Background context: A range of morphologic and psychosocial variables has been suggested as risk factors for serious low back pain (LBP) illness. Although the relative contributions of structural and psychosocial variables are intensely debated, the validity of differing hypotheses has proven difficult to test because the incidence of serious disabling LBP illness is low in healthy subjects. These factors dictate the requirement for large sample sizes, extensive structural imaging and extended longitudinal study. Previous studies included either small cohorts with intensive imaging testing or large population studies that do not establish a detailed morphologic baseline.
Purpose: To establish, using a strict patient sample design, the relative contribution of structural and psychosocial determinants of serious LBP illness among subjects with no previous LBP disability or clinical LBP illness.
Study design/setting: A prospective, longitudinal study of subjects with high risk factors for serious LBP as determined by structural and psychosocial characteristics.
Patient sample: One hundred subjects with known mild persistent low back pain and a 2:1 ratio of chronic (non-lumbar) pain syndrome were recruited from a study population with a predisposition to disc degenerative disease, to undergo baseline examination, testing and 5-year follow-up.
Outcome measures: Observations were made at 6-month intervals over 4 to 6 years (mean, 5.3) for the after primary outcomes measures: episodes of serious back pain (visual analogue scale [VAS] > or =6), episodes of occupational disability less than 1 week, episodes of occupational disability for 1 week, remission episodes of all back pain symptoms at least 6 months and medical visits primarily for LBP evaluation and treatment.
Methods: Lumbar magnetic resonance imaging (MRI), lumbar provocative discography (in psychometrically normal subjects), physical examinations, medical and work histories and psychometric testing were performed at baseline. Imaging and psychometric testing were graded by blinded examiners. A scripted interview was conducted every 6 months during follow-up by independent research assistants who also were blinded to patient baseline data. The interview covered interval medical, occupational and accident or injury histories.
Results: Psychosocial variables strongly predicted both long- and short-term disability events, duration and health-care visits for LBP problems (p<0.0001-0.004). The likelihood of a sustained remission from the baseline persistent (subclinical) LBP appeared to be linked to occupation factors (leaving a heavy labor occupation; p=0.0001), neurophysiologic variables (chronic nonlumbar pain; p=0.0002) and psychometric profiles at baseline (DRAM and FABQ-PA; p=0.003-0.002). Of the structural findings measured only moderate or severe Modic changes of the vertebral end plate were weakly associated with an adverse outcome. A positive provocative discogram at baseline did not predict any future adverse event.
Conclusion: The development of serious LBP disability in a cohort of subjects with both structural and psychosocial risk factors was strongly predicted by baseline psychosocial variables. Structural variables on both MRI and discography testing at baseline had only weak association with back pain episodes and no association with disability or future medical care.