Other primary headaches—thunderclap-, cough-, exertional-, and sexual headache

Thunderclap headache (TCH) describes a sudden severe explosive onset headache. The severity itself is not diagnostically helpful [ 8– 10], it is the rapidity of onset, with pain developing from zero to maximum in seconds to minutes.

Landtblom has provided the only population-based data citing an incidence of 43 per 100,000 persons (> 18 years) per year of all sudden onset headache (within 10 s) [ 9] with primary TCH (PTCH) being cited as occurring in 38 per 100,000 persons per year.

It is best to understand TCH from a historical perspective. Isolated thunderclap onset headache is the most consistently reported presentation of a secondary headache, with the time frame of the pathology and onset of the headache supporting an association [ 5]. The term ‘isolated’ infers the headache to be in neurologically and systemically intact patients. The most dramatic presentation is subarachnoid haemorrhage (SAH); one third present with isolated headache. It is from this literature that our knowledge of TCH has evolved. The prevalence of spontaneous SAH is about nine cases per 100,000 people a year, thus a rare disease [ 11]. The disproportionate attention it has garnered lies in the high mortality. A meta-analysis of 33 studies reported a case-fatality rate of 8∙3–66∙7% (variability depending on country of data acquisition). Also reported was a 17% reduction in case-fatality over three decades, an observation supported in subsequent populations [ 12]. The improvement has been attributed to better management, primarily rapidity of access to medical care.

Of all spontaneous SAH, 85% are aneurysmal while 10% are perimesencephalic, thus venous and carrying a generally benign prognosis. The most definitive method of diagnosis is unenhanced computerised tomography (CT), which will show SAH in 99% of cases if imaged within 6 h of ictus [ 13, 14]. In those with negative CT, examination of cerebrospinal fluid (CSF) for altered blood products remains positive from 12 h to 2 weeks in 100% of cases [ 15– 17]. The risk of angiography in lieu of CSF examination, or in late presentations, is detection of an incidental aneurysm and potential unnecessary and cost- ineffective intervention [ 18, 19]. It has been argued that using third generation CT scans to image patients within 6 h of ictus, the sensitivity and specificity for detecting aneurysmal subarachnoid haemorrhage is 100%. While the detection of subarachnoid blood on CT- negative patients is about 7%, the cause for the CSF xanthochromia in these patients runs a benign course [ 20, 21].

The existence of a benign form of thunderclap headache is supported by longitudinal observations of patients investigated with normal CT and CSF examinations who did not develop any subsequent adverse sequalae and continued to experience recurrent attacks (Table 3). Moreover, it was shown that clinically it is not possible to reliably differentiate those who have experienced a subarachnoid haemorrhage from those with subsequently normal investigations [ 8, 9, 22– 25]. In both groups the onset of the headache is within seconds to a few minutes in most; in 20% escalation of severity can be greater than 5 min [ 8, 10]. There is no difference in severity, past history of headache, precipitating event (exertion or valsalva) nor additional neurological symptoms. Moreover all groups had previously experienced similar TCH. Clinical guidelines have been developed for diagnosing SAH in patients presenting with acute headache with almost 100% sensitivity but poor specificity of 13.6% [ 26].

Thunderclap onset headache has also been associated with a multitude of secondary pathologies. A systematic review of the literature from 2014 identified 119 causes in a total of 2345 cases reported in isolation, case series or cohorts [ 27]. The diagnoses relied on the combination of ‘sudden’ and ‘severe’. The definition of ‘sudden’ was not detailed and it was acknowledged that the majority had not used ICHD criteria. Within each pathological category are a number of likely incidental findings.

By far the largest contribution was primary headache in 459 cases, 213 primary thunderclap headaches, followed by primary sexual headache in 182, bath-related headache in 37 and exertional headache and combinations of the aforementioned. Three-hundred and ninety-eight cases were precipitated by cerebrovascular disorders, 206 from SAH, 46 from other sources of intracranial haemorrhage, venous and arterial thromboembolism, intracranial dissection, stroke, hypertensive encephalopathy and vasculitis. Included were 18 cases likely to be incidental findings, primarily unruptured cerebral aneurysms. One hundred and sixty cases of sudden and severe headache were reported in association with infection, 44% affecting the CNS and the remainder systemic with likely CNS involvement or, encompassed within the terminology of a ‘viral illness’ without further elaboration. One hundred and seventy-three patients were diagnosed with reversible vasoconstrictive ‘syndrome’ (RCVS). The largest contribution to the 119 non-vascular precipitants was from pituitary apoplexy in 43 cases and 32 related to alterations in CSF pressure. Less than 20 cases were related to the peripartum state and drugs.

The current ICHD of primary thunderclap headache defines the rapidity of onset as being within 1 min (Table 2). This will capture up to 75% of primary and secondary thunderclap headache, while altering this definition to escalation within 5 min will capture 95% of cases [ 8]. The subsequent headache, following the initial thunderclap, most frequently has clinical features consistent with migraine (Table 3).

In summary, TCH is the most common headache syndrome associated with a secondary precipitating pathology. The majority will have PTCH [ 9]. Primary and secondary thunderclap headache, however, cannot be reliably differentiated clinically thus, all patients should be investigated. Within 6 h of ictus, it is likely that aneurysmal SAH can be identified by CT alone. However, late presentations and less common causes, will require further investigation.

The most consistent treatment is Indometacin, effective over a dose range between 25–250 mg. Mathew first reported the response to Indomethacin in a double-blind placebo controlled manner in two patients with cough headache resistant to other tried preventatives [ 40]. Response was achieved in 1–4 weeks and maintained over an 18 month follow-up period.

The aim is to achieve an optimal dose which successfully suppresses the symptoms and then periodically reduce the dose to see whether the symptoms have become quiescent [ 30, 32, 33].

There are also anecdotal reports of treatment responses to acetazolamide [ 41], naproxen [ 42], propranolol [ 43] and lumbar puncture [ 30, 44]. The latter involved removal of 40 ml of cerebrospinal fluid. Constituents were normal.

Despite the prevalence data of 1%, the published cohorts of cough headache remain small. The assumption is that this is because the disorder tends to be self-limiting and remissions prolonged. The primary cases tend to be more common than secondary cases (Tables 4 and 5). There is a preponderance of published posterior fossa lesions purported to be secondary causes, with ACM I disproportionately represented [ 45]. The natural history of ACM I in both asymptomatic and symptomatic patients (which includes those with a syrinx) is benign and without progression in the majority [ 46]. Analysis of ACM I cohorts show that the prevalence of headache is generally in line with population prevalences, although that for cough headache has been reported higher than the 1% population prevalence [ 47– 49]. Although outcome of headache in those operated upon is reported to be good this is based on a relatively limited follow-up period. In one cohort of 96 patients 75% had resolution or improvement, the follow-up period was a mean of 3.6 years; the 30 day post-operative complication rate was 27% and included one death [ 49]. In another cohort a comparison was made between conservative (68 patients) and operated groups (109 patients). The improvement in cough headache was 40% and 94.6%, respectively, and for migraine 61.5% and 92.9%. The mean follow-up period for the conservative group was 4.8 years and that of the operated group only 1.3 years which, does not take into account the natural history of the primary disorder [ 47]. Moreover, none of the reports can account for the significant impact of surgical placebo [ 50]. Risks of intervention include wound sepsis, meningitis, stroke, CSF fistula and hydrocephalus [ 46].

Secondary cough headache has also been reported to respond to Indometacin [ 51, 52]. In the case report of Buzzi, a 54 year-old patient presented with a 10 year history of cough headache as an isolated presentation. She was found to have an abnormal examination and imaging showed a Chiari 1 malformation and a syrinx extending from C7-T7. The episodes of Valsalva-precipitated headache responded to Indomethacin 25 mg bd.

This case highlights the importance of adhering to the Hippocratic principle to ‘Do No Harm’. This is particularly pertinent in the cases of patients presenting with valsalva headache, normal neurological examination and a Chiari I. It is clear that not all patients with a Chiari I develop cough headache [ 45]. The onset of the Valsalva headache occurs in the 5–6th decades while the malformation is considered congenital. Conservative management of paediatric populations, looking at natural history, suggest that the outcome in most patients is benign [ 53, 54]. There are occasional cases where there is both worsening and improvement; of the latter 3 of 147 patients had spontaneous improvement of the associated syrinx [ 54]. In Symond’s original cohort of cough headache patients, one patient developed symptoms after successful treatment for an acoustic neuroma. Another gave a 10 years history of cough headache with a 4 years history of spontaneous remission; when seen at 10 years she was found to have basilar compression from Paget’s disease with abnormal cranial nerve examination and a mild spastic paraparesis [ 31]. There is a single case report of cough headache, without precipitating thunderclap headache, associated with bilateral acute parietal infarcts and reversible vasoconstriction [ 55]. Suppression of the headache was achieved with the use of an antitussive, benproperine.

Diamond reported on the complete resolution of symptoms with Indomethacin in 13 of his 15 patients within 1–4 weeks; medication was withdrawn after three to 12 months with all but one patient remaining asymptomatic, suggesting natural remission [ 62]. In Pascual’s series of 16 patients 4 seemed to respond to pre-emptive ergotamine, one patient with regular attacks to propranolol and one each to Indomethacin and Flunarizine [ 32]. Five of nine patients in a later cohort, responded to prophylactic nadolol or propranolol over a period ranging from two to six months [ 33]

The main forms of exertional headache associated with a secondary pathology are those where exertion precipitates a TCH. Management remains as detailed in the section TCH.

The most consistently reported responses in sexual headache are for Indomethacin and betablockers, most commonly cited propranolol, but also metoprolol and nadolol [ 32, 42, 66, 69, 74, 75]. Raskin recommended the use of pre-emptive Indomethacin 50 mg after dinner [ 76], successful in four out of five patients. Prophylactic success with Indomethacin, oral treatment 25 to 100 mg given 30–60 min prior to sexual activity, has corroborated Raskin’s experience [ 77, 78]. Frese reported benefit with triptans for the headache precipitated by sexual activity if attacks usually lasted longer than 2 h. The paper also reported on the preventative benefit of rizatriptan, almotriptan and sumatriptan, 30 min before sexual intercourse [ 77].

The majority of patients with sexual headache present with thunderclap onset headache and therefore will need to be managed for a possible secondary precipitant. Lundberg reported on 4–12% of patients, confirmed to have subarachnoid haemorrhage, presenting with sexual thunderclap headache [ 79]. Headache secondary to spontaneous intracranial hypotension, precipitated by sexual intercourse, is no longer part of the classification for primary sexual headache but is a reported precipitant in spontaneous intracranial hypotension [ 80].