Response

Editor – We thank Dr Thompson for the interest in our paper and he raises some valid points.¹ Indeed, we have not strictly adhered to the inclusion criteria of the original Ottawa rule study from Perry et al.² Using the clinical information provided on the computed tomography (CT) request we attempted to ascertain, in so far as possible, those patients undergoing a CT of the head for the investigation of subarachnoid haemorrhage (SAH). We included patients whose requests included a working diagnosis of SAH or clinical information such as sudden onset headache, thunderclap headache or ‘worst headache of life’. As a retrospective study this represented our best estimation of the patient cohort undergoing CT of the head for the investigation of SAH, although as the author rightly states these patients may not truly have been suspected of this diagnosis. This is evident in the subsequently low proportion of patients in whom a lumbar puncture was performed (32%). We are, as radiologists and as researchers, limited by the clinical information that has been provided in the request. However, a subgroup analysis of patients (n=65; 18%), who do meet the strict inclusion criteria has been performed and detailed in our article. In short, the Ottawa rule was 100% sensitive in this cohort and missed no cases of SAH.

Rather than being misleading, the results of our article may in fact be hypothesis generating. In the larger cohort of patients, using the less stringent inclusion criteria, the Ottawa rule was still 100% sensitive. Although not described in our paper, the Ottawa rule did not miss any important intracranial diagnosis in this cohort, including viral meningitis (n=6; 1.6%), subdural haematoma (n=3; 0.8%), intraparenchymal haemorrhage (n=5; 1.4%), arteriovenous malformation (n=2; 0.6%) and primary brain neoplasm (n=1; 0.3%). The scope of the Ottawa rule therefore may be wider than previously described and its clinical applicability may not be limited only to patients with a thunderclap headache. This study, however, is retrospective. These results should therefore be considered as hypothesis generating rather than confirmatory and would require validation within the context of a prospective study.

The results of our study add to the growing body of evidence for the use of this tool and although it will not usurp the opinion of the clinician it may be helpful for risk stratification and to facilitate the discussion with radiological colleagues when requesting CT in patients with acute non-traumatic headache presenting to the emergency department.