Oral Presentation 50th International Society for the Study of the Lumbar Spine Annual Meeting 2024

Introduction:

Modic changes (MC) are associated with endplate damage, neoinnervation, and chronic low back pain (cLBP). Nonetheless, treatment outcomes in patients with MC are discrepant, and there is evidence that discrepant treatment outcomes in patients with MC reflect poor patient stratification. The need for better patient stratification also suggests uncertainty in the etiology of MC. Two MC etiologies have been proposed — one positing disc infection with bacteria, e.g., Cutibacterium acnes (C. acnes), and the other auto-immune. Despite evidence for these two etiologies, non-invasive biomarkers do not exist, which makes it difficult to stratify MC patients and interpret their treatment outcomes. In this work, we propose that MR spectroscopy (MRS) measurement of intradiscal propionic acid (PA), a metabolite of C. acnes, can clarify MC etiology and help stratify MC patients.

Methods:

MRI and MRS data from 88 patients with cLBP (>3-months) were included in this prospective study. 3T MRI of the lumbar spine included sagittal T₁- and T₂-weighted sequences. MRS (2-3 discs per patient) were acquired using a point resolved spectroscopy sequence. Water signals were suppressed using three chemical selective suppression RF/gradient pulse pairs with fat suppression enabled. Acquired spectra were filtered, baselined, and artifact-corrected. The area-under-the-curve of the PA peak was normalized to voxel size (normalized AUC). MC type of the adjacent endplates was graded on MRI: no MC; MC1; MC2. We used a mixed effects model accounting for multiple measurements per patient to estimate the mean AUC, and a Tukey HSD post-hoc test to test for group differences. p < 0.05 was considered significant.

Results:

MCs were present in 56% of patients (49/88; 33/55 male/female; mean±SD 52.6±14.6 years; PROMIS T-Score 59.1±7.6; PEG-3 4.52±2.20). MRS data were acquired in 222 discs adjacent to endplates with no MC (180 discs), MC1 (19 discs), and MC2 (23 discs). Discs adjacent to MC1 had significantly higher PA levels (normalized AUC 3.57±0.64 [mean±SEM]), on average, than discs adjacent to no MC (1.94±0.21, p=0.039) and MC2 (1.22±0.58, p=0.018, Figure 1). PA levels varied amongst discs, and the likelihood of a measurable PA level, i.e., signal of the PA peak exceeded the background level, was higher for discs adjacent to MC1 (13/19 discs) than no MC (100/180 discs) and MC2 (10/23 discs).

Discussion:

Results indicated that discs adjacent to MC1 had higher PA concentrations, on average, than discs adjacent to no MC and MC2. MC-adjacent discs also had a wide range of PA concentrations, including very low concentrations. This suggests that some MC may be bacterial in origin, and it is consistent with histopathology findings suggesting two different etiologies for MC. These data are notable because they are the first non-invasive evidence of differences in C. acnes load amongst MC-adjacent discs, and because they challenge the belief that the disc space is sterile. One implication relates to patient stratification: PA levels may be useful for understanding why some patients with MC improve or respond to treatment while others do not.