Innervation of pathologies in the lumbar vertebral end plate and intervertebral disc

Abstract

Background context: Magnetic resonance imaging (MRI) has limited diagnostic value for chronic low back pain because of the unclear relationship between any anatomic abnormalities on MRI and pain reported by the patient. Assessing the innervation of end plate and disc pathologies-and determining the relationship between these pathologies and any abnormalities seen on MRI-could clarify the sources of back pain and help identify abnormalities with enhanced diagnostic value.

Purpose: To quantify innervation in the vertebral end plate and intervertebral disc and to relate variation in innervation to the presence of pathologic features observed by histology and conventional MRI.

Study design/setting: A cross-sectional histology and imaging study of vertebral end plates and intervertebral discs harvested from human cadaver spines.

Methods: We collected 92 end plates and 46 intervertebral discs from seven cadaver spines (ages 51-67 years). Before dissection, the spines were scanned with MRI to grade for Modic changes and high-intensity zones (HIZ). Standard immunohistochemical techniques were used to localize the general nerve marker protein gene product 9.5. We quantified innervation in the following pathologies: fibrovascular end-plate marrow, fatty end-plate marrow, end-plate defects, and annular tears.

Results: Nerves were present in the majority of end plates with fibrovascular marrow, fatty marrow, and defects. Nerve density was significantly higher in fibrovascular end-plate marrow than in normal end-plate marrow (p<.001). Of the end plates with fibrovascular and fatty marrow, less than 40% were Modic on MRI. Innervated marrow pathologies collocated with more than 75% of the end plate defects; hence, innervation was significantly higher in end plate defects than in normal end plates (p<.0001). In the disc, nerves were observed in only 35% of the annular tears; in particular, innervation in radial tears tended to be higher than in normal discs (p=.07). Of the discs with radial tears, less than 13% had HIZ on T2 MRI. Innervation was significantly less in radial tears than in fibrovascular end-plate marrow (p=.05) and end-plate defects (p=.02).

Conclusions: These findings indicate that vertebral end-plate pathologies are more innervated than intervertebral disc pathologies and that many innervated end-plate pathologies are not detectable on MRI. Taken together, these findings suggest that improved visualization of end-plate pathologies could enhance the diagnostic value of MRI for chronic low back pain.

Keywords: End plate; Innervation; Intervertebral disc; Low back pain; Modic change.

Figure 1

Sub-regions of motion segments and various types of endplate and disc pathologies (see Methods for histological descriptions). (A) Sub-regions used to classify the location of endplate nerves and pathologies. (B) Fibrovascular endplate marrow. (C) Fatty endplate marrow. (D) Avulsion-type endplate defect. (E) Nodule-type endplate defect. (F) Fracture-type endplate defect. (G) Concentric annulus tear. (H) Radial annulus tear. Note: arrows indicate specified pathology in images where multiple pathologies are present; Heidenhain trichrome stain.

Figure 2

Comparison of nerve density between cranial and caudal endplates and among anterior, central, and posterior regions. (A) Cranial and caudal endplates had similar nerve density (p = 0.30, Mann-Whitney U test). (B) Anterior, central, and posterior endplate regions had similar nerve density (p = 0.60, Kruskal-Wallis test). However, the nerve density in the anterior endplate region was significantly greater than that of the central endplate region when the nerve densities of the regions were compared pair-wise for each endplate (p = 0.01, Wilcoxon signed rank test). Note: data points in (B) represent the combined densities of cranial and caudal endplates for a given region. Horizontal lines indicate group means.

Figure 3

Examples of PGP 9.5-positive nerves in the vertebral endplate and intervertebral disc. (A) Typical perivascular nerve in hematopoietic marrow adjacent to the endplate. (B) Nerves within fibrovascular endplate marrow. (C) Nerves within fatty endplate marrow. (D) Nerves in avulsion-type endplate defect. (E) Typical nerves in outer annulus of the disc. (F) Nerves in the outer annulus of a disc with a radial tear. Note: red box indicates location of high-magnification region of interest; arrows indicate PGP 9.5-positive nerve fibers.

Figure 4

Endplate cartilage avulsion with fibrovascular marrow (left) and corresponding MR images (right). Donor information: 63-year-old woman; L3–L4 disc.

Figure 5

Comparison of nerve density among endplates with abnormal marrow and between endplates with and without endplate defects (see Methods for classification). (A) Abnormal endplate marrow had significantly greater nerve density than did normal endplate marrow (p < 0.005, Kruskal-Wallis test). The nerve density in fibrovascular marrow (p < 0.001), but not fatty marrow (p = 0.15), was significantly higher than that in normal marrow (Mann-Whitney U post-hoc tests). (B) The nerve density in and adjacent to endplate defects was significantly greater than that of normal endplates (p < 0.0001, Mann-Whitney U test). Note: darkened circles indicate the defected endplates that also contained either fibrovascular or fatty endplate marrow. Horizontal lines indicate group means.

Figure 6

Comparison of the number of nerve fibers between discs with and without annulus tears. The number of nerve fibers tended to be greater in discs with radial tears than in discs without tears (p = 0.07) and was similar between discs with concentric tears and discs with no tears (p = 0.94, Mann Whitney U tests). Note: darkened circles indicate discs with both radial and concentric tears. Horizontal lines indicate group means.

Figure 7

Comparison of nerve fibers between discs with radial tears and endplates with pathologies. (A) The number of nerve fibers was significantly lower in discs with radial tears than in endplates with fibrovascular marrow (p = 0.05, Mann-Whitney U test). (B) The number of nerve fibers was significantly lower in discs with radial tears than in defected endplates (p = 0.02, Mann-Whitney U test). Note: darkened circles indicate the defected endplates that also contained either fibrovascular or fatty endplate marrow. Horizontal lines indicate group means.