In this issue the paper by Cumurciuc et al¹ (pp 1084–7) reports 17 cases from 123 consecutive patients with cerebral venous thrombosis (CVT) in whom the only neurological sign was headache. The series is biased by the fact that the Department of Neurology of the Lariboisiere Hospital in Paris has particular competence both in headache and stroke, and runs an emergency headache centre where patients can consult a headache specialist without long waiting times. What are the lessons to be learned from this paper?

1. Computed tomography (CT) without contrast is not sensitive enough to rule out CVT in patients with progressive headache.

2. Most patients with CVT had risk factors for venous thrombosis. The combination of unexplained headaches with risk factors like oral contraceptives,² systemic lupus erythematosus (SLE), or recurrent venous thrombosis justifies the use of CT with contrast or magnetic resonance imaging (MRI).

3. The new classification of the International Headache Society still includes the entity of primary thunderclap headache.³ The authors disagreed within themselves whether a primary thunderclap headache exists. I am convinced that all patients with thunderclap headache have an underlying cerebral disease, which has to be identified. Symptoms include warning leaks of aneurysms, other cerebral bleeds, meningeal inflammation, and, as shown in this paper, cerebral venous thrombosis.

4. An unexplained question is whether the prognosis of patients with CVT and headache as the only symptom is similar to patients with focal neurological signs or epileptic seizures. The overall prognosis of CVT is much better than previously suspected.^4–6 Therefore it would be interesting to know whether these patients should be treated with iv heparin, as was the case in the study from Paris, and how long anticoagulation should be sustained.

5. A particular problem is lateral sinus thrombosis. The group led by Bousser is experienced enough to be able to distinguish between sinus thrombosis and sinus dysplasia or aplasia. Less experienced neurologists might mistake sinus dysplasia for thrombosis and patients are exposed unnecessarily to long term anticoagulation.

6. Patients in this study were identified as patients with CVT and asked for symptoms. It would be of interest to perform a prospective study the other way around. The question is how many patients with new and progressive headache or patients with thunderclap headache have CVT. In this way one could perform a cost-benefit analysis of more extended imaging with computed tomography angiography (CTA), magnetic resonance angiography (MRA), magnetic resonance venography (MRV), or traditional angiography.

In summary this important paper has sharpened my view of patients with progressive or thunderclap headache and will prompt me more often than in the past to consider CVT.