- Email: [email protected]
Reproduction of the Lumbar Lordosis: A Comparison of Standing Radiographs Versus Supine Magnetic Resonance Imaging Obtained with Straightened Lower Extremities
REPRODUCTION OF THE LUMBAR LORDOSIS: A COMPARISON OF STANDING RADIOGRAPHS VERSUS SUPINE MAGNETIC RESONANCE IMAGING OBTAINED WITH STRAIGHTENED LOWER EXTREMITIES Marianne Løgtholt Andreasen, MSc,a Lotte Langhoff, MSc,b Tue Secher Jensen, MSc,c and Hanne B. Albert, PT, MPH, PhDd
ABSTRACT Objective: This study investigates whether it is possible to reproduce the lumbar lordosis in the upright position during
magnetic resonance imaging (MRI) by positioning the patient supine with straightened lower extremities and investigates intra- and interexaminer reliability of measurements of the lumbar lordosis on radiographs and MRI. Methods: This was an observational study, which included an intra- and interexaminer reliability study. The lumbar lordosis was measured digitally on radiographs taken from 22 patients in an upright standing position, and 22 MRI scans of the same patients lying supine with straightened lower extremities. These measurements were compared statistically. Intraand interexaminer reliability was calculated applying the Bland and Altman method. Results: The lumbar lordosis in the standing position was reproduced in the straightened supine position with a median deviation of 38. Intra- and interexaminer reliability was better for MRI than for radiographs. The mean differences were close to 0, especially for interexaminer reliability during MRI. On radiographs, there was a higher agreement on interexaminer than on intra-examiner reliability. Conclusion: The findings of this study show that lumbar lordosis in the upright position can be reproduced by positioning the patient supine with straightened lower extremities. (J Manipulative Physiol Ther 2007;30:26-30) Key Indexing Terms: Lordosis; Reproducibility of Results; Magnetic Resonance Imaging; Radiography
T
he use of diagnostic imaging is common in patients with sciatica. This most commonly entails conventional radiographs, computed tomography, or magnetic resonance imaging (MRI). For maximum value to the clinician and for optimal diagnosis and management, it is important that these images visualize all structures that could be causing the pain. The procedure of obtaining diagnostic images includes optimal positioning of the patient. a Chiropractor, Back Research Centre, Clinical Locomotion Science, Ringe, Denmark. b Chiropractor, Back Research Centre, Clinical Locomotion Science, Ringe, Denmark. c Chiropractor, Back Research Centre, Clinical Locomotion Science, Ringe, Denmark. d Researcher, Back Research Centre, Clinical Locomotion Science, Ringe, Denmark. Submit requests for reprints to: Marianne Løgtholt Andreasen, MSc, The Back Research Centre, Clinical Locomotion Science, Sygehus Fyn, Lindevej 5, 5750 Ringe, Denmark (e-mail: [email protected]). Paper submitted February 8, 2006; in revised form September 5, 2006; accepted October 7, 2006. 0161-4754/$32.00 Copyright D 2007 by National University of Health Sciences. doi:10.1016/j.jmpt.2006.11.009
26
The traditional supine position used during MRI, with slight flexion of hips and knees, creates a hypolordosis of the lumbar spine relative to the standing position. This supine position is often similar to the one patients with radicular pain spontaneously adopt to relieve their symptoms.1 In patients with pain originating from nerve root compromise, disc herniations or protrusions may be invisible on traditionally positioned supine images.2 The relief of leg pain, in the supine flexed position, could be caused by less tension or pressure on the sciatic nerve. A likely explanation for this could be that the lordosis causes changes in the morphology of the spinal canal.3-10 This change is likely to affect the severity of symptoms.2 When a nerve root is irritated, tension in this root will result in radiating symptoms.11-14 In addition to changes in the tension of the sciatic nerve, flexion and extension of the lumbar spine have also been shown to cause changes in the morphology of the spinal canal and intervertebral foramina.2,15,16 A narrowing of the spinal canal by applying axial loading has been documented, especially when extension was also added.3,4,7,10,15 During extension, the sagittal diameter of the spinal canal and dural sac was found to decrease, mainly owing to a thickening of the ligamentum flavum. A posterior bulge of the disc with
Journal of Manipulative and Physiological Therapeutics Volume 30, Number 1
Andreasen et al MRI Lumbar Lordosis
extension has also been implicated, although conflicting opinions exist.6,17 Results from previous studies indicate that the hypolordotic position could reduce disc herniation and increase the spinal canal diameter. It is therefore worthwhile to attempt to retain the lumbar lordosis in the upright position when positioning the patient supine.18-20 Previous studies have attempted to reproduce the standing lumbar lordosis by adding axial compression during supine lumbar MRI or by using vertical magnetic resonance scanners.21-23 As of yet, no studies have been carried out to examine whether the lumbar lordosis could be reproduced simply by positioning the patient supine with straightened lower extremities. If this is the case, an easy, quick, and inexpensive way will have been identified. There were 2 objectives of this study. The first was to investigate whether the lumbar lordosis in the standing position in radiographs can be reproduced by positioning the supine patient with straightened hips and knees during MRI scanning. The second objective was to investigate the intra- and interexaminer reliability of measurements of the lumbar lordosis by radiograph and MRI.
MATERIALS This study is an observational study, including an intraand interexaminer reliability study. The material originated from a doctoral thesis by Albert.24 The subjects were between 18 and 65 years of age, with radicular pain below the knee and a duration of pain between 2 weeks and 1 year. Patients with cauda equine syndrome, previous back surgery, and spinal tumors were excluded. In this study, written consent was obtained from all participants, and approval was given by the local ethics committee of the County of Funen and Vejle before the study began. The project was completed in accordance with the Helsinki Declaration II. Subjects received no treatment during the study. Twenty-two random and consecutively chosen patients were examined by MRI and conventional radiography. Measurements of the lumbar lordosis were carried out on all images. The lumbar lordosis on MRI was compared to that on conventional radiographs to determine whether the lordosis of the radiographs was reproduced. The applied radiograph technology was the Ultra SPARC 5-10, Optimus RAD (Philips, Copenhagen, Denmark). Patient position for the conventional radiographs was the upright standing position. A lateral projection of the lumbar spine was used. The MRI scanner was an open, low-field 0.2T magnetic resonance unit with a body spine surface coil. The most central sagittal T1-weighted image, where the spinous process and the conus medullaris were most clearly identified, was used. Patients were positioned supine with straightened lower extremities. The examiners of the images did not take part in the actual radiograph or MRI procedures. A staff member at the Back Research Centre in Ringe, Denmark, blinded all pictures for any patient information,
Fig 1. For measurement of the lumbar lordosis, lines were drawn parallel to the superior endplates of S1 and L1. The angle of the intersecting lines (AB and DC) was calculated.
making it impossible to match radiograph and MRI during measurements. The pictures were analyzed independently by 2 senior students at the Institute of Sports Science and Clinical Biomechanics, University of Southern Denmark. The applied computer software was Analyze version 5.0 (AnalyzeDirect, Overland Park, Kansas). The angle of the lumbar lordosis was measured. The lumbar lordosis was defined as the angle between a line parallel to the superior endplate of S1 and L1, respectively (Fig 1). For practical reasons, this method is often depicted using the intersecting angle of the perpendicular lines.25 After placing the lines in Analyze, 4 angles appeared. Of these, the 2 at the diverging end of the triangle made up by the 2 lines were added and subtracted from 1808 whereby the angle of the lumbar lordosis was found. These angles and following calculations should be seen as a practical geometrical adaptation. Using a random selection of pictures, we achieved a consensus beforehand regarding the placement of the lines along the superior endplates. On endplates with osteophytes, the lines were placed as close to the original superior angles on the body of the vertebra as possible. Any depressions in the endplates were ignored. On dome-shaped sacral endplates, the lines were placed trans-
27
28
Andreasen et al MRI Lumbar Lordosis
Fig 2. The intra-examiner reliability for radiograph and MRI for both examiners is visualized. Boxes represent 95% LOA for each examiner. Horizontal lines represent the mean differences of an examiner’s first and second measurements. ecting the dome. Before including the images in the study, it was determined whether S1 and L1 were sufficiently visualized to be able to measure the lumbar lordosis. Both examiners completed all of the measurements twice, with approximately 3 weeks between the sets. In each set, the examiners completed the measurements in the same order. Between the sets, the order of the pictures was changed by a staff member. The examiners were blinded to the results of the other examiner, and the first set of their own measurements, until all measurements were completed. All measurements were made in degrees and noted with 1 decimal directly in Excel spreadsheets (Microsoft, Inc, Redmond, Wash). Separate Excel spreadsheets were used for each examiner and for each set of measurements. Also noted was the slice number chosen for analysis in each series of MRI.
DATA ANALYSIS Intra- and interexaminer reliability was described by Bland and Altman’s limit of agreement (LOA).26 The interpretation is that any new measurements with 95% confidence will be within the LOA. The mean differences were also calculated. Deviations of the mean difference from 0 indicate a systematic error. Negative values therefore show the values of the second set of measurements to be larger than the first set and vice versa for the positive values. The reproducibility of the lumbar lordosis was presented as a median value. The median value was found as a mean value of the differences between each examiner’s first and second measurements, on both radiograph and MRI. The difference between the mean differences of radiograph and MRI was calculated and applied. Data analysis was completed using STATA version 8.0 (StataCorp, LLC, College Station, NC).
Journal of Manipulative and Physiological Therapeutics January 2007
Fig 3. The interexaminer reliability for radiograph and MRI between examiners 1 and 2 is visualized. The second measurements of the examiners were applied. Boxes represent 95% LOA for each examiner. Horizontal lines represent the mean differences of an examiner’s first and second measurements.
RESULTS Of the patients in this study, 8 (36%) were women and 14 (64%) were men, with a median age of 49.5 years (range, 27-69 years). In 4 of the MRI series, the examiners chose to measure different slices for their measurements. The chosen images were at most 1 slice apart in the sequence of an MRI series. Insufficient visualization of S1 and L1 caused 2 radiographs to be excluded from the picture material. The median value of the lumbar lordosis on radiograph and MRI was 538 and 488, respectively. The LOA (95% confidence interval) for radiograph and MRI was 478 to 598and 448 to 538, respectively. The overall difference between the angle of the lumbar lordosis in the standing position and the lumbar lordosis in the supine position with straightened lower extremities was 38 (median value). From Figures 2 and 3, it is seen that the LOAs of the MRI measurements were considerably smaller than those of the radiographs. This shows the agreement on MRI measurements to be higher than on radiographs. The mean differences of the intra- and interexaminer reliability were all close to 0, indicating a high level of precision and agreement, which for MRI during the interexaminer reliability study was almost perfect. The majority of the mean differences were negative; the only positive mean difference being the radiograph measurements of examiner 2 in the intra-examiner study.
DISCUSSION This study is the first that we are aware of to investigate whether it is possible to reproduce the lordosis in the upright position by placing patients supine with straightened hips and knees. The difference between the angle of the lumbar lordosis in the standing position and that in the supine position with
Journal of Manipulative and Physiological Therapeutics Volume 30, Number 1
straight legs was 38 (median value). The normal range of the lumbar lordosis is 508 to 608,25 and 38 would amount to a change in the lordosis of only 5% to 6%. We therefore conclude that the lumbar lordosis in the standing position has been essentially reproduced by placing the patient supine with straightened lower extremities. When the lumbar lordosis in the standing position was compared to the straightened supine position, all measurements by the 2 examiners were used, giving a relatively precise estimate of the exact value. The median of 38 of change from the standing to the supine position shows that the 2 positions are comparable. For the 2 positions to be equal, this value should have been 0. However, the deviation from 0 of 38 can be ascribed to human errors and poor radiograph quality. The lumbar lordosis was measured using a method described by Yochum and Rowe.25 In some literature, this method is called the Cobb method.27-29 Some authors have found other methods to be slightly more precise,27,28 but the method described by Yochum and Rowe was chosen because it is easy to apply, and because it is the most widely used method, making it easier to compare results. The applied method only provides information pertaining to the end of the curve without revealing any information about the position of rest of the lumbar vertebrae. This, however, was not considered a problem, because only the angles of the lumbar lordosis relative to one another were studied. The irregularities in the superior endplates of L1 and S1 can provide the examiner with several options when placing the lines. Anticipating that this problem could arise, the examiners discussed the placement of the lines before any measurements were made. Even so, placing the lines on the radiographs caused great difficulties. Especially, the superior endplate of S1 was often poorly visualized, mainly because of the overlying iliac bones or tilted vertebrae. Furthermore, the outlines of structures on the radiographs were often relatively blurred. The LOAs of the intra- and interexaminer reliability of this study span an interval of approximately F128 for radiograph and F78 for MRI. The larger interval for the radiographs is probably explained by the poor quality of the radiographs. The LOAs of MRI can be considered clinically insignificant relative to the normal lumbar lordosis of 508 to 608.25 The results fall within the range of other studies with values ranging from 2.88 to 108.29-32 The poorer quality of the radiographs also shows in the mean differences of the intra- and interexaminer reliability. The values for MRI were all closer to 0 than those for the radiographs. The mean difference of the MRI interexaminer reliability of 0.058 shows almost perfect agreement. The deviation from 0 in regard to the mean differences shows a minor systematic error. Furthermore, to find a mean difference of 0 in a study of this character is rare. The LOA is smaller in the interexaminer study than in the intra-examiner study for radiograph. This is highly unusual as it shows that
Andreasen et al MRI Lumbar Lordosis
the 2 examiners found it easier to reproduce each other’s results than their own. A possible explanation could be that the examiners improved their accuracy from the first set of measurements to the second. To avoid some of the potential human errors, measurements were done digitally, which has been shown to have a slightly lower variability compared to manual methods.29 In a few cases, the examiners chose different MRI slices for their measurements, possibly leading to errors. When commencing this study, the examiners were inexperienced in measuring the lumbar lordosis. But the applied method was quite simple, and the preparations for this study offered the opportunity to practice. The level of experience is therefore unlikely to have had a great influence on the results of this study. It was decided to use the examiners’ second measurements for the interexaminer reliability study as they, at this point, would have the maximum amount of experience. To improve this study, a third set of pictures with the patients’ legs in the traditional flexed position could have been included. This would have offered an opportunity to study the differences between the 2 supine positions, and those in comparison with the upright position. Beattie et al33 showed a mean difference of 12.28 between the flexed and straightened supine position. This result is likely to be similar to what could have been found if a third group had been included in this study. We suggest that future studies include a third set of pictures. Furthermore, the quality of the included radiographs should be optimized. Future studies should be done specifically looking into the diagnostic implications of the reproduced lumbar lordosis.
CONCLUSIONS The results of this study show that it is possible to reproduce the lumbar lordosis in the standing position by positioning the patient supine with straightened lower extremities. This was done without adding axial compression to the patient. It could therefore be recommended that patients in the future are positioned with straight legs during supine lumbar MRI examinations. If future research confirms the diagnostic value of reproducing the lumbar lordosis of the standing position, this underlines the need to modify patient positioning during lumbar MRI scans.
Practical Applications ! The lumbar lordosis in the standing position was reproduced in the supine straightened position with a median deviation of 38. ! Intra- and interexaminer reliability was better for MRI than for x-rays. The mean differences were close to 0, especially for interexaminer reliability during MRI.
29
30
Andreasen et al MRI Lumbar Lordosis
ACKNOWLEDGMENT The authors wish to thank the staff at the radiographic department at the Back Research Centre, Clinical Locomotion Science, in Ringe, Denmark, for assistance in obtaining the material for this project. This study was funded by the regional Institute of Health Sciences Research and the Medical Council of the County of Funen, Denmark. There is no conflict of interest to declare.
REFERENCES 1. Williams MM, Hawley JA, McKenzie RA, Van Wijmen PM. A comparison of the effect of two sitting postures on back and referred pain. Spine 1991;16:1185-91. 2. Weishaupt D, Scmid MR, Zanetti M, Boos N, Romanowski B, Kissling RO, et al. Positional MR imaging of the lumbar spine: does it demonstrate nerve root compromise not visible at conventional MR imaging? Radiology 2000;215:247-53. 3. Danielson B, Wille´n J. Axially loaded magnetic resonance image of the lumbar spine in asymptomatic individuals. Spine 2001;26:2601-6. 4. Wille´n J, Danielson B, Gaulitz A, Niklason T, Schfnstrfm N, Hansson T. Dynamic effects on the lumbar spinal canal: axially loaded CT-myelography and MRI in patients with sciatica and/ or neurogenic claudication. Spine 1997;22:2968-76. 5. Wille´n J, Danielson B. The diagnostic effect from axial loading of the lumbar spine during computed tomography and magnetic resonance imaging in patients with degenerative disorders. Spine 2001;26:2607-14. 6. Fredericson M, Lee SU, Welsh J, Botts K, Norbash A, Carragee EJ. Changes in posterior disc bulging and intervertebral foraminal size associated with flexion-extension movement: a comparison between L4-5 and L5-S1 levels in normal subjects. Spine J 2001;1:10-7. 7. Chung SS, Lee CS, Kim SH, Chung MW, Ahn JM. Effect of low back posture on the morphology of the spinal canal. Skeletal Radiol 2000;29:217-23. 8. Saifuddin A, Blease S, Macsweeney E. Axial loaded MRI of the lumbar spine. Clin Radiol 2003;58:661-71. 9. Fujiwara A, An HS, Lim T, Haughton VM. Morphologic changes in the lumbar intervertebral foramen due to flexionextension, lateral bending and axial rotation. Spine 2001;2: 876-82. 10. Penning L, Wilmink JT. Biomechanics of the lumbosacral dural sac: a study of flexion-extension myelography. Spine 1981;6:398-408. 11. Omura T, Sano M, Omura K, Hasegawa T, Nagano A. A mild acute compression induces neuropraxia in rat sciatic nerve. Int J Neurosci 2004;114:1561-72. 12. Ikeda K, Yokoyama M, Tomita K, Tanaka S. Vulnerability of the gradually elongated nerve to compression injury. Hand Surg 2001;6:29-35. 13. Fern R, Harrisson PJ. The contribution of ischeamia and deformation to conduction block generated by compression of the cat sciatic nerve. Exp Physiol 1994;79:583-92. 14. de Medianiceli L, Leblanc AL, Merle M. Functional consequences of isolated nerve stretch: experimental long-term static loading. J Reconstr Microsurg 1997;13:185-92. 15. Inufusa A, Howard AS, Lim TH, Hasagawa T, Haughton VM, Nowicki BH. Anatomic changes of the spinal canal and intervertebral foramen associated with flexion-extension movement. Spine 1996;21:2412-20.
Journal of Manipulative and Physiological Therapeutics January 2007
16. Scmid MR, Stucki G, Duewell S, Wildermoth S, Romanowski B, Hodler J. Changes in cross-sectional measurements of the spinal canal and intervertebral foramina as a function of body position. AJR Am J Roentgenol 1999;172: 1095-102. 17. Zamani AA, Moriarty T, Hsu L, Winalski CS, Schaffer JL, Isbister H, et al. Functional MRI of the lumbar spine in erect position in a superconducting open-configuration MR system: preliminary results. J Magn Reson Imaging 1998;8: 1329-33. 18. Cho K, Rah U. Study of the lumbar curvature with various factors of pelvic inclination—changes of radiographic lumbar curvature according to hip joint flexion. Yonsei Med J 1995;6: 153-60. 19. Levine D, Wittle MV. The effect of pelvic movement on lumbar lordosis in the standing position. J Orthop Sports Phys Ther 1996;24:130-6. 20. Lord M, Small JM, Dinsay JM, Jocylane M, Watkins R. Lumbar lordosis: effects of sitting and standing. Spine 1997; 22:2571-4. 21. Wisleder D, Smith MB, Mosher TJ, Zatsiorsky V. Lumbar spine mechanical response to axial compression load in vivo. Spine 2001;26:E40329. 22. Madsen R. Simulating spinal canal morphology of upright standing by means of axial loading during supine MRI: a study of spinal stenosis. Undergraduate Research Report, University of Southern Denmark; 2004. 23. Kimura S, Steinbach GC, Watenpaugh DE, Hargens AR. Lumbar spine disk height and curvature response to an axial load generated by a compression device compatible with magnetic resonance imaging. Spine 2001;26: 2596-600. 24. Albert HB. The short term efficacy of active conservative treatment for patients with severe sciatica: a randomized clinical controlled trial. PhD thesis, University of Southern Denmark; 2004. 25. Yochum TR, Rowe LJ. Essentials of skeletal radiology. 3rd ed. Baltimore7 William and Wilkins; 2005. p. 217. 26. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;8476:307-10. 27. Harrison DE, Harrison DD, Cailliet R, Janik TJ, Holland B. Radiographic analysis of lumbar lordosis: Centroid, Cobb, TRALL, and Harrison posterior tangent methods. Spine 2001; 26:E235-42. 28. Chernukha KV, Daffner RH, Reigel DH. Lumbar lordosis measurement: a new method versus Cobb technique. Spine 1998;23:74-9. 29. Shea KG, Stevens PM, Nelson M, Smith JT, Masters KS, Yandow S. A comparison of manual versus computer-assisted radiographic measurement: intraobserver measurement variability for Cobb angles. Spine 1998;23:551-5. 30. Polly DW, Kilkelly FX, McHale KA, Asplund LM, Mulligan M, Chang AS. Measurements of the lumbar lordosis: evaluation of intraobserver, interobserver and technique variability. Spine 1996;21:1530-5. 31. Carman DL, Browne RH, Birch JG. Measurement of scoliosis and kyphosis radiographs: intraobserver and interobserver variation. J Bone Joint Surg 1990;72:328-33. 32. Morrissy RT, Goldsmith GS, Hall EC, Kehl D, Cowie GH. Measurement of the Cobb angle on radiographs of patients with scoliosis. J Bone Joint Surg 1990;72:320-7. 33. Beattie PF, Brooks WM, Rothstein JM, Sibbitt WL, Robergs RA, MacLean T, et al. Effect of lordosis on the position of the nucleus pulposus in supine subjects: a study using magnetic resonance imaging. Spine 1994;19:2096-102.
ABSTRACT Objective: This study investigates whether it is possible to reproduce the lumbar lordosis in the upright position during
magnetic resonance imaging (MRI) by positioning the patient supine with straightened lower extremities and investigates intra- and interexaminer reliability of measurements of the lumbar lordosis on radiographs and MRI. Methods: This was an observational study, which included an intra- and interexaminer reliability study. The lumbar lordosis was measured digitally on radiographs taken from 22 patients in an upright standing position, and 22 MRI scans of the same patients lying supine with straightened lower extremities. These measurements were compared statistically. Intraand interexaminer reliability was calculated applying the Bland and Altman method. Results: The lumbar lordosis in the standing position was reproduced in the straightened supine position with a median deviation of 38. Intra- and interexaminer reliability was better for MRI than for radiographs. The mean differences were close to 0, especially for interexaminer reliability during MRI. On radiographs, there was a higher agreement on interexaminer than on intra-examiner reliability. Conclusion: The findings of this study show that lumbar lordosis in the upright position can be reproduced by positioning the patient supine with straightened lower extremities. (J Manipulative Physiol Ther 2007;30:26-30) Key Indexing Terms: Lordosis; Reproducibility of Results; Magnetic Resonance Imaging; Radiography
T
he use of diagnostic imaging is common in patients with sciatica. This most commonly entails conventional radiographs, computed tomography, or magnetic resonance imaging (MRI). For maximum value to the clinician and for optimal diagnosis and management, it is important that these images visualize all structures that could be causing the pain. The procedure of obtaining diagnostic images includes optimal positioning of the patient. a Chiropractor, Back Research Centre, Clinical Locomotion Science, Ringe, Denmark. b Chiropractor, Back Research Centre, Clinical Locomotion Science, Ringe, Denmark. c Chiropractor, Back Research Centre, Clinical Locomotion Science, Ringe, Denmark. d Researcher, Back Research Centre, Clinical Locomotion Science, Ringe, Denmark. Submit requests for reprints to: Marianne Løgtholt Andreasen, MSc, The Back Research Centre, Clinical Locomotion Science, Sygehus Fyn, Lindevej 5, 5750 Ringe, Denmark (e-mail: [email protected]). Paper submitted February 8, 2006; in revised form September 5, 2006; accepted October 7, 2006. 0161-4754/$32.00 Copyright D 2007 by National University of Health Sciences. doi:10.1016/j.jmpt.2006.11.009
26
The traditional supine position used during MRI, with slight flexion of hips and knees, creates a hypolordosis of the lumbar spine relative to the standing position. This supine position is often similar to the one patients with radicular pain spontaneously adopt to relieve their symptoms.1 In patients with pain originating from nerve root compromise, disc herniations or protrusions may be invisible on traditionally positioned supine images.2 The relief of leg pain, in the supine flexed position, could be caused by less tension or pressure on the sciatic nerve. A likely explanation for this could be that the lordosis causes changes in the morphology of the spinal canal.3-10 This change is likely to affect the severity of symptoms.2 When a nerve root is irritated, tension in this root will result in radiating symptoms.11-14 In addition to changes in the tension of the sciatic nerve, flexion and extension of the lumbar spine have also been shown to cause changes in the morphology of the spinal canal and intervertebral foramina.2,15,16 A narrowing of the spinal canal by applying axial loading has been documented, especially when extension was also added.3,4,7,10,15 During extension, the sagittal diameter of the spinal canal and dural sac was found to decrease, mainly owing to a thickening of the ligamentum flavum. A posterior bulge of the disc with
Journal of Manipulative and Physiological Therapeutics Volume 30, Number 1
Andreasen et al MRI Lumbar Lordosis
extension has also been implicated, although conflicting opinions exist.6,17 Results from previous studies indicate that the hypolordotic position could reduce disc herniation and increase the spinal canal diameter. It is therefore worthwhile to attempt to retain the lumbar lordosis in the upright position when positioning the patient supine.18-20 Previous studies have attempted to reproduce the standing lumbar lordosis by adding axial compression during supine lumbar MRI or by using vertical magnetic resonance scanners.21-23 As of yet, no studies have been carried out to examine whether the lumbar lordosis could be reproduced simply by positioning the patient supine with straightened lower extremities. If this is the case, an easy, quick, and inexpensive way will have been identified. There were 2 objectives of this study. The first was to investigate whether the lumbar lordosis in the standing position in radiographs can be reproduced by positioning the supine patient with straightened hips and knees during MRI scanning. The second objective was to investigate the intra- and interexaminer reliability of measurements of the lumbar lordosis by radiograph and MRI.
MATERIALS This study is an observational study, including an intraand interexaminer reliability study. The material originated from a doctoral thesis by Albert.24 The subjects were between 18 and 65 years of age, with radicular pain below the knee and a duration of pain between 2 weeks and 1 year. Patients with cauda equine syndrome, previous back surgery, and spinal tumors were excluded. In this study, written consent was obtained from all participants, and approval was given by the local ethics committee of the County of Funen and Vejle before the study began. The project was completed in accordance with the Helsinki Declaration II. Subjects received no treatment during the study. Twenty-two random and consecutively chosen patients were examined by MRI and conventional radiography. Measurements of the lumbar lordosis were carried out on all images. The lumbar lordosis on MRI was compared to that on conventional radiographs to determine whether the lordosis of the radiographs was reproduced. The applied radiograph technology was the Ultra SPARC 5-10, Optimus RAD (Philips, Copenhagen, Denmark). Patient position for the conventional radiographs was the upright standing position. A lateral projection of the lumbar spine was used. The MRI scanner was an open, low-field 0.2T magnetic resonance unit with a body spine surface coil. The most central sagittal T1-weighted image, where the spinous process and the conus medullaris were most clearly identified, was used. Patients were positioned supine with straightened lower extremities. The examiners of the images did not take part in the actual radiograph or MRI procedures. A staff member at the Back Research Centre in Ringe, Denmark, blinded all pictures for any patient information,
Fig 1. For measurement of the lumbar lordosis, lines were drawn parallel to the superior endplates of S1 and L1. The angle of the intersecting lines (AB and DC) was calculated.
making it impossible to match radiograph and MRI during measurements. The pictures were analyzed independently by 2 senior students at the Institute of Sports Science and Clinical Biomechanics, University of Southern Denmark. The applied computer software was Analyze version 5.0 (AnalyzeDirect, Overland Park, Kansas). The angle of the lumbar lordosis was measured. The lumbar lordosis was defined as the angle between a line parallel to the superior endplate of S1 and L1, respectively (Fig 1). For practical reasons, this method is often depicted using the intersecting angle of the perpendicular lines.25 After placing the lines in Analyze, 4 angles appeared. Of these, the 2 at the diverging end of the triangle made up by the 2 lines were added and subtracted from 1808 whereby the angle of the lumbar lordosis was found. These angles and following calculations should be seen as a practical geometrical adaptation. Using a random selection of pictures, we achieved a consensus beforehand regarding the placement of the lines along the superior endplates. On endplates with osteophytes, the lines were placed as close to the original superior angles on the body of the vertebra as possible. Any depressions in the endplates were ignored. On dome-shaped sacral endplates, the lines were placed trans-
27
28
Andreasen et al MRI Lumbar Lordosis
Fig 2. The intra-examiner reliability for radiograph and MRI for both examiners is visualized. Boxes represent 95% LOA for each examiner. Horizontal lines represent the mean differences of an examiner’s first and second measurements. ecting the dome. Before including the images in the study, it was determined whether S1 and L1 were sufficiently visualized to be able to measure the lumbar lordosis. Both examiners completed all of the measurements twice, with approximately 3 weeks between the sets. In each set, the examiners completed the measurements in the same order. Between the sets, the order of the pictures was changed by a staff member. The examiners were blinded to the results of the other examiner, and the first set of their own measurements, until all measurements were completed. All measurements were made in degrees and noted with 1 decimal directly in Excel spreadsheets (Microsoft, Inc, Redmond, Wash). Separate Excel spreadsheets were used for each examiner and for each set of measurements. Also noted was the slice number chosen for analysis in each series of MRI.
DATA ANALYSIS Intra- and interexaminer reliability was described by Bland and Altman’s limit of agreement (LOA).26 The interpretation is that any new measurements with 95% confidence will be within the LOA. The mean differences were also calculated. Deviations of the mean difference from 0 indicate a systematic error. Negative values therefore show the values of the second set of measurements to be larger than the first set and vice versa for the positive values. The reproducibility of the lumbar lordosis was presented as a median value. The median value was found as a mean value of the differences between each examiner’s first and second measurements, on both radiograph and MRI. The difference between the mean differences of radiograph and MRI was calculated and applied. Data analysis was completed using STATA version 8.0 (StataCorp, LLC, College Station, NC).
Journal of Manipulative and Physiological Therapeutics January 2007
Fig 3. The interexaminer reliability for radiograph and MRI between examiners 1 and 2 is visualized. The second measurements of the examiners were applied. Boxes represent 95% LOA for each examiner. Horizontal lines represent the mean differences of an examiner’s first and second measurements.
RESULTS Of the patients in this study, 8 (36%) were women and 14 (64%) were men, with a median age of 49.5 years (range, 27-69 years). In 4 of the MRI series, the examiners chose to measure different slices for their measurements. The chosen images were at most 1 slice apart in the sequence of an MRI series. Insufficient visualization of S1 and L1 caused 2 radiographs to be excluded from the picture material. The median value of the lumbar lordosis on radiograph and MRI was 538 and 488, respectively. The LOA (95% confidence interval) for radiograph and MRI was 478 to 598and 448 to 538, respectively. The overall difference between the angle of the lumbar lordosis in the standing position and the lumbar lordosis in the supine position with straightened lower extremities was 38 (median value). From Figures 2 and 3, it is seen that the LOAs of the MRI measurements were considerably smaller than those of the radiographs. This shows the agreement on MRI measurements to be higher than on radiographs. The mean differences of the intra- and interexaminer reliability were all close to 0, indicating a high level of precision and agreement, which for MRI during the interexaminer reliability study was almost perfect. The majority of the mean differences were negative; the only positive mean difference being the radiograph measurements of examiner 2 in the intra-examiner study.
DISCUSSION This study is the first that we are aware of to investigate whether it is possible to reproduce the lordosis in the upright position by placing patients supine with straightened hips and knees. The difference between the angle of the lumbar lordosis in the standing position and that in the supine position with
Journal of Manipulative and Physiological Therapeutics Volume 30, Number 1
straight legs was 38 (median value). The normal range of the lumbar lordosis is 508 to 608,25 and 38 would amount to a change in the lordosis of only 5% to 6%. We therefore conclude that the lumbar lordosis in the standing position has been essentially reproduced by placing the patient supine with straightened lower extremities. When the lumbar lordosis in the standing position was compared to the straightened supine position, all measurements by the 2 examiners were used, giving a relatively precise estimate of the exact value. The median of 38 of change from the standing to the supine position shows that the 2 positions are comparable. For the 2 positions to be equal, this value should have been 0. However, the deviation from 0 of 38 can be ascribed to human errors and poor radiograph quality. The lumbar lordosis was measured using a method described by Yochum and Rowe.25 In some literature, this method is called the Cobb method.27-29 Some authors have found other methods to be slightly more precise,27,28 but the method described by Yochum and Rowe was chosen because it is easy to apply, and because it is the most widely used method, making it easier to compare results. The applied method only provides information pertaining to the end of the curve without revealing any information about the position of rest of the lumbar vertebrae. This, however, was not considered a problem, because only the angles of the lumbar lordosis relative to one another were studied. The irregularities in the superior endplates of L1 and S1 can provide the examiner with several options when placing the lines. Anticipating that this problem could arise, the examiners discussed the placement of the lines before any measurements were made. Even so, placing the lines on the radiographs caused great difficulties. Especially, the superior endplate of S1 was often poorly visualized, mainly because of the overlying iliac bones or tilted vertebrae. Furthermore, the outlines of structures on the radiographs were often relatively blurred. The LOAs of the intra- and interexaminer reliability of this study span an interval of approximately F128 for radiograph and F78 for MRI. The larger interval for the radiographs is probably explained by the poor quality of the radiographs. The LOAs of MRI can be considered clinically insignificant relative to the normal lumbar lordosis of 508 to 608.25 The results fall within the range of other studies with values ranging from 2.88 to 108.29-32 The poorer quality of the radiographs also shows in the mean differences of the intra- and interexaminer reliability. The values for MRI were all closer to 0 than those for the radiographs. The mean difference of the MRI interexaminer reliability of 0.058 shows almost perfect agreement. The deviation from 0 in regard to the mean differences shows a minor systematic error. Furthermore, to find a mean difference of 0 in a study of this character is rare. The LOA is smaller in the interexaminer study than in the intra-examiner study for radiograph. This is highly unusual as it shows that
Andreasen et al MRI Lumbar Lordosis
the 2 examiners found it easier to reproduce each other’s results than their own. A possible explanation could be that the examiners improved their accuracy from the first set of measurements to the second. To avoid some of the potential human errors, measurements were done digitally, which has been shown to have a slightly lower variability compared to manual methods.29 In a few cases, the examiners chose different MRI slices for their measurements, possibly leading to errors. When commencing this study, the examiners were inexperienced in measuring the lumbar lordosis. But the applied method was quite simple, and the preparations for this study offered the opportunity to practice. The level of experience is therefore unlikely to have had a great influence on the results of this study. It was decided to use the examiners’ second measurements for the interexaminer reliability study as they, at this point, would have the maximum amount of experience. To improve this study, a third set of pictures with the patients’ legs in the traditional flexed position could have been included. This would have offered an opportunity to study the differences between the 2 supine positions, and those in comparison with the upright position. Beattie et al33 showed a mean difference of 12.28 between the flexed and straightened supine position. This result is likely to be similar to what could have been found if a third group had been included in this study. We suggest that future studies include a third set of pictures. Furthermore, the quality of the included radiographs should be optimized. Future studies should be done specifically looking into the diagnostic implications of the reproduced lumbar lordosis.
CONCLUSIONS The results of this study show that it is possible to reproduce the lumbar lordosis in the standing position by positioning the patient supine with straightened lower extremities. This was done without adding axial compression to the patient. It could therefore be recommended that patients in the future are positioned with straight legs during supine lumbar MRI examinations. If future research confirms the diagnostic value of reproducing the lumbar lordosis of the standing position, this underlines the need to modify patient positioning during lumbar MRI scans.
Practical Applications ! The lumbar lordosis in the standing position was reproduced in the supine straightened position with a median deviation of 38. ! Intra- and interexaminer reliability was better for MRI than for x-rays. The mean differences were close to 0, especially for interexaminer reliability during MRI.
29
30
Andreasen et al MRI Lumbar Lordosis
ACKNOWLEDGMENT The authors wish to thank the staff at the radiographic department at the Back Research Centre, Clinical Locomotion Science, in Ringe, Denmark, for assistance in obtaining the material for this project. This study was funded by the regional Institute of Health Sciences Research and the Medical Council of the County of Funen, Denmark. There is no conflict of interest to declare.
REFERENCES 1. Williams MM, Hawley JA, McKenzie RA, Van Wijmen PM. A comparison of the effect of two sitting postures on back and referred pain. Spine 1991;16:1185-91. 2. Weishaupt D, Scmid MR, Zanetti M, Boos N, Romanowski B, Kissling RO, et al. Positional MR imaging of the lumbar spine: does it demonstrate nerve root compromise not visible at conventional MR imaging? Radiology 2000;215:247-53. 3. Danielson B, Wille´n J. Axially loaded magnetic resonance image of the lumbar spine in asymptomatic individuals. Spine 2001;26:2601-6. 4. Wille´n J, Danielson B, Gaulitz A, Niklason T, Schfnstrfm N, Hansson T. Dynamic effects on the lumbar spinal canal: axially loaded CT-myelography and MRI in patients with sciatica and/ or neurogenic claudication. Spine 1997;22:2968-76. 5. Wille´n J, Danielson B. The diagnostic effect from axial loading of the lumbar spine during computed tomography and magnetic resonance imaging in patients with degenerative disorders. Spine 2001;26:2607-14. 6. Fredericson M, Lee SU, Welsh J, Botts K, Norbash A, Carragee EJ. Changes in posterior disc bulging and intervertebral foraminal size associated with flexion-extension movement: a comparison between L4-5 and L5-S1 levels in normal subjects. Spine J 2001;1:10-7. 7. Chung SS, Lee CS, Kim SH, Chung MW, Ahn JM. Effect of low back posture on the morphology of the spinal canal. Skeletal Radiol 2000;29:217-23. 8. Saifuddin A, Blease S, Macsweeney E. Axial loaded MRI of the lumbar spine. Clin Radiol 2003;58:661-71. 9. Fujiwara A, An HS, Lim T, Haughton VM. Morphologic changes in the lumbar intervertebral foramen due to flexionextension, lateral bending and axial rotation. Spine 2001;2: 876-82. 10. Penning L, Wilmink JT. Biomechanics of the lumbosacral dural sac: a study of flexion-extension myelography. Spine 1981;6:398-408. 11. Omura T, Sano M, Omura K, Hasegawa T, Nagano A. A mild acute compression induces neuropraxia in rat sciatic nerve. Int J Neurosci 2004;114:1561-72. 12. Ikeda K, Yokoyama M, Tomita K, Tanaka S. Vulnerability of the gradually elongated nerve to compression injury. Hand Surg 2001;6:29-35. 13. Fern R, Harrisson PJ. The contribution of ischeamia and deformation to conduction block generated by compression of the cat sciatic nerve. Exp Physiol 1994;79:583-92. 14. de Medianiceli L, Leblanc AL, Merle M. Functional consequences of isolated nerve stretch: experimental long-term static loading. J Reconstr Microsurg 1997;13:185-92. 15. Inufusa A, Howard AS, Lim TH, Hasagawa T, Haughton VM, Nowicki BH. Anatomic changes of the spinal canal and intervertebral foramen associated with flexion-extension movement. Spine 1996;21:2412-20.
Journal of Manipulative and Physiological Therapeutics January 2007
16. Scmid MR, Stucki G, Duewell S, Wildermoth S, Romanowski B, Hodler J. Changes in cross-sectional measurements of the spinal canal and intervertebral foramina as a function of body position. AJR Am J Roentgenol 1999;172: 1095-102. 17. Zamani AA, Moriarty T, Hsu L, Winalski CS, Schaffer JL, Isbister H, et al. Functional MRI of the lumbar spine in erect position in a superconducting open-configuration MR system: preliminary results. J Magn Reson Imaging 1998;8: 1329-33. 18. Cho K, Rah U. Study of the lumbar curvature with various factors of pelvic inclination—changes of radiographic lumbar curvature according to hip joint flexion. Yonsei Med J 1995;6: 153-60. 19. Levine D, Wittle MV. The effect of pelvic movement on lumbar lordosis in the standing position. J Orthop Sports Phys Ther 1996;24:130-6. 20. Lord M, Small JM, Dinsay JM, Jocylane M, Watkins R. Lumbar lordosis: effects of sitting and standing. Spine 1997; 22:2571-4. 21. Wisleder D, Smith MB, Mosher TJ, Zatsiorsky V. Lumbar spine mechanical response to axial compression load in vivo. Spine 2001;26:E40329. 22. Madsen R. Simulating spinal canal morphology of upright standing by means of axial loading during supine MRI: a study of spinal stenosis. Undergraduate Research Report, University of Southern Denmark; 2004. 23. Kimura S, Steinbach GC, Watenpaugh DE, Hargens AR. Lumbar spine disk height and curvature response to an axial load generated by a compression device compatible with magnetic resonance imaging. Spine 2001;26: 2596-600. 24. Albert HB. The short term efficacy of active conservative treatment for patients with severe sciatica: a randomized clinical controlled trial. PhD thesis, University of Southern Denmark; 2004. 25. Yochum TR, Rowe LJ. Essentials of skeletal radiology. 3rd ed. Baltimore7 William and Wilkins; 2005. p. 217. 26. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;8476:307-10. 27. Harrison DE, Harrison DD, Cailliet R, Janik TJ, Holland B. Radiographic analysis of lumbar lordosis: Centroid, Cobb, TRALL, and Harrison posterior tangent methods. Spine 2001; 26:E235-42. 28. Chernukha KV, Daffner RH, Reigel DH. Lumbar lordosis measurement: a new method versus Cobb technique. Spine 1998;23:74-9. 29. Shea KG, Stevens PM, Nelson M, Smith JT, Masters KS, Yandow S. A comparison of manual versus computer-assisted radiographic measurement: intraobserver measurement variability for Cobb angles. Spine 1998;23:551-5. 30. Polly DW, Kilkelly FX, McHale KA, Asplund LM, Mulligan M, Chang AS. Measurements of the lumbar lordosis: evaluation of intraobserver, interobserver and technique variability. Spine 1996;21:1530-5. 31. Carman DL, Browne RH, Birch JG. Measurement of scoliosis and kyphosis radiographs: intraobserver and interobserver variation. J Bone Joint Surg 1990;72:328-33. 32. Morrissy RT, Goldsmith GS, Hall EC, Kehl D, Cowie GH. Measurement of the Cobb angle on radiographs of patients with scoliosis. J Bone Joint Surg 1990;72:320-7. 33. Beattie PF, Brooks WM, Rothstein JM, Sibbitt WL, Robergs RA, MacLean T, et al. Effect of lordosis on the position of the nucleus pulposus in supine subjects: a study using magnetic resonance imaging. Spine 1994;19:2096-102.