Japanese Orthopaedic Association (JOA) clinical practice guidelines on the management of lumbar spinal stenosis, 2021 - Secondary publication

Using the Swedish International Statistical Classification of Diseases and Related Health Problems (ICD) code, Knutsson et al. prospectively investigated if smoking was a risk factor leading to surgery for LSS. Of a total of 331,941 participants classified according to smoking category, 16% were former smokers, 26% were moderate smokers, and 14% were heavy smokers. Over the 30.7-month follow-up period, compared with non-smokers, all smoking categories were associated with the surgery rate. Heavy smokers had a higher risk of undergoing surgery rate than moderate or former smokers [].

Adamova et al. examined 56 patients who were followed up for an average of 88 months of 151 patients with mild-to-moderate lumbar spinal stenosis. Within this period, various conservative treatments were given, and 34 patients (60.7%) achieved satisfactory results (improvement or no change). Six patients (10.7%) underwent surgery due to the exacerbation of symptoms in an average of 28 months (range 5–50 months). In electromyography, an abnormal H-wave finding of the soleus muscle was a poor prognostic factor [].

Fukushima et al. examined predictors of surgery in 274 patients with LSS diagnosed based on clinical symptoms and MRI findings. Of the 185 patients underwent various conservative treatments and could be followed up for 3 years (follow-up rate of 67.3%), 82 patients (44.3%) underwent surgery. Multivariate analysis revealed that cauda equina syndrome (OR = 3.38) and degenerative spondylolisthesis/scoliosis (OR = 2.00) were significant predictors leading to surgery []. Minamide et al. conducted a longitudinal study over a minimum period of 10 years in 34 patients with LSS who underwent conservative treatment. Twenty-nine patients could be followed up (average follow-up period of 11.1 years), and evaluation of the clinical course by the Japanese Orthopaedic Association (JOA) score rating system for low back pain showed that LSS was exacerbated in 38%, showed no change in 31%, and was alleviated in 31%. The cross-sectional area of the spinal canal of the exacerbation group was significantly smaller than that in the no-change group and the alleviation group. Of the nine patients in the exacerbation group, six who underwent surgery had an initial cross-sectional area of the dural sac of <50 mm² []. Wessberg et al. investigated 107 patients who responded to a questionnaire (follow-up rate of 89%, average follow-up period of 3.3 years) among 129 patients who could be followed up of 146 patients diagnosed with moderate LSS based on clinical symptoms and imaging findings (MRI and CT). The minimal clinically important differences of the visual analog scale (VAS) for leg pain and low back pain improved in 32% and 36%, remained unchanged in 55% and 54%, and deteriorated in 13% and 10%, respectively. Gait disorder was unchanged in 49%, improved in 29%, and deteriorated in 22%. In addition, 10 patients underwent surgery (7%). Patients with cross-sectional area of the dural sac <0.5 cm² did not show any improvement in clinical outcomes []. Adamova et al. investigated 53 of 153 patients with mild to moderate LSS (numerical rating scale, 0–6, Oswestry Disability Index, 0%–60%, and walking distance 21 m or more). The average follow-up period was 139 months. During this period, various conservative treatments were given according to the symptoms, and 54.7% of the patients showed satisfactory results (improvement or no change). Seven patients (13.2%) underwent surgery due to exacerbation of symptoms over an average of 34 months. Multivariate analysis revealed that a spinal canal transverse diameter on CT of 13.6 mm or less was correlated with unsatisfactory results. However, abnormal findings on electromyography, which were significant poor prognostic factors over the average follow-up period of 88 months mentioned above, were not significant poor prognostic factors [].

Medical history useful for the diagnosis of LSS for which consensus has been reached includes gluteal or lower extremity symptom exacerbated by walking or standing which improves or resolves with sitting or bending forward in the middle-aged and elderly. The North American Spine Society (NASS) guidelines also specify these findings as medical history and examination findings that are useful for the diagnosis of LSS and state that if the symptoms are not exacerbated by walking, it is highly possible that it is not LSS []. An international survey by the International Society for the Study of the Lumbar Spine (ISSLS) Taskforce reported that the medical history that spine specialists consider important in the diagnosis of LSS includes the following seven items: whether the patient 1) experiences leg or buttock pain while walking, 2) flexes forward to relieve symptoms, 3) experiences relief when using a shopping cart or bicycle, 4) experiences motor or sensory disturbance while walking, 5) experiences normal and symmetric foot arterial pulses, 6) experiences lower limb muscle weakness, or 7) has low back pain []. In an observational study that included 1448 patients with low back pain or low back pain and leg pain, Chad et al. reported that if 1) the symptoms are bilateral, 2) leg pain is stronger than low back pain, 3) pain is experienced while walking and in the standing position, 4) sitting is more comfortable, and 5) the patients’ age is 48 years or more, LSS is highly possible (sensitivity, 0.96), whereas the probability of another condition is very low (likelihood ratio, LR: 0.19) []. On the other hand, Ogata et al. retrospectively investigated the interview sheets at the first visit of 531 patients diagnosed with LSS and reported that there were many complaints of low back pain (77.8%), leg pain (69.3%), leg numbness (77.2%), and gait disturbance (62.3%) but few typical symptoms of exacerbation when the back was stretched (27.5%) and flexed forward to relieve symptoms (19.2%) [].

Although neurogenic intermittent claudication is a characteristic finding of LSS, there are patients who can continue walking despite the manifestation of symptoms in the lower extremities; therefore, intermittent claudication cannot be considered as an essential medical history. Okoro et al. reported that the accuracy of patient-reported walking distances was low; in particular, patients aged 60 years or more tended to underreport [].

The presence or absence of low back pain is not emphasized in the diagnostic definition of the NASS guidelines. Hara et al. compared 274 patients with LSS with 212 individuals with no symptom in the lower extremities and reported that the LSS group had a higher complaint rate of low back pain (88%) and degree of pain. They reported that the incidence of low back pain with intermittent claudication (low back pain that becomes painful after walking for a while and is alleviated after sitting and resting) was significantly higher in the LSS group (85%) than that in the control group (38%) [].

With regard to other medical histories that would be helpful, it has been reported that patients with LSS experience frequent nocturnal leg cramps [].

Neurological examination findings in the lower extremities and the Kemp's test are useful, but they are not considered highly specific for LSS. Kemp's test has been reported to be positive in 60.8% of patients for L5 nerve root disorder caused by spinal stenosis at the L4-5 level and in 79.6% of patients for L5 nerve root disorder caused by foraminal stenosis at the L5-S level []. Adachi et al. indicated that tenderness of tibial nerves in the popliteal region in the tibial nerve compression test is useful for the diagnosis of LSS and reported that 92.6% of 108 patients with LSS were positive, which was a significantly higher rate than that among healthy controls [].

Based on subjective symptoms and objective findings caused by walking, Kikuchi et al. classified neurogenic intermittent claudication into three types, namely, the cauda equina type, which presents multi-radicular disorder characterized by sensory disorders of the lower extremities, buttocks, and perineal region; nerve root type, which presents mono-radicular disorder characterized by pain in the lower extremities and buttocks; and the combined type, which combines both types []. It has been reported that in the cauda equina type, intermittent claudication is not improved by selective nerve root block, but in the nerve root type, intermittent claudication temporarily disappears after a block.

It is important to differentiate intermittent claudication, a characteristic symptom of LSS, from vascular intermittent claudication, such as peripheral arterial disease (PAD). In the latter, lower limb pain is alleviated by just stopping walking, regardless of the posture. Diagnosis is made in combination with examination findings, such as the presence or absence of pulses of the dorsalis pedis artery or posterior tibial artery, while keeping in mind that it may be impalpable even if normal, and the measurement of ankle brachial pressure index (ABI). Attention is required because patients with poor pulsation of the dorsalis pedis artery, low ABI, and difficulty in responding to block treatment are at risk of arteriosclerosis obliterans []. In addition, it should be noted that they may occur in combination. It has been reported that 6.7% of patients with LSS experience LSS with PAD (LSSPAD), and the factors correlated with LSSPAD were found to be old age, diabetes mellitus, cerebrovascular disease, and ischemic heart disease [].

With regard to comorbidities, multiple studies have reported hypertension and diabetes mellitus as risk factors for symptomatic LSS. A large-scale multicenter study in Japan reported a high complication rate of heart diseases and hypertension in patients with LSS []. In addition, a population surveillance study involving 2666 respondents out of 4400 individuals randomly selected from the Japanese population nationwide stated that old age (OR 5.38, 95% confidence interval 2.03–14.21), diabetes mellitus (OR 2.05, 95% confidence interval 1.14–3.67), urination disorder (OR 2.17, 95% confidence interval 1.10–4.29), a combination of osteoarthritis and fracture (OR 2.71, 95% confidence interval 1.53–4.82) and severe depression (OR 3.55, 95% confidence interval 1.97–6.40) were correlated with symptomatic LSS []. Lotan et al. compared 537 patients with LSS with the control group data of the national statistical survey database of Israel and reported that the LSS had a significantly higher prevalence of hypertension, i.e., 23.2% (7.8% for the control group), and diabetes mellitus, i.e., 13.6% (5.9% for the control group), and there was no significant difference for ischemic heart disease (11.9%) and hyperlipidemia (4.4%) []. Uesugi et al. compared 526 patients with LSS aged ≥50 years who visited 64 institutions nationwide with a control group composed of 1218 men and 1636 women aged ≥50 years extracted by stratified random sampling from the National Health and Nutrition Survey 2006 in Japan. They reported that for both men and women aged 50–69 years, the LSS group had a higher complication rate of hypertension and diabetes mellitus than the control group []. Maeda et al. examined 968 individuals in the Wakayama Spine Study, a population-based cohort study, and reported that moderate spinal stenosis on imaging with symptoms was associated with diabetes mellitus (OR 3.9, 95% confidence interval 1.52–9.34) and low ABI level (OR 1.36, 95% confidence interval 1.04–1.81), whereas no item was associated with severe spinal stenosis on imaging with symptoms []. In addition, obesity, abdominal circumference, and family history of low back pain [] have also been reported to be correlated with LSS.

Many patients with LSS experience a decline in physical functions due to decreased activity and experience great stress. Ishimoto et al. studied 1009 individuals in the abovementioned Wakayama Spine Study and compared them based on the presence or absence of lumbar spinal stenosis. They found that the fastest 6-m walking time was significantly longer in individuals who complained of LSS and was correlated with symptomatic LSS (OR 1.17, 95% confidence interval 1.01–1.34). The fastest 6-m walking time has been reported to be a useful finding for the evaluation of decreased physical abilities due to LSS []. Park et al. compared 77 patients with LSS with 385 age- and sex-matched controls and reported that the prevalence of sarcopenia evaluated by grip strength of the patients with LSS (25%) was significantly higher than that of the control group (12%) []. Sekiguchi et al. investigated the association between the prevalence of LSS based on age and psychosocial factors in a large-scale national multicenter study (18,642 patients who received treatment for spinal diseases at 2177 hospitals nationwide) and reported that the prevalence of LSS increased with age, that subjective stress (OR 1.69, 95% confidence interval 1.57–1.82) and the load on lower extremities and low back (OR 1.41, 95% confidence interval 1.31–1.52) were associated with LSS, and that patients with LSS had a low work satisfaction level [].

It has been reported that if leg pain is strong at rest, there is a higher possibility of foraminal stenosis than spinal canal stenosis. Of 172 patients whose symptoms temporarily disappeared following an L5 nerve root block for the chief complaint of unilateral leg pain, 98 patients who underwent surgery were analyzed. Compared with L4-5 spinal canal stenosis (60 patients), L5-S foraminal stenosis (38 patients) was associated with a significantly higher complaint rate of leg pain at rest (76 vs. 35%) and stronger degree of pain (VAS 66 ± 31 vs. 13 ± 19 mm) []. In a study of 100 patients who had a definitive diagnosis by selective nerve root infiltration for L5 nerve root, including L4-5 spinal canal stenosis (51 patients) and L5-S foraminal stenosis (49 patients), Yamada et al. developed a diagnostic support tool for symptomatic foraminal stenosis. With a total of 20 points, consisting of 3 points for Bonnet's test, 5 points for Freiberg's test, 3 points for pain in the sitting position, and 9 points for pain in the supine position, this tool had a sensitivity of 75.5% and a specificity of 82.3% with a cutoff value of 5 points and is useful as a diagnostic support tool for symptomatic foraminal stenosis []. In the future, the verification of the reliability and validity of this support tool is expected.

Tandem spinal stenosis (TSS) is the narrowing of the spinal canal in tandem with the cervical and lumbar spines. Lee et al. analyzed 440 human skeletal specimens collected from the late 19th century to the early 20th century and reported that the prevalence of TSS was 0.9%–5.4%. It was associated with the narrowing of the cervical and lumbar spines, and the positive predictive value, i.e., that narrowing in either region predicted concurrent narrowing in the other, was 15.3%–32.4% []. A systematic review on TSS reported that, clinically, TSS is present in 7.85%–60% of patients with spinal stenosis [].

In 2363 patients with LSS requiring surgical treatment, the prevalence of diffuse idiopathic skeletal hyperostosis (DISH) in the middle and lower thoracic spine was 31.7%, which is significantly higher than the prevalence for those without LSS (16.7%), and DISH (OR 1.65, 95% confidence interval 1.32–2.07) was correlated with LSS [].

Many physician-administered or patient self-administered diagnostic support tools for LSS that possess high sensitivity and specificity have been developed. The use of these diagnostic support tools is considered useful for not only the diagnosis of LSS in primary care but also for self-diagnosis by patients as a self-administered tool. However, these tools should be used for screening, and detailed examination including imaging tests by specialists is required for definitive diagnosis.

Plain X-ray is a basic examination in the treatment for lumbar diseases. It is difficult to evaluate spinal stenosis using plain X-ray, but alignment evaluation, including dynamic imaging, has added value to the findings of MRI and CTM. Based on a comparison between standing lateral-view X-ray and supine MRI, it has been reported that if there is an anterior slippage of ≥5 mm, lumbar spinal stenosis is highly possible, so MRI is recommended []. It has also been reported that L5/S1 intervertebral mobility and L5 posterior slip in standing lateral-view X-ray are correlated with symptomatic L5/S1 foraminal stenosis [], and the vacuum phenomenon of the L5/S1 disc in lateral-view X-ray in the standing extension position is also correlated with symptomatic L5/S1 foraminal stenosis [].

• 2)
  Plain CT

In a comparison between 67 patients with LSS and 100 control individuals in Israel, it was found that the lumbar lordotic angle, sacral slope, and anteroposterior diameter of the bony canal were significantly smaller in patients with LSS (p < 0.001) []. In addition, the development and widespread use of multidetector CT (MDCT, also known as multi-slice CT) in recent years has made imaging evaluation of LSS easy. In particular, three-dimensional composite images of MDCT are useful for diagnosing foraminal stenosis [,]. Moreover, it was reported that MDCT after L5 radiculography may be useful for diagnosing L5/S foraminal stenosis [], and MDCT after epidurography was found to be superior to CT alone or MRI for the diagnosis of foraminal stenosis [].

• 3)
  Myelography and CTM

A few reports have stated that myelography and CTM have superior sensitivity to MRI as they can evaluate dynamic factors [,]. According to a study that compared MRI and CTM, CTM can measure the thickening of the ligamentum flavum in cross-sectional images more accurately, but there is only a slight difference between the two in the measurement of the cross-sectional area of the spinal canal []. Whether or not this difference affects clinical symptoms and treatment outcomes is an issue to be examined in the future.

• 4)
  MRI

Many reports have described the usefulness of MRI in the diagnosis of LSS. A systematic review of 46 reports concluded that MRI is the most reliable diagnostic imaging method for LSS and that myelography should be avoided considering its invasiveness, despite its superiority in diagnostic accuracy []. In 54 patients with LSS, the stenosis was classified as mild, moderate, and severe using MRI and CT, and in 20%–35% of patients, the degree of stenosis was evaluated as more severe using CT than using MRI, whereas in 2%–11% of patients, the degree of stenosis was evaluated as severe using MRI than using CT. In other words, MRI was less likely to overestimate stenosis than CT. MRI had a higher inter-rater reliability than CT. Therefore, it was concluded that MRI is more highly recommended for evaluating the degree of stenosis [].

In a comparison between dural sac area and spinal canal area, it was reported that dural tube sac was a more sensitive measurement parameter in LSS []. The reliability of grading for the dural sac area was examined, and it was concluded to be useful as an easy and practical evaluation for central stenosis []. Another study compared dural sac area measurement and morphological classification and reported that both methods had acceptable inter- and intra-rater reliability and could be used for determining LSS []. In addition, it has been reported that the rate of combination with redundant nerve roots of the cauda equina was high when the dural tube area was 55 mm² or less [].

A study evaluated cross-sectional images of the ligamentum flavum and revealed that the thickness of the ligamentum flavum was not associated with clinical diagnosis, pain, or functions []. Another study indicated that the ligamentum flavum area was more sensitive than the ligamentum flavum thickness and that the sensitivity for diagnosis was higher [].

There have been many reports on the association between imaging evaluation by MRI and clinical symptoms. It has been reported that clinical symptoms and the cross-sectional area and thickness of ligamentum flavum, which are morphological parameters of MRI, were weakly correlated with Oswestry Disability Index (ODI) (r = 0.249, p = 0.007 and r = 0.250, p = 0.007, respectively) []. Among 1586 patients with lumbar disorders recruited in a retrospective study, spinal stenosis on MRI (OR 1.61, 95% CI 1.26–2.05) and spondylolisthesis (OR 2.83, 95% CI 2.08–3.88) were independent predictors of surgical indication in 722 patients who underwent surgery []. In a study comparing 100 patients who underwent surgery for LSS and 100 asymptomatic age- and sex-matched controls, the anteroposterior diameter of the spinal canal on cross-sectional MR images was significantly smaller in the LSS group, and the cutoff value of the anteroposterior cross-sectional diameter of the spinal canal in patients requiring surgery was L4 <14 mm, L5 < 14 mm, and S1 <12 mm []. On the other hand, in 202 patients with LSS who underwent surgery, there was no association between the severity of spinal stenosis on MRI and the ODI, intensity of low back pain or leg pain either preoperatively or at 1 year postoperatively []. An analysis of 109 surgically treated patients with LSS reported that the minimum dural sac area (mm²) was not correlated with walking distance, ODI, the Health-related QOL Score (SF-36), EuroQol-5 Dimension (EQ-5D), and the intensity of low back pain or leg pain (VAS) []. A study of 63 patients with LSS reported that the dural sac area (mm2) was not correlated with walking distance []. In terms of other MRI findings associated with LSS, facet effusion has been reported to be a predictor of dynamic stenosis []. It has been reported that three-dimensional reconstructed MR images was superior to conventional cross-sectional images in the evaluation of spinal stenosis, especially at the lower lumbar region [].

• 5)
  Axial-loaded MRI

Since the symptoms of LSS appear or exacerbate in the standing position or when walking, it is insufficient to evaluate the responsible level using MRI, which is performed at rest while lying down. In particular, when intervertebral instability associated with spondylolisthesis is involved, there are limitations to the evaluation []. Recently, there have been increasing reports that axial-loaded MRI, in which axial load is applied to the spine and scanned under the axial-loading conditions or using a compression device, is useful for the diagnosis of central stenosis [, , , , , , , ]. MRI in the loaded extension position has also been reported to be useful for the diagnosis of foraminal stenosis [].

• 6)
  Sedimentation sign

Sedimentation sign, which is defined as the absence of nerve root sedimentation in cross-sectional images during imaging in the supine position, was reported to have a high positive rate for severe spinal stenosis [,]. It was also reported that the sedimentation sign is common in stenosis at the L2/3 or L3/4 levels, and is associated with central stenosis or multilevel stenosis, and patients with a positive sign had surgical improvement []. A study of the sedimentation sign on MRI in 72 patients with LSS (50 patients with central stenosis or combined stenosis and 22 patients with lateral recess stenosis) and 43 patients with disc herniation reported that the sedimentation sign was more common in patients with central stenosis or combined stenosis than in patients with only lateral recess stenosis and patients with disc herniation []. It has been reported that the post-decompression negative conversion of the sedimentation sign that was present before surgery was associated with improved clinical symptoms, and the retention of a positive sedimentation sign could be the result of incomplete decompression or surgical complications []. It has been reported that although the sedimentation sign is not a prognostic factor for surgical outcomes of patients undergoing decompression surgery, it was associated with limited effects in patients undergoing conservative treatment []. On the other hand, in a study comparing the morphological classification of four grade of spinal stenosis and the presence or absence of the sedimentation sign in MRI of 110 symptomatic LSS patients with neurogenic intermittent claudication (73 treated surgically and 37 treated conservatively), the sedimentation sign was negative in one-third of the surgically treated patients and it was concluded that the sedimentation sign did not exceed the prediction of treatment contents compared with the severity of the morphological classification for stenosis []. It has been reported that the sedimentation sign demonstrated high intra- and inter-rater reliability and was useful for differentiating patients with LSS from asymptomatic controls, but it could not differentiate patients with LSS from patients with low back pain and those with vascular intermittent claudication []. A retrospective study of 105 patients with LSS and 215 patients with nonspecific low back pain reported that the sedimentation sign was not useful for differentiating between the two conditions []. A meta-analysis of seven appropriately designed studies on the sedimentation sign concluded that its usefulness for diagnosing mild and moderate LSS was not established yet []. Thus, the diagnostic value of the sedimentation sign remains uncertain.

• 7)
  Foraminal stenosis

In the past, foraminal stenosis was considered difficult to diagnose, since the intervertebral foramen was proposed to be a hidden zone by Macnab in 1971. In recent years, the number of studies on foraminal stenosis has increased with advances in diagnostic imaging technologies. Multiple reports have indicated that nerve root visualization using three-dimensional MRI is useful for the auxiliary diagnosis of foraminal stenosis [, , ]. It was also reported that when the bilateral difference of the foraminal spinal nerve angle (FSNA, the angle between the L5 nerve just below the L5 pedicle and L5/S intervertebral disc) in oblique coronal T2-weighted imaging was ≥10°, it was possible to diagnose foraminal stenosis with a sensitivity of 94% and specificity of 91% [].

In addition, diffusion tensor imaging (DTI) has been reported to be useful for the diagnosis of foraminal stenosis as it enables not only morphological evaluation but also quantitative evaluation of the lumbar spinal nerve [,]. DTI has been reported to be useful for the diagnosis of double crush lesion (the pathological condition in which the L5 nerve is compressed at two sites, namely, in the L4/5 spinal canal and at the L5/S intervertebral foramen) []. Moreover, it has been reported that the addition of paraspinal mapping (PM) improved the diagnostic sensitivity [,]. Furthermore, to improve the diagnostic accuracy of MRI, a new grading evaluation focusing on perineural adipose tissue and nerve compression at the intervertebral foramen in sagittal images (Lee system, 0: no foraminal stenosis, 1: fat loss in two opposite directions, 2: fat loss in four opposite directions, 3: rupture or morphological change of the nerve root) has been reported [,]. An MRI automated diagnostic tool that achieved diagnosis at a high accuracy using T1/T2-weighted images of 200 cases has also been reported [].

While the electrophysiological examination of the lower extremities has been reported to have a low association with MRI changes and clinical symptoms [], it is useful for diagnosing LSS with unclear stenosis on images [], and it has been reported that H wave and F wave were correlated with the site of stenosis []. It has been reported that delay in cauda equina conduction time (CECT) by magnetic stimulation motor evoked potentials [] and Lp potential latency by tibial nerve stimulation somatosensory evoked potentials [] were useful for the functional diagnosis of cauda equina syndrome. It can be inferred that the electrophysiological examination is more useful than imaging tests for the diagnosis of foraminal stenosis []. It has been reported that dermatomal somatosensory evoked potentials (SEPs) are superior for the visualization of multi-root abnormality, and somatosensory evoked potentials (SSEPs) were correlated with leg numbness and were found to be useful for diagnosing nerve root disorders [,]. Superficial peroneal nerve sensory nerve action potentials (SPNSNAPs) have also been reported to be useful for the diagnosis of L5/S foraminal stenosis (sensitivity 59.6%, specificity 93.5%) [].

In 109 patients with LSS who underwent surgery, there were 21 cases (19%) in which either the responsible level or the neurological disorder type judgment, or both, changed after the gait loading test. Therefore, the gait loading test was considered useful for the diagnosis of responsible levels and neurological disorder type in LSS []. On the other hand, a study of 25 patients with LSS reported that the walkable distance by the treadmill test was correlated with the degree of stenosis on MRI and ODI but not with the degree of clinical low back and leg pain []. In a study comparing the usefulness of the gait loading test and lumbar extension-loading test (maintenance of lumbar extension at 10°–30° until symptoms worsen or fatigue was felt) in 116 patients with LSS, the extension-loading test was reported to be as useful as the gait loading test for diagnosing the responsible levels [].

• 2)
  Biomarkers

There are some reports on biomarkers for LSS; however, they are preliminary reports. The expression level of microRNA (miR)-29a in the plasma and intervertebral disc tissues of patients with LSS was found to be significantly lower than that in patients with disc herniation and healthy controls, and miR-29a was reported as a potential biomarker for LSS []. The concentration of phosphorylated neurofilament heavy subunits (pNfHs) in the cerebrospinal fluid has also been reported to be a potential biomarker for LSS []. In a study on 67 female patients with LSS (LSS group) and 67 age- and weight-matched controls, both bone resorption marker (μ-NTx) and bone formation marker (ALP) levels were increased in the LSS group, which suggested that the increase in bone turnover associated with physical activity restriction might be involved in patients with LSS [].

MRI enables understanding of the state and degree of spinal stenosis, but there are reports of large inter-rater variability and differences in reproducibility even among evaluations by the same examiner [,]. There was a report that even diffusion MRI was not useful for the determination of decompressed intervertebral space [], and some reports indicated that myelography and CTM were advantageous despite their invasiveness [,]. In addition, the severity of stenosis on MRI was not associated with clinical symptoms [,]. A cross-sectional study that included 938 participants from among the local population in Wakayama Prefecture (The Wakayama Spine Study) reported that 285 participants (30.4%) had severe central stenosis (exceeding two-thirds of the spinal canal area), and the odds ratio of having clinical symptoms (4.41 for males and 2.50 for females) was significantly higher than that of mild or no stenosis, whereas only 17.5% of patients with severe central stenosis were symptomatic []. Thus, MRI remains a noninvasive examination suitable for the imaging diagnosis of LSS at present. However, it is important to recognize that LSS cannot be diagnosed by imaging findings alone. Therefore, the most important tools for diagnosis of LSS are the clinical symptoms and physical examination, and various diagnostic methods described in this guideline are no more than auxiliary. In LSS, besides determining the responsible level of clinical symptoms and the level and location of the lumbar spine to be decompressed when planning a surgery, in addition to MRI and plain X-ray including dynamic imaging, it is recommended to combine multiple auxiliary examinations, including imaging tests such as myelography, CTM, 3D MRI and DTI, as well as electrophysiological examinations such as SEP.

When developing the recommendation statements, four RCTs that evaluated the usefulness of limaprost were adopted. Among these RCTs, two conducted research on patients with cauda equina type or combined type, one on patients with nerve root type, and one on patients with no description of the disease type. The target disease types, controls, and endpoints differed among the studies; thus, it was determined that a meta-analysis would be difficult to perform. Therefore, a qualitative integration was performed, and the final recommendation was decided.

In the two RCTs on patients with the cauda equina type or combined type of LSS [,], improvement of leg numbness, walking distance, and the Health-related QOL Score (SF-36) was significantly superior in the limaprost administration group. In the RCT on patients with the nerve root type [], no improvement of pain was noted in the limaprost group, and the improvement of low back/lower limb pain and QOL following the combination therapy with NSAIDs was superior to that following limaprost administration alone. In the RCT on patients with all disease types [], there was no significant difference in the improvement of pain, walking distance, and QOL, compared with those after pregabalin administration.

On the other hand, there is no evidence that the limaprost group had a higher incidence of adverse events than other drug groups; thus, limaprost is considered a highly safe drug.

In summary, there is strong evidence of its usefulness in patients with the cauda equina type or combined type, and the balance between benefits and harms is secured and recommended, but evidence of its efficacy against the nerve root type and pain is considered insufficient.

When developing the recommendation statements, two RCTs that evaluated the usefulness of NSAIDs were adopted. One RCT each was conducted on the cauda equina type and nerve root type, and both RCTs compared them with limaprost. The target disease types, endpoints, and evaluation methods differed between the studies; thus, it was determined that a meta-analysis would be difficult to perform. A qualitative integration was performed, and the final recommendation was decided.

In the RCT that examined the efficacy against the cauda equina type [], leg numbness, walking distance and SF-36 score significantly improved in the limaprost group compared with those in the NSAID group. Thus, the use of NSAIDs cannot be recommended. On the other hand, in the RCT on the nerve root type [], the improvement of low back pain and leg pain tended to be better in the NSAID group than in the limaprost group, and low back pain, leg pain, and QOL (RDQ) significantly improved in the NSAID + limaprost group than in the limaprost alone group. Thus, they can be expected to be effective against nerve root pain and low back pain.

On the other hand, with regard to adverse events, no issue has been indicated in short-term RCTs, but it is a well-known fact that it is necessary to take renal dysfunction and digestive disorders, among others, into consideration. Thus, short-term use is desirable. In summary, it is suggested to use NSAID in the short term for LSS with nerve root type and low back pain. On the other hand, they have limited usefulness against the cauda equina type, and their administration is not recommended.

When developing the recommendation statements, four RCTs and one observational study that evaluated the usefulness of gabapentinoids were adopted. Two RCTs each evaluated the usefulness of gabapentin and pregabalin. The evaluation methods, evaluation period, and controls differed among the studies; thus, it was determined that a meta-analysis would be difficult to perform. A qualitative integration was performed, and the final recommendation was decided.

There was no consistency with regard to its efficacy against pain and numbness, with one study stating that gabapentinoids were effective [], one study stating them as ineffective [], and one study stating them as equivalent to limaprost []. There was no consistency with regard to its efficacy for walking distance, with one study stating gabapentinoids as effective [] and two studies stating them as ineffective [,]. With regard to QOL, two studies stated them as ineffective [,] and one study stated them as equivalent to limaprost [].

On the other hand, the incidence of adverse events tended to be high in the gabapentinoid group in all studies. The observational study on Japanese patients with spinal stenosis reported pregabalin-related adverse events in 12.5% of patients (12/96 patients) [].

In summary, there was no consistency in the evidence related to the efficacy of gabapentinoids, and the incidence of adverse events was higher than that for placebo or other drugs; thus, evidence about its usefulness is considered insufficient.

One RCT that evaluated the usefulness of methylcobalamin and one RCT that evaluated the usefulness of Kampo medicine were adopted.

The RCT that evaluated the usefulness of methylcobalamin [] was a 24-month single blinded follow-up study that categorized 152 patients into the methylcobalamin group or control group. Pain (yes/no), walking distance (<1000 m/≥1000 m), and adverse events (blood and urine tests) were investigated. No significant difference was found regarding pain, but the percentage of patients who walked ≥1000 m was significantly higher in the methylcobalamin group, whereas no significant changes in blood and urine tests were noted (no specific data presented). However, in these two groups, patient education, exercise, physiotherapy, NSAIDs, analgesics, muscle relaxants, vitamin B1 and B6 preparations, or epidural steroid injections (ESIs) were used in combination and is doubtful whether the study really determined the treatment effects of methylcobalamin. Thus, it is desirable to only consider the research results as reference.

An RCT that evaluated the usefulness of Kampo medicine [] was an 8-week follow-up study that categorized 27 patients into the Hachimijiogan group and propionic acid group. Low back pain, numbness, intermittent claudication appearance time, and adverse events were investigated. Low back pain, numbness, and intermittent claudication appearance time significantly improved in the Hachimijiogan group, whereas adverse events did not occur in any patient. However, this research included a small sample size, had no description regarding blinding, was a subjective evaluation, and had significant differences in patient background, so the evidence is not necessarily strong. Thus, it is desirable to only consider the research results as reference.

There are five systematic reviews, including a Cochrane review, on exercise therapy, but all reviewed papers were published before 2012. Although exercise therapy may be effective in mitigating low back, buttock, and leg pain and improving physical functions, currently, there is insufficient evidence for deciding recommendations [, , , , ]. However, six RCTs published after 2012 related to exercise therapy reported that exercise therapy was effective in mitigating pain and improving physical functions and ADL/QOL [, , , , , ]. In these RCTs [, , , ,], randomization and blinded evaluation was performed, but it is necessary to consider bias. Due to the nature of exercise therapy, blinded treatment is not possible.

Multiple RCTs have proven that exercise therapy and physiotherapy under the guidance of specialists, such as physiotherapists, are more effective in alleviating pain and improving physical functions and ADL/QOL than self-training. In an RCT that compared the outcomes of a comprehensive training program that combined low back pain education, exercise therapy and manipulation under the guidance of specialists, and self-training, in which patients received exercise guidance only at the first session, a 2 year follow-up survey showed that alleviation of pain and improvement of physical (walking) function was significantly superior in the comprehensive training group [] 7). In an RCT that compared the effectiveness of three types of conservative treatment, namely, medications and epidural injection, group exercise, and manual therapy/individualized exercise, all the methods were found to improve gait ability, but manual therapy/individualized exercise was superior to the other two methods in terms of alleviating pain and improving physical functions and gait ability []. In Japan, an RCT showed that exercise therapy and manual treatment under the guidance of physiotherapists alleviated pain and improved physical functions and ADL/QOL better than home exercise [].

One systematic review focused on the exercise type, but the evidence is insufficient and no conclusion was reached regarding what type of exercise is effective []. An RCT on exercise type published after the systematic review reported the usefulness of core stability exercise []. However, this RCT involved selection bias, such as non-blindness of the evaluators, and evaluated only the score immediately after the 4 week treatment. In the future, high-quality research on the type of exercise therapy optimal for the treatment of LSS is required.

An RCT that compared the effects of epidural steroid injection (ESI) therapy alone or combination treatment with physiotherapy revealed that ESI therapy significantly improved physical functions, but the addition of physiotherapy resulted in better improvement []. However, it has been reported that the addition of physiotherapy significantly improved QOL (emotional function, mental stability, and health awareness) for up to 10 weeks (immediately after the completion of treatment). A three group RCT that compared groups with the addition of ultrasound treatment or pseudo ultrasound treatment to exercise therapy, with no-treatment group, reported that 3-week exercise therapy was effective in improving pain and functional disorders and that the addition of ultrasound treatment reduced analgesic usage []. However, this research evaluated outcomes only immediately after treatment (3 weeks later), so it was unclear if the effects lasted over the long term.

A systematic review that compared improvement in physical function by exercise therapy and decompression surgery reported that exercise therapy was less effective than decompression []. However, in this review, only one RCT was adopted []. Data regarding improvement in physical function extracted from 12 observational studies on exercise therapy and 10 observational studies on decompression surgery were compared, and a conclusion was drawn. Another RCT that compared exercise therapy and decompression surgery published in 2015 reported that exercise therapy led to the alleviation of pain and the improvement of physical function effects, ADL, and QOL, equivalent to surgical treatment 2 years later []. However, it is undeniable that selection bias might have influenced the result because 65% of the eligible participants did not participate in the research out of the fear of being allocated to the exercise therapy group. In addition, the effects of exercise therapy and decompression surgery are estimated to be different for severe and mild patients, so future research is anticipated.

A survey of participant preferences for treatment methods was conducted as a subanalysis of the above-mentioned RCT that compared three types of conservative treatments [,]. The older the participants, the more they recognize the effects and advantages of exercise therapy. Patients tend to prefer individualized care according to their medical conditions and background, and conservative treatments assigned in RCTs are effective in improving pain and walking in many patients. It has also been reported that they recognize the fact that the effects are temporary, and that the obstacles to continuation of conservative treatment were transportation and the portion of costs not covered by insurance.

A study in Japan reported that the cost required for exercise therapy and manual treatment under the guidance of physiotherapists twice a week for a period of 6 weeks was 331 dollars (self-paid cost of 63 dollars) and 100 dollars (self-paid cost of 20 dollars) for home exercise []. Thus, exercise therapy can be considered a treatment that can be performed at a relatively lower cost.

There are very few reports on severe adverse events caused by exercise therapy, and the incidence of transient muscle pain and joint pain was reported to be 0%–54% [,,,]. A report indicated that exacerbation of the symptoms of LSS occurred in 10% of cases due to exercise therapy [], but this was rarely reported in other studies, and the frequency of occurrence may differ based on the presence or absence of guidance by specialists and exercise type and load.

There was no RCT on the use of a “flexion brace”, that is, a brace that restricts extension or maintains anteflexion of the lumbar spine. With regard to lumbar corsets, according to an RCT that assessed standard lumbar corsets and prototype LSS belts in 104 patients with LSS [], the use of prototype LSS belts significantly extended walking distance, but their effect was not significantly different from that of standard lumbar corsets. Corset and belt use extended the walking distance in both groups, suggesting that orthosis therapy may be effective for improving gait ability. However, to clarify the usefulness of orthosis therapy, a comparison with a placebo (no treatment) group is required.

With regard to walking sticks, an RCT assessed the use of walking sticks for 2 weeks in 40 patients with LSS with intermittent claudication aged ≥55 years [] and compared them with patients who did not use walking sticks. The Zurich Claudication Questionnaire, VAS for low back and, leg pain, Oswestry Disability Index, and Hospital Anxiety and Depression Scale revealed no significant differences, and a gait test performed as a part of a subsequent self-controlled study also revealed that the use or non-use of walking sticks did not result in differences in walking distance and walking posture.

There is no study on the treatment of LSS with hyperthermia alone, but an RCT that categorized 39 patients with LSS into three groups, i.e., hyperthermia/TENS/therapeutic ultrasound combination, ESI, and no treatment [] revealed that hyperthermia/TENS/therapeutic ultrasound combination therapy did not show significant differences in pain, ADL, QOL, and gait ability compared with no treatment. Therefore, these interventions cannot be considered useful for management of LSS.

In an RCT that evaluated therapeutic ultrasound in 45 patients with LSS [], the usefulness of therapeutic ultrasound for alleviating pain and improving ADL, QOL, and gait ability was not observed.

With regard to TENS, in a CCT of 44 patients who underwent decompression surgery for LSS [], those who received TENS showed significantly better results in terms of leg pain and numbness and walking satisfaction than those who did not receive TENS. However, the study assessed the effects of TENS on postoperative residual symptoms of LSS, not on the symptoms of LSS itself, so it is necessary to consider the indirectness.

Since the first edition of these guidelines, multiple interventional studies have been conducted on orthosis therapy and physical therapy, but the accumulated evidence is insufficient to formulate a definitive recommendation. In the future, it will be necessary to verify the usefulness of these therapies by conducting high-quality interventional studies.