Diagnosis of subarachnoid haemorrhage: Systematic evaluation of CT head diagnostic accuracy and comparison with the 2022 NICE guidelines

Abstract

Introduction

Aneurysmal subarachnoid haemorrhage has a high incidence, and morbidity. It has been suggested that a negative non-contrast CT head can rule out SAH if performed within 6 h of symptom onset.

Research question

What is the sensitivity of CT head at ruling out SAH stratified by time-point, and what is the potential impact of omitting Lumbar Puncture (LP) from the diagnostic pathway?

Material and methods

Systematic review and meta-analysis (PROSPEROID CRD42022379929). Three databases were searched, and articles published between January 2000–May 2022 included (Search date 27^th November 2022). Primary objective was diagnostic accuracy of CT scans for detecting SAH at <6 h from symptom onset, including reported sensitivity, and specificity values.

Results

63 articles were included (38,237 patients, 7673 with SAH). Pooled CT head sensitivity was 0.94 for excluding SAH (22 studies, 95% Confidence Interval [CI] 0.90–0.97). At <6 h, CT head sensitivity was 0.995 (6 studies, 95% CI 0.941–1.000). Most studies (57.1%, n = 36/63) were classified as high risk of bias. If LP was removed from the diagnostic pathway in the UK, assuming an incidence of 4800 SAH per-year, 336 SAH would be missed per-year, 24 per-year if LP was removed for negative CT < 6 h (95% CI 0–278) and 58 per-year if mean sensitivity is used (95% CI 0–240).

Discussion and conclusion

CT head appears to be highly sensitive at excluding SAH <6 h from symptom onset. High quality, prospective data is required to further established the utility of early (<6 h) negative CT head. We recommend that if there is strong clinical suspicion of SAH, yet CT head is reported negative <6 h of symptom onset, that a LP be performed.