The Syndrome of Transient Headache with Neurological Deficits and CSF Lymphocytosis (HaNDL): Electrophysiological Findings Suggesting a Migrainous Pathophysiology

Episodes of headache with transient neurological deficits and cerebrospinal fluid (CSF) lymphocytosis are suggestive of a benign and self-limiting syndrome which was first delineated by Bartleson et al. (1) and given the acronym HaNDL in 1995 by Berg and Williams (2). Gómez-Aronda et al. (3), who reported the hitherto largest series of patients (n = 50), preferred to call the syndrome pseudomigraine with temporary neurological symptoms and lymphocytic pleocytosis (PMP). The syndrome has been classified (code 7.8) in the 2nd edition of the International Headache Society classification of headache disorders (ICHD-II) (4) as a secondary headache attributed to a non-vascular intracranial disorder with the following diagnostic criteria: A, episodes of moderate or severe headache lasting hours before resolving fully and fulfilling criteria C and D; B, CSF pleocytosis with lymphocytic predominance (>15 cells/μl) and normal neuroimaging, CSF culture and other tests for aetiology; C, episodes of headache are accompanied by or shortly follow transient neurological deficits and commence in close temporal relation to the development of CSF pleocytosis; D, episodes of headache and neurological deficits recur over < 3 months. The precise pathogenesis of HaNDL, however, is unknown and its phenotypic similarities with migraine with aura have been underlined by some authors (3, 5–8).

Most migraineurs present interictally a deficit of the habituation of evoked cortical responses (9), for instance of pattern-reversal visual evoked potentials (PR-VEP) (10). Lack of habituation of the auditory evoked cortical potentials (11) leads in migraineurs to an increased intensity dependence (IDAP) of these potentials (12). Occipital high-frequency (10 Hz) repetitive transcranial magnetic stimulation (rTMS), which activates the underlying visual cortex in most subjects, is able to normalize PR-VEP habituation in migraine patients, which suggests that the habituation deficit is due to a reduced preactivation level (13).

Mild subclinical abnormalities of neuromuscular transmission can be detected with single-fibre electromyography (SFEMG) in migraine patients with complex neurological (14) and/or prolonged auras (15). In HaNDL neurological symptoms are indeed complex in nature and of long duration. We have therefore examined whether the electrophysiological phenotype found in migraine with aura would be present in a patient presenting with a typical HaNDL syndrome.

Case report

A 16-year-old, right-handed woman without any personal history of migraine experienced in the early morning hours left-sided numbness and weakness for 20 min and thereafter a throbbing right temporal headache persisting all day long. The headache was accompanied by photophobia, phonophobia and nausea. There was no history of previous infection or fever. Neurological examination performed during the headache phase was normal. Lumbar puncture revealed a CSF lymphocytic pleocytosis with a white blood cell count of 126/mm³ with 95% lymphocytes, a glucose level of 62 mg/dl, and an elevated protein content of 77 mg/dl. On EEG there was slow activity over the right fronto-temporal region which disappeared 3 days later. Computed tomography scan and magnetic resonance imaging (MRI) (with angiographic sequences) of the brain were normal. An extensive microbiological, serological and immunological work-up remained negative. The patient presented three similar episodes before being referred to our department, each separated by a headache-free period of 1 week. A second lumbar puncture performed 2 days after the fifth and last episode again showed a CSF lymphocytosis (white blood cell count of 260/mm³ with 69% of lymphocytes), with an elevated protein content of 90 mg/dl. At follow-up at 1 year, the patient had not experienced further episodes of headache. She had a family history of headache: her mother suffers from migraine without aura (ICHD-II code 1.1) and her father from migraine with simple visual aura (ICHD-II code 1.2.1).

Methods

All electrophysiological recordings were made twice. The first session of recording was planned during the HaDNL episode, but in the headache-free interval at least 3 days before and after an attack. The second session was performed 1 year after resolution of HaNDL. The two sessions were similar in terms of schedule except the absence of rTMS in the second session due to difficulty of timing. We began in the morning by SFEMG, followed in the afternoon by cortical auditory evoked potentials. We waited for 1 h before recording visual evoked potentials (before and just after transcranial magnetic stimulation).

Single-fibre electromyography

Single muscle fibre activity was recorded as previously described (14, 15) with a Nicolet^® Viking IV device in right m. extensor digitorum communis (EDC) during stimulation of the corresponding motor branch of the radial nerve. Results were expressed as the ‘mean value of consecutive differences’ (MCD) in μs of successive interpotential intervals. Our normative values in 16 healthy controls with a mean age of 34.7 ± 11.2 years (14, 15) are 17.1 ± 2.6 μs for mean MCD in 20 analysed fibres and the highest MCD for single fibres in healthy controls was 39 μs.

Intensity dependence of cortical auditory evoked potentials (IDAP)

IDAP was measured as described previously (11) by delivering binaural 1000-Hz tones randomly at four intensities (50, 60, 70 and 80 dB) above sensation level. The N1 (between 50 and 150 ms poststimulus) and P2 (between 90 and 230 ms poststimulus) components were identified. Peak-to-peak amplitude of N1-P2 was measured at each stimulus intensity and the amplitude/stimulus intensity function (ASF) slope was calculated. The normative value for IDAP in our laboratory is 0.64 ± 0.45 μV/10 dB tested in a group of 14 healthy volunteers without personal or family history of migraine and a mean age of 35.7 ± 14.7 years (11).

Pattern-reversal visual evoked potentials (PR-VEP)

We recorded PR-VEP before and immediately after 10-Hz rTMS over the occipital scalp (total of 900 pulses delivered in 18 trains with an intertrain interval of 10 s, intensity set to the phosphene threshold) as described in our previous study (13). We measured the amplitude of the N1-P1 component in 12 blocks of 100 averaged PR-VEP responses (six before, six after rTMS). PR-VEP potentiation or habituation were defined as percentage amplitude increase or decrease between the first and the sixth block of 100 averaged responses. We also calculated the linear regression to the six amplitudes of consecutive blocks.

Results (see Table 1)

Table 1

Results

SF-EMG	HaNDL patient during HaNDL phase	HaNDL patient after HaNDL phase	Healthy subjects in the literature	Healthy subjects in our laboratory
Mean MCD	37.1	36.2	Upper limit = 34.9	17.1 ± 2.6
Per cent of fibres with jitter > 50 μs	20	20	<10	0

IDAP	HaNDL patient during HaNDL phase	HaNDL patient after HaNDL phase	Healthy subjects in our laboratory
50 dB	12.78	10.9	5.22 ± 2.27
60 dB	12.04	10.91	6.35 ± 3.07
70 dB	17.48	16.37	6.75 ± 2.63
80 dB	22.61	20.09	7.23 ± 3.11
Slope	3.49	3.3	0.64 ± 0.45

PR-VEP	HaNDL patientduring HaNDL phasebefore rTMS	HaNDL patient during HaNDL phaseafter rTMS	HaNDL patientafter HaNDL phasewithout rTMS	Healthy subjectsin our laboratorywithout rTMS
Block 1	4.4	4.3	5.5	6.5 ± 2
Block 2	3.1	3.3	6.0	6.6 ± 2.2
Block 3	3.6	3.4	6.1	6.0 ± 2.2
Block 4	6.6	3.5	6.3	5.7 ± 2
Block 5	4.1	2.5	6.0	5.4 ± 1.8
Block 6	5.4	3.1	6.2	5.4 ± 2.2
Slope	0.32	−0.25	0.11	−0.27 ± 0.23
Habituation	24%	−28%	13%	−17.8 ± 20%

SF-EMG, Single-fibre electromyography: mean MCD and percentage of fibres with a jitter superior to 50 μs in our headache with neurological deficits and CSF lymphocytosis (HaNDL) patient (during the HaNDL phase and 1 year after its resolution), normative data in the literature for subjects younger than 20 years (extensor digitorum communis muscle) (16) and in our laboratory (n = 16 healthy subjects, mean age 34.7 ± 11.2 years) (14).

IDAP, Increased intensity dependence: mean N1-P2 amplitudes (μV; mean ± SD) at increasing stimulation intensities and mean amplitude–stimulus function slopes in our HaNDL subject (during the HaNDL phase and 1 year after its resolution) and 14 healthy volunteers (11).

PR-VEP, Pattern-reversal visual evoked potentials: blocks amplitude (μV), percentage habituation between first and sixth block of 100 averaged responses and linear regression slope of habituation over six blocks before and after 10 Hz repetitive transcranial magnetic stimulation (rTMS) in the HaNDL subject during HaNDL phase and without rTMS 1 year after its resolution. Normative data obtained in 24 healthy subjects (mean age 23.5 ± 2.5 years) in our laboratory (13) without rTMS.

First session (during the HaNDL episode but in the headache-free interval)

On SFEMG the mean MCD was abnormal (37.1 μs) and 20% of fibres had a jitter superior to 50 μs.

IDAP expressed as the amplitude/stimulus intensity function slope was clearly superior (3.49 μV/10 dB) to normal controls.

Before 10 Hz rTMS, PR-VEP N1-P1 amplitude increased between the first and subsequent blocks of averagings in our subject, indicating PR-VEP potentiation. After rTMS there was a clear habituation (−28%). Computing the linear regression over the sequential six blocks of PR-VEP averaging the slope was 0.32 before rTMS, but – 0.25 after rTMS.

Second session (after resolution of HaNDL)

On SFEMG the mean MCD was abnormal (36.2 μs) and 20% of fibres had a jitter superior to 50 μs.

IDAP expressed as the amplitude/stimulus intensity function slope was also clearly superior (3.3 μV/10 dB) to normal controls. The PR-VEP also showed a clear potentiation (13%),

Discussion

This patient with a headache with neurologic deficits and cerebrospinal fluid lymphocytosis (HaNDL) had an electrophysiological pattern similar to that found in subtypes of migraine with aura, i.e. subtle but definite SFEMG abnormalities as found in patients suffering from migraine with complex and/or prolonged aura (14, 15), increased IDAP and potentiation on PR-VEP, the latter two abnormalities being the hallmark of cortical information processing between attacks in migraine both with and without aura (9). Moreover, as recently found in migraineurs (13) a normalization of PR-VEP habituation was induced by 10 Hz rTMS of the visual cortex. This electrophysiological pattern stays stable after the resolution of the HaNDL episodes.

A possible controversy could be the age of our HaNDL subject. She is only 16 years old and few published data are available with normative data in teenagers. Nevertheless, extensive normal SFEMG values with respect to that age are published (16) and IDAP and PR-VEP data from teenagers have been obtained in our laboratory (17), showing no significant differences from adults.

The aetiopathogenesis of HaNDL is still unknown. A transcranial Doppler sonography performed during and after attacks (8) showed asymmetrical fluctuations in middle cerebral arterial blood flow velocity and pulsatility indicating that intracranial vasomotor changes play a role in the pathophysiology of HaNDL. These findings closely resemble those observed during attacks of migraine with or without aura (18, 19). Using single-photon emission computed tomography (SPECT) (5–7), a reduction of cortical blood flow was found in six out of eight patients on the side of origin of the neurological deficits during an episode of HaNDL with a progressive recovery of normal blood flow during the following days. The authors suggested that their SPECT findings combined with EEG data, transcranial Doppler sonography and the clinical progression of neurological symptoms were compatible with spreading depression, a mechanism that is thought to be responsible for the migrainous aura (20). Another indirect argument favouring a common pathophysiological denominator between migraine with aura and HaNDL is the absence of diffusion changes when diffusion-weighted MRI is performed during an episode of HaNDL (21). Beside migraine, viral infection has been proposed, though never proven, as the culprit in HaNDL. It was hypothesized that such an infection could produce via the immune system an aseptic inflammation of the leptomeningeal vasculature (3).

A heteroduplex analysis and DNA sequencing of the familial hemiplegic migraine (FHM1) gene CACNA1A (19p13) in 10 HaNDL patients failed to identify any sporadic mutation or shared polymorphism in the exons or the intro/exon boundaries (22). These results do thus not support a role of the CACNA1A gene in HaNDL. To our knowledge, a search for mutations in the ATP1A2 gene on chromsome 1q23, which encodes the α2 subunit of the Na⁺ K⁺ ATPase pump and is mutated in FHM2 (23), has not yet been performed.

One might argue that the electrophysiological abnormalities found in our patient could reflect her predisposition to migraine, as her mother has migraine without aura and her father migraine with aura. Contingent negative variation (CNV) amplitude and habituation abnormalities have indeed been found in non-symptomatic members ‘at risk’ in migraine families (24). This has not been reported yet for other electrophysiological tests, but similarities of PR-VEP habituation and IDAP were greater within related than non-related pairs of migraineurs (17). No data are available on SFEMG in non-affected family members of migraineurs. As far as SFEMG abnormalities are concerned, they were pronounced only in patients suffering from migraine with complex and/or prolonged auras (14, 15). Our patient's father has aura with typical and exclusive visual auras. Thus, it seems unlikely that he presents the subtype of migraine with aura that is associated with SFEMG abnormalities and that his daughter would have inherited the neuromuscular junction dysfunction at a rather severe degree without having presented migraine attacks before.

The SFEMG abnormalities reported here can be reversed by treatment with the carbonic anhydrase inhibitor acetazolamide in certain migraine with aura patients (25). This may suggest that they reflect dysfunctioning ion channels, as acetazolamide was found effective in several central channelopathies (26, 27). Lack of habituation on PR-VEP and increased IDAP, which is also due to an habituation deficit (11), could be due to a reduced preactivation level of sensory cortices, which can be reversed by high-frequency rTMS (13). The cause of the reduced preactivation excitability in migraine is not known, but it could be due to impaired monoaminergic subcortico-cortical pathways. Taking all data together, HaNDL seems to be associated between attacks, but also after its resolution, with cortical excitability changes and with incidental abnormalities of quantal release of acetylcholine at the neuromuscular junction, which suggests that it could be a subform of migraine with aura possibly due to channelopathy features. Although our patient did not suffer from migraine before or after (follow-up of 1 year) her HaNDL episodes, she has clear migrainous endophenotypic vulnerability markers. Considering the relatively high prevalence of migraine in a large HaNDL series (3), she might thus be at risk of developing a more common form of migraine over the years.