Pooled analysis of patients with thunderclap headache evaluated by CT and LP: Is angiography necessary in patients with negative evaluations?

Sean I. Savitz*, Emily B. Levitan, Robert Wears, Jonathan A. Edlow

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

24 Citations (Scopus)

Abstract

Background: Severe, abrupt onset headache raises concern for aneurysmal subarachnoid hemorrhage (SAH). The current standard work-up is brain CT scan followed by LP if the CT is non-diagnostic in patients with a normal neurological exam. Some have suggested that angiography is also indicated in this common clinical situation. Is evaluation with brain CT and LP for thunderclap headache to rule out SAH sufficient and is angiography needed? Methods: We systematically searched for studies that followed neurologically-intact patients with thunderclap headache and normal CT and LP for at least 1 year. The primary outcome was SAH. We estimated the proportion of patients who developed SAH and the one-sided upper 95% confidence bound. Results: Seven studies including 813 patients were identified. None of the patients developed SAH during follow-up (pooled proportion = 0, upper 95% confidence bound = 0.004). Conclusion: Although our methods have important limitations, we believe that this analysis will give clinicians better tools to decide whether or not to pursue further work-up with angiography in patients with thunderclap headache and normal neurological exam, CT, and LP.

Original languageEnglish
Pages (from-to)123-125
Number of pages3
JournalJournal of the Neurological Sciences
Volume276
Issue number1-2
DOIs
Publication statusPublished - 15 Jan 2009
Externally publishedYes

Keywords

  • Angiography
  • Headache
  • Imaging
  • Lumbar puncture
  • Subarachnoid hemorrhage

Fingerprint Dive into the research topics of 'Pooled analysis of patients with thunderclap headache evaluated by CT and LP: Is angiography necessary in patients with negative evaluations?'. Together they form a unique fingerprint.

Cite this