Abstract
Benign cough headache is an uncommon primary headache disorder marked by short- lasting attacks of pain triggered by coughing. Magnetic resonance imaging of the brain is required to assure that the cough headache is truly benign. The aetiology of the pain is unclear, but is probably associated with the brief increased intracranial pressure that attends coughing. We have reviewed the clinical features, aetiology, differential diagnosis, management, and prognosis of benign cough headache.
Introduction
Headache aggravated or brought on by cough has long been known to be a symptom of intracranial disease (1). This review will focus on benign cough headache, which by definition is unassociated with an intracranial disorder (2) (Table 1). For clarity, we will separate benign cough headache from benign exertional headache (Table 2), although obviously some overlap occurs in these two conditions. Benign cough headache can be triggered by other sudden Valsalva manoeuvres, but is typically not triggered by sustained physical exercise (3). This is in contradiction to the typical circumstance in benign exertional headache.
International Headache Society diagnostic criteria for benign cough headache (2)
International Headache Society diagnostic criteria for benign exertional headache (2)
In 1932, Tinel described four patients with headache brought on by coughing, nose-blowing, breath-holding, and bending the head forward (4). Before the reports of Symonds and Rooke, however, cough and exertional headaches were always considered ominous symptoms, and there was no clear recognition that benign types of these headaches existed. In 1968, Rooke noted that ‘in every patient with this complaint, an intracranial lesion of potentially serious nature, such as brain tumour, aneurysm or vascular anomaly, has been suspected; and even when no such lesion could be identified, an uneasy uncertainty usually has remained’ (5).
The landmark paper entitled ‘Cough headache’ by Sir Charles Symonds in 1956 brought attention to this disorder (6). Symonds clearly described cases of both secondary and primary (or benign) cough headache. He presented patients with headache provoked by coughing, and noted that in these same patients, sneezing, straining at stool, laughing, or stooping could also provoke the headache. He did not describe headache precipitated by physical exercise. He outlined the clinical course of benign cough headache, and suggested a pathophysiological mechanism for the disorder.
In 1968, Rooke reviewed 93 patients with benign exertional headache (5). He did not separate cough headache from headaches caused by running. However, his data underscored Symonds’ concept that cough headache could be benign.
Epidemiology
In a population-based study, benign cough headache had a lifetime prevalence of around 1% (7).
Clinical features
Benign cough headache is typically bilateral, of sudden onset, and lasts < 1 min (2). It is precipitated rather than aggravated by coughing, and can be prevented by avoiding coughing. The International Headache Society (IHS) criteria specify that structural lesions must be ruled out by neuroimaging before the diagnosis of benign cough headache can be made. It is not associated with nausea, vomiting, conjunctival injection, lacrimation, nasal congestion, or rhinorrhoea (1).
Because of the historical importance of Symonds’ cases, they will be reviewed in detail (6). Eighteen of the 21 cases of benign cough headache reported were males. The age range was 37–77 years, with the average age being 55 years. Symonds’ patients generally complained of severe bilateral head pain, lasting 2–10 min. Two patients had severe pain for a few minutes followed by a dull ache that lasted 1–2 h. The quality of the pain was usually bursting. Most patients had pain at the vertex spreading to both frontal regions. As noted previously, other manoeuvres besides cough could trigger the pain, but physical exercise was not mentioned as a precipitant. Two patients could trigger headache by quick rotation of the head. Two patients could cough without pain while lying down. Two patients noted that when they awoke in the morning they could cough without pain, but that the liability appeared after they assumed the vertical posture. One patient noted decreased pain severity if he coughed with his neck extended. Two patients had relief of cough headaches after lumbar puncture; one had relief for a few weeks and then had complete relief after air encephalography, and the second had complete relief after lumbar puncture when seen 6 months later. Two patients recovered after infected tooth extraction. Nine patients had eventual recovery, six had significant spontaneous improvement over a period of 18 months to 12 years, and the remaining six either had the disorder continue without change or died of an unrelated disease. Five patients had a past history of migraine.
Rooke considered cough headache to be a form of exertional headache, and reviewed the clinical features of 93 patients with benign exertional headache (5). The male to female ratio was 4:1, and exertional headache was twice as common in those over 40 compared with those aged 10–40 years. Rooke outlined the generally favourable course of benign exertional headache. Eleven patients had complete headache relief by 1 year, 30 had complete relief of headache within 5 years, and 73 either improved or were free of headache after 10 years. Seven of 31 patients noted a respiratory infection before their exertional headaches began, and four patients had unequivocal relief after dental extractions. One patient had complete headache relief after pneumoencephalography.
Pascual et al. studied 13 benign cough headache patients in detail (3) (Table 3) While other Valsalva manoeuvres could trigger headache in four patients, none had headache precipitated by sustained physical exercise. The pain was moderately severe, and usually lasted seconds to < 30 min. In one case the headache remained for up to 1 day after being triggered, and in one patient with a normal carotid ultrasound the headache was unilateral. None of the benign cough headache patients in this study had concomitant migraine. Computed tomography (CT) was normal in 11 patients, and magnetic resonance imaging (MRI) was normal in four.
Clinical characteristics of cough and exertional headache patients (after (3))
Raskin has reviewed the clinical features of benign cough headache (8, 9). Many patients note headache onset during lower respiratory tract infections accompanied by cough, or during vigorous weightlifting programmes. The headache arises moments after the cough, reaches its peak almost immediately, and then subsides over several seconds to a few minutes. Sometimes the pain remains at the peak for several seconds before decreasing. Most patients are pain-free between attacks, but some have a dull headache following the triggered attack that may persist for hours. Raskin warns that these patients often complain of continuous headaches, thus emphasizing the need to directly ask about triggering by cough. Nausea and other migrainous features are uncommon, and although the headache is usually bilateral, it can occasionally be unilateral.
Over the years some rare clinical characteristics have been reported. Bruyn described a patient who could trigger headache by coughing or yawning (10). A case of cough headache presenting as unilateral toothache completely responsive to indomethacin has been reported (11). This patient's pain was in the right maxillary region and radiated to the ipsilateral temple, ear, and occiput. Unilateral cough headache with circannual periodicity has been reported (12). This patient's cough headaches were often induced by exertion as well. Unilateral benign cough headache coexisting with chronic paroxysmal hemicrania has been reported (13).
Aetiology
The aetiology of cough headache is not completely clear. It seems intuitive that it is associated with increased intracranial pressure (ICP), as coughing increases ICP. What causes the pain is not clear, however, as elucidated by Sir Charles Symonds in 1956:
In the normal person, coughing, however violent, does not cause headache, and perhaps the first question to be answered is why it does not. That it causes an increase in intracranial pressure from venous dilatation we know from observing the response when measuring the cerebrospinal fluid pressure. The intracranial venous dilatation is presumably obtained at the expense of cerebral capillaries and expulsion from the cranial cavity of such a proportion of the cerebrospinal fluid as can be accommodated by the spinal subarachnoid space. Under these conditions it has been argued … that there should be a displacement of the brain towards the foramen magnum. This would involve traction upon the structures, vascular and dural, anchoring the brain to the skull which are known to be sensitive to this form of stimulation. Yet no headache results, presumably because the degree and direction of traction are inadequate to cause it. (6)
Symonds later hypothesized that in benign cough headache, ‘the pain is due to stretching of a pain-sensitive structure within the posterior fossa, which may … be due to an adhesive arachnoiditis’ (6).
Williams measured CSF pressures from the lumbar region and the cisterna magna in 16 patients during coughing in the sitting position (14). All patients had ‘disease in the cervical region’ requiring myelography. None of the cases had a complete blockage of the spinal subarachnoid space. During a cough, there was a phase during which the lumbar pressure exceeded the cisternal, followed by a phase in which the cisternal pressure exceeded the lumbar. Thus on coughing, the intrathoracic and intra-abdominal pressure was felt to be transmitted through the valveless veins around the vertebrae to the epidural veins, which then distended with blood (15). Once distended, these veins compressed the spinal dura, causing a pressure wave that passed into the head and than rapidly downwards again (15). Williams commented that the upward passage of fluid was relatively easy, but that the downward rebound from the head towards the spine might cause tissue to jam in the foramen magnum. This would create a pressure difference between the head and spine, which he termed the ‘craniospinal pressure dissociation’ (14). He postulated that in Chiari I malformations, the ebb and flow of fluid through the foramen magnum could progressively impact the tonsils, leading to pain. Williams subsequently investigated the aetiology of cough headache in two patients with Chiari I malformations (15). In these two patients, he verified the presence of a craniospinal pressure dissociation preoperatively. Decompression of the cerebellar tonsils relieved the cough headache and eliminated the craniospinal pressure dissociation. Nightingale and Williams later described four more similar patients successfully treated with surgical decompression (16).
Patients with benign cough headache do not have tonsilar herniation, however, so Raskin hypothesized that the pain in these circumstances was due to a heightened sensitivity of yet unidentified receptors (8). Raskin successfully treated four cough headache patients with repetitive intravenous dihydroergotamine, and offered that unstable serotonergic neurotransmission might be important in the aetiology of cough headache (8). One case of cough headache has occurred after periaqueductal gray (PAG) electrode implantation (17). The phenotypic cough headache developed immediately after PAG implantation and was relieved by indomethacin. The patient used indomethacin for 4 months, then remained headache-free after stopping the medication. She remained headache-free at 5-year follow-up. After noting that benign cough headache could be provoked by a sudden increase in ICP and could remit after the sudden decrement in ICP that attended lumbar puncture, Raskin later hypothesized that the nature of the receptors sensitive to ICP alterations was probably key to understanding benign cough headache (9).
Wolff thought that cough headache often occurred around the time of a systemic infection. The systemic illness was thought to alter the vascular tone in the head (18). In Wolff's experience, most patients with cough headache had a prior history of migraine.
A case of a man with a large goitre who could trigger headache by raising his arms has been reported (19). This manoeuvre would decrease venous return from the head, increasing intracranial pressure. Presumably in this patient, the sudden increase in venous pressure was enough to cause headache.
Wang et al. postulated that benign cough headache was due to CSF hypervolaemia, leading to an increased craniospinal pressure dissociation during coughing, thus explaining the response to acetazolamide, indomethacin, and lumbar puncture (20).
Knappertz noted that some individuals have incompetent or absent jugular venous valves, and postulated that this subgroup might be more prone to increases in intracranial pressure with increases in intrathoracic pressure (21). He hypothesized that patients with cough headache might have inhibited functional jugular venous return. To the best of our knowledge, competency of jugular venous valves has not been evaluated in patients with benign cough headache.
The aetiology of benign cough headache can be best summed up by Symonds himself, who in 1970 wrote ‘as far as I am aware its origin remains a mystery’ (22). Perhaps in a benign cough headache patient there is a lowered threshold for the development of pain associated with the increased intracranial pressure normally caused by coughing.
Differential diagnosis
Before a diagnosis of benign cough headache can be made, intracranial masses and specifically posterior fossa lesions must be ruled out (Fig. 1). Table 4 lists some secondary causes of cough headache. Pascual et al. found clinical differences between patients with benign and symptomatic cough headache (3) (Table 3) Symptomatic cough headache started earlier in life, had a longer attack duration, and did not respond to indomethacin. However, indomethacin response should not be used to differentiate benign from symptomatic cough headache, as a patient with a Chiari I malformation-associated cough headache that responded completely to indomethacin has been reported (23).
Magnetic resonance imaging of the brain and cervical spine in a patient presenting with cough headache and bilateral upper extremity sensory symptoms. (a) Sagittal T1 image showing descent of the cerebellar tonsils into the foramen magnum and associated cervical syrinx. (b) Axial T2 image showing cross-section of the cervical syrinx. (c) Sagittal T1 image showing extent of the cervical syrinx.
Unilateral cough headache has been reported with carotid stenosis (24, 25), but this has not been seen by the authors. In the case reported by Britton and Guiloff, the patient initially had right-sided cough-induced headaches without focal neurological deficits. Two years later he developed episodic left-hand sensory symptoms. Later that same year he developed left-hand weakness and numbness of the left face. He was admitted to hospital and found to have pyramidal weakness of the left hand. CT of the head was normal and a myelogram showed no abnormality of the cord or foramen magnum. Twenty-four hours later he developed a continuous right-sided headache, ‘similar to the cough headache’, and the next day developed a severe left hemiparesis. CT showed a large right ACA/MCA infarct. Cerebral angiogram showed stenosis of the right ICA close to its origin, and the possibility of dissection was raised. Echocardiography was normal. Angiogram 5 weeks later revealed an occluded right internal carotid artery. The patient was left with hemiparesis, but the cough headache had not recurred at the time of the case report. The authors admitted that the relation of the cough headache to the abnormality of the right internal carotid artery was uncertain. They reasonably recommended studying the carotid arteries in cough headache patients when the headache was associated with focal neurological symptoms or signs and no intracranial or foramen magnum lesion could be found (24). The patient described by Rivera et al. initially had right-sided cough headache alone (25). Four months later the cough headache was associated with focal neurological symptoms (right ‘blunted vision’, weakness of the left arm, facial paresis, dysarthria, abnormal spontaneous movements of the left arm). The focal symptoms occurred after coughing and also with episodes of hypotension during dialysis, and lasted 10–15 min. He was admitted to the hospital and found to have bilateral carotid bruits with an otherwise normal neurological examination. CT of the head was normal, and carotid ultrasound showed significant left common carotid stenosis. The patient was treated with anti-platelet agents and codeine with ‘transient improvement’, and died 1.5 years after his cough headache started (25). Whether the cough headache was actually caused by the carotid stenosis is unclear in this case.
Cough headache secondary to an unruptured cerebral aneurysm has been reported (26). The patient was a 42-year-old female who for 24 days complained of a right temporal, severe pain induced by coughing or bending forward. The pain lasted 1–5 min, followed by a dull ache lasting 1 h. Neurological examination was normal 2 weeks after the onset of symptoms, and indomethacin failed to give her relief. Twenty-four days after headache onset, she complained of a continuous right-sided head pain worsened by cough, strain, or bending forward. On examination she was found to have right-sided ptosis, and the next day developed a more complete third nerve palsy. CSF showed 16 erythrocytes/mm3 and 11 leucocytes/mm3, and cerebral angiogram revealed an 8-mm right-sided aneurysm at the junction of the posterior communicating artery and the internal carotid artery. She was free of cough headache post-operatively at 1-year follow-up (26). Rooke, however, reported that none of the 14 patients with unruptured aneurysms that he studied complained of exertional headache (5). A patient with a large venous angioma of the posterior fossa presenting with cough headache has been reported (27). However, close inspection of this case reveals that the headache was exacerbated, not elicited, by coughing, and thus would not readily be confused with benign cough headache.
Pascual et al. found differences between patients with benign cough headache and benign exertional headache, lending credence to their separate classification (3) (Table 3). Cough headache was triggered by Valsalva manoeuvres, while exertional headache was triggered by sustained physical exercise. Benign cough headache affected a much older population than that affected by benign exertional headache. The average age of patients afflicted with benign cough headache was 67 ± 11 years, while the average age of patients with benign exertional headache was 24 ± 11 years. Benign cough headache was shorter than benign exertional headache, and treatment responses were different (Table 3). Benign cough headache was typically sharp and stabbing while benign exertional headache was described as pulsating.
Migraine, cluster headache, post-lumbar puncture headache, and idiopathic intracranial hypertension can be aggravated by cough, but only benign cough headache is elicited by cough (1, 28).
Diagnostic evaluation
Given the differential diagnosis outlined above, every patient with cough headache should have an MRI of the brain to rule out a posterior fossa lesion. Whether a patient with an unruptured aneurysm can present with cough headache is not clear, but it seems reasonable to obtain an MRA of the intracranial circulation in most cases. We do not typically do carotid ultrasounds or MRAs of the extracranial circulation in the evaluation of cough headache, unless the patient gives a history consistent with transient ischaemic attacks.
Management
Any chest disease that may be causing the cough should be identified and treated (29). Because of the typical short duration of benign cough headache, preventative rather than abortive treatment is used. Mathew established the efficacy of indomethacin at a dose of 150 mg/day in a double-blind study involving two patients with benign cough headache (30). Of 16 patients treated with indomethacin, Raskin reported that 10 responded completely, four had moderate improvement, and two had no response (31). The effective dosage ranged from 50 to 200 mg, with an average of 78 mg. The duration of treatment was 6 months to 4 years (31). Raskin has noted that indomethacin rarely fails, but that the dosage sometimes must approach 250 mg daily (8). We have had success in using indomethacin in the 25–150 mg/day range, and usually combine this with a proton pump inhibitor in those patients who require long-term treatment. Indomethacin decreases intracranial pressure (32), and that may be why it is effective in this condition compared with other non-steroidal anti-inflammatories. Since some patients lose the liability to benign cough headache over time, treatment should be withdrawn periodically.
Acetazolamide (20) and methysergide (33, 34) have been reported to be effective in open-label trials. In the Wang study, four of five patients with indomethacin-responsive benign cough headache responded favourably to acetazolamide at maximum doses of 1125–2000 mg/day. Two of these patients responded completely to acetazolamide. The patients were allowed to adjust the maintenance dose to treatment effectiveness or side effects, and the mean maintenance dose was 656 mg. One patient withdrew from the study because of intolerable distal limb numbness (20). The methysergide-responsive patient described by Calandre et al. had unilateral, throbbing, cough-induced headaches lasting 30 min to a ‘few hours’ (33). This patient only had headache with coughing, straining, and stooping, and responded completely to methysergide at an unknown dose after failing nicardipine and propranolol and after partially responding to amitriptyline. The methysergide-responsive patient described by Bahra and Goadsby had unilateral cough-induced headaches lasting 15–30 min (34). She required 2 mg daily and then lower doses for 9 weeks total, and was then headache-free after cessation.
Several authors have written of the occasional efficacy of lumbar puncture (5, 6, 31). Raskin noted the effectiveness of 40 ml lumbar puncture in six of 14 patients (31). Three had immediate relief after the procedure, and in the other three relief came over 2 days. One of the responders redeveloped cough headache 6 weeks after initial lumbar puncture, but responded completely to repeat spinal tap. Six of the eight who failed lumbar puncture responded to indomethacin.
Raskin has written of the effectiveness of naproxen, ergonovine, IV dihydroergotamine, and phenelzine, but found propranolol ineffective (8). Mateo and Pascual found naproxen partially effective in one case (13). Of the six patients with cough headache and normal MRIs described by Calandre et al., one found propranolol effective, while two found it ineffective (33). Aside from Raskin's observations (8), there is one other report of the efficacy of IV dihydroergotamine in treating cough headache (35). Although unreported, topiramate might be tried in refractory cases, given that it inhibits carbonic anhydrase.
Prognosis
Of the 21 patients reported by Symonds, nine had eventual recovery, six had significant spontaneous improvement over a period of 18 months to 12 years, and the remaining six either had the disorder continue without change or died of an unrelated disease (6). The disorder persisted for 2 months to 2 years in the 13 patients reviewed by Pascual et al. (3).
Conclusions
Benign cough headache is relatively rare, but important to recognize as effective treatment is available. The aetiology of the pain is unclear, although it is probably related to the transient increased intracranial pressure that attends coughing. MRI of the brain must be performed in every patient to confirm that the cough headache is truly benign.