Download Paroxysmal hemicrania: a prospective clinical

doi:10.1093/brain/awn010
Brain (2008), 131, 1142^1155
Paroxysmal hemicrania: a prospective clinical study
of 31 cases
Elisabetta Cittadini,1 Manjit S. Matharu1 and Peter J. Goadsby1,2
1
Institute of Neurology, The National Hospital for Neurology and Neurosurgery, Queen Square, London, UK and
Headache Group, Department of Neurology, University of California, San Francisco, San Francisco CA, USA
2
Correspondence to: Professor Peter J. Goadsby, Headache Group, Department of Neurology, University of California,
San Francisco, Box 0114, 505 Parnassus Avenue, San Francisco, CA 94 143- 0114, USA
E-mail: [email protected]
Paroxysmal hemicrania is a rare syndrome characterized by repeated attacks of strictly unilateral, severe,
short-lasting pain occurring with cranial autonomic features. The hallmarks of this syndrome are the relatively
short attacks and the exquisite response to indometacin.We describe the phenotype of this condition in a series
of 31 patients.The mean duration of attack was 17 min.The mean attack frequency was 11.The distribution of the
pain was orbital and temporal in 77% of the patients, retro-orbital in 61%, frontal in 55%, occipital in 42%;
although pain was also reported in the vertex, second division of trigeminal nerve, neck, nose, jaw, parietal
region, ear, teeth, eyebrow, shoulder (ipsilateral and bilateral), arm and third division of trigeminal nerve. Of
the cohort, 87% had lacrimation, 68% had conjunctival injection, 58% rhinorrhoea, 54% nasal congestion, ptosis
and facial flushing. Other cranial autonomic features include eyelid oedema, forehead/facial sweating, sense of
aural fullness and periaural swelling, miosis, mydriasis and swelling of the cheek. The majority of the patients
(80%) were agitated or restless, or both, with the pain and 26% were aggressive. All patients had positive placebo
control indometacin test (100^200 mg intramuscularly), or a positive oral indometacin trial or both. We suggest
the International Headache Society criteria be revised to remove specification of attack site, and to include the
full range of cranial autonomic features. Currently, the sine qua non for paroxysmal hemicrania is a response to
indometacin. Since there is no reliable clinical marker of that response we recommend an indometacin test,
either orally or by injection for any patient with lateralized discrete attacks of head pain with associated cranial
autonomic symptoms.
Keywords: paroxysmal hemicrania; trigeminal autonomic cephalalgias; indometacin; cluster headache
Abbreviations: CGRP = calcitonin gene-related peptide; HIS = International Headache Society; NO = nitric oxide;
PH = paroxysmal hemicrania; REM = rapid eye movement; SUNCT/SUNA = short-lasting unilateral neuralgiform headache
attacks with conjunctival injection and tearing syndrome; TAC = trigeminal autonomic cephalalgia; VRS = verbal rating scale.
Received August 18, 2007. Revised December 10, 2007. Accepted January 10, 2008. Advance Access publication February 5, 2008
Introduction
Paroxysmal hemicrania (PH) is a relatively rare primary
headache disorder first described by Sjaastad and Dale
(1974) and christened ‘chronic paroxysmal hemicrania’ 2
years later (Sjaastad and Dale, 1976). Patients have been
described in several countries (Joubert et al., 1987; Antonaci
and Sjaastad, 1989; Sjaastad, 1992). The incidence and
prevalence of PH has been estimated to be about 1–3% that
of cluster headache (Antonaci and Sjaastad, 1989), or about
1 in 50 000 (Matharu et al., 2003a), although it may be
rarer (Sjaastad and Bakketeig, 2007). None of these
estimates are based on substantial series. PH is classified
as a trigeminal autonomic cephalalgia (TACs) by the
International Headache Society (IHS) (2004), and defined
by at least 20 attacks of severe unilateral orbital, supraorbital or temporal pain, lasting 2–30 min, accompanied
by ipsilateral cranial autonomic features such as ptosis,
eyelid oedema, conjunctival injection, lacrimation, nasal
blockage or rhinorrhoea. Attacks usually have a frequency
of above five a day, and respond completely to indometacin. In the first edition of the classification (1988), all
cases were referred to as chronic PH. However, with
validation of the episodic variant over subsequent years
(Kudrow et al., 1987; Newman et al., 1992; Goadsby
and Lipton, 1997), the second edition of the classification recognized both an episodic and a chronic form
ß The Author (2008). Published by Oxford University Press on behalf of the Guarantors of Brain. All rights reserved. For Permissions, please email: [email protected]
Paroxysmal hemicrania
(Headache Classification Committee of The International
Headache Society, 2004).
Case series of PH have been presented with 84 cases
reviewed in from the literature for the largest one (Antonaci
and Sjaastad, 1989). Russell (1984) described 105 attacks,
although sample from only five patients with chronic
paroxysmal hemicrania to assess severity, duration and
circadian pattern. In 2002, Boes and Dodick (2002)
retrospectively reviewed 74 cases from the Mayo Clinic in
an attempt to further refine the clinical spectrum of this
condition. More recently, Zidverc-Trajkovic and colleagues
(2005) reported eight consecutive patients affected by PH.
Until now, the rarity of the syndrome has prevented
substantial cohorts to be collected by one group over a
short enough period to provide a broadly based prospective
description of the syndrome with the same persons
conducting the clinical interviews.
Therefore, the aim of this study was to describe in detail
the clinical picture of PH by assessing prospectively a large
group of patients. The work was initially reported in
preliminary form at the 16th Migraine Trust International
Symposium (London, 18–20 September 2006, Cittadini and
Goadsby, 2006).
Methods
Thirty-one patients (mean age 37 years, range 5–68) with possible
PH, were identified at the National Hospital for Neurology and
Neurosurgery and the Hospital for Sick Children, Great Ormond
Street from May 1995 to January 2007 and attended the outpatient departments between January 2004 and January 2007. The
data used for this study were obtained from the clinical notes and
by information received on the phone by one of us (E.C.).
Patients were seen by at least two specialists in the out-patient
clinic, and some were seen more than once if admitted for
investigation. Twenty-eight patients were contacted by telephone
after giving written consent, or consent during the clinic
appointment. Two subjects were not contacted by telephone
because they were children; their history had been taken directly
by one author and they were followed up through their parents
(P.J.G.). One patient was lost to follow-up and we used his
medical notes to obtain the information. Again, this patient had
had the history taken directly by one of us (P.J.G.). The study was
approved by the National Hospital for Neurology and
Neurosurgery Joint Research Ethics Committee.
A standard proforma of questions, including side, site, type of
pain, severity, duration, frequency and periodicity of the attacks;
presence or absence of autonomic features; possible triggers and
behaviour during the attacks was completed for each patient. The
response to oral indometacin and the placebo-controlled Indotest
(Matharu et al., 2004) was assessed. In addition, the patient’s
response to other medications, and personal and family history of
headache were recorded. Information regarding response to other
medications and neuroimaging reports were collected from the
clinical notes.
The results were collated and summarized using Excel
(Microsoft).
Brain (2008), 131, 1142^1155
1143
Results
Of 31 patients, 26 patients had PH as defined by the IHS
Classification Committee (2004). Five patients had shortlasting head pain with cranial autonomic features responding
absolutely and robustly to indometacin but not fulfilling the
current IHS criteria (Headache Classification Committee of
The International Headache Society, 2004). In particular, one
patient (#13) reported as an autonomic feature a sense of
aural fullness and had a low frequency of attacks (three per
day), and three patients (#14, #18, #29) had a low frequency
of attacks over a 24-h period, two, three and four per day,
respectively. Another patient reported unilateral pain localized over the occipital region (#17).
Clinical characteristics used by the IHS
to define PH
Pain
Laterality. Fifteen patients (48%) had exclusively rightsided attacks; 15 (48%) patients had exclusively left-sided
attacks and one (3%) patient had side alternating attacks,
which were more frequent on the right side.
Site of attacks. The majority of the patients complained of
pain in the distribution recognized by the IHS (Headache
Classification Committee of The International Headache
Society, 2004), which included: 24 (77%) in the temporal
region and orbital region and 19 (61%) in the retro-orbital
region. Their pain was also reported in the frontal region in
17 (55%), the occipital region in 13 (42%), the vertex in 11
(36%), the maxillary (second) division of the trigeminal nerve
in 10 (32%), the neck in eight (26%), the nose and jaw in six
(19%), the parietal region in five (16%) and the ear and teeth
in four (13%) patients. Furthermore, the pain was over the
eyebrow, ipsilateral shoulder and ipsilateral arm in three
(10%) patients each. One patient had pain over the shoulder
bilaterally and another one had pain over the mandibular
(third) division of the trigeminal nerve (Table 1).
Severity of the pain. Patients were asked to rate the severity
of their attacks on a verbal rating scale (VRS) of 0 to 10, 0
being no pain at all and 10 the most severe pain imaginable.
Twenty patients (65%) rated their most painful attacks at 10.
Five patients (16%) rated their most painful attacks at nine.
One patient rated the most painful attack between 8 and 10
and another one rated the most painful attack between 8
and 9. Two (6%) patients rated their most painful attacks at
seven. The pain was reported severe/very severe in two
children. Twenty (65%) patients said that this was the most
painful condition they had ever experienced comparing it to
childbirth and perforation of duodenal ulcer.
Cranial autonomic features
According to the IHS classification criteria (Headache
Classification Committee of The International Headache
Society, 2004), patients with PH are required to have at
least one of the cranial autonomic feature. In our cohort,
1144
Brain (2008), 131, 1142^1155
E. Cittadini et al.
Table 1 Site of attacks
Site
n (%)
Temporal region
Orbital region
Retro-orbital region
Frontal region
Occipital region
Vertex
V2
Neck
Nose
Jaw
Parietal region
Ear
Teeth
Eyebrow
Shoulder
Arm
V3
24 (77)
24 (77)
19 (61)
17 (55)
13 (42)
11 (36)
10 (32)
8 (32)
6 (19)
6 (19)
5 (16)
4 (13)
4 (13)
3 (10)
3 Ipsilateral (10%) and 1 bilateral (3%)
3 Ipsilateral (10%)
1 (3)
Table 2 Associated cranial autonomic symptoms
Autonomic feature
n (%)
Lacrimation
Conjunctival injection
Rhinorrhoea
Nasal congestion
Ptosis
Facial flushing
Eyelid oedema
Forehead/facial sweating
Aural fullness
Aural swelling
Miosis
Mydriasis
Swelling of the cheek
27 (87)
21 (68)
18 (58)
17 (55)
17 (55)
17 (55)
13 (42)
10 (32)
8 (26)
4 (13)
3 (10)
2 (6)
1 (3)
we found autonomic features at the following frequencies:
27 (87%) patients had lacrimation, 21 (68%) had
conjunctival injection, 18 (58%) had rhinorrhoea, 17
(55%) had nasal congestion, 17 (55%) had ptosis, 13
(42%) had eyelid oedema, 10 (32%) had forehead and
facial sweating and three (10%) had miosis. In addition,
other autonomic features were reported such as facial
flushing in 17 (55%) patients, a sense of aural fullness in
eight (26%) and a sense of aural swelling in four (13%)
patients. Two (6%) patients reported ipsilateral mydriasis
and one (3%) patient reported swelling of the cheek. One
of our patients had only a sense of ear fullness during the
attacks and denied all the other autonomic symptoms.
Another patient sometimes had cranial autonomic features
without pain (Table 2; Pareja, 1995).
Duration and frequency of attacks
Length of attacks. The IHS criteria for the length of attacks
in PH are between 2 and 30 min (Headache Classification
Committee of The International Headache Society, 2004).
In our cohort, the range of the attacks was between 10 s and
4 h. In 17 (55%) patients, the longest attack duration was
more than 30 min. One patient (#9) had one attack that
lasted 48 h; therefore, this patient’s data were not included
in the final analysis of length of attacks. The mean length of
the attacks was 17 min and the median was 19 min.
Number of attacks per day. The IHS criteria require the
number of attacks per day to be more than five for more
than half of the time (Headache Classification Committee
of The International Headache Society, 2004). The number
of attacks per day ranged from 2 to 50. The mean was 11
attacks in 24 h and the median was nine. In our cohort,
four (13%) patients had fewer than four attacks a day for
more than half of the time. In particular, in two patients
(#13, #29) the range of the attacks in 24 h was between
three and five, in one patient (#14) the range was between
two and three, and in one patient (#18) the range was
between two and four attacks per day. One patient (#26)
reported a range of attacks between no attack and eight
attacks per day with fewer than five attacks per day for
more then half of the time when we saw and diagnosed the
condition. However, during one of his previous bouts he
had had seven to eight attacks per day.
Indometacin response
The IHS criteria require an absolute response to indometacin (Headache Classification Committee of The
International Headache Society, 2004). In our cohort, 17
(55%) patients had the modified Indotest (Antonaci et al.,
1998a); a placebo-controlled indometacin test with intramuscular indometacin at the dose of 100 or 200 mg
(Matharu et al., 2004). Out of these 17 patients, 14
(82%) had a positive response. They subsequently underwent a trial of oral indometacin that was positive in 13
(93%) cases, with one case not assessable due to
indometacin side-effects after a few days. Of the remaining
three patients who had the Indotest, one had an equivocal
response to the modified Indotest and two patients had a
negative response. Each of these had a positive response to
oral indometacin. Finally, 14 (45%) patients had only the
oral trial with indometacin which was positive (Fig. 1). The
mean daily dose of indometacin used in the oral trial based
on 30 patients was 137 mg (range 30–300 mg). The median
dose was 150 mg/day. Table 3 shows the daily effective dose
for each patient.
Clinical features not currently used by the
IHS to define PH
Age and gender of patients and duration of symptoms
Seventeen patients were male and 14 were female (M:F ratio
of 1:1). The mean age at the onset of the symptoms was
37 years (range 5–68).
Paroxysmal hemicrania
Brain (2008), 131, 1142^1155
31 patients
14 patients had positive
oral indometacin
response
Indotest
17 patients
Negative Indotest
2 patients
Equivocal Indotest
1 patient
Positive Indotest:
14 patients
13 pts had positive oral
indometacin trials
Positive oral
indometacin trial
1 pt stopped oral trial
due to side effects
Positive Oral
indometacin
trial
Fig. 1 Flow chart of patients who had the placebo-controlled
Indotest and open label oral trials with Indometacin.
1145
Table 4 Characteristics of the pain
Characteristics
n (%)
Sharp
Stabbing
Throbbing
Shooting
Burning sensation
Constant
Boring
Pressure
Knife in the head
Tightening
Dull
Aching
Sticking
Pushing
Cramped
Toothache pain
Tearing
Twisting
Screaming
Stiffness
Red hot poker
Power drilling
16 (51)
15 (49)
10 (32)
4 (13)
4 (13)
3 (10)
3 (10)
2 (6)
2 (6)
2 (6)
2 (6)
2 (6)
1 (3)
1 (3)
1 (3)
1 (3)
1 (3)
1 (3)
1 (3)
1 (3)
1 (3)
1 (3)
Table 3 The effective dose of indometacin per day
Patient
Daily dose
of indometacin
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
200 ^225 mg
200 mg
50 mg
150 mg
200 mg
150 ^225 mg
150 ^300 mg
100 mg
150 mg
50 mg
50 ^75 mg
75^150 mg
50 mga
75 mg
150 mg
75 mg
250 mg
75 mg
150 mg
200 mg
50 mg
200 ^300 mg
30 mgb
50 mg
150 ^200 mg
300 mg
75 mg
150 mg
100 mg
200 mg
100 ^150 mg
a
Patient stopped the intake of indometacin after a few days due to
side effects. bPatient 5½ years old.
Type of pain
In our group, the majority of the patients 16 (52%)
described their pain as sharp and 15 (48%) referred to it as
stabbing. In 10 (32%) patients, the pain was reported as
throbbing. Other types of pain are reported in the Table 4.
Triggering of attacks
In our group, the attacks could be either triggered or
spontaneous. Common triggers were stress or relaxation after
stress in eight (26%) patients; exercise in seven (23%)
patients; alcohol and neck movement in six (19%) patients, a
warm environment, cold weather, a strong smell, bending
down and coughing, were triggers in five (16%) patients each;
sneezing in four (13%) patients; and, tiredness and straining
in three (10%) patients each. Other less frequent triggers
including cutaneous triggers, weather change, irregular sleep
or lack of sleep, the perimenstrual or menstrual period and
lifting were recognized by two (6%) patients each. Additional
triggers were environment temperature change, cold–heat
wind, dietary products (cream, cheese, chocolates, coffee,
citrus fruits), skipping meals, flickering lights, sleep and
pressure over the greater occipital nerve. These triggers were
indicated by one (3%) patient in each category. We also used
glyceryl trinitrate in four patients and this triggered a typical
attack of PH in one. One patient noted that some bouts were
precipitated by infections such as ear infection, tonsillitis and
chest infection (Table 5).
Agitation
Agitation or a sense of restlessness is typical behaviour
during cluster headache attacks, occurring in 490% of cases
1146
Brain (2008), 131, 1142^1155
E. Cittadini et al.
Table 5 Triggers to attacks
Characteristics
n (%)
Stress/relaxation after stress
Exercise
Alcohol
Warm environment
Strong smell
Neck movement
Cold weather
Bending down
Coughing
Sneezing
Tiredness
Straining
Cutaneous triggers
Weather Change
Irregular sleep/lack of sleep
Period/week before the period
Lifting
Temperature change
Cold/heat wind.
Hot weather
Skipping meal
Dietary products (cream, chocolates,
cheese, coffee, citrus fruits)
Flickering lights
sleep
Pressure over the GON
NTG
Ear/chest infection, tonsillitis
8 (26)
7 (23)
6 (19)
5 (16)
5 (16)
6 (19)
5 (16)
5 (16)
5 (16)
4 (13)
3 (10)
3 (10)
2 (6)a
2 (6)
2 (6)
2 (6)
2 (6)
1 (3)
1 (3)
1 (3)
1 (3)
1 (3)
1 (3)
1 (3)
1 (3)
1 (25)b
1 (3)
a
Touching around the jaw, talking, chewing, resting the right side of
the face on the pillow in one patient (#12). Touching the lower lip,
talking, chewing, cleaning teeth, wind blowing against face in
another patient (Boes et al., 2003). bNTG was used only in four
patients.
(Bahra et al., 2002). In a retrospective study of PH, 50% of
the patients were agitated (Antonaci and Sjaastad, 1989),
while in a more recent prospective study (Zidverc-Trajkovic
et al., 2005) 63% were agitated. In a recent prospective
study, agitation was described in 62% of SUNCT patients
and in 55% of SUNA patients (Cohen et al., 2006a). In our
cohort, 25 (80%) patients were agitated or restless, or both
and eight (26%) patients were aggressive, mainly verbally
rather than physically, during the attack.
Migrainous symptoms
Phonophobia and photophobia, which are usually bilateral,
are typical features of migraine (Headache Classification
Committee of The International Headache Society, 2004),
although they can be unilateral in some cases (Drummond,
1986; Vingen et al., 1998). Unilateral phonophobia and
photophobia have been reported to be more common
in PH, hemicrania continua (Irimia et al., 2008),
SUNCT (Cohen et al., 2006a; Irimia et al., 2008) and
cluster headache (Vingen et al., 1998; Irimia et al., 2008).
We found that 20 (65%) of our patients had phonophobia,
which was unilateral in five (25%) patients. Out of these
20 patients, 16 (80%) had personal or family history for
migraine, or headache not otherwise specified or both.
Twenty (65%) patients had photophobia, which was
unilateral in eight (40%) cases. Out of these 20 patients,
13 (42%) patients had personal or family history of
migraine, or headache not otherwise specified or both.
Motion sensitivity is a common symptom in migraine. We
found 16 (52%) patients had motion sensitivity and 12
(39%) of these had a personal of migraine or family history
of headache, or both. Osmophobia is well recognized in
migraine (Kelman, 2004; Zanchin et al., 2005). In our
cohort, seven patients had osmophobia and all had personal
or family history of migraine, or headache not otherwise
specified or both. Twelve (39%) patients in our cohort had
nausea or vomiting during the attacks, or both. Each of
these patients had a personal or family history of migraine,
or headache not otherwise specified or both. It seems likely
given the high population prevalence of migraine (Steiner
et al., 2003) and data on the family history of migraine and
headache severity (Stewart et al., 2006), that most severe
headache reported here in families was migraine.
Circadian and circannual periodicity
It is reported in PH that the attacks occur regularly
throughout the day, without a nocturnal preponderance
(Russell, 1984; Antonaci and Sjaastad, 1989; Boes and
Dodick, 2002; Zidverc-Trajkovic et al., 2005). However,
nocturnal attacks associated with the rapid eye movement
(REM) phase of sleep have also been described (Kayed et
al., 1978). In our cohort, there was no clear preponderance
of nocturnal attacks. Two (6%) patients (#2 and #26) had
clear nighttime preponderance. In contrast with cluster
headache, we did not find a strictly circadian and
circannual periodicity, noting a considerable individual
variation in attack periodicity. Fifteen (48%) patients
reported their attacks were generally random over a 24-h
period. Within this group, however, two patients (#1 and
#4) specified that the attacks experienced at the onset of
their headache used to occur at the same time during the
24 h, which was 04:00 h in one case, and afternoon and
night for the other case. Another patient (#10) reported a
period lasting 18 months during which one attack occurred
at 11:00 h and he also had another period lasting three
months during which he had one attack at 16:00 h. One
patient (#17) reported that the attacks that occurred
at night were more severe and longer than those during
the day.
The remaining 14 (45%) patients could have predictable
attacks during the 24 h. Three patients (#3, #27, #29) had
attacks only during the waking hours. One of these patients
had one predictable attack at 19:30 h. Five patients (#6, #18,
#20, #28, #30) had attacks mainly during waking hours. In
one case one to two attacks tended to arrive at the same
time each day and the others are random, whereas they
Paroxysmal hemicrania
were all predictable for the first 6 months. In another case,
they occurred at the same time every morning and at hourly
intervals and rarely during the night. One patient (#11)
specified that although their attacks were mainly random,
some tended to occur at the same time in the morning. One
patient (#12) had attacks both in the morning and night;
one patient (#13) had attacks during the waking hours
(70% of the attacks) and in the early morning (30%); one
patient (#21) stated that the first attack occurred around
09:00–10:00 h and then the other attacks were random
during the day, though the worst attacks were in the evening.
One patient (#25) reported attacks mainly during the
waking hours and only occasionally at night, although
they used to occur in the past either during the sleep hours
or the waking hours; one patient (#19) had attacks during
the day, whereas at the beginning they were present during
both sleeping and waking hours and the worst attack was
at 17:00 h.
With regard to the circannual periodicity, one patient
(#1) experienced a worsening of their attacks in spring and
autumn, one (#3) during autumn and winter and another
one (#17) had bouts tending to occur in February.
Background pain, migraine and analgesic overuse
Typically, PH is not characterized by interictal pain. In one
prospective study (Zidverc-Trajkovic et al., 2005), one out
of eight patients had tenderness between the attacks, while
in a retrospective study interparoxysmal discomfort was
reported in 28 out of 84 patients. We found that 18 (58%)
patients had background pain between each attack. The
presence of interparoxysmal pain raised the differential
diagnosis of hemicrania continua. Careful interview and the
use of headache diary tracking hourly pain allowed us to
clarify the clinical presentation. Out of these 18 patients,
eight (44%) had medication overuse. Seven (88%) of these
eight patients had personal or a family history of migraine,
or headache not otherwise specified, or both. The
remaining 10 (56%) patients had background pain without
analgesic overuse, and seven (70%) of them had personal
history positive for headache, either migraine or headache
the details of which were not available. The interparoxysmal
pain in our cohort was relatively mild in comparison to the
attack pain, a factor which serves to contrast these patients
with those with hemicrania continua.
We evaluated the relationship between migraine and PH
and found that 16 (51%) patients had a personal history
positive for migraine and 18 (58%) patients had a family
history positive for migraine, or headache not otherwise
specified or both.
Medication overuse is associated with chronic daily
headache in migraine (Bigal et al., 2004), and in cluster
headache (Paemeleire et al., 2006) and SUNCT patients
(Cohen et al., 2006a), with the common theme being
background migrainousness (Goadsby, 2006). We assessed
the presence of medication overuse, defined by the use of
Brain (2008), 131, 1142^1155
1147
analgesic on 15 or more days per month, and found that it
was present at some point in 10 (32%) patients. The
analgesics used were codeine and paracetamol (acetaminophen) combinations, paracetamol, aspirin and caffeine
combinations, ibuprofen, aspirin, paracetamol–caffeine–
codeine–doxylamine combinations, caffeine and paracetamol
combinations, ergotamine and caffeine combinations.
Of this group, eight (80%) patients had positive personal
history for migraine, or positive family history for headache
or both.
Periodicity and chronicity of PH
PH can be classified as episodic or chronic depending on
the presence of a remission period. About 20% of patients
suffered from episodic PH, which is diagnosed when there
is a remission period between bouts of attacks of at least
1 month (Headache Classification Committee of The
International Headache Society, 2004). The remaining
80% of patients have chronic PH, which is diagnosed
when patients have daily attacks without a remission of at
least 1 month. In contrast to cluster headache, in PH the
chronic form is more frequent than the episodic form.
Interestingly, in a recent prospective study of SUNA/
SUNCT the chronic form was also more frequent (Cohen
et al., 2006a). We found that 20 (65%) patients had
primary chronic PH at onset and six (19%) of these
patients were completely in remission at interview.
The mean duration of the chronic phase was 4.3 years
(range 3–7 years). The mean duration of remission was 4.3
years (range 3–6 years).
One patient with primary chronic PH had first a chronic
phase lasting 3 years, followed by a remission lasting about
4 years, followed by an episodic phase lasting 1 year,
followed by a remission of 1 year to date. Another patient
with primary chronic PH initially had a chronic phase
lasting about 3 years, followed by an episodic form lasting
about 6 years, followed by a chronic form lasting about 3
years, followed by a remission lasting about 15 months and
followed finally by a chronic form. Two (6%) patients had
secondary chronic PH with a mean interval between
the onset of the disease to the chronic form of about
34 months. Six (19%) patients had a primary episodic
form, with their bouts lasting for a mean of 16 weeks
(range 4–24 weeks), and a remission phase that lasted for a
mean of 100 weeks (range of 12–376 weeks; 7.3 years). One
patient with primary episodic PH had an unusually long
remission, which lasted 3 years and another unusual short
remission which lasted only 1 month. One (3%) patient
had secondary episodic PH: he had a chronic form which
lasted 3 years, followed by 9 months of remission and then
a recurrence of two to three attacks over a period of
4 months. Two (6%) patients had post-traumatic chronic
PH. One of them had first a chronic phase lasting 2 years,
followed by an episodic phase lasting 1 year, followed
by a remission period lasting another 1 year, followed by
1148
Brain (2008), 131, 1142^1155
a recurrence of his attacks for about 3 months and followed
finally by a remission lasting 4 years to date.
Medication responses
In our cohort, 17 (55%) patients had oxygen with none
having a good response, five (30%) a mild response and
twelve (71%) patients no response. Twelve (39%) patients
tried subcutaneous sumatriptan 6 mg with a good response
in two cases (17%), a mild response in one (8%) patient
and no response in nine (75%).
Comorbidities, examination and MRI findings
Four patients (13%) had abnormal findings on neurological
examination. Three patients had sensory changes, which
were hyperaesthesia to pinprick in V1-V2-V3 in one case,
allodynia in V1 in the second case and there was loss of all
sensory modalities over the medial aspect of the right upper
and lower eyelid, and medial aspect of the area above the
right supra-orbital ridge in the third case. One patient with
a diagnosis of ophthalmic arterio-venous malformation,
had impairment of the visual function in the corresponding
eye which included: (i) ipsilateral ptosis, (ii) severely
impaired visual acuity, (iii) large central scotoma, (iv)
venous stasis on fundoscopy and (v) defect in the eye
abduction and afferent papillary defect.
Twenty-five (80%) patients had brain imaging, which
included MRI or CT scan. Sixteen (64%) patients had normal
intracranial appearances, while in nine (36%) cases radiological abnormalities were found. They were as follows:
(i) Vascular loop compressing the trigeminal nerve
ipsilateral to the pain and generalized cerebral
involution in keeping with the patient’s age (#25);
(ii) bilateral vascular loops not compressing the trigeminal nerve and some non-specific lesions in the
periventricular and subcortical white matter of both
frontotemporal areas (#12);
(iii) ischaemic lesions in the basal ganglia and pons
(#19);
(iv) ophthalmic arterio-venous malformation (#14);
(v) sphenoid wing meningioma, which was subsequently
removed with only a temporary remission of the
headache (#20);
(vi) megacisterna magna (#6);
(vii) bulky pituitary gland associated with a deviation
towards the left side of the pituitary stalk. In this
case, there was also a solitary and non-specific T2/
FLAIR hyperintense lesion within the subcortical
white matter of the right frontal lobe (#18);
(viii) several non-specific, non-enhancing foci of abnormal
signal within the cerebral white matter and mild
mucosa thickening throughout the paranasal sinuses
and a 10-mm pineal cyst (#26);
(ix) a prominent left vertebral artery, and an ectatic
vertebro-basilar junction/basilar artery in close
E. Cittadini et al.
relation to the left trigeminal root entry zone
ipsilateral to the pain. The patient underwent
microvascular decompression with resolution of the
pain since (#30).
Discussion
We described a large case series of patients affected by
episodes of short-lasting unilateral headache with cranial
autonomic symptoms responsive to indomeathcin: PH. Our
data suggest that it is an excruciating syndrome with a high
frequency and brief duration of attacks that is exquisitely
sensitive to indometacin. In our cohort, some patients with
otherwise typical attack phenotypes had frequencies below
five per day, aural fullness or discomfort as a cranial
autonomic feature and a wider cranial distribution of pain
than has been previously recognized. Our case mix does not
confirm the reported female preponderance. PH is a distinct
relatively short-lasting, strictly unilateral primary headache
with cranial autonomic features that responds to indometacin. Our data suggest some revision to the current diagnostic
criteria (Table 6) and serve to remind clinicians of the
existence of, and clinical features of this unique primary
headache.
Table 6 Diagnostic criteria for paroxysmal hemicrania
Current diagnostic criteria
3.2 Diagnostic criteria
A. At least 20 attacks fulfilling criteria B ^E.
B. Attacks of severe unilateral, orbital, supraorbital or temporal
pain last 2^30 min.
C. Headache is accompanied by at least one of the following:
1. ipsilateral conjunctival injection and/or lacrimation;
2. ipsilateral nasal congestion and/or rhinorrhoea;
3. ipsilateral eyelid oedema;
4. ipsilateral forehead and facial sweating;
5. ipsilateral miosis and/or ptosis.
D. Attacks have a frequency above 5 per day for more than half of
the time, although periods with lower frequency may occur.
E. Attacks are prevented completely by therapeutic doses of
indometacin.
F. Not attributed to another disorder.
Proposed diagnostic criteria
3.2 R Diagnostic criteria
A. At least 20 attacks fulfilling criteria B ^E.
B. Attacks of unilateral pain lasting 2^30 min.
C. Headache is accompanied by at least one of the following:
1. ipsilateral conjunctival injection and/or lacrimation;
2 ipsilateral nasal congestion and/or rhinorrhoea;
3. ipsilateral eyelid oedema;
4. ipsilateral forehead and facial sweating;
5. ipsilateral miosis and/or ptosis;
6. ipsilatreal forehead and facial flushing;
7. ipsilateral sense of aural fullness and/or periaural swelling.
D. Attacks have a frequency above 5 per day for more than half of
the time, although periods with lower frequency may occur.
E. Attacks are prevented completely by indometacin.
F. Not attributed to another disorder.
Paroxysmal hemicrania
Epidemiology
PH is a rare syndrome. The relatively small number of
patients that we diagnosed at the National Hospital for
Neurology and Neurosurgery, and the Hospital for Sick
Children, Great Ormond Street, given our referral base and
interests, suggests this is the case. Considering we have seen
more than 800 patients affected by TACs in the same
period, our cohort is certainly small. In our group, the
youngest patient with PH was 4-years old and the oldest
was 68-years old. Our findings confirm the existing
literature that states this condition can begin at any age
(Kudrow and Kudrow, 1989; Broeske et al., 1993; Gladstein
et al., 1994; de Almeida et al., 2004), but the mean age at
onset is in the 30s (Antonaci and Sjaastad, 1989). The Vaga
study suggests that the prevalence of PH is no more than 1
in 1838 (Sjaastad and Bakketeig, 2007), and is in line with
our experience. Given nearly all the TACs seen in the
period in our unit have been seen by at least one of us
(P.J.G.), our data (Cohen et al., 2006a) and that again of
the Vaga study (Sjaastad and Bakketeig, 2007), suggest that
Short-lasting Unilateral Neuralgiform headache attacks with
Conjunctival injection and Tearing (SUNCT/SUNA) syndrome is at least as common as PH. This may be a valuable
clinical lesson. Our impression is that SUNCT is misdiagnosed more often as atypical trigeminal neuralgia than
PH is misdiagnosed as cluster headache. Given the
therapeutic implications neither mistake is helpful to the
patients.
Brain (2008), 131, 1142^1155
1149
widely in the face and head. This suggests a modification
to the current criteria to acknowledge the side distribution,
and is probably a finding that follows detailed questioning
of a larger cohort of patients. Patients need to be asked
about pain distribution, and conversely the distribution
should not put one off the diagnosis.
Autonomic symptoms
By current definition, PH patients require at least one
cranial autonomic feature with attacks. Lacrimation, nasal
congestion, conjunctival injection and rhinorrhoea are
the most frequent symptoms and signs accompanying the
attack (Antonaci and Sjaastad, 1989). In our group,
the majority of the patients (97%) had at least one of
the autonomic features currently specified (Headache
Classification Committee of The International Headache
Society, 2004). However, a wider range of autonomic
features was reported, such as facial flushing (55%), a sense
of aural fullness (26%) or of aural swelling (13%). In
particular, one patient only had the sense of aural fullness.
Boes and colleagues (1998) recognized a similar symptom
in a small series of patients, and we have found it clinically
useful to enquire of this symptom. It can be found in each
of the TACs, including cluster headache and SUNCT/SUNA
and can be bothersome to some patients. Our findings
suggest this symptom should be included in the list of
cranial autonomic features in any revision of the diagnostic
criteria for PH, and probably for TACs more broadly.
Sex distribution
Laterality and severity of attacks
PH has been considered to be predominantly a problem of
females. In the initial reported female:male ratio was 7 : 1
(Sjaastad and Dale, 1974), and in a subsequent review of 84
patients (Antonaci and Sjaastad, 1989) the female:male ratio
was 2.36 : 1. In our case series, we did not find a clear
female preponderance. This is in mark contrast to cluster
headache where there is a clear male preponderance (Bahra
et al., 2002) and rather more similar to SUNCT at 1.5 male
to 1 female (Cohen et al., 2006a). The prevailing view of
PH as a condition of females may arise from mis-diagnosis
of males with PH as cluster headache, as it is the more
common problem and recognized to have a male
preponderance.
Typically the attacks of PH are strictly side-locked, although
three cases of bilateral pain have been reported (Pollmann
and Pfaffenrath, 1986; Mulder and Spierings, 2004;
Bingel and Weiller, 2005). In our series, no patients had
convincing bilateral attacks, and we have speculated that
such patients may be a separate group (Matharu and
Goadsby, 2005). Typically, cluster headache is described as
an excruciating syndrome and is often called ‘suicide
headache’ (Torelli and Manzoni, 2003), although it is now
reasonably established that other TACs, as SUNCT/SUNA
are indeed excruciating conditions (Cohen et al., 2006a).
We found that the majority of our patients considered their
pain to be the worst pain they had ever experienced with a
severity of 10 out of 10 in a VRS. These findings support
the general impression that TACs are indeed all extremely
painful conditions and underscore the importance of
accurate diagnosis and swift management to alleviate the
patients’ very considerable suffering.
Site of the pain
The IHS (Headache Classification Committee of The
International Headache Society, 2004) considers the site
of the pain to be unilateral, supraorbital or temporal, and a
previous case series reported frequent localization in the
ocular and temporal regions (Antonaci and Sjaastad, 1989).
In our cohort, we also found that pain was mainly reported
over the temporal (77%) region, orbital (77%) and
supraorbital/frontal region (55%) consistent with the
current criteria. However, the pain was also reported
Duration and frequency
Current criteria for the diagnosis of PH mandate attack
duration of 2 and 30 min and a frequency of more than five
attacks per day for more than half of the time (Headache
Classification Committee of The International Headache
1150
Brain (2008), 131, 1142^1155
Society, 2004). In a prospective study of 105 attacks in five
patients (Russell and Storstein, 1984), the mean attack
duration was 13 min with a range of 3–46 min and the
mean attack frequency was 14 attacks in 24 h with a range
of 4–38. We have found a range of the attack duration
between 10 s and 4 h, with a mean of 17 min, although of
the patients with very short attacks had longer attacks of
several minutes. We noted in about half of the patients
(55%) the attack duration was more than 30 min, although
within each individual attacks of 30 min or less would
occur. One patient even had an attack that lasted 48 h.
Whether this was one attack or not can never be established
retrospectively, although patient histories at the outset are
always retrospective so it is valuable to be aware of this
variability. The longer attack length can lead to the misdiagnosis of cluster headache, although the lack of response
to oxygen and sumatriptan are against that conclusion, as is
the positive response to indomethacin. Moreover, one
would expect variability over time with bouts not being
identical in episodic PH and some variability in chronic PH
from year to year until it is treated.
Interictal painçPH versus hemicrania
continua
An interictal discomfort or pain is reported in up to onethird of patients (Antonaci and Sjaastad, 1989). We found
that 18 patients had background pain between each attack.
The main differential diagnosis here is hemicrania continua.
A headache diary is very useful during the work-up of PH
as it can provide important information regarding temporal
aspects of the pain. In our experience PH typically has
prominent cranial autonomic features, whereas in hemicrania continua these features are modest or inconstant.
Further, in PH the attacks are much more severe and the
attack length is relatively short, compared with longer less
severe worsenings of hemicrania continua. It is also our
general impression that background pain in PH is generally
less severe than it is in hemicrania continua (Matharu et al.,
2003a). A careful history supplemented with a headache
diary allows these distinctions to be made. It must be said
that a weakness of this analysis might be the view that both
disorders are on a spectrum. Given the clinical distinctions
above and the very different pattern of brain imaging seen
when comparing PH (Matharu et al., 2006) to hemicrania
continua (Matharu et al., 2004), it would seem reasonable
to continue to endeavour to study these conditions
separately.
Concomitant headache, migraine and
background pain
In our cohort, 16 (52%) patients had a personal history of
migraine. Migraine is common (Steiner et al., 2003), and so
commonly seen in patients with other trigeminal autonomic
cephalalgias, such as cluster headache (Bahra et al., 2002)
E. Cittadini et al.
and SUNCT/SUNA (Cohen et al., 2006a). The high
prevalence of migraine in our group may suggest a
predisposition for primary headaches, patients with a
disabling headache seeking medical advice more often or
perhaps reveals that the biological predisposition to migraine
is more common than we have hitherto considered. If one
considers the peak prevalence of migraine about one-third of
adult Western females, using criteria (Headache Classification
Committee of The International Headache Society, 2004) that
would under-diagnose migraine by excluding patients with
so-called tension-type headache with one migraine feature,
such as photophobia or phonophobia, or patients with now
accepted chronic migraine (Olesen et al., 2006), is the
background genetic rate of migraine much higher than
previously considered? It may simply be that in patients with
two genetic hits—PH and migraine, the migraine biology is
more likely to be activated than without the extraPH-problem. Indeed, patients with migraine (Wilkinson
et al., 2001; Bahra et al., 2003), cluster headache (Paemeleire
et al., 2006) or SUNCT/SUNA (Cohen et al., 2006a) with
underlying migraine biology tend to dominate cohorts who
develop medication overuse headache. Here, seven of eight
patients with medication overuse had migraine biology,
which supports the emerging principle that migraine, active
or latent, may be the crucial factor in predisposing to
medication overuse problems (Goadsby, 2006).
Behaviour
Typically, patients with cluster headaches are agitated and
restless (Torelli and Manzoni, 2003), indeed aggressive
behaviour is seen (Bahra et al., 2002). Recently, it has been
reported that 62% of the patients with SUNCT were
agitated during attacks (Cohen et al., 2006a). In both
conditions, activations in the region of the posterior
hypothalamic grey have been reported using functional
imaging techniques (May et al., 1998, 1999; Cohen et al.,
2006b). Importantly, posteromedial hypothalamotomy has
been used as a therapeutic treatment in patients with
behavioural disorders that include aggression (Sano et al.,
1970; Schvarcz et al., 1972; Sano and Mayanagi, 1988).
Recent functional imaging results with PH identify again
activations in the region of the posterior hypothalamic grey
(Matharu et al., 2006), as for cluster headache and SUNCT.
Whether the area is indeed hypothalamic or sub-thalamic,
such as the calcitonin gene-related peptide (CGRP) neurons
of the sub-parafasicular thalamus (de Lacalle and Saper,
2000) is yet to be determined. Our finding here that 80% of
the patients were agitated, or restless or both, and one
quarter were described as being aggressive during the pain
is consistent with the broad correlation of the functional
imaging findings with the phenotype of the attack.
Family history
A family history of PH has been reported (Cohen et al.,
2006a), and this case is in our reported cohort (#6).
Paroxysmal hemicrania
Brain (2008), 131, 1142^1155
1151
Table 7 Comparison on the trigeminal autonomic cephalalgias based on cohorts we have studied (Bahra et al., 2002; Cohen
et al., 2005; Cohen et al., 2006a) and patients we have reviewed (Cohen et al., 2007)
Sex
Pain
Quality
Severity
Distribution
Attacks
Frequency (per day)
Length (minutes)
Triggers
Alcohol
NTG
Cutaneous
Agitation/restlessness
Episodic versus chronic
Circadian/circannual periodicity
Treatment effects
Oxygen
Sumatriptan 6 mg
Indomethacin
Migraine features with attacks
Nausea
Photophobia/phonophobia
Cluster headache
Paroxysmal hemicrania
SUNCT/SUNA
3M to 1F
M=F
1.5M to 1F
Sharp/stab/throb
Very severe
V14C24V24V3
Sharp/stab/throb
Very severe
V14C24V24V3
Sharp/stab/throb
Severe
V14C24V24V3
1^ 8
30 ^180
20
2^30
100
1^5
+++
+++
+
+
90%
90:10
Present
80%
35:65
Absent
+++
65%
10:90
Absent
70%
90%
No effect
No effect
20%
100%
No effect
510%
No effect
50%
65%
40%
65%
25%
25%
Although there are exceptions to all rules the responses and effects listed reflect the behaviour of the largest parts of the cohorts we
have studied.
SUNCT = SUNA, Short-lasting unilateral neuralgiform headache attacks with conjunctival injection and tearing/short-lasting unilateral
neuralgiform headache attacks with cranial autonomic features; C = cervical; V = trigeminal.
Typically, migraine is described as a genetic headache
disorder (Ferrari, 1998) and mutations have been discovered in patients affected by familial hemiplegic migraine
(Ophoff et al., 1996; De Fusco et al., 2003; Dichgans et al.,
2005). Among the TAC syndromes, a family history of CH
(Leone et al., 2001) and SUNCT has been reported
(Gantenbein and Goadsby, 2005). A large group of patients
need to be evaluated in order to establish a genetic
component, although the rarity of this disorder will make
this a challenge. It certainly seems likely that the underlying
determinant of the disorder is genetically determined.
Triggers
Most attacks are spontaneous, while it is reported that 10%
of patients can trigger an attack with neck movement and
about 7% of the patients can trigger an attack with alcohol
(Antonaci and Sjaastad, 1989). In our cohort, the majority
of the attacks were spontaneous; when identified triggers
included stress or relaxation after stress, alcohol, exercise
and neck movement. In comparison, cluster headache
attacks can be typically and easily triggered by alcohol,
nitroglycerine (NTG) (Ekbom, 1968; Fanciullacci et al.,
1995), exercise, strong smells and increased body heat
(Bahra et al., 2002), whereas in SUNCT/SUNA the attacks
are generally triggered by cutaneous triggers such as
touching the face or scalp, washing, shaving, eating brushing the teeth, talking and coughing (Pareja and Sjaastad,
1997; Cohen et al., 2006a). One patient out of four
reported a typical attack following use of a NTG spray. In
general terms, lack of response to alcohol, NTG and lack of
cutaneous triggers to attacks are clinical pointers to PH
when considering the differential diagnosis among the TAC
syndromes (Table 7).
Indometacin response
A response to indometacin has been said to be a sine qua
non in PH and we have adopted that position to define the
cohort here. The mean daily dosage is 100 mg with a range
of 25–300 mg. However, some patients needed only
12.5 mg/day (Antonaci and Sjaastad, 1989). We found
that all patients (97%) who could tolerate oral indometacin
had an absolute response to the oral trial. The range of
effective doses in our cohort varied between 30 mg in a
5-year old to 300 mg/day. Interestingly, we noted that one
patient had (#26) had a highly effective response during
two consecutive trials with indometacin. However, a third
trial at the highest dose of 375 mg/day for 2 weeks was
not beneficial and subsequently fourth trial at the dose
300 mg/day was beneficial. Otherwise, responses and doses
tended to be stable with time so that it is our practice to
1152
Brain (2008), 131, 1142^1155
re-examine and re-image the brain in patients who stop
responding. Some two-thirds of patients report at some
point side-effects, mainly gastrointestinal problems, which
may lead to a discontinuation of the medicine. In this
context, where gastrointestinal side-effects have been a
problem employing a placebo-controlled Indotest by
injection was very helpful (Matharu et al., 2004). In our
cohort, the majority of the patients (82%) had a positive
test, which confirmed the diagnosis of PH. We also assessed
the response to high-flow oxygen and sumatriptan subcutaneously, which are typically positive in cluster headaches (Lance and Goadsby, 2005). Open label use of
subcutaneous sumatriptan 6 mg has been reported to be
reported to be both ineffective (Dahlof, 1993; Antonaci
et al., 1998b) and helpful (Pascual and Quijano, 1998) in
patients with PH. If one considers a good or better
response as an outcome, in our patients oxygen was no
effective. Sumatriptan subcutaneously was either clearly
useful in about 20% or not useful at all; this would explain
the controversy in the literature which is simply a sampling
issue. A response to sumatriptan should not distract from a
trial of indometacin.
Neurological examination, secondary
headache and associations
As a rule, patients with PH tend to have normal
neurological examinations. In our group, three (10%)
patients had sensory impairment: hyperaesthesia to pinprick
in V1-V2-V3 in one case (#12), allodynia in V1 in a second
case (#3) and in the third case, sensory loss in all modalities
over the medial aspect of the right upper and lower eyelid,
and medial aspect of the area above the right supra-orbital
ridge. One patient (#12) of this group had bilateral vascular
loops not compressing the trigeminal nerve and the sensory
impairment was on the same side of the attack. One patient
had post-traumatic PH and the sensory impairment was
attributed to damage of the trigeminal nerve following a
traumatic injury (#29, Matharu and Goadsby, 2001). One
patient (#14) with an arterio-venous malformation had
impairment of visual acuity in the corresponding eye. The
headache was on the same side of the malformation,
although it started about 18 years after the initial
presentation of the proptosis. When he developed PH, the
proptosis had progressed significantly. Three years later, he
underwent a trial with indometacin with complete resolution of the headache. It seems unlikely that the malformation was causal. Overall, in our case series there is only
infrequent impairment of the trigeminal nerve. This is
substantially different form the recent data regarding
SUNCT/SUNA syndrome in which the reduced facial
sensation may be more common (Cohen et al., 2006a).
Probably as a result of publication bias, secondary PH is
relatively common in the literature in comparison to the
primary variety, and has been suggested to be caused by a
diverse set of pathological processes (Boes and Dodick, 2002).
E. Cittadini et al.
One of our patients had an ischaemic lesion in the basal
ganglia and pons, which may account for the problem
(#19). PH has been described in patients with macroprolactinoma (Boes et al., 1998; Sarov et al., 2006), prolactinsecreting adenoma (Raskin, 1988) and microadenoma
(Gatzonis et al., 1996). Pituitary adenoma have also been
reported in patients with SUNCT (Ferrari et al., 1988;
Massiou et al., 2002; Levy et al., 2003; Matharu et al., 2003b)
and in patients with cluster headache (Greve and Mai,
1988; Milos et al., 1996; Porta-Etessam et al., 2001; Negoro
et al., 2005).
Several hypotheses have been proposed to explain the
association between pituitary adenomas and TAC syndromes. First, these tumours may have a mechanical effect.
Second, an increased intrasellar pressure may play a role in
the genesis of the headache (Arafah et al., 2000). Finally, it
has been proposed that the pathogenesis of the attacks is
predominately neurohormonally mediated rather than by
the size or invasiveness of the tumour (Levy et al., 2004).
Activation of the headache phenotype by treatments aimed
at biochemical change in smaller tumours and the early
recrudescence of symptoms with biochemical changes (Levy
et al., 2005) both favour a non-mechanical explanation.
PH has been reported in association with migraine,
cluster headache, trigeminal neuralgia and cough headaches
(Matharu et al., 2003a). We had two patients with PH-tic
syndrome, that we have reported previously (Boes et al.,
2003). Both patients had abnormal brain MRI findings: one
patient (#25) had vascular loop compressing the trigeminal
nerve on the same side as the pain and the other patient
(#30) had an ectatic basilar artery close to the trigeminal
root entry zone, ipsilateral to the side of his head pains.
The latter underwent microvascular decompression with a
resolution of both pains and he has been pain-free for the
last 5 years. So far, there have been no other reports of
vascular decompression procedures in PH-tic-syndrome
(Hannerz, 1993; Caminero et al., 1998; Martinez-Salio
et al., 2000; Zukerman et al., 2000; Boes et al., 2003; Leone
et al., 2006). Occasionally, cluster-tic syndrome can be also
associated with trigeminal nerve compression (Solomon
et al., 1985), although in those four cases the vascular
decompression leads only to relief of the tic component.
PH has been described in association with cough headache
(Mateo and Pascual, 1999) and both headaches responded
well to indometacin. We had one patient (# 9) with PH and
a cough headache. Interestingly, only the paroxysmal pain
responded satisfactory to indometacin whereas the cough
pain did not improve. Meningioma of cavernous sinus has
been described in association with PH (Sjaastad et al., 1995)
with only a moderate and transient improvement of the
headache following excision. We had one patient (#20) with
a history of sphenoid wing meningioma. The removal of
the tumour led to a transitory remission lasting about 9
months. After this period, the pain came back with the
same pattern that was evident preoperatively.
Paroxysmal hemicrania
Pathogenesis
The pathogenesis of PH is poorly understood in relation to
other primary headaches, such as migraine (Goadsby and
Oshinsky, 2008). CGRP and vasoactive intestinal polypeptide (Goadsby and Edvinsson, 1996) are elevated during
acute attacks of PH. This release is likely to mark
trigeminovascular system and cranial parasympathetic
activation, and is similar to that observed in cluster
headache (Goadsby and Edvinsson, 1994). Trigeminal and
autonomic activation in concert seems a marker of this
group of syndromes—the trigeminal—autonomic cephalalgias (TACs, Goadsby and Lipton, 1997). The physiology of
this system is relatively well described (May and Goadsby,
1999), indeed a degree of cranial autonomic activation is a
normal response to a cranial nociceptive input (May et al.,
2001). An important new dimension to understanding these
syndromes, and PH in particular, comes with the application of functional imaging techniques. It has been shown
during PH there is activation in the posterior hypothalamic
region (Matharu et al., 2006), in addition to ventral
midbrain activation. The same region near the posterior
hypothalamus is also activated in cluster headache and
SUNCT (Cohen and Goadsby, 2006), while the ventrolateral midbrain area is activated in hemicrania continua
(Matharu et al., 2004). Whether this later area holds the key
to understanding the indometacin effect in these syndromes
remains one of the more fascinating questions for the next
few years. Indometacin inhibits the production of nitric
oxide (NO) by endothelial and inducible nitric oxide
synthase (Beasley et al., 1998; Hrabak et al., 2001) and
perhaps works by antagonizing one or more steps in the
NO pathway, although given the relatively small proportion
of PH patients sensitive to nitroglycerin in our series there
must be other aspects to this problem. Certainly posterior
hypothalamic region activation is the common theme of
TACs (Cohen and Goadsby, 2006) and with the clinical
picture makes a useful case for the utility of the grouping.
Conclusion
We have described 31 patients affected by PH and have
characterized the clinical picture of this disorder in great
detail. An advantage of the study is that all patients have
been seen by at least two clinicians with extensive
experience of TACs, that the cases have been contemporaneously seen and that the clinical aspects have been done
prospectively. A limitation of all hospital-based studies is
the bias of referral; however, for such a rare condition there
seems no other reasonable way to collect an appropriate
cohort. We have found the current IHS classification
(Headache Classification Committee of The International
Headache Society, 2004) to be helpful, and provide
suggestions based on our data for a revision. We propose
removing specifying the site of pain accepting that it can be
anywhere on the head, and to include other autonomic
Brain (2008), 131, 1142^1155
1153
features, such as ipsilateral forehead/facial flushing, ipsilateral sense of aural fullness and periaural swelling in the
range of cranial autonomic features present during the
attacks (Table 6). We emphasize that the absolute response
to indometacin is essential for the diagnosis, and suggest
the placebo-controlled Indometacin test as a suitable aid in
the diagnostic evaluation. We note the frequency of the
attacks is more than five per day, although low attack
frequency can be present with other typical attacks.
Moreover, while the typical attack is short at 20 min,
longer attacks occur. Interestingly, the majority of the
patients showed signs of agitation, or restlessness or both
and about one-third were aggressive during the attacks
consistent with the neuroimaging of the TACs locating
some important part of the pathophysiology to the
posterior hypothalamic region. PH is a treatable primary
headache that is very rewarding to diagnose and manage
that can be differentiated on clinical grounds from the
other TACs (Table 7) in order to provide optimal
management.
References
Antonaci F, Pareja JA, Caminero AB, Sjaastad O. Chronic paroxysmal
hemicrania and hemicrania continua. Parenteral indomethacin: the
‘Indotest’. Headache 1998a; 38: 122–8.
Antonaci F, Pareja JA, Caminero AB, Sjaastad O. Chronic paroxysmal
hemicrania and hemicrania continua: lack of efficacy of sumatriptan.
Headache 1998b; 38: 197–200.
Antonaci F, Sjaastad O. Chronic paroxysmal hemicrania (CPH): a review
of the clinical manifestations. Headache 1989; 29: 648–56.
Arafah BM, Prunty D, Ybarra J, Hlavin ML, Selman WR. The dominant
role of increased intrasellar pressure in the pathogenesis of hypopituitarism, hyperprolactinemia, and headaches in patients with pituitary
adenomas. J Endocrinol Metab 2000; 85: 1789–93.
Bahra A, May A, Goadsby PJ. Cluster headache: a prospective clinical
study in 230 patients with diagnostic implications. Neurology 2002; 58:
354–61.
Bahra A, Walsh M, Menon S, Goadsby PJ. Does chronic daily headache
arise de novo in association with regular analgesic use? Headache 2003;
43: 179–90.
Beasley TC, Bari F, Thore C, Thirkawala N, Louis T, Busija D. Cerebral
ischemia/reperfusion increases endothelial nitric oxide synthase levels by
an indomethacin-sensitive mechanism. J Cereb Blood Flow Metab 1998;
18: 88–96.
Bigal ME, Rapoport AM, Sheftell FD, Tepper SJ, Lipton RB. Transformed
migraine and medication overuse in a tertiary headache centre—clinical
characteristics and treatment outcomes. Cephalalgia 2004; 24: 483–90.
Bingel U, Weiller C. An unusual indomethacin sensitive headache: a case
of bilateral episodic paroxysmal hemicrania without autonomic
symptoms? Cephalalgia 2005; 25: 148–50.
Boes CJ, Dodick DW. Refining the clinical spectrum of chronic
paroxysmal hemicrania: a review of 74 patients. Headache 2002; 42:
699–708.
Boes CJ, Matharu MS, Goadsby PJ. The paroxysmal hemicrania-tic
syndrome. Cephalalgia 2003; 23: 24–8.
Boes CJ, Swanson JW, Dodick DW. Chronic paroxysmal hemicrania
presenting as otalgia with a sensation of external acoustic meatus
obstruction: two cases and a pathophysiologic hypothesis. Headache
1998; 38: 787–91.
Broeske D, Lenn NJ, Cantos E. Chronic paroxysmal hemicrania in a young
child: possible relation to ipsilateral occipital infarction. Journal Child
Neurol 1993; 8: 235–6.
1154
Brain (2008), 131, 1142^1155
Caminero AB, Pareja JA, Dobato JL. Chronic paroxysmal hemicrania-tic
syndrome. Cephalalgia 1998; 18: 159–61.
Cittadini E, Goadsby PJ. Revisiting the International Headache Society
criteria for paroxysmal hemicrania: a case series of 21 patients.
Cephalalgia 2006; 26: 1401–2.
Cohen AS, Goadsby PJ. Functional neuroimaging of primary headache
disorders. Expert Rev Neurother 2006; 6: 1159–72.
Cohen AS, Matharu MS, Goadsby PJ. Suggested guidelines for treating
SUNCT and SUNA. Cephalalgia 2005; 25: 1200.
Cohen AS, Matharu MS, Goadsby PJ. Short-lasting Unilateral
Neuralgiform Headache Attacks with Conjunctival injection and
Tearing (SUNCT) or cranial Autonomic features (SUNA). A prospective
clinical study of SUNCT and SUNA. Brain 2006a; 129: 2746–60.
Cohen AS, Matharu MS, Goadsby PJ. Trigeminal autonomic cephalalgias:
current and future treatments. Headache 2007; 47: 969–80.
Cohen AS, Matharu MS, Kalisch R, Friston K, Goadsby PJ. Functional
MRI in SUNCT (Short-lasting Unilateral Neuralgiform headache attacks
with Conjunctival injection and Tearing) and SUNA (Short-lasting
Unilateral Neuralgiform headache attacks with cranial Autonomic
symptoms) shows differential hypothalamic activation with increasing
pain. Cephalalgia 2006b; 26: 1402–3.
Dahlof C. Subcutaneous sumatriptan does not abort attacks of chronic
paroxysmal hemicrania (CPH). Headache 1993; 33: 201–2.
de Almeida DB, Cunali PA, Santos HL, Brioschi M, Prandini M. Chronic
paroxysmal hemicrania in early childhood: case report. Cephalalgia
2004; 24: 608–9.
De Fusco M, Marconi R, Silvestri L, Atorino L, Rampoldi L, Morgante L,
et al. Haploinsufficiency of ATP1A2 encoding the Na+/K+ pump 2
subunit associated with familial hemiplegic migraine type 2. Nat Genet
2003; 33: 192–6.
de Lacalle S, Saper CB. Calcitonin gene-related peptide-like immunoreactivity marks putative visceral sensory pathways in human brain.
Neuroscience 2000; 100: 115–30.
Dichgans M, Freilinger T, Eckstein G, Babini E, Lorenz-Depiereux B,
Biskup S, et al. Mutation in the neuronal voltage-gated sodium channel
SCN1A causes familial hemiplegic migraine. Lancet 2005; 366: 371–7.
Drummond PD. A quantitative assessment of photophobia in migraine
and tension headache. Headache 1986; 26: 465–9.
Ekbom K. Nitroglycerin as a provocative agent in cluster headache. Arch
Neurol 1968; 19: 487–93.
Fanciullacci M, Alessandri M, Figini M, Geppetti P, Michelacci S. Increase
in plasma calcitonin gene-related peptide from extracerebral circulation
during nitroglycerin-induced cluster headache attack. Pain 1995; 60:
119–23.
Ferrari MD. Migraine. Lancet 1998; 351: 1043–51.
Ferrari MD, Haan J, van Seters AP. Bromocriptine-induced trigeminal
neuralgia attacks in a patient with pituitary tumor. Neurology 1988; 38:
1482–4.
Gantenbein A, Goadsby PJ. Familial SUNCT. Cephalalgia 2005; 25: 457–9.
Gatzonis S, Mitsikostas DD, Ilias A, Zournas CH, Papageorgiou C. Two
more secondary headaches mimicking chronic paroxysmal hemicrania.
Is this the exception or the rule? Headache 1996; 36: 511–3.
Gladstein J, Holden EW, Peralta L. Chronic paroxysmal hemicrania in a
child. Headache 1994; 34: 519–20.
Goadsby PJ. Is medication-overuse headache a distinct biological entity?
Nat Clin Pract Neurol 2006; 2: 401.
Goadsby PJ, Edvinsson L. Human in vivo evidence for trigeminovascular
activation in cluster headache. Brain 1994; 117: 427–34.
Goadsby PJ, Edvinsson L. Neuropeptide changes in a case of chronic
paroxysmal hemicrania- evidence for trigemino-parasympathetic activation. Cephalalgia 1996; 16: 448–50.
Goadsby PJ, Lipton RB. A review of paroxysmal hemicranias, SUNCT
syndrome and other short-lasting headaches with autonomic features,
including new cases. Brain 1997; 120: 193–209.
Goadsby PJ, Oshinsky M. The pathophysiology of migraine. In: Silberstein
SD, Lipton RB and Dodick D, editors. Wolff’s headache and other head
pain. New York: Oxford; 2008. p. 105–19.
E. Cittadini et al.
Greve E, Mai J. Cluster headache-like headaches: a symptomatic feature?
A report of three patients with intracranial pathologic findings.
Cephalalgia 1988; 8: 79–82.
Hannerz J. Trigeminal neuralgia with chronic paroxysmal hemicrania: the
CPH-tic syndrome. Cephalalgia 1993; 13: 361–4.
Headache Classification Committee of The International Headache Society.
Classification and diagnostic criteria for headache disorders, cranial
neuralgias and facial pain. Cephalalgia 1988; 8: 1–96.
Headache Classification Committee of The International Headache Society.
The International Classification of Headache Disorders (second edition).
Cephalalgia 2004; 24: 1–160.
Hrabak A, Vercruysse V, Kahan IL, Vray B. Indomethacin prevents the
induction of inducible nitric oxide synthase in murine peritoneal
macrophages and decreases their nitric oxide production. Life Sci 2001;
68: 1923–30.
Irimia P, Cittadini E, Paemeleire K, Cohen AS, Goadsby PJ. Unilateral
photophobia or phonophobia in migraine compared with trigeminal
autonomic cephalalgias. Cephalalgia 2008; 28: in press.
Joubert J, Powell D, Djikowski J. Chronic paroxysmal hemicrania in a
South African black. A case report. Cephalalgia 1987; 7: 193–6.
Kayed K, Godtlibsen OB, Sjaastad O. Chronic paroxysmal hemicrania. 4.
‘‘REM sleep locked’’ nocturnal headache attacks. Sleep 1978; 1: 91–5.
Kelman L. The place of osmophobia and taste abnormalities in migraine
classification: a tertiary care study of 1237 patients. Cephalalgia 2004;
24: 940–6.
Kudrow DB, Kudrow L. Successful aspirin prophylaxis in a child with
chronic paroxysmal hemicrania. Headache 1989; 29: 280–1.
Kudrow L, Esperanca P, Vijayan N. Episodic paroxysmal hemicrania?
Cephalalgia 1987; 7: 197–201.
Lance JW, Goadsby PJ. Mechanism and management of headache.
New York: Elsevier; 2005.
Leone M, Mea E, Genco S, Bussone G. Coexistence of TACS and
trigeminal neuralgia: pathophysiological conjectures. Headache 2006; 46:
1565–70.
Leone M, Russell MB, Rigamonti A, Attanasio A, Grazzi L, D’Amico D,
et al. Increased familial risk of cluster headache. Neurology 2001; 56:
1233–6.
Levy M, Jager HR, Powell MP, Matharu MS, Meeran K, Goadsby PJ.
Pituitary volume and headache: size is not everything. Arch Neurol
2004; 61: 721–5.
Levy M, Matharu MS, Meeran K, Powell M, Goadsby PJ. The clinical
characteristics of headache in patients with pituitary tumours. Brain
2005; 128: 1921–30.
Levy MJ, Matharu MS, Powell MP, Meeran K, Goadsby PJ. Phenoytpic
characterisation of the headache associated with pituitary tumours.
Cephalalgia 2003; 23: 621.
Martinez-Salio A, Porta-Etessam J, Perez-Martinez D, Balseiro J, GutierrezRivas E. Chronic paroxysmal hemicrania-tic syndrome. Headache 2000;
40: 682–5.
Massiou H, Launay JM, Levy C, El Amran M, Emperauger B, Bousser M-G.
SUNCT syndrome in two patients with prolactinomas and bromocriptineinduced attacks. Neurology 2002; 58: 1698–9.
Mateo I, Pascual J. Coexistence of chronic paroxysmal hemicrania and
benign cough headache. Headache 1999; 39: 437–8.
Matharu MS, Boes CJ, Goadsby PJ. Management of trigeminal autonomic
cephalalgias and hemicrania continua. Drugs 2003a; 63: 1637–77.
Matharu MS, Cohen AS, Frackowiak RSJ, Goadsby PJ. Posterior
hypothalamic activation in paroxysmal hemicrania. Annals of
Neurology 2006; 59: 535–45.
Matharu MS, Cohen AS, McGonigle DJ, Ward N, Frackowiak RSJ,
Goadsby PJ. Posterior hypothalamic and brainstem activation in
hemicrania continua. Headache 2004; 44: 747–61.
Matharu MS, Goadsby PJ. Post-traumatic chronic paroxysmal hemicrania
(CPH) with aura. Neurology 2001; 56: 273–5.
Matharu MS, Goadsby PJ. Bilateral paroxysmal hemicrania or bilateral
paroxysmal cephalalgia, another novel indomethacin-responsive primary
headache syndrome? Cephalalgia 2005; 25: 79–81.
Paroxysmal hemicrania
Matharu MS, Levy MJ, Merry RT, Goadsby PJ. SUNCT syndrome
secondary to prolactinoma. Journal Neurol NeurosurgPsychiatry 2003b;
74: 1590–2.
May A, Bahra A, Buchel C, Frackowiak RS, Goadsby PJ. Hypothalamic
activation in cluster headache attacks. Lancet 1998; 352: 275–8.
May A, Bahra A, Buchel C, Turner R, Goadsby PJ. Functional MRI in
spontaneous attacks of SUNCT: short-lasting neuralgiform headache
with conjunctival injection and tearing. Ann Neurol 1999; 46: 791–3.
May A, Buchel C, Turner R, Goadsby PJ. MR-angiography in facial and
other pain: neurovascular mechanisms of trigeminal sensation. J Cereb
Blood Flow Metab 2001; 21: 1171–6.
May A, Goadsby PJ. The trigeminovascular system in humans: pathophysiological implications for primary headache syndromes of the neural
influences on the cerebral circulation. J Cere Blood Flow Metab 1999;
19: 115–27.
Milos P, Havelius U, Hindfelt B. Cluster-like headache in a patient with a
pituitary adenoma. With a review of the literature. Headache 1996; 36:
184–8.
Mulder LJ, Spierings EL. Non-lateralized pain in a case of chronic
paroxysmal hemicrania? Cephalalgia 2004; 24: 52–53.
Negoro K, Kawai M, Tada Y, Ogasawara J, Misumi S, Morimatsu M. A
case of postprandial cluster-like headache with prolactinoma: dramatic
response to cabergoline. Headache 2005; 45: 604–6.
Newman LC, Gordon ML, Lipton RB, Kanner R, Solomon S. Episodic
paroxysmal hemicrania: two new cases and a literature review.
Neurology 1992; 42: 964–66.
Olesen J, Bousser MG, Diener HC, Dodick D, First M, Goadsby P, et al.
New appendix criteria open for a broader concept of chronic migraine.
Cephalalgia 2006; 26: 742–6.
Ophoff RA, Terwindt GM, Vergouwe MN, van Eijk R, Oefner PJ,
Hoffman SMG, et al. Familial hemiplegic migraine and episodic ataxia
type-2 are caused by mutations in the Ca2+ channel gene CACNL1A4.
Cell 1996; 87: 543–52.
Paemeleire K, Bahra A, Evers S, Matharu MS, Goadsby PJ. Medicationoveruse headache in cluster headache patients. Neurology 2006; 67:
109–13.
Pareja JA. Chronic paroxysmal hemicrania: dissociation of the pain and
autonomic features. Headache 1995; 35: 111–3.
Pareja JA, Sjaastad O. SUNCT syndrome. A clinical review. Headache
1997; 37: 195–202.
Pascual J, Quijano J. A case of chronic paroxysmal hemicrania
responding to subcutaneous sumatriptan. J Neurol Neurosurg
Psychiatry 1998; 65: 407.
Pollmann W, Pfaffenrath V. Chronic paroxysmal hemicrania: the first
possible bilateral case. Cephalalgia 1986; 6: 55–7.
Porta-Etessam J, Ramos-Carrasco A, Berbel-Garcia A, Martinez-Salio A,
Benito-Leon J. Clusterlike headache as first manifestation of a
prolactinoma. Headache 2001; 41: 723–5.
Raskin NH. Headache. New York: Churchill-Livingstone; 1988.
Brain (2008), 131, 1142^1155
1155
Russell D. Chronic paroxysmal hemicrania: severity, duration and time of
occurrence of attacks. Cephalalgia 1984; 4: 53–6.
Russell D, Storstein L. Chronic paroxysmal hemicrania: heart rate changes
and ECG rhythm disturbances. A computerized analysis of 24h
ambulatory ECG recordings. Cephalalgia 1984; 4: 135–44.
Sano K, Mayanagi Y. Posteromedial hypothalamotomy in the treatment of
violent, aggressive behaviour. Acta Neurochir Suppl (Wien) 1988; 44:
145–51.
Sano K, Mayanagi Y, Sekino H, Ogashiwa M, Ishijima B. Results of
stimulation and destruction of the posterior hypothalamus in man.
J Neurosurg 1970; 33: 689–707.
Sarov M, Valade D, Jublanc C, Ducros A. Chronic paroxysmal hemicrania
in a patient with a macroprolactinoma. Cephalalgia 2006; 26: 738–41.
Schvarcz JR, Driollet R, Rios E, Betti O. Stereotactic hypothalamotomy for
behaviour disorders. J Neurol Neurosurg Psychiatry 1972; 35: 356–9.
Sjaastad O. Cluster headache syndrome. London: W. B. Saunders; 1992.
Sjaastad O, Bakketeig LS. The rare, unilateral headaches. Vaga study of
headache epidemiology. J Headache Pain 2007; 8: 19–27.
Sjaastad O, Dale I. Evidence for a new (?) treatable headache entity.
Headache 1974; 14: 105–8.
Sjaastad O, Dale I. A new (?) clinical headache entity ‘‘chronic paroxysmal
hemicrania’’. Acta Neurologica Scandinavica 1976; 54: 140–59.
Sjaastad O, Stovner LJ, Stolt-Nielsen A, Antonaci F, Fredriksen TA. CPH
and hemicrania continua: requirements of high dose indomethacin
dosages- an ominous sign? Headache 1995; 35: 363–7.
Solomon S, Apfelbaum RI, Guglielmo KM. The cluster-tic syndrome and
its surgical therapy. Cephalalgia 1985; 5: 83–9.
Steiner TJ, Scher AI, Stewart WF, Kolodner K, Liberman J, Lipton RB.
The prevalence and disability burden of adult migraine in England and
their relationships to age, gender and ethnicity. Cephalalgia 2003; 23:
519–27.
Stewart WF, Bigal ME, Kolodner K, Dowson A, Liberman JN, Lipton RB.
Familial risk of migraine: variation by proband age at onset and
headache severity. Neurology 2006; 66: 344–8.
Torelli P, Manzoni GC. Pain and behaviour in cluster headache. A
prospective study and review of the literature. Funct Neurol 2003; 18:
205–10.
Vingen JV, Pareja JA, Stovner LJ. Quantitative evaluation of photophobia
and phonophobia in cluster headache. Cephalalgia 1998; 18: 250–6.
Wilkinson SM, Becker WJ, Heine JA. Opiate use to control bowel motility
may induce chronic daily headache in patients with migraine. Headache
2001; 41: 303–9.
Zanchin G, Dainese F, Mainardi F, Mampreso E, Perin C, Maggioni F.
Osmophobia in primary headaches. J Headache Pain 2005; 6: 213–5.
Zidverc-Trajkovic J, Pavlovic A, Mijajlovic M, Jovanovic Z, Sternic N,
Kostic V. Cluster headache and paroxysmal hemicrania: differential
diagnosis. Cephalalgia 2005; 25: 244–8.
Zukerman E, Peres MFP, Kaup AO, Monzillo PH, Costa AR. Chronic
paroxysmal hemicrania-tic syndrome. Neurology 2000; 54: 1524–6.