BASH Guideline
Section 1: The Clinical Approach
Section 2: Primary Headaches
1.3 When to consider secondary headache
1.3.1 Identifying headache due to secondary causes
Serious secondary headaches are uncommon (Alter et al., 1994; Hamilton & Kernick, 2007; Locker et al., 2006). Very few secondary causes of headache have been reliably shown to have unique headache symptoms.
The most consistent indicators for secondary headache are:
- Thunderclap (sudden onset) headache (Landtblom et al., 2002; Linn et al., 1998; Locker et al., 2006)
- Associated focal neurological deficit (Hamilton & Kernick, 2007; Locker et al., 2006; Sempere et al., 2005)
- Associated systemic features (Hamilton & Kernick, 2007; Locker et al., 2006; Sempere et al., 2005)
- Patients over the age over 50 years (Edlow et al., 2008; Ramirez-Lassepas et al., 1997)
There is currently poor evidence that different disease processes associated with headache have unique or specific clinical presentations not covered by the clinical indicators cited above. For example:
- Giant Cell Arteritis – onset age >50 years plus systemic features
- Idiopathic intracranial hypertension – abnormal neurological examination with papilloedema
- CNS infection – systemic features (e.g. pyrexia) +/- focal neurological features. Features that do not help to differentiate primary from secondary headaches are:
- Severity
- Clinical characteristics
- Treatment response
Neither severity (Ramirez-Lassepas et al., 1997) nor response to drug treatment differentiates between primary and secondary headache. For example, headaches associated with intracerebral haemorrhage (Seymour et al., 1995), subarachnoid haemorrhage (Pfadenhauer et al., 2006; Rothrock, 2005), venous sinus thrombosis (Agostoni, 2004), carotid dissection (Abisaab et al., 2004; Leira et al., 2001), and carbon monoxide poisoning (Lipton et al., 1997) have all been reported to respond to simple analgesics or triptans.
There is no specific ‘brain tumour’ type headache. The most frequent phenotype of headache associated with brain tumours is that of episodic tension-type headache, or migraine without aura (Vázquez-Barquero et al., 1994).
Patients with the clinical syndrome of migraine (with or without typical aura) or tension type headache and a normal neurological examination do not have an increased risk of a secondary precipitant (Alter et al., 1994; Kernick et al., 2008; Kurth et al., 2015).
The following may be of reassurance to busy clinicians: data from the Landmark study suggest that a new clinic diagnosis of migraine was almost always correct - 98% of patients with a clinic diagnosis of migraine met international headache society criteria for migraine (87%, or probable migraine 11%) (Tepper et al., 2004).
For other isolated headache syndromes with normal neurological examination there are insufficient data to enable a definitive conclusion.
As the presence of focal or systemic symptoms and/or abnormal neurological signs significantly increases the chance of there being an abnormality, an appropriate neurological examination – including fundoscopy – is required when assessing the patient presenting with headache.
Useful and brief ways to perform the neurological examination are found at:
http://www.youtube.com/watch?v=wyBNYB0RLvU http://www.youtube.com/watch?v=q56WgXvn0iU
1.3.2 Thunderclap headache
Thunderclap headache is the most common isolated headache associated with a secondary precipitant (Landtblom et al., 2002; Linn et al., 1998; Locker et al., 2006).
The primary concern in thunderclap headache is to exclude a subarachnoid haemorrhage. In prospective studies of thunderclap headache (primary and secondary care), subarachnoid haemorrhage was present in 19.5-25%, and in 12% headache was the only symptom (Linn et al., 1994; Morgenstern et al., 1998).
Investigation of Thunderclap Headache
- Refer immediately to hospital
- Urgent CT brain imaging
In patients presenting with isolated thunderclap headache with no other associated general or neurological symptoms, a normal neurological examination and a clear time of onset, the sensitivity of high resolution CT performed within 6 hours of onset is 98.5% - 99.85% for diagnosis of subarachnoid haemorrhage (Backes et al., 2012; Blok et al., 2015; Mark et al., 2013; Perry et al., 2011; Stewart et al., 2014).
This sensitivity drops to 90% after more than 6 hours (Backes et al., 2012).
Lumbar puncture
If CT brain is normal, a lumbar puncture for examination of cerebrospinal fluid should be performed.
The current consensus is based on the guidelines for the analysis of CSF for bilirubin in suspected SAH (Cruickshank et al., 2008):
- Measure CSF pressure when performing the LP
- Send the CSF for standard constituents – protein, glucose, microbiology, and bilirubin/xanthochromia
- Send a simultaneous blood sample for glucose, bilirubin and total protein
- The specimen for spectrophotometry should be the least blood-stained fraction of CSF to be taken. The reason is that oxyhaemoglobin may interfere with the detection of bilirubin, and is a confounding element in analysis
- Protect the CSF sample for spectrophotometry from light to reduce the degradation of bilirubin, in order to minimise false-negative results
- Current consensus is that CSF should not be examined for bilirubin earlier than 12 hours after the ictus. Bilirubin in CSF forms 9-15 hours after a bleed (Cruickshank et al., 2008; Fishman, 1980; Morgan et al., 2004).
- If negative results are obtained from both CT brain and CSF analysis from 12 hours to within 2 weeks of the onset of thunderclap headache, it can be considered that SAH is excluded as a diagnosis (Cruickshank et al., 2008; Perry et al., 2008).
Other pathologies associated with thunderclap headache
Thunderclap headache has been associated with other secondary pathologies, including infective causes, cerebrovascular causes such as intracranial haemorrhage (Luda et al., 1995), cerebral venous sinus thrombosis (Cortez et al., 2009; de Bruijn et al., 1996; Jaiser et al., 2008), malignant hypertension (Heo et al., 2009; Tang-Wai et al., 2001) and arterial dissection (Arnold et al., 2008; Biousse et al., 1992; Buyle et al., 2001; Cox et al., 1991; Fisher, 1982; Joo & Lee, 2005; Shibata et al., 2006), non-vascular neurological causes (such as pituitary apoplexy) (Cryer & Ferriman, 1971; Dodick & Wijdicks, 1998; Epstein et al., 1971; Espinosa et al., 2007; Garza & Kirsch, 2007; He, 1983; Kuzma & Goodman, 1999) and spontaneous intracranial hypertension (Asakura et al., 2001; Balgera et al., 2009; Broadley et al., 2005; Devenney et al., 2014; Famularo et al., 2005; Ferrante & Savino, 2005; Schievink et al., 2001; Vaidhyanath et al., 2007; Wang & Fuh, 2005). Spontaneous intracranial hypotension classically presents as an orthostatic headache (Mokri, 1999; Schievink, 2006).
If CT and CSF are normal, additional investigations may be indicated if warranted by the clinical presentation. There are a significant number of guidelines for thunderclap headache from different countries and different subspecialty organisations but there is no universal consensus as to when additional investigations are indicated. If CT or MR angiography is considered, the fact that this may identify incidental intracranial aneurysms or other abnormalities should be borne in mind (Vlak et al., 2011).