Association between cervical MRI findings and patient-reported severity of headache in patients with persistent neck pain: a cross-sectional study.

Background: Neck pain and headaches often co-occur, and the presence of degenerative cervical Magnetic Resonance Imaging (MRI) findings has been associated with the presence of headaches. However, previous studies have not provided conclusive evidence about their association, and imaging studies examining the associations between headache severity and MRI findings have been suggested. This study aims to investigate the associations between independent variables, single MRI findings, and an aggregate score of MRI findings, and the outcome variable, headache severity.

Methods: This cross-sectional study examined patients with neck pain and headaches in specialist care. MRI findings and outcome measures were collected at the time of clinical entrance between 2011 and 2014. The headache severity was assessed using the Neck Disability Index questionnaire. Ten degenerative MRI findings were routinely evaluated, and an overall score was derived by aggregating single findings across levels C2-C7. Univariate and multivariable ordinal logistic regression analyses assessed the associations expressed as odds ratios (OR) and 95% confidence interval (95% CI).

Results: A total of 574 patients were included. Higher headache severity was significantly associated with female sex and younger age. The presence of single cervical MRI findings was linked to lower odds of severe headaches (ORs < 1), and having two or three findings further decreased the likelihood (OR 0.40, 95% CI 0.23-0.68) compared to having none. A sensitivity analysis assessed the OR estimates for the aggregate score as robust.

Conclusions: This study showed that, among patients with persistent neck pain referred to secondary care, degenerative MRI findings in the cervical spine were inversely associated with headache severity. The association between an aggregated score of MRI findings and headache severity was stronger than that of single findings. These findings reflect associations observed within a selected clinical population and warrant further investigation in populations with differing symptom profiles.