A diagnostic support tool for lumbar spinal stenosis: a self-administered, self-reported history questionnaire

A series of 137 patients with LSS and 97 with lumbar disc herniation who successfully recovered following surgical treatment in our department during 2000 and 2003 were included in this study (Table 1). Patients with cervical myelopathy, diabetic neuropathy, previous surgery, peripheral vascular disease, inflammatory disorders, and degenerative scoliosis (defined as lateral tilting of more than 10 degrees) were excluded. Each patient was evaluated by the study investigators using a standard protocol. Operative and follow-up visit notes were reviewed to determine if stenosis was confirmed intraoperatively and if symptoms improved following surgery. Nerve root compression resulting exclusively from a herniated nucleus pulposus was not considered as a symptom of LSS.

Assessment of history included questions on location, frequency and severity of pain, and on symptoms including numbness, tingling, and provocative factors. The physical examination included a gait-loading test to confirm neurogenic intermittent claudication; this test involves assessment of walking capacity and symptoms, and a neurological examination of motor, sensory, and reflex activity [8]. We investigated symptoms during gait loading and neurological findings just after gait loading. Reflexes were graded from 0 (no response) to 4 (clonus) at the Achilles tendon and patellar tendon, and strength was graded from 0 (no movement) to 5 (normal) at the knee flexors and extensors, ankle dorsiflexors and plantar flexors, and extensor hallucis longus. A pinprick sensation was graded as absent, decreased, or normal at the dorsomedial foot, dorsolateral foot, medial calf, and lateral calf. MRI radiographical reports were abstracted from patient records.

The LSS patients were categorized into radicular and cauda equina types based on history, physical examination, and MRI findings. The radicular type was characterized by symptoms of pain, burning, numbness, and paresthesias following a specific dermatome with radiological evidence of the responsible nerve root compression, which was confirmed if intermittent claudication was abolished following single nerve root infiltration. Patients of the cauda equina type presented some bilateral symptoms related to cauda equina compression syndrome with less dermatomal-specific neurogenic claudication and radiological evidence of cauda equina compression. The cauda euqina type spinal stenosis is distinct from the cauda equina syndrome. A full blown cauda equina syndrome occurs in rare instances in the cauda equina type spinal stenosis. Therefore, urgent surgery is not required in the cauda equina type spinal stenosis. Predictive factors for overlap symptoms between the two types of LSS were derived from the data and factors for predicting the cauda equina type were also determined. Based on the results of univariate analysis for predictors of LSS, a self-administered, self-reported history questionnaire (SSHQ) was developed as a diagnostic support tool for LSS.

This study was performed in six university hospitals, ten medical centers, and thirty one hospitals and clinics affiliated with university hospitals or medical centers during January and March in 2004. A series of 115 patients with LSS gave informed consent to participate in the study and answered the SSHQ before surgery. All these patients recovered following surgery. Patients with cervical myelopathy, diabetic neuropathy, previous surgery, inflammatory disorders, and degenerative scoliosis were excluded. All patients were evaluated by study investigators using the same protocol as that in derivation study 1. Operative and follow-up visit notes were reviewed to determine if stenosis was confirmed intraoperatively and if symptoms improved following surgery. Nerve root compression resulting exclusively from a herniated nucleus pulposus was not considered as a part of LSS syndrome. All LSS patients were categorized into radicular or cauda equina types based on history, physical examination, and MRI findings using the same criteria as those in derivation study 1. There were 55 patients with radicular type LSS and 60 patients with the cauda equina type (Table 2). A responsible nerve root was confirmed if intermittent claudication was abolished following single nerve root infiltration. The sensitivity of each question on the SSHQ was calculated and compared between the radicular and cauda equina types. To assess the cut-off point to distinguish between the types, one point was assigned to each question on the SSHQ, and the clinical prediction rule was defined based on the scores.

We prospectively evaluated the association between the diagnosis of LSS and clinical information, including the history and physical examination of patients with leg symptoms. This study was performed in six university hospitals, ten medical centers, and sixty eight hospitals and clinics affiliated with university hospitals or medical centers during July and September in 2004. We enrolled consecutive patients older than 20 years of age with primary symptoms of pain or numbness in the legs. We excluded patients who have been treated by some medical practices within one year before examination. Patients with cervical myelopathy, previous surgery, degenerative scoliosis (defined as lateral tilting of more than 10 degrees) and inflammatory disorders were also excluded. This study included 250 patients who complained of leg symptoms, including cases of LSS (n = 165), lumbar disc herniation (n = 61), diabetic neuropathy (n = 13), and peripheral vascular disease (n = 11) (Table 3). The study was approved by the institutional review board of each study institution as necessary. Written informed consent was obtained from the all patients. The patients gave informed consent and then answered the SSHQ. The following steps were taken to reach a final diagnosis for each of the enrolled patients (Figure 1). In the first step, at each institution the orthopedic physician who saw a patient made the clinical diagnosis based on the history, physical examination, and radiographic findings. In addition, to verify the diagnosis made by each physician, six board-certified spine surgeons approved by the Japanese Board of Spine Surgery also made a diagnosis for each patient based on the clinical information and findings of the MRI. The opinions of six board-certified spine surgeons approved by the Japanese Board of Spine Surgery were used as the gold standard for diagnosis of LSS. The radicular type was characterized by symptoms of pain, burning, numbness, and paresthesias following a specific dermatome with radiological evidence of the responsible nerve root compression, which was confirmed if intermittent claudication was abolished following single nerve root infiltration. Patients of the cauda equina type presented some bilateral symptoms related to cauda equina compression syndrome with less dermatomal-specific neurogenic claudication and radiological evidence of cauda equina compression.

The sensitivity, specificity, likelihood ratio, and area under the receiver operating characteristic (ROC) curve were estimated. 217 patients classified by investigators as suffering from lower back pain without a significant change in symptoms were given the SSHQ two weeks later during an outpatient visit, and the test-retest reliability over two weeks was investigated in these patients.

History and physical examination variables were dichotomized at clinically sensible cut-off values. Pinprick, strength, and Achilles reflexes were each classified as always normal or with at least 1 abnormal finding. Univariate analyses were performed to derive predictors of LSS using logistic regression analysis. Two-by-two contingency tables were prepared to calculate the sensitivity, specificity, and likelihood ratio of the SSHQ. The area under the ROC curve for the derivation data set was estimated to investigate the internal validity of the clinical prediction rule, and the area under the ROC curve for the validation data set was estimated to examine the external validity. Reliability was investigated based on the reproducibility in the test-retest method. Test-retest analysis was performed in 217 patients with a 14-day period between the first and second tests. Test-retest data were examined graphically by plotting the difference between tests against the mean of the 2 tests [9]. The intraclass correlation coefficient of the SSHQ score for the first and second tests was calculated to confirm reproducibility. The κ coefficient was calculated to examine conformity for each item, based on the following criteria: 0 to < 0.2, poor; 0.2 to < 0.4, fair; 0.4 to < 0.6, moderate; 0.6 to < 0.8, substantial; and > 0.8, almost perfect [10]. All the studies were approved by the ethics committee of Fukushima Medical University.

Key factors for predicting overlapping symptoms between the two types of LSS are shown in Table 4. Five history findings had an odds ratio ≥ 2 or p < 0.05: age > 50, lower-extremity pain or numbness, increased pain when walking, increased pain when standing, and improvement of symptoms on bending forward. No physical examination finding had an odds ratio ≥ 2 or p < 0.05. There was no significant difference in the odds ratio of all the predictive factors except for age. Key factors for predicting the cauda equina type of LSS are shown in Table 5. Six history findings had an odds ratio ≥ 2 or p < 0.05: numbness around the buttocks, walking nearly causes urination, a burning sensation around the buttocks, numbness in the soles of both feet, numbness in both legs, and numbness without pain. Physical examination findings with an odds ratio ≥ 2 or p < 0.05 included the absence of or a weak Achilles reflex response.

Based on the results of univariate analysis for predictors of LSS, we developed the SSHQ as a diagnostic support tool for LSS (see Additional File 1 and 2). The SSHQ included the following questions:

Q1: Numbness and/or pain in the thighs down to the calves and shins.

Q2: Numbness and/or pain increase in intensity after walking for a while, but are relieved by taking a rest.

Q3: Standing for a while brings on numbness and/or pain in the thighs down to the calves and shins.

Q4: Numbness and/or pain are reduced by bending forward.

The key questions for diagnosis of cauda equina symptoms were as follows:

Q5: Numbness is present in both legs.

Q6: Numbness is present in the soles of both feet.

Q7: Numbness arises around the buttocks.

Q8: Numbness is present, but pain is absent.

Q9: A burning sensation arises around the buttocks.

Q10: Walking nearly causes urination.

The sensitivity of each question in the derivation study was calculated for the radicular and cauda equina types of LSS. The sensitivity differed significantly between the categories (Figure 2). To assess the cut-off point to distinguish between the two types, each question was assigned one point. The scores of predictors of cauda equina symptoms (Q5-Q10) were significantly different between the categories, and the cut-off point was two (Figure 3). Based on these results, a clinical prediction rule was defined based on the total scores: a score of 4 points on Q1–Q4 indicates the presence of LSS; a score of 4 on Q1–Q4 and < 1 on Q5–Q10 indicates the radicular type of LSS; and a score of > 1 on Q1–Q4 and > 2 on Q5–Q10 indicates the cauda equina type of LSS.

It took respondents an average of about 1 minute to answer the 10 questions on the SSHQ. Performance indices for the clinical prediction rule are shown in Table 6. The area under the ROC curve was 0.797 in the derivation set and 0.782 in the validation data set (Figure 4). These findings indicate that the SSHQ has both internal and external validity as a diagnostic tool for LSS. The difference between tests plotted against the mean of the tests indicated no obvious relationship or bias (Figure 5). The intraclass correlation coefficient of the SSHQ score for the first and second tests was 0.85, which indicates sufficient reproducibility. One item of the κ coefficient was found to be "fair" (question 8), and all other items were rated as having a conformity of moderate or above.