Background
Lumbar spinal stenosis (LSS) is a common spinal condition and the most frequent indication for spinal surgery in elderly people. The prevalence of LSS is estimated to be 9% in the general population and up to 47% in people older than 60 years [
1]. LSS may occur on a congenital narrow lumbar canal or may result from degenerative processes. Acquired LSS is secondary to facet joint osteoarthritis,
ligamentum flavum hypertrophy and/or bulging of the intervertebral disc [
2] and leads to the narrowing of the spaces around neurovascular structures of the spine [
3]. LSS has a dynamic component: central stenosis increases in lumbar spine extension and decreases lumbar spine flexion [
2]. Limitation in walking distance impacts patients’ functioning and quality of life.
People with an anatomic LSS can remain asymptomatic for several years. Pain in the back and in the lower limbs is the most frequent symptom. The clinical sign most evocative of LSS is neurogenic claudication described as leg pain during walking, relieved by lumbar flexion or by sitting, while the association with low back pain (LBP) is inconsistent. Physicians usually consider a combination of clinical signs and imaging findings [
3]. However, in an asymptomatic population, up to 20% of subjects have an imaging result consistent with anatomical LSS. Therefore a correlation between clinical signs and imaging findings is important to make the diagnosis and to offer the proper treatment [
2]. The 1
st-line treatment of LSS is conservative and includes analgesics, corticosteroid spinal injections, exercise therapy and physical activity [
3]. Exercise therapy seems better than no treatment for leg pain [
4,
5]. However, evidence of efficacy of conservative treatment is of low quality.
GPs are on the 1
st line for the diagnosis and the management of musculoskeletal disorders, especially because they follow-up elderly people at long term and because elderly people are less eager to consult specialists than younger people [
6,
7]. In 2017, 88,137 GPs were board-registered in France. Most GPs did not receive a specific training in the management of musculoskeletal disorders. As a result, many lack confidence in the management of musculoskeletal disorders in daily practice [
8]. Specific guidelines (such as those for LBP) may help GPs’ decision-making in clinical practice. However, they are not fully implemented in primary care yet. A review published in 2016 by the Cochrane Collaboration suggested that training interventions or recommendations made by tertiary care specialists designed to promote behavioral changes and improve care were not appropriate for primary care, because they did not take into account the specific burden of primary care practice (e.g. lack of time, comorbidities) [
8]. In the case of LSS, despite a high prevalence in elderly people, largely-disseminated diagnosis criteria and national or international guidelines are lacking.
In the present study, we aimed to assess how GPs diagnose and treat people with LSS in France.
Discussion
Overall, French GPs lack confidence with diagnosing LSS and prescribing pharmacological and non-pharmacological treatments for people with LSS. Our findings may be explained by the lack of consensual national and international guidelines in primary care or of specific training during medical studies in France.
In our survey, GPs rated their confidence to diagnose people with LSS 5.6 (2.4)/10. A set of clinical diagnosis criteria was proposed by the International Society for Study of the Lumbar Spine (ISSLS) in 2016. The 7 most relevant items according to the ISSLS were: 1/ leg or buttock pain while walking, 2/ flex forward to relieve symptoms, 3/ feel relief when using a shopping cart or a bicycle, 4/ motor or sensory disturbance while walking, 5/ pulses in the foot present and symmetric, 6/ lower extremity weakness, and 7/ LBP [
1]. Recently, the N-CLASS criteria were proposed for the diagnosis of neurogenic claudication caused by LSS [
10]. This set includes: 1/ age > 60 years, 2/ positive 30-s extension test (typical leg symptoms reproduced during active spine extension performed in standing position for 30s), 3/ pain in both legs, 4/ leg pain relieved by sitting, 5/ leg pain decreased by leaning forward or flexing the spine, and 6/ negative SLR-60 test (Straight Leg Raise test: positive if leg pain is produced below 60°). In the present study, the clinical signs most frequently cited by French GPs differed from those included in these 2 published datasets. An explanation is that these datasets may lack applicability to primary care [
11]. In the ISSLS study, GPs represented only 1% of participants and were not included in Genevay’s survey. It would be interesting to build a set of specific LSS criteria by spine experts in collaboration with GPs, specifically designed to the constraints of primary care practice. In Japan, two diagnostic tools for LSS have been validated (the self-administrated, self-reported history questionnaire [SSHQ], and the developmental clinical diagnosis support tool [ST]). A survey evaluated the degree of awareness and use of these tools in 1,811 Japanese physicians [
12]. Among GPs, the degree of awareness for both tools was less than 30%, and their implementation ranged from 31 to 36%. Improving the knowledge of validated diagnostic tools by practitioners could help to improve the management of LSS. In our study, 85/86 (98.8%) of GPs would prescribe imaging, especially MRI (63/86 [73.3%]). Imaging is often reserved for diagnostic confirmation, especially during pre-surgical evaluation [
3]. MRI is the recommended test for confirming LSS [
7]. A study showed 96% of sensitivity, 67% of specificity, 4% of positive predictive value and 100% of negative predictive value for the diagnosis of symptomatic LSS with MRI [
13]. Electromyogram is not recommended [
3] but can be used to rule out differential diagnosis in atypical symptoms [
14].
LSS-related impairments are associated to activity limitation dominated by reduced pain free and maximal walking distances. In an interview of 33 patients with LSS, 88% reported “experiencing pain/discomfort”, 85% “problems with physical function”, 73% “difficulty exercising” and 55% “difficulty participating in hobbies and leisure activities” [
15]. Therefore, a combination of pharmacological and non-pharmacological treatments should be offered for an optimal management of people with LSS. In our survey, GPs rated their confidence with prescribing pharmacological treatments in people with LSS 5.5 (2.5)/10. Non-opioids analgesics and NSAIDs were prescribed as 1
st-line treatments by most GPs. In a large cohort assessing the current treatment strategies by GPs for the management chronic pain of 1,379 elderly outpatients [
6], prescriptions of analgesics by GPs followed national and international recommendations. However, in a population over 65 years, with comorbidities, these prescriptions must be limited in time [
16]. Pregabalin and gabapentin have also been used. However, the efficacy of these drugs in people with LSS has not been proven yet. 69/81 (85.1%) GPs would prescribe steroid injections. GPs who answered “other” indicated that this prescription was left to the specialist. Previous studies found limited evidence of a lack of effectiveness of epidural steroid injections in LSS [
3]. In 2015, a meta-analysis showed the effectiveness of epidural corticosteroid injection, in a context of radiculopathy, on pain (-7.55 [95%CI, -11.4 to 3.74]), function (-0.33 [CI, -0.56 to -0.09]) and surgery risk (0.62 [0.41 to 0.92]) at short term. But no specific effect was shown in the treatment of LSS [
17].
GPs rated their confidence with prescribing non-pharmacological treatments only 4.8 (2.6)/10. Most GPs would advise their patients to practice a regular physical activity but not specifically cycling, a common modality of lumbar flexion-based exercises, or endurance training. Given the pathogenesis of neurogenic claudication in LSS related to the narrowing of the spinal canal in lumbar extension and its widening with the relief of the nerve root in lumbar flexion [
18], a lumbar-flexion-based training program is usually recommended. In 2003, Iversen showed that a cycling program was feasible in a population of elderly people with LBP [
19]. A recent pilot study [
20] described the barriers (pain, fatigue, too large bicycle, burden of hospital follow-up, lack of time and motivation) and facilitators (clinical improvement, surveillance, ease-of-use) to home-based cycling in elderly people with LSS and found that adherence was stable over the 3-month follow-up. However, nearly all GPs (77/82 [93.9%]) would refer people with LSS to a physiotherapist. An interpretation of this latter finding is that GPs would not pre-judge the non-pharmacological treatment to be prescribed but would rather leave that to the physiotherapists, which seems to be a reasonable therapeutic strategy.
Optimal health care journey of people with LSS is underreported in literature. In our survey, GPs would mostly refer their patients to a surgeon, suggesting that they would rather consider surgical than conservative treatment for people with LSS. A study assessing the management of LBP in primary care found similar results [
21]: rheumatologists were consulted in 93% and surgeons in 60%. Several GPs reported that they would not refer their patients to a specialist in PMR because of the lack of accessibility to this specialty. In 2016, specialists in PMR were only 2,114 in France. A limitation of our survey on referral is that we did not specifically assess the potential relationship between reasons for referral and different specialists.
Several educational reasons may have contributed to the lack of confidence reported by French GPs with the management of people with LSS. During the 2
nd cycle of medical studies in France, LSS diagnosis and treatments are taught during rheumatology courses. In the reference textbook, only a short section is dedicated to LSS. Diagnosis and treatment key points could be summarized as follows: 1/ average age is 60 years, 2/ clinical signs are pain increases in lumbar lordosis, paraesthesia, motor or sensory deficit, sphincter disorders, pain while walking and relieved by lumbar flexion (shopping cart sign), 3/ diagnosis is confirmed by MRI, and 4/ therapeutic options include symptomatic treatment, epidural steroid injections, lumbar flexion-based rehabilitation and surgery [
22]. Residents in general practice do not receive a specific training for the management of elderly people with spinal disorders during the 3
rd cycle of medical studies. Furthermore, continuing medical education of GPs do not include specific training on the management of people with spinal disorders. In the present study, GPs reported they would be interested in receiving educational material on the management of LSS and the results of our survey.
Our study has limitations. The response rate was low but comparable to those reported from previous studies conducted in primary care [
23]. Fundamental weakness are the purely descriptive nature of the report, the small sample size, and the low survey response rate which, while similar to other survey type studies, limits conclusions because of the high risk of bias in the survey. Our population of GPs was randomly selected from all over France, but we did not evaluate practices in other countries. Only volunteer GPs answered the questionnaire and one can assume that GPs feeling the least confident with the questionnaire may have not responded, which could have led to a selection bias and an overestimation of self-reported confidence scores. To assess GPs’ confidence with prescribing pharmacological and non-pharmacological treatments for people with LSS, we used a self-administered VAS. However, there is no validated scale to assess this outcome. For example, in another study, participants were asked to rate their confidence with diagnosing the condition using a 5-class Likert scale (“definitely yes”, “most likely”, “likely”, “not sure” and “definitely sure”) [
24]. The very low exposure to LSS in this sample is unexpected and certainly had an impact on our results. We can wonder whether there is confusion about the definition of LSS itself or whether LSS is underdiagnosed in primary care. Finally, because of slow accrual, we had to change the method of data collection after study commencement. Answers may have been different between participants who responded online and those who responded by mail.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.