The relationships between low back pain and lumbar lordosis: a systematic review and meta-analysis

Abstract

Background ContextClinicians regard lumbar lordotic curvature (LLC) with respect to low back pain (LBP) in a contradictory fashion. The time-honored point of view is that LLC itself, or its increment, causes LBP. On the other hand, recently, the biomechanical role of LLC has been emphasized, and loss of lordosis is considered a possible cause of LBP. The relationship between LLC and LBP has immense clinical significance, because it serves as the basis of therapeutic exercises for treating and preventing LBP. PurposeThis study aimed to (1) determine the difference in LLC in those with and without LBP and (2) investigate confounding factors that might affect the association between LLC and LBP. Study DesignSystematic review and meta-analysis. Patient SampleThe inclusion criteria consisted of observational studies that included information on lumbar lordotic angle (LLA) assessed by radiological image, in both patients with LBP and healthy controls. Studies solely involving pediatric populations, or addressing spinal conditions of nondegenerative causes, were excluded. MethodsA systematic electronic search of Medline, Embase, Cochrane Library, CINAHL, Scopus, PEDro, and Web of Science using terms related to lumbar alignment and Boolean logic was performed: (lumbar lordo*) or (lumbar alignment) or (sagittal alignment) or (sagittal balance). Standardized mean differences (SMD) and 95% confidence intervals (CI) were estimated, and chi-square and I2 statistics were used to assess within-group heterogeneity by random effects model. Additionally, the age and gender of participants, spinal disease entity, and the severity and duration of LBP were evaluated as possible confounding factors. ResultsA total of 13 studies consisting of 796 patients with LBP and 927 healthy controls were identified. Overall, patients with LBP tended to have smaller LLA than healthy controls. However, the studies were heterogeneous. In the meta-regression analysis, the factors of age, severity of LBP, and spinal disease entity were revealed to contribute significantly to variance between studies. In the subgroup analysis of the five studies that compared patients with disc herniation or degeneration with healthy controls, patients with LBP had smaller LLA (SMD: −0.94, 95% CI: −1.19 to −0.69), with sufficient homogeneity based on significance level of .1 (I2=45.7%, p=.118). In the six age-matched studies, patients with LBP had smaller LLA than healthy controls (SMD: −0.33, 95% CI: −0.46 to −0.21), without statistical heterogeneity (I2=0%, p=.916). ConclusionsThis meta-analysis demonstrates a strong relationship between LBP and decreased LLC, especially when compared with age-matched healthy controls. Among specific diseases, LBP by disc herniation or degeneration was shown to be substantially associated with the loss of LLC.