Oral Presentation 51st International Society for the Study of the Lumbar Spine Annual Meeting 2025

Introduction

Mechanical pain low back pain (mLBP) is defined as pain that varies in intensity with movement and/or posture.  Currently, non-specific chronic LBP (nsCLBP) is classified into four mLBP categories (Figure 1):  non-mechanical (non-mLBP), mechanical pain without directional preference (mLBP-no DP), mechanical pain with flexion dysfunction (mLBP-FD), and mechanical pain with extension dysfunction (mLBP-ED).  These categories, which are based on physical exam, may not capture the complexity of mechanical pain.  The objective of this study was to use a data-driven approach to identifying mLBP phenotypes.

 Methods

Participants in a longitudinal cohort study of adults with nsCLBP (ComeBack) assessed the effect of each of the following on their LBP:  sitting, getting up from sitting, standing, walking, lifting, bending forward, and bending backward.   Each variable was scored on a 5 point Likert scale, ranging from “cannot, due to pain” to “no pain”.  Spearman’s correlation was used to determine correlation between the variables, and a non-linear principal component analysis (NLPCA) to uncover latent mLBP phenotypes. Each participant also underwent a physical examination by a physical therapist (PT), who assigned them to one of the four mLBP categories, or to Other.  A one-way Analysis of Variance (ANOVA) was conducted followed by a Tukey post-hoc test.

 Results

450 participants were included in the analysis. Correlation coefficients were generally low but statistically significant (Figure 2).  Two principal components (PC) captured the most informative features (Figure 3). PC1 (29.2% of total variance) was strongly correlated with variables associated with extension postures - ‘walking’ and ‘bending backwards’- and labelled the “Extension Dysfunction Phenotype” (ED-Phenotype).  PC2 (18.2% of total variance), was strongly correlated with variables associated with lumbar flexion- ‘sitting’, ‘lifting’, ‘getting up’, and ‘bending forwards’- and labelled the “Flexion Dysfunction Phenotype” (FD-Phenotype).  Each phenotype is a continuous variable with higher values indicating greater dysfunction.  The ANOVA found that the mean values for the phenotypes differed between mLBP categories (Figures 4 and 5), with moderate effect sizes.  Significant results from the Tukey post-hoc test  are in Table 1. Notably, the mean value for the ED-Phenotype was significantly higher in the mLBP-ED category than the mLBP-FD and mLBP-noDP categories, while the mean value of the FD-Phenotype was significantly higher in the mLBP-FD category than the mLBP-ED and mLBP-noDP categories.

 Discussion

The key finding from this exploratory analysis is that the current method for categorizing mLBP does not capture the complexity of the phenomenon.  PCA demonstrates two dominant latent mLBP phenotypes (flexion dysfunction and extension dysfunction) that underlie the observed mLBP categories, with variable degrees of each of these phenotypes within each category.  This novel finding may have implications for identifying clinically relevant nsCLBP subgroups.   Future research should focus on improving methods for measuring mLBP, correlating mLBP parameters with imaging and other biomarkers, and ultimately conducting clinical trials to determine if improved measurement of mLBP translates to improved matching of mLBP subgroups to interventions.