesis seen for the first time at the spine clinic of a paediatric hospital between July 1st 2001 and March 1st 2006 were reviewed. Patients were defined as potential subjects for inclusion in the study if they had postero-anterior (PA) and lateral (LAT) standing radiographs of the spine and pelvis showing both femoral heads. All subjects with a history or clinical signs of hip, pelvic or lower limb disorder were excluded. For patients who had undergone surgical treatment of spondylolisthesis, only the preoperative radiographs were considered. All 18 patients with high-grade spondylolisthesis were included. Twenty-two patients were randomly selected from the remaining 108 patients with low-grade spondylolisthesis. Thus a total of 40 subjects with spondylolisthesis (18 high-grade and 22 low-grade) were available for this study. All radiographs used in the current study were retrieved from the PACS system by an independent observer and there was no header or any information on the radiographs in order to minimize potential sources of bias. The mean percentage of slip was 16 ± 8% (range: 4–44%) for subjects with low-grade spondylolisthesis and 80 ± 17% (range: 53–100%) for subjects with high-grade slips. The mean age was 14.7 ± 2.9 years (range: 7.9–20.0 years).
Six spine surgeons from four different institutions classified all 40 subjects twice (at a one week interval), based on digital standing PA and LAT radiographs of the spine and pelvis viewed on a computer screen. The observers were allowed to view the radiographs with the software of their choice. They classified the spondylolisthesis for each subject into one of eight types described by the classification system provided in Figure 2. In this classification system, the slip grade, the degree of dysplasia and the sacro-pelvic balance are determined visually. Quantitative criteria for slip grade and sacro-pelvic balance are given as a reference, while qualitative criteria are provided for the degree of dysplasia. However, observers were not allowed to make any measurements on the radiographs but only made visual estimates in order to classify each case.
Statistical analysis was performed by a biostatistician from PhDx Systems Inc (Albuquerque, NM). Classification reliability was assessed by calculating the intra- and inter-observer percentage of agreement, as well as the kappa coefficients. The resulting kappa values were interpreted based on the recommendations of Landis and Koch  (< 0: poor agreement; 0–0.20: slight agreement; 0.21–0.40: fair agreement; 0.41–0.60: moderate agreement; 0.61–0.80: substantial agreement; 0.81–1: almost perfect agreement).
Number of cases identified by each observer for each spondylolisthesis type
Intra- and inter-observer reliability
Degree of dysplasia
The lowest reliability was associated with the degree of dysplasia (low- vs. high-dysplastic) wherein intra-observer agreement was substantial (kappa: 0.69) but inter-observer agreement was only moderate (kappa: 0.43). Seventeen cases resulted in agreement between all six observers, and eight cases had agreement for five observers. However, 15 subjects resulted in agreement for only three or four of the six observers, indicating that in these cases agreement was only related to chance. Of those 15 subjects, nine had low-grade spondylolisthesis and six had high-grade spondylolisthesis.
This study evaluated the reliability of a classification previously proposed for lumbosacral spondylolisthesis [21, 22]. First, it confirmed that surgeons can classify radiographic findings into all eight types in a random sample of subjects with spondylolisthesis. Second, the classification system yielded substantial intra-observer reliability (kappa: 0.72) and a moderate inter-observer reliability (kappa: 0.49). These values are similar to those reported in comparing the reliability of the King et al.  and Lenke et al.  classification systems for adolescent idiopathic scoliosis , although the current classification is a "visual semi-quantitative" classification system. Richards et al.  found intra- and inter-observer kappa coefficients of 0.81 and 0.61 respectively for the King et al. classification, while the intra- and inter-observer kappa values were respectively 0.60 and 0.50 for the Lenke et al. classification. Other studies have also reported kappa values in the same range for the King et al. classification (intra-observer kappa: 0.64; inter-observer kappa: 0.44)  and the Lenke et al. classification (intra-observer kappa: 0.73; inter-observer kappa: 0.62) .
It is expected that the clinical reliability of the classification could be improved if direct measurements on the radiographs were made. Indeed, the fact that no measurements were allowed in this study may explain most of the cases for which there was disagreement concerning slip grades around 50% (Figure 3), although the intra- and inter-observer reliability was almost perfect for slip grade assessment. Measurement of pelvic tilt and sacral slope could undoubtedly improve the reliability associated with sagittal sacro-pelvic balance, as the associated intra-observer reliability was almost perfect but the inter-observer reliability was substantial. Consequently, future development of the classification should attempt to increase the reliability concerning the assessment of the slip grade and of the sacro-pelvic balance, while keeping the classification simple to use clinically. In the future, classification of cases will be made using a software that requires identification of only a limited number of anatomical landmarks and that automatically provides the slip grade and sacro-pelvic balance.
Furthermore, this study demonstrated that most of the disagreement in the classification centered in determination of the degree of dysplasia when using only qualitative criteria. Although intra-observer agreement for the degree of dysplasia was substantial (kappa: 0.69), inter-observer agreement was only moderate (kappa: 0.43). As seen in Figure 6, each observer developed his own interpretation to discriminate low- from high-dysplasia. The problem with determination of the degree of dysplasia is therefore related to the high variability between observers concerning the interpretation of the criteria. This source of inter-observer disagreement could surely be reduced by the implementation of quantitative criteria such as those proposed in the original classification , although these could be difficult to apply clinically in their present form due to their complexity. It is possible that determination of the degree of dysplasia with the aid of CT scan images and/or MRI could have improved the reliability of the classification.
The reliability for classifying patients could also be improved by using Ferguson and lateral lumbosacral junction views, in addition to the PA and LAT radiographs of the complete spine and pelvis that were used in this study. The reduced visibility of the lumbosacral junction on PA and LAT radiographs of the complete spine and pelvis might explain part of the discrepancies in this study, especially for high-grade subjects.
Proposal of the refined SDSG classification of lumbosacral spondylolisthesis (Table 3)
SDSG classification system of lumbosacral spondylolisthesis
Normal pelvic incidence
High pelvic incidence
In the perspective of developing a surgical treatment algorithm for lumbosacral spondylolisthesis, it is important to propose a classification system that is simple to use clinically and highly reliable. Due to the moderate inter-observer reliability for the assessment of the degree of dysplasia and also due to the difficulty to define reliable quantitative criteria for dysplasia, we have decided to exclude the assessment of dysplasia from the classification system.
In addition, recent work from the SDSG has led us to modify the determination of spino-pelvic balance in low-grade spondylolisthesis. A study (unpublished data, manuscript under preparation) conducted on 257 patients with low-grade spondylolisthesis and using K-means cluster analysis of sacro-pelvic parameters (pelvic incidence, sacral slope and pelvic tilt) showed that these patients were divided into two distinct groups: a group with normal or near normal pelvic incidence (< 60°) and a group with high pelvic incidence (≥ 60°). The low PI/low SS and high PI/high SS groups mentioned in the original classification system are in fact subtypes of the two groups recently described (pelvic incidence < 60° vs. pelvic incidence ≥ 60°).
Also, it is now recognized that preservation or restoration of an adequate global sagittal balance is of prime importance in the management of spinal deformity, so that assessment of global balance has been introduced into the SDSG classification. Two studies [29, 30] have demonstrated that sagittal balance is related to the health-related quality of life in spinal deformity. Furthermore, from their study on 120 controls, 91 subjects with low-grade spondylolisthesis and 40 subjects with high-grade spondylolisthesis, Mac-Thiong et al.  have shown the importance of evaluating the global sagittal balance with respect to the position of the femoral heads (spino-pelvic balance).
The revised classification of lumbosacral spondylolisthesis supported by the SDSG is based on three important characteristics that can be assessesd from the preoperative imaging studies: 1) the grade of slip, 2) the sacro-pelvic balance, and 3) the global spino-pelvic balance. Accordingly, five different types of spondylolisthesis have been identified (Table 3). To classify a patient, the degree of slip is quantified first from the lateral radiograph, in order to determine if it is low-grade (grades 0, 1 and 2, or < 50% slip) or high-grade (grades 3, 4 and spondyloptosis, or ≥ 50% slip). Next, the sagittal balance is measured by determining sacro-pelvic and global spino-pelvic balance, using measurements of pelvic incidence (PI), sacral slope (SS), pelvic tilt (PT) and the C7 plumbline.
This study evaluated the reliability of a classification previously proposed for lumbosacral spondylolisthesis. The reliability study showed suboptimal inter-observer reliability regarding the assessment of dysplasia, but excellent reliability for the slip grade and sacro-pelvic balance. A refined classification excluding the assessment of dysplasia, while incorporating the assessment of the slip grade, sacro-pelvic balance and global spino-pelvic balance is proposed, and now includes five types of lumbosacral spondylolisthesis. In addition, the reliability of the classification is expected to increase with direct measurements from the radiographs.
- Wiltse LL, Newman PH, Macnab I: Classification of spondylolysis and spondylolisthesis. Clin Orthop. 1976, 117: 23-29.PubMed
- Marchetti PC, Bartolozzi P: Classification of spondylolisthesis as a guideline for treatment. The Textbook of Spinal Surgery. Edited by: Bridwell KH, DeWald RL, Hammerberg KW, et al. 1997, Philadelphia: Lippincott-Raven, 1211-1254. 2
- Herman MJ, Pizzutillo PD, Cavalier R: Spondylolysis and spondylolisthesis in the child and adolescent athlete. Orthop Clin N Am. 2003, 34: 461-467. 10.1016/S0030-5898(03)00034-8.View Article
- Lonstein JE: Spondylolisthesis in children. Cause, natural history, and management. Spine. 1999, 24: 2640-2648. 10.1097/00007632-199912150-00011.PubMedView Article
- Newton PO, Johnston CE: Analysis and treatment of poor outcomes following in situ arthrodesis in adolescent spondylolisthesis. J Pediatr Orthop. 1997, 17: 754-761. 10.1097/00004694-199711000-00010.PubMed
- Smith JA, Hu SS: Management of spondylolysis and spondylolisthesis in the pediatric and adolescent population. Orthop Clin North Am. 1999, 30: 487-499. 10.1016/S0030-5898(05)70101-2.PubMedView Article
- Curylo LJ, Edwards C, DeWald RW: Radiographic markers in spondyloptosis. Implications for spondylolisthesis progression. Spine. 2002, 27: 2021-2025. 10.1097/00007632-200209150-00010.PubMedView Article
- During J, Goudfrooij H, Keessen W: Towards standards for posture. Postural characteristics of the lower back system in normal and pathologic conditions. Spine. 1985, 10: 83-87. 10.1097/00007632-198501000-00013.PubMedView Article
- Hanson DS, Bridwell KH, Rhee JM: Correlation of pelvic incidence with low- and high-grade isthmic spondylolisthesis. Spine. 2002, 27: 2026-2029. 10.1097/00007632-200209150-00011.PubMedView Article
- Hresko MT, Labelle H, Roussouly P: Classification of high-grade spondylolisthesis based on pelvic version and spine balance: possible rationale for reduction. Spine. 2007, 32: 2208-2213.PubMedView Article
- Jackson RP, Phipps T, Hales C: Pelvic lordosis and alignment in spondylolisthesis. Spine. 2003, 28: 151-160. 10.1097/00007632-200301150-00011.PubMedView Article
- Labelle H, Roussouly P, Berthonnaud É: Spondylolisthesis, pelvic incidence, and spinopelvic balance. A correlation study. Spine. 2004, 29: 2049-2954. 10.1097/01.brs.0000138279.53439.cc.PubMedView Article
- Labelle H, Roussouly P, Berthonnaud É: The importance of spino-pelvic balance in L5-S1 developmental spondylolisthesis. A review of pertinent radiological measurements. Spine. 2005, 30: S27-S34. 10.1097/01.brs.0000155560.92580.90.PubMedView Article
- Marty C, Boisaubert B, Descamps H: The sagittal anatomy of the sacrum among young adults, infants, and spondylolisthesis patients. Eur Spine J. 2002, 11: 119-125. 10.1007/s00586-001-0349-7.PubMedPubMed CentralView Article
- Rajnics P, Templier A, Skalli W: The association of sagittal spinal and pelvic parameters in asymptomatic persons and patients with isthmic spondylolisthesis. J Spinal Disord Tech. 2002, 15 (1): 24-30.PubMedView Article
- Roussouly P, Berthonnaud É, Schwender J: Changes in spinal and pelvic sagittal parameters following surgery for high-grade isthmic spondylolisthesis. Eur Spine J. 2001, 10: S19-
- Roussouly P, Gollogly S, Berthonnaud É: Sagittal alignment of the spine and pelvis in the presence of L5-S1 isthmic lysis and low-grade spondylolisthesis. Spine. 2006, 31: 2484-2490. 10.1097/01.brs.0000239155.37261.69.PubMedView Article
- Schwab FJ, Farcy JPC, Roye DP: The sagittal pelvic tilt index as a criterion in the evaluation of spondylolisthesis: Preliminary observations. Spine. 1997, 22: 1661-1667. 10.1097/00007632-199707150-00026.PubMedView Article
- Vidal J, Marnay T: Morphology and anteroposterior body equilibrium in spondylolisthesis L5-S1 [in French]. Rev Chir Orthop Reparatrice Appar Mot. 1983, 69: 17-28.PubMed
- Whitesides TE, Horton WC, Hutton WC: Spondylotic spondylolisthesis. A study of pelvic and lumbosacral parameters of possible etiologic effect in two genetically and geographically distinct groups with high occurrence. Spine. 2005, 30: S12-S21. 10.1097/01.brs.0000155574.33693.60.PubMedView Article
- Mac-Thiong J-M, Labelle H: A proposal for a surgical classification of pediatric lumbosacral spondylolisthesis based on current literature. Eur Spine J. 2006, 15: 1425-1435. 10.1007/s00586-006-0101-4.PubMedView Article
- Mac-Thiong J-M, Labelle H: Classification of pediatric lumbosacral spondylolisthesis. Stud Health Technol Inform. 2006, 123: 141-145.PubMed
- Landis JR, Koch GG: The measurement of observer agreement for categorical data. Spondylotic spondylolisthesis. Biometrics. 1977, 33: 159-174. 10.2307/2529310.PubMedView Article
- King HA, Moe JH, Bradford DS: The selection of fusion levels in thoracic idiopathic scoliosis. J Bone Joint Surg Am. 1983, 65: 1302-1303.PubMed
- Lenke LG, Betz RR, Harms J: Adolescent idiopathic scoliosis. A new classification to determine extent of spinal arthrodesis. J Bone Joint Surg Am. 2001, 83: 1169-1181.PubMed
- Richards BS, Sucato DJ, Konigsberg DE: Comparison of reliability between the Lenke and King classification systems for adolescent idiopathic scoliosis using radiographs that were not premeasured. Spine. 2003, 28: 1148-1157. 10.1097/00007632-200306010-00012.PubMed
- Cummings RJ, Loveless EA, Campbell J: Interobserver reliability and intraobserver reproducibility of the system of King et al. for the classification adolescent idiopathic scoliosis. J Bone Joint Surg Am. 1998, 80: 1107-1111.PubMed
- Ogon M, Giesinger K, Behensky H: Interobserver and intraobserver reliability of Lenke's new scoliosis classification system. Spine. 2002, 27: 858-863. 10.1097/00007632-200204150-00014.PubMedView Article
- Glassman SD, Bridwell K, Dimar JR: The impact of positive sagittal balance in adult spinal deformity. Spine. 2005, 30: 2024-2029. 10.1097/01.brs.0000179086.30449.96.PubMedView Article
- Mac-Thiong J-M, Transfeldt EE, Mehbod AA: Can C7 plumbline and gravity line predict health related quality of life in adult scoliosis?. Spine. in revision
- Mac-Thiong J-M, Wang Z, de Guise JA: Postural model of sagittal spino-pelvic balance and its relevance for lumbosacral developmental spondylolisthesis. Spine. 33 (21): 2316-2325.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.