Background
The density of the paraspinal muscles and their cross-sectional area (CSA) size are known to be associated with variables such as age, gender and weight [
1‐
6]. Current evidence suggests that these muscles are smaller in patients with chronic low back pain as compared to healthy individuals of a similar age [
3,
5,
7]. It is generally agreed that muscles’ CSA and density reflect muscle performance and physical function [
8‐
10] of individuals. Condition of muscle such as density, CSA size and fatty infiltration can be attained via medical imaging techniques that provide non-invasive and reproducible information [
4,
11]. Computed tomography (CT) and magnetic resonance imaging (MRI) have been used for measuring CSA and the degeneration rate of muscles in patients with muscular dystrophy [
12,
13].
The paraspinal muscles play an important role in the stability and functional movements of the lumbar vertebral column [
14,
15]. Previous in vitro studies have reported that the lumbar spine is inherently unstable [
16] and that much of its stability is dependent on the integrated function of the active, passive and neural sub-systems [
17,
18]. Degenerative lumbar spinal stenosis (DLSS) is a clinical entity, an age-dependent phenomenon that is associated with degenerative changes of the three-joint complex and ligament flavum hypertrophy [
19‐
21]. Although spinal instability has been shown to play an essential role in DLSS [
22], the role of paraspinal muscles remains elusive.
The aim of this study was to examine via medical imaging techniques whether an association between the condition of paraspinal muscles and symptomatic DLSS exists.
Discussion
The present study is, to the best of our knowledge, among the largest to examine the relationship between paraspinal muscles density and cross-sectional area (CSA) and symptomatic degenerative lumbar spinal stenosis (DLSS).
The results of our study show that both males and females in our stenotic group manifested significantly greater density of the paraspinal muscles and higher CSA (for the erector spinae muscle) at the L3 level compared to the control group. Additionally, the density of the multifidus (both sexes) and erector spinae (males) muscles was found to increase the likelihood of symptomatic DLSS. However, no point of discrimination could be established to differentiate between individuals with and without symptomatic DLSS.
Our findings are in contrast to the general view that paraspinal muscle density and size tend to show lower CSA and density in individuals with low back pain [
36,
38,
39], disc pathology [
40], degenerative lumbar flat back [
41], degenerative kyphosis [
42] and spinal stenosis [
43]. Yarjanian and colleagues reported that CSA of paraspinal muscles at L5-S1 level had decreased in individuals with clinical stenosis (
n = 15) compared with the asymptomatic group. The observed muscle atrophy was attributed to muscle denervation and/or disuse [
43]. This result may seem to contradict ours as we have found the reverse phenomenon, but not so. Indeed individuals with lumbar spinal stenosis manifest evidence for paraspinal denervation [
44‐
46], mainly at the L5 level [
47] hence, muscle atrophy at L5-S1 level is expected. This results in inadequate lower lumbar segmental stabilization, which in turn requires compensation by the paraspinal muscles at higher levels. Increase in density of paraspinal muscles at L3 level and above in DLSS individuals is therefore anticipated. Support to the above notion can be found in Leinonen et al. [
48] study which reported that individuals with clinical lumbar spinal stenosis manifested (“unexpectedly”) good back muscle endurance and significantly lower paraspinal muscle fatigue compared to healthy subjects. These authors claimed that the poor relationship between muscle endurance and denervation observed in individuals with clinical stenosis may indicate the compensatory use of other trunk extensor muscles.
The fact that degenerative lumbar spinal stenosis is triggered by segmental instability [
22], explains why this disease is commonly accompanied by facet-joints arthrosis, intervertebral disc disease, ligamentum flavum hypertrophy and osteophyte formation [
49,
50].
Instability was determined by Pope and Panjabi [
51] as a mechanical entity that is related to loss of stiffness. The paraspinal muscles act to support the spine and maintain its stability [
52]. Multifidus and erector spinae muscles also act as back extensors [
53]. The psoas muscle is considered a major flexor for the hip joint, an intersegmental extensor in the mid lumbar region, and generally functions as an active postural muscle [
54‐
56]. The different functions of the three muscles in stabilizing the lower spine explain why multifidus and erector spinae were more responsive to DLSS than the psoas.
Two parameters were used in the current study to evaluate muscle status: first, muscle density that reflects quantity of muscle fibers, as well as the area of a muscle’s fiber and the general packing of the contractile material [
57]. Second, CSA, which is determined mainly by the total number of muscle fibers and, to a lesser degree, by the size of the fibers [
58]. According to our study, density is the more sensitive parameter of the two. We here argue that during the initial phase of spinal stenosis cascade, when the spine segment is prone to repetitive high loading and shearing forces that affect its stiffness (stability), the paravertebral muscles may enhance activities and contractions in order to compensate for increased segmental mobility. Only at a later stage additional mechanisms such as the thickening of ligamentum flavum and hypertrophy of facet-joints are involved. Supporting studies have reported that in short-term spine instability, muscles can respond actively and reduce spinal movements [
59,
60], whereas in the long term, tissue remodeling in the form of osteophyte formation and ligament hypertrophy may help to restore stability [
22,
61]. It has also been claimed that the human spine responds to changes in stability by utilizing its own passive and active preventive mechanisms, i.e., muscles, ligaments and bone structures [
59]. It has recently been shown [
62] that patients with low-back pain (LBP) have significantly larger CSA of the psoas at the levels of L3-4 and L4-5 than the control group. The authors hypothesized that hypertrophy of the psoas was the result of its increased activity in maintaining stability of the lumbar spine.
Based on the above findings, we argue that the paraspinal muscles work to control mobility and to achieve stability in symptomatic DLSS individuals, thus resulting in thicker and denser muscles. This muscle behavior, essentially a recruitment strategy to compensate for reduced stability, was noted in other parts of the body, for example the trunk [
63,
64]. Additionally, in vitro studies [
16,
59] demonstrate that the multifidus muscle has the capacity to restore control of segmental motion following injury. Positive correlation between training programs and increased paraspinal muscle features (density and CSA) was also reported [
65‐
68]. Some studies have also connected the condition of paraspinal muscles with physical activities [
9,
10,
68] and spine-fusion [
5,
7,
69]. It can be argued that the greater density and CSA of the paraspinal muscles in the DLSS group is due to hyperactivity (spasm) of the muscles to limit motion and control for back pain [
70]. However, this hypothesis can be rejected as no association has been found between severity of low back pain and the CSA of paraspinal muscles [
71]. Moreover, paraspinal muscles atrophy has been noted in patients with acute or subacute LBP compared to controls [
4,
68].
Our results are not in agreement with those of Kalichman et al. who found no association between paraspinal muscles density and radiological spinal stenosis [
6]. Conflicting data have also been presented by others with regard to degenerative lumbar spine (e.g., facet-joints arthrosis, disc space narrowing) [
6,
41,
42] and have demonstrated an attenuation in paraspinal features. This may be attributed to the fact that different inclusion criteria were used for the study groups (e.g., clinical vs. radiological stenosis) and not all degenerated spinal segments necessarily develop instability [
72].
Several studies have shown a negative correlation between the condition of paraspinal muscle and LBP and/or disc pathology [
4,
36,
37,
39,
69,
73]. Others have disputed this and reported no correlation between the condition of the muscles and LBP [
6,
10,
71,
74]. These conflicting data can perhaps be explained by one of the following: (a) since LBP disorder is recognized as a multi-factorial origin [
75], muscle condition in these individuals can vary, (b) different methods were applied for measurements (e.g., MRI, CT and ultrasonography), (c) various measurements (e.g., density, total CSA and/or free-fat CSA) and different lumbar spinal levels and locations were used.
A correlation between symptomatic DLSS and paraspinal muscle density was noted in our study. Changes in CSA were less significant. A possible explanation for this (changes in densities not in CSA) is that under intense muscle activation, changes in density will precede changes in CSA [
58,
76]. Furthermore, after training, the magnitude of changes in density will be 2–3 times greater than changes in CSA [
77].
Our findings indicate that the greatest change in muscle density was notable for multifidus muscle (less for psoas and erector spinae); it was also a significant factor to increase the likelihood for symptomatic DLSS in both sexes. This result is not surprising as the multifidus is the main muscle controlling spinal motion and also contributes to nearly 2/3 of the stiffness at L4-5 [
78].
Approximately one-third of the elderly population manifests radiological stenosis without symptoms [
79,
80]. Therefore, diagnosis of clinical syndrome of spinal stenosis must be carried out based on the combination of symptoms and signs together with the imaging findings [
81]. Others have also underscored the importance of history and physical examination in determining the clinical syndrome of LSS [
82] and the caution that physician should take when evaluating older patients with suspected spinal stenosis [
83].
Since the increased paraspinal density in symptomatic DLSS individuals was a specific radiological finding for this disorder and not related to other degenerative lumbar spine disease, we suggest that this measurement can be used as a radiological marker for detecting the clinical syndrome of DLSS. Nevertheless, to establish reliable standards for DLSS, a much larger sample is required.
Study limitation
The outcomes of this study warrant further investigations and verifications to determine whether our results are reproducible on different lumbar spine levels (e.g., L1, L4 and L5 rather than L3), locations and populations. An MRI study may provide higher resolution and clearer images of the soft tissue; this technique may be preferable to CT. Finally, one might argue that our finding may be due to the negative effects of renal colic in the control population (as some of them may indeed suffer from short episodes of low back pain in the past) not the ‘positive’ effect of the stenosis. However, considerable changes in the mass of muscles in individuals with a short episode of low back rather than a chronic condition, are not expected.