Research Report

A New Evaluation Method for Lumbar Spinal Instability: Passive Lumbar Extension Test
Background and Purpose. Although many studies have described clinical examination measures for the diagnosis of lumbar spinal instability, few of them have investigated the sensitivity and specificity of the measures that were used. The authors devised a passive lumbar extension (PLE) test for assessing lumbar spinal instability. The purpose of this study was to investigate the sensitivity, specificity, and positive likelihood ratio of this test. Subjects and Methods. The PLE test as well as the instability catch sign, painful catch sign, and apprehension sign tests were done for 122 subjects with lumbar degenerative diseases. The subjects were divided into 2 groupsinstability positive and instability negative on the basis of findings on flexion-extension films of the lumbar spine. The sensitivity, specificity, predictive values, and positive likelihood ratio of each test were investigated. Results. The sensitivity and specificity of the PLE test were 84.2% and 90.4%, respectively. These values were higher than those of other signs. The positive likelihood ratio of the PLE test was 8.84 (95% confidence interval 4.5117.33). Discussion and Conclusion. The PLE test is an effective method for examining patients for lumbar spinal instability and can be performed easily in an outpatient clinic. [Kasai Y, Morishita K, Kawakita E, et al. A new evaluation method for lumbar spinal instability: passive lumbar extension test. Phys Ther. 2006;86:16611667.]

Key Words: Functional radiographs, Lumbar degenerative diseases, Lumbar spine, Physical examinations, Segmental instability. Yuichi Kasai, Koichiro Morishita, Eiji Kawakita, Tetsushi Kondo, Atsumasa Uchida

Physical Therapy . Volume 86 . Number 12 . December 2006

1661

umbar spinal instability is defined as the loss of ability of the spine to maintain its pattern of displacement under physiologic loads with no initial or additional neurological deficit, no major deformity, and no incapacitating pain.1 At present, lumbar spinal instability is diagnosed on the basis of findings on flexion-extension films obtained by lateral lumbar radiography, but there is no consensus regarding the diagnostic radiographic criteria, and the clinical definition of lumbar spinal instability is ambiguous.2 4 Although many studies510 have described clinical examination measures for the diagnosis of lumbar spinal instability, few of them9,10 have investigated the sensitivity and specificity of the measures that were used. Among those few studies, Abbott et al9 reported a passive accessory intervertebral motion test with a sensitivity of 29% and a specificity of 89% and a flexion passive physiological intervertebral motion test with a sensitivity of 5% and a specificity of 99.5% for the diagnosis of translational lumbar spinal instability. These studies showed that there are no clinical examination measures for assessing lumbar spinal instability with both high sensitivity and high specificity. In this study, we investigated the sensitivity, specificity, and positive likelihood ratio of a newly devised passive lumbar extension (PLE) test originating principally from prone instability tests reported by Wadsworth et al11 and Magee.12 Method Subjects The subjects enrolled in this study were 122 consecutive patients who visited the spine clinic of our hospital between January and June 2001 and who were diagnosed as having lumbar spinal canal stenosis (89 patients; 27 patients had the central type, 15 had the lateral type, and 47 had both lateral and central types), lumbar spon-

dylolisthesis (21 patients), or lumbar degenerative scoliosis (12 patients). The subjects, 43 men and 79 women, had a mean age of 68.9 years (range 39 88 years) at the time of the initial consultation. The duration of illness was between 1 month and 5 years (X 11.2 months). The mean Japanese Orthopedic Association ( JOA) score (perfect score 29 points), which was developed to clinically assess the efficacy of treatment for lumbar spine diseases, was 22.6 points (range 529 points). The assessment items of the JOA scoring system evaluated in the interview and clinical examination are pain, gait disturbance, sensory disturbance, muscular power, activities of daily living, and bladder function. Eighty-six patients (70.5%) had lumbago, 74 (60.7%) had intermittent claudication, and 52 (42.6%) had neurological symptoms in the lower legs. Many of our patients are referred to us by other clinics for further evaluation and surgical treatment; of the 122 patients enrolled in this study, 45 (36.9%) underwent spinal decompression and fusion within 1 year after the initial visit. Radiological Evaluation of Lumbar Spinal Instability Several radiographic diagnostic criteria have been proposed for lumbar spinal instability1316; however, at present, there is no consensus in this regard. Therefore, we reviewed the literature to check the cutoff values for angular motion and translational motion used in the evaluation of lumbar spinal instability. The reported cutoff values for angular motion were 10 degrees (Dupuis et al2), 15 degrees (White and Panjabi1 and Nachemson4), and 20 degrees (Hayes et al17); we adopted the highest cutoff value, 20 degrees, for angular motion. The reported cutoff values for translational motion were 3 mm (Dvorak et al18 and Knutsson19), 4 mm (Dupuis et al2), and 5 mm (Shaffer et al15 and Hayes et al17); we used the highest cutoff value, 5 mm,

Y Kasai, MD, is Associate Professor, Department of Orthopaedic Surgery, Mie University Graduate School of Medicine, Tsu City, Mie Prefecture, Japan. Address all correspondence to Dr Kasai at: ykasai@clin.medic.mie-u.ac.jp. K Morishita, MD, is Spine Surgery Fellow, Department of Orthopaedic Surgery, Mie University Graduate School of Medicine. E Kawakita, MD, is Spine Surgery Fellow, Department of Orthopaedic Surgery, Mie University Graduate School of Medicine. T Kondo, MD, is Spine Surgery Fellow, Department of Orthopaedic Surgery, Mie University Graduate School of Medicine. A Uchida, MD, is Professor and Chairman, Department of Orthopaedic Surgery, Mie University Graduate School of Medicine. Dr Kasai provided concept/idea/research design and writing. Dr Kawakita and Dr Kondo provided data collection, and Dr Morishita provided data analysis. Dr Uchida provided project management. An oral presentation of the results of this study was made at the 19th Annual Meeting of the North American Spine Society; October 2530, 2004; Chicago, Ill. This article was received September 5, 2005, and was accepted August 7, 2006. DOI: 10.2522/ptj.20050281

1662 . Kasai et al

Physical Therapy . Volume 86 . Number 12 . December 2006

Figure 2. Figure 1.
The end-plate angle was defined as the angle generated by 1 line drawn from the inferior margin of the superior vertebral body and another line drawn from the superior margin of the inferior vertebral body. Translation was calculated according to the method of Stokes and Frymoyer20; the distance between the 2 arrows was measured, the end-plate angle was obtained from 2 lines drawn from the posterior margins of the superior and inferior vertebral bodies, and then the end-plate angle bisector was drawn.

for this parameter. With respect to angular motion, Maigne et al10 reported that patients showing an intervertebral end-plate angle of less than 5 degrees on the flexion film had significant clinical symptoms relevant to lumbar spinal instability; therefore, we also adopted a cutoff value of 5 degrees for the intervertebral endplate angle on the flexion film. Thus, we used the following 3 criteria to assess radiological instability of the lumbar spine: angular motion of 20 degrees, translational motion of 5 mm, and intervertebral end-plate angle on the flexion film of 5 degrees. We have no evidence justifying the use of these 3 criteria for the assessment of radiological instability of the lumbar spine; however, because the cutoff values adopted for the criteria are the highest among those previously reported, we believe that they constitute a valid method. For practical reasons, we distributed subjects who met 1 or more of the 3 criteria into the lumbar spinal instabilitypositive group and subjects who did not meet any of those criteria into the lumbar spinal instability-negative group. The relationship between 2 vertebrae was assessed on radiographic films for every lumbar vertebra from L12 to L5S1 by 2 independent observers who were orthopedists and had 8 and 14 years of clinical experience. With regard to the measurement methods, the end-plate angle (Fig. 1) was defined as the angle generated by 1 line drawn from the inferior margin of the superior vertebral body and another line drawn from the superior margin of the inferior vertebral body; angular motion was defined as the difference between the end-plate angle obtained from the extension film and that obtained from the flexion film. Translation was calculated according to the
Physical Therapy . Volume 86 . Number 12 . December 2006

method of Stokes and Frymoyer20; the distance between the 2 arrows shown in Figure 2 was measured, the end-plate angle was obtained from 2 lines drawn from the posterior margins of the superior and inferior vertebral bodies, and then the end-plate angle bisector was drawn. The difference between the values measured by the 2 observers was 0.3 0.2 mm (X SD) for translational motion, showing little measurement deviation between the observers and very few errors introduced by magnification. The differences between the values measured by the 2 observers were 1.2 0.6 degrees (X SD) for angle motion, 0.3 0.2 mm (X SD) for translational motion, and 0.2 0.1 mm (X SD) for the intervertebral end-plate angle on the flexion film, showing little measurement deviation between the observers. Eventually, the radiographic assessments of lumbar spinal instability by the 2 physicians coincided for all 122 subjects, revealing a concordance rate of 100%. The evaluation based on these criteria revealed that 38 subjects were instability positive and 84 subjects were instability negative, as shown in Table 1. There were no significant differences between the 2 groups with regard to age, sex, diagnosis, JOA score, or the numbers of subjects who had been surgically treated. PLE Test In the PLE test (Fig. 3) that we have devised, the subject is in the prone position; both lower extremities then are elevated concurrently to a height of about 30 cm from the bed while maintaining the knees extended and gently pulling the legs. The PLE test was implemented by 2 independent orthopedists who had 12 and 15 years of clinical experience. The physicians were unaware of the
Kasai et al . 1663

Table 1.

Clinical Data for Lumbar Spinal Instability-Positive and Lumbar Spinal Instability-Negative Groupsa Lumbar Spinal Instability-Positive Group (n 38) 68.3 12.3 9 men and 29 women LSCS (26), LS (8), LDS (4) 20.8 7.3 18 (47.4) Lumbar Spinal Instability-Negative Group (n 84) 69.1 10.5 34 men and 50 women LSCS (63), LS (13), LDS (8) 23.6 9.6 27 (32.1)

Parameter Age, y, X SD Sex Diagnosis (no. of subjects) JOA score, points, X SD No. (%) of surgically treated subjects
a

Diagnosis: LSCS lumbar spinal canal stenosis, LS lumbar spondylolisthesis, LDS lumbar degenerative scoliosis. JOA Japanese Orthopedic Association.

no abnormality was detected in either test, then the subjects were judged to have negative PLE test results. Instability Catch Sign, Painful Catch Sign, and Apprehension Sign Tests The instability catch sign, painful catch sign, and apprehension sign tests were performed by an orthopedist who had 20 years of clinical experience and who had not implemented the PLE test. These 3 tests were done prior to the PLE test. The PLE test and the other 3 tests for lumbar spinal instability were assessed by different physicians because assessments may be influenced by preconceived ideas if these tests are assessed by same physician. In addition, the tester for these 3 tests also was unaware of the results of the radiological evaluation of lumbar spinal instability. These 3 tests for lumbar spinal instability were described in detail by Kotilainen and Valtonen.5 For the instability catch sign test, subjects were asked to bend their bodies forward as much as possible and then return to the erect position; subjects who were not able to return to the erect position because of sudden low back pain were judged to have lumbar spinal instability. For the painful catch sign test, subjects were asked to lift both lower legs in the knee extension position and then return their legs slowly to the examination table; subjects whose legs fell down instantly to the examination table because of sudden low back pain were judged to have lumbar spinal instability. For the apprehension sign test, subjects were asked whether they had felt a sensation of lumbar collapse because of sudden low back pain when they performed ordinary acts, including bending back and forth or from side to side and sitting down or standing up; subjects who had experienced such a sensation were judged to have lumbar spinal instability. As for other tests, there are no publications reporting the sensitivity and specificity of these 3 clinical examinations for assessing lumbar spinal instability. Because these 3 tests are commonly used in clinical practice to assess lumbar spinal instability at our clinic, we investigated their sensitivity and specificity and compared the results with those of the PLE test.

Figure 3.
Passive lumbar extension test.

results of the radiological assessment of lumbar spinal instability. We surmised that hypermobility derived from lumbar spinal instability would cause low back pain and that because the PLE test was associated with severe hypermobility of the lumbar region, it would induce low back pain. The lumbar region was judged to be abnormal when, during elevation of both lower legs during the PLE test, the subjects complained of strong pain in the lumbar region, including low back pain, very heavy feeling on the low back, and feeling as if the low back was coming off, and such pain disappeared when they returned the lower legs to the initial position. In contrast, subjects complaints of an abnormal sensation, such as mild numbness or a prickling sensation, during this test was not considered abnormal. Because initially the judgment of a positive result (complaint of pain in the lumbar region) was thought to be ambiguous in the PLE test, the test was conducted twice to examine the reproducibility and reliability. The test was repeated 2 to 4 weeks after the first test for the convenience of a follow-up visit. If the subjects complained of strong pain or any abnormal sensation in the lumbar region during the second PLE test, like they did during the initial visit, then they were judged to have positive PLE test results. If the results were evaluated as abnormal in 1 of the 2 PLE tests, either at the initial or at the second visit, then the subjects were judged to have equivocal PLE test results. If
1664 . Kasai et al

Physical Therapy . Volume 86 . Number 12 . December 2006

Table 2.
Data from Passive Lumbar Extension (PLE) Test and Instability Catch Sign, Painful Catch Sign, and Apprehension Sign Tests No. of Subjects With Indicated Test Result: PLE Group Instability positive Instability negative Positive 32 8 Negative 6 76 Instability Catch Sign Positive 10 12 Negative 28 72 Painful Catch Sign Positive 14 23 Negative 24 61 Apprehension Sign Positive 7 10 Negative 31 74

Table 3.
Sensitivity, Specificity, Predictive Values, and Likelihood Ratios of Each Test for Lumbar Spinal Instability Positive Predictive Value (%) 80.0 45.5 37.8 41.2 Negative Predictive Value (%) 92.7 65.5 71.8 70.5 Positive Likelihood Ratio (95% Confidence Interval) 8.84 1.84 1.35 1.55 (4.5117.33) (0.873.89) (0.782.32) (0.643.76)

Test Passive lumbar extension Instability catch sign Painful catch sign Apprehension sign

Sensitivity (%) 84.2 26.3 36.8 18.4

Specificity (%) 90.4 85.7 72.6 88.1

Data Analysis The numbers of subjects who had positive results in the PLE test or the 3 clinical tests for lumbar spinal instability in the instability-positive and instability-negative groups were determined. The sensitivity, specificity, positive and negative predictive values, and positive likelihood ratios of the PLE test and the 3 clinical tests for lumbar spinal instability also were determined. For the data analysis, we created a 2 2 table from the data obtained and calculated sensitivity, specificity, prevalence, positive predictive value, and negative predictive value. The positive likelihood ratio was calculated with the following formula: sensitivity/(1 specificity). The 95% confidence interval of the natural logarithm of the positive likelihood ratio was calculated with the following formula: standard deviation of natural logarithm of the positive likelihood ratio natural logarithm of the positive likelihood ratio. Results PLE Test Of the 38 subjects in the lumbar spinal instabilitypositive group, 32 had positive PLE test results and 6 had negative results (Tab. 2). Of the 84 subjects in the lumbar spinal instability-negative group, 8 had positive PLE test results and 76 had negative results. No subjects had equivocal results, for instance, normal results in the first evaluation and abnormal results in the second evaluation. Table 3 shows the sensitivity, specificity, positive and negative predictive values, and likelihood ratio of the PLE test.

Instability Catch Sign, Painful Catch Sign, and Apprehension Sign Tests Table 2 shows the numbers of subjects who had positive or negative results in the 3 clinical tests for lumbar spinal instability. The sensitivity, specificity, predictive values, and positive likelihood ratio of the PLE test were higher than those of the instability catch sign, painful catch sign, and apprehension sign tests, as shown in Table 3. Discussion The results of this study revealed that as a test for evaluating lumbar spinal instability, the PLE test was more sensitive and specific than the instability catch sign, painful catch sign, and apprehension sign tests. For the PLE test, we believe that the judgment was valid because contradictory results, such as a positive assessment in the first test and a negative assessment in the second test, were not obtained, and the results were assessed similarly. Thus, this test was considered to be highly reproducible. Many of the subjects with positive PLE test results were women, probably because many of the subjects with lumbar spondylolisthesis were women, many of whom were lumbar spinal instability positive. The results of the PLE test did not correlate with those of the JOA score and the presence or absence of neurological symptoms and intermittent limping. Because low back pain probably is attributable to lumbar spinal instability in subjects with positive PLE test results, the data from the PLE test for subjects with lumbar degenerative diseases can be very useful for determining treatment strategy, that is, whether to provide conservative treatment with a corset or to perform procedures such as spinal fusion and
Kasai et al . 1665

Physical Therapy . Volume 86 . Number 12 . December 2006

surgery with spinal instrumentation such as a pedicle screw system and plate fixation. Twenty-five (62.5%) of 40 subjects with positive PLE test results in this study eventually underwent spinal fusion surgery. We cannot conclude only from the results of this study that surgery always is indicated for subjects with positive PLE test results, but we suggest that not simply spinal decompression but spinal fusion should be performed for subjects with positive PLE test results. The radiological assessment of lumbar spinal instability with the criteria established in this study showed that 38 of the 122 subjects had lumbar spinal instability, with a pretest probability (prevalence) of 31.1%. This high prevalence may have occurred because the number of patients at our clinic who were referred from other clinics or requested surgical treatment for lumbar spinal instability at the initial visit was higher than the number of patients with common lumbar degenerative diseases. If a similar study were conducted with general patients with lumbar degenerative diseases, it is thought that the prevalence and the positive predictive value would decrease and the negative predictive value would increase. To assess the external validity, therefore, we suggest that a study needs to be conducted with general patients with lumbar degenerative diseases in the future. On the basis of the sensitivity, specificity, and positive likelihood ratio of the PLE test conducted in this study, however, we believe that the PLE test has high validity. With regard to the evaluation of lumbar spinal instability, some studies have reported that radiographic findings are not always consistent with clinical symptoms.10,2123 The reason may be that when patients perform flexion and extension of the lumbar spine, they use their discretion, fearing the occurrence of lumbar pain, and thus may not actually perform maximal flexion or extension. When the PLE test is conducted, lumbar spinal extension is strong, leading to tension in the anterior longitudinal ligament or the fibrous ring of the intervertebral disk and relaxation in the zygapophyseal capsule, to which surrounding mechanoreceptors or nociceptors react strongly and induce lumbar pain. In this respect, the PLE test serves to assess the reappearance of painful clinical symptoms at the time of simple extension of the lumbar spine because pain is caused by passive extension of the lumbar spine. The limitations of this report are as follows. The sampling was unique because many of the subjects of this study had relatively severe clinical symptoms; the assessment of the pain in the lumbar region by the PLE test was ambiguous; the subjects elevated their lower extremities to a height of about 30 cm in the PLE test, but the height was somewhat variable; and the mechanism of a

positive PLE test result in lumbar spinal instability-positive subjects was not clarified. Therefore, we suggest that studies to solve these limitations are needed in the future. Conclusion The sensitivity and specificity of the PLE test were 84.2% and 90.4%, respectively. These values were higher than those of other tests. The positive likelihood ratio for the PLE test was 8.84; therefore, this test is an effective method for evaluating lumbar spinal instability and can be performed easily in an outpatient clinic.
References
1 White AA, Panjabi MM. The problem of clinical instability in the human spine: a systematic approach, part 4: the lumbar and lumbosacral spine. In: White AA, Panjabi MM, eds. Clinical Biomechanics of the Spine. 2nd ed. New York, NY: JB Lippincott Co; 1990:342361. 2 Dupuis PR, Yong-Hing K, Cassidy JD, et al. Radiologic diagnosis of degenerative lumbar spinal instability. Spine. 1985;10:262276. 3 Kirkaldy-Willis WH, Farfan HF. Instability of the lumbar spine. Clin Orthop. 1982;165:110 123. 4 Nachemson A. Lumbar spine instability: a critical update and symposium summary. Spine. 1985;10:290 291. 5 Kotilainen E, Valtonen S. Clinical instability of the lumbar spine after microdiscectomy. Acta Neurochir. 1993;125:120 126. 6 Tokuhashi Y, Matsuzaki H, Sano S. Evaluation of clinical lumbar instability using the treadmill. Spine. 1993;18:23212324. 7 OSullivan PB. Lumbar segmental instability: clinical presentation and specific stabilizing exercise management. Manual Therapy. 2000;5: 212. 8 Hicks GE, Fritz JM, Delitto A, et al. Interrater reliability of clinical examination measures for identification of lumbar segmental instability. Arch Phys Med Rehabil. 2003;84:1858 1864. 9 Abbott JH, McCane B, Herbison P, et al. Lumbar segmental instability: a criterion-related validity study of manual therapy assessment. BMC Musculoskelet Disord. 2005;6:56 64. 10 Maigne J, Lapeyre E, Morvan G, et al. Pain immediately upon sitting down and relieved by standing up is often associated with radiologic lumbar instability or marked anterior loss of disc space. Spine. 2003; 28:13271334. 11 Wadsworth CT, Di Fabio R, Johnson D. The spine. In: Wadsworth CT, ed. Manual Examination and Treatment of the Spine and Extremities. Baltimore, Md: Williams & Wilkins; 1988:70 71. 12 Magee DJ. Orthopaedic Physical Assessment. 3rd ed. Philadelphia, Pa: WB Saunders; 1997:399. 13 Pitkanen MT, Manninen HI, Lindgren KA, et al. Segmental lumbar spine instability at flexion-extension radiography can be predicted by conventional radiography. Clin Radiol. 2002;57:632 639. 14 Posner I, White AA, Edwards WT, et al. A biomechanical analysis of the clinical stability of the lumbar and lumbosacral spine. Spine. 1982;7:374 389. 15 Shaffer WO, Spratt KF, Weinstein J, et al. The consistency and accuracy of roentgenograms for measuring sagittal translation in the lumbar vertebral motion segment: an experimental model. Spine. 1990;15:741750. 16 Yone K, Sakou T. Usefulness of Posners definition of spinal instability for selection of surgical treatment for lumbar spinal stenosis. J Spinal Disord. 1999;12:40 44.

1666 . Kasai et al

Physical Therapy . Volume 86 . Number 12 . December 2006

17 Hayes MA, Howard TC, Gruel CR, et al. Roentgenographic evaluation of lumbar spine flexion-extension in asymptomatic individuals. Spine. 1989;14:327331. 18 Dvorak J, Panjabi M, Chang DG, et al. Functional radiographic diagnosis of the lumbar spine: flexion-extension and lateral bending. Spine. 1991;16:562571. 19 Knutsson F. The instability associated with disc degeneration in the lumbar spine. Acta Radiol. 1944;25:593 609. 20 Stokes IAF, Frymoyer JW. Segmental motion and instability. Spine. 1987;14:688 691.

21 Nachemson AL. Instability of the lumbar spine: pathology, treatment and clinical evaluation. Neurosurg Clin North Am. 1991;2:785790. 22 Sano S, Yokokura S, Nagata Y, et al. Unstable lumbar spine without hypermobility in postlaminectomy cases: mechanism of symptoms and effect of spinal fusion with and without spinal instrumentation. Spine. 1990;15:1190 1197. 23 Sihvonen T, Lindgren KA, Airaksinen O, et al. Movement disturbances of the lumbar spine and abnormal back muscle electromyographic findings in recurrent low back pain. Spine. 1997;22:289 295.

Physical Therapy . Volume 86 . Number 12 . December 2006

Kasai et al . 1667