Contrary to our hypothesis, there were no significant differences between the superficial and deep regions of the multifidus for histologically measured area of fat, muscle area, fiber cross sectional area, muscle regeneration markers, muscle degeneration markers, and vascularization. While total collagen content between the two regions was the same, the superficial region of the multifidus was found to have less loose and more dense collagen than the deep region. These findings suggest that the functional differences that have been previously observed between the superficial and deep regions of the multifidus are not a result of muscle composition, and that muscle degeneration does not occur preferentially in a given region of the multifidus.
Differences in paraspinal muscle activation during motion and muscle composition have been highlighted as targets in assessing both muscle health and potential for improvement with physical therapy[22, 42]. Lumbar multifidus has been especially implicated given its assumed role in spinal stability and common pathological changes in LBP[43–47]. The role of the multifidus has further been parsed out based on its distinct superficial and deep layers, with the superficial multifidus thought to be involved primarily in lumbar extension and the deep multifidus focused on control of dynamic intersegmental stability via intervertebral compression, thus providing spinal stability[22, 25, 48]. Given these differences in activation patterns, it has previously been hypothesized that there would be differences in proportion of muscle fiber types between layers. Specifically, the deep multifidus would have a higher proportion of type 1 fibers compared to the superficial multifidus[22, 48], since type I fibers are fatigue resistant and most suitable to providing the constant low load activity necessary to provide dynamic lumbar stability. Importantly, this supposedly higher proportion of type I fibers in deep multifidus has been used as a target for the low load, tonic exercises often implemented in LBP exercise rehabilitation regimens[22, 49]. Studies comparing the proportion of these fibers in patients have demonstrated conflicting results. One study examining both autopsy specimens and biopsies from patients undergoing surgery for disc herniation found a significantly greater proportion of type I fibers in deep multifidus[50]. However, the majority of the literature has demonstrated no difference in fiber-type distribution based on multifidus depth across a range of ages, consistent with the results in this study[8, 9, 51–55].
Regional differences in collagen within the lumbar multifidus have not been thoroughly explored. Our histological results found that although the superfical and deep muscle biopsies both had an elevated area fraction of collagen (~25%) and superficial mutifidus biopsies had less dense collagen and more loose collagen than deep multifidus biopsies. Increased collagen expression within muscle tissue has been associated with chronic inflammatory changes and subsequent irregular remodeling thought to be related to LSP[56]. Such changes often accompany muscle atrophy and increased fat deposition to comprise what is defined as tissue fibrosis. Progression of paraspinal muscle fibrosis is associated with decline in muscle functionality and recovery at these late stages is prolonged. Previous animal studies examining thoracolumbar fascia (TLF) in rats have shown an organized distribution of collagen based on tissue depth, with more superficial layers having a higher percentage of densely packed collagen fibers compared to deeper layers of the TLF, which were composed of loose, irregularly oriented collagen fibers[57]. Increases in connective tissue within multifidus appear to be related to duration of LBP, with injury models only demonstrating increases in collagen area after several months[14]. Additionally, fibrotic gene expression has been shown to be upregulated in individuals with chronic symptoms compared to those with acute symptoms[58]. These results suggest that fibrotic deposition within muscle may affect muscle function in patients with LSP, further predisposing patients to maladaptive muscle changes and fatty infiltration. The mechanisms underlying the increase in collagen deposition in multifidus are not fully understood at this time, but physical activity appears to reduce the inflammation associated with this tissue fibrosis, which may improve whole muscle function[56, 59–64].
Markers of muscle vascularity, damage, and degeneration did not significantly differ between superficial and deep samples Although the average percentage of centralized nuclei for each biopsy level was higher than expected for healthy skeletal muscle (3%), this percentage was lower than values reported for other studies examining lumbar spine pathology[10, 11]. Despite the high proportion of markers for muscle recovery (centralized nuclei for regeneration and Pax7+ for satellite cell density), the number of muscle fibers that exhibited degenerative characteristics remained high. This is consistent with previous results from our group that suggest the amount of muscle degeneration in chronic LBP patients outpaces the ability to restore a healthy muscle phenotype[10, 65].
Although this study provides the first documented comparison between superficial versus deep histology and MR imaging of the lumbar multifidus in patients with LSP, there were several limitations that must be considered in assessing our results. First, the nature of biopsy collection was such that identifying and matching the exact corresponding region to the MRI was not possible. We defined an ROI on MRI as close to the biopsy location as possible, based on a standardized biopsy region in our protocol and biopsy level identification via intraoperative fluoroscopy. By involving a limited number orthopedic surgeons and training them on biopsy technique and location, we attempted to mitigate variability in biopsy location. Second, it is impossible to measure collagen from routine clinical imaging (i.e. T1-weighted, T2-weighted MRIs) due to the short transverse relaxation time of fibrotic tissues. While MRI pulse sequenceses sensitive to collagen in tissues exist - such as ultrashort echo time - they are not routinely used to assess patients with LSP. Third, this was a prospective study and only 16 patients were included. However from this sample population, it was unlikely that differences in muscle microstructure/macrostructure between the superficial and deep multifidus would be evident with a larger sample size. Fourth, the majority of patients in this study all had an identifiable lumbar spine pathology such as stenosis or disc degeneration resulting in radicular symptoms, which may differentially impact superficial and deep regions of the multifidus as compared to nonspecific LBP. While biopsies from subjects with nonspecific LBP would allow for the differentiation of histological changes as a result of potential denervation, this patient population is less likely to receive surgical intervention without an identified anatomical pain generator and ethical issues often preclude obtaining regional biopsies. Furthermore, this study did not take into account some patient factors that have been implicated in muscle quality, such as spinal alignment[66, 67], or level of physical activity[56, 59].