Neuroimaging findings in headache with normal neurologic examination: Systematic review and meta-analysis.

OBJECTIVE: To determine if pooled estimates of the prevalence of unexpected findings in patients with headache and normal neurologic examination support current expert opinion-based neuroimaging guidelines. METHODS: We searched PubMed and EMBASE for studies reporting neuroimaging findings in patients with headache and normal neurologic examination up to September 30, 2017. The overall and disease-specific prevalence of unexpected findings were pooled through random-effects meta-analysis. This study is registered with PROSPERO, registration number CRD42017079714. RESULTS: In forty-one studies including 15,760 participants, the overall prevalence of unexpected findings and normal variants was 17.5% (95% CI: 13.1-22.3). The prevalence was 26.6% (95% CI: 15.5-39.4) in studies using MRI only. The prevalence of vascular, neoplastic, and non-neoplastic findings was 6.6%, 1.4%, and 9.6%. The pooled disease-specific prevalence was 2.0% for stroke, 1.8% for aneurysms, 0.8% for subdural hematoma, 0.7% for hydrocephalus, 0.2% for glioma, and 0.1% for meningioma. In secondary analysis, there was 0.4% increase in the prevalence of vascular unexpected findings with each 1% increase in the proportion of migraine with aura (p-value for meta-regression = 0.005). CONCLUSIONS: In patients with headache and normal neurologic examination, important vascular and neoplastic unexpected findings are rare and better detected with MRI. This supports current American College of Radiology and European Headache Federation recommendations to avoid systematic imaging in such patients and prefer MRI when imaging is needed.

Neuroimaging of headaches in patients with normal neurological examination: protocol for a systematic review.

INTRODUCTION: Headache disorders (HD) are among the most frequent neurological disorders seen in neurology practice. Because secondary HD are rare, patients' examination is most often unremarkable. However, the will to relieve patients' anxiety and the fear of prosecutions lead to overuse of neuroimaging thus resulting in the discovery of incidental findings (IF) or normal variants that can lead to futile or harmful procedures. Knowing the probability of identifying a potentially clinically significant lesion in patients with isolated headache could facilitate decision-making and reduce health costs. This review aims to determine the prevalence of incidental findings and normal anatomic variants (NAV) on neuroimaging studies performed in patients presenting with headache and normal neurological examination. METHOD AND ANALYSIS: Studies reporting neuroimaging findings in patients with headache and normal neurological examination and published before the 30 September 2017 will be identified by searching PubMed, Medline and EMBASE (Excerpta Medica Database). Relevant unpublished papers and conference proceedings will also be checked. Full texts of eligible studies will then be accessed and data extracted using a standard data extraction sheet. Studies will be assessed for quality and risk of bias. Heterogeneity of studies will be evaluated by the χ2 test on Cochrane's Q statistic. The prevalence of NAV and IF across studies and in relevant subgroups will be estimated by pooling the study-specific estimates using a random-effects meta-analysis. Visual analysis of funnel plot and Egger's test will be used to detect publication bias. The report of this systematic review will be compliant with the Meta-analysis of Observational Studies in Epidemiology guidelines. ETHICS AND DISSEMINATION: The current study is based on published data; ethical approval is, therefore, not required. The final report of this systematic review will be published in a peer-reviewed journal. Furthermore, findings will be presented at conferences and submitted to relevant health authorities. TRIAL REGISTRATION NUMBER: CRD42017079714.