DOI: 10.1097/bn9.0000000000000129 ISSN: 3066-8905

Sagittal Pelvic Morphology and Postural Alignment in Chronic Low Back Pain: A Radiographic Discriminant Analysis With Multivariate Predictive Modeling

Deed E. Harrison, Ibrahim Moustafa, Paul A. Oakley, Christopher J. Colloca
Show PDF Cite

Study Design:

Cross-sectional observational radiographic study.

Objective:

To identify radiographic variables that discriminate chronic low back pain (CLBP) patients and characterize bivariate correlations between pelvic morphology (PM) indices and sagittal alignment using a parsimonious logistic model.

Summary of Background Data:

Sagittal displacement changes how intervertebral discs share loads and may increase sensitivity in spinal tissues. The posterior tangent pelvic incidence angle (PTPIA) and angle of pelvic incidence (API) are two common radiographic reference points used for biomechanical analysis, each offering different geometric perspectives.

Methods:

Standing lateral full-spine radiographs were obtained in 72 asymptomatic volunteers (49 male, 23 female; mean age: 28.3 yr; NRS 0.8) and 72 CLBP patients (28 male, 44 female; mean age: 52.0 yr; NRS 4.5; pain >3 mo). Twenty-eight radiographic variables were digitized using a validated quadrilateral vertebral body model. Statistical analyses, including ANOVA, Cohen d, point-biserial correlations, Pearson correlations, ANCOVA, multivariate OLS regression, and logistic regression with five-fold cross-validated ROC analysis, were performed. In a prespecified sensitivity analysis, age and sex were added to test whether the findings reflected demographic rather than CLBP-related postural differences.

Results:

Eighteen of 28 variables discriminated CLBP from normal, with 4 large effect sizes ( d ≥0.80): S1-hip axis distance ( d =1.276), PR_inf angle ( d =0.977), pelvic tilt angle (PTA) ( d =0.922), and sacral base angle (SBA; d =0.864). API did not significantly differ ( d =0.185; P =0.268), while PTPIA was increased in CLBP ( d =0.405; P =0.016). In volunteers, PTPIA correlated with lordosis ( r =−0.428; P <0.001); in CLBP, this relationship was absent (r=−0.137; P =0.251). API and SBA were a geometrically appropriate pairing (API referencing the sacral endplate; r =−0.596 normals; P <0.001). The groups differed in age and sex ( P <0.001). After jointly adjusting for age and sex, two large-effect variables (S1-HA, partial r =+0.267, P =0.002; SBA, partial r =−0.299, P <0.001) and PR_inf (partial r =−0.182, P =0.032) retained independent associations with CLBP group membership, while pelvic tilt angle did not reach significance (partial r =+0.122, P =0.155).

Conclusions:

Four key postural variables: S1-hip axis distance, PR_inf angle, PTA, and SBA effectively distinguish CLBP from normal, and all are modifiable. PTPIA and API are not interchangeable as PM indices; they require appropriate geometric counterparts. A three-variable logistic model predicts CLBP accurately (AUC=0.854).

Level of Evidence:

Level III—diagnostic study.

More from our Archive