Abstract
Background
Although olfactory hallucination (OH) has been reported in patients with primary headaches, olfactory aura has not been recognised by the International Classification of Headache Disorders (ICHD-2). In this study, we examined the frequency and characteristics of OH among children and adolescents with primary headaches.
Methods
839 neurologically normal patients with primary headaches (537 migraine) were eligible for the assessment of olfactory hallucination. Headache diagnosis was based on the ICHD. Data were prospectively collected during clinic sessions and using headache diaries.
Results
Olfactory hallucination was reported exclusively during headache attacks by 21/839 (2.5%) patients, all of whom had migraine. The prevalence of olfactory hallucination was 3.9% among migraineurs (6.5% among those with migraine aura). Olfactory hallucination shortly followed the onset of headaches and lasted from 15 to 50 minutes. Of those with MA, 10 patients had visual aura; two had somatosensory aura; one had motor aura; and two had a combination of visual and somatosensory aura. Using the ICHD-2, both OH and migraine aura occurred in the same headache attacks. In 12/15 patients, OH occurred simultaneously with migraine aura, whereas in 3/12 patients, it preceded aura.
Conclusion
Our findings show that olfactory hallucination occurs in migraine and it has similarities to migraine aura.
Background
Olfactory disturbances include a variety of conditions such as parosmia, osmophobia, phantosomia, anosmia, olfactory agnosia and olfactory hypersensitivity. In parosmia there is distortion of the natural smell of an odour, which is transcribed differently, for example the patient sniffs a perfume and complains it smells of tar or smoke. Anosmia implies insensitivity to odour while olfactory agnosia indicates inability to recognise an odour sensation. Patients with osmophobia dislike smell in the presence of an odour. Hyperosmia or olfactory hypersensitivity refers to an increase in sensitivity to smell.
Olfactory hallucination (OH), termed phantosmia, refers to perceived smell in the absence of odorant stimulation, and it can be a symptom of severe head injury, brain tumour, epilepsy, drugs, infection and psychiatric illness (1–3). Olfactory hallucination has been reported among patients with primary headaches (4–19); however, olfactory symptoms continue to be omitted from the International Classification of Headache Disorders (ICHD-3) (20).
There is a paucity of data on OH in children with primary headache, and no readily available prospective studies describing it. In the paediatric population, where aura incidence and prevalence may be higher than in adults, a higher rate of OH is expected if truly an aura phenomenon.
In this study, our aim was to examine the frequency and characteristics of OH among children and adolescents with primary headaches.
Methods
Subjects and study design
Subjects comprised children and adolescents with primary headaches who were referred to the paediatric headache clinic at Queen’s University Hospital between August 2009 and July 2012. Our hospital is a busy hospital that provides a secondary care level to a population of about 200,000 under the age of 17 years. Data including headache characteristics, associated symptoms (photophobia, phonophobia, nausea, vomiting), migraine aura (visual, somatosensory, language, motor, and brainstem symptoms) and characteristics of OH were prospectively collected on initial and subsequent hospital clinics. Standard 6-monthly hospital follow-up was arranged for all patients, during which they kept their own headache diaries. Data were documented in the medical notes and kept confidentially in a database using a Microsoft Excel spreadsheet (Microsoft Windows, 2007). Data were also stored in an encrypted computer in a secure location. The study was approved by the Audit Department at Barking, Havering and Redbridge University Trust and informed consent was obtained from patients before taking part in the study. Brain MRI scans were arranged for all patients with OH. Other investigations including EEG were carried out on individual cases to exclude causes of OH.
Exclusion and inclusion criteria
Patients in whom a diagnosis of epilepsy or seizures was suspected were specifically excluded. Those with specific learning difficulties, language barriers, secondary headaches and systemic illnesses such as psychiatric disorders or neurological conditions were also disqualified. Case eligibility included: i) age 4 years or older; ii) headache history of 6 months or more; iii) five separate headache attacks; iv) at least two episodes of OH on separate headache attacks; and v) no abnormal neurological findings.
Definition
In this study, OH refers to perceived smell in the absence of odorant stimulation. Hence, the patient smells something when there is nothing in the environment producing the odour being experienced. Headache diagnoses were made on the basis of the ICHD-2. We use the term ‘frequency’ to refer to a mean number of headache attacks per month, separate headache attacks to refer to a headache free interval of at least 24 hours and chronic headache to refer to at least 15 headache days per month for 3 months or more. A family history was considered positive if a first-degree relative (father, mother, brother, sister) was affected by OH.
Olfactory hallucination
The key point was to study in detail the characteristics of OH among eligible patients with primary headaches. All patients were asked whether they had experienced any pleasant, unpleasant, usual or unusual odour during or in between headache attacks. The carers were also asked if they supported such observations. Patients who reported an odour were then asked to identify the smell and whether a possible source was present in the surrounding.
Patients with OH were asked further questions including: i) If OH has occurred before, during or after the onset of headache attacks?; ii) How long olfactory hallucination lasted?, iii) Whether OH developed or subsided suddenly or gradually?; iv) Whether olfactory stimulants triggered headache attacks? We recorded the presence of associated symptoms (photophobia, phonophobia, nausea and vomiting) and migraine aura (vision, somatosensory; motor, language). We also enquired about family history of OH and headaches.
Results
Of the 1118 patients with headaches, 279 patients were excluded from the study. Of these, there were 155 patients with less than a 6-month history of headaches; 62 patients with suspected psychiatric and social issues; 13 with abnormal soft neurological findings; 12 with significant brain abnormalities; 10 with a history of seizures; nine patients were under 4 years of age; seven had sinusitis; five had idiopathic intracranial hypertension; two had learning difficulties; two had been in road traffic accidents; and two had dysmorphic features. On initial assessment, none of the non-eligible patients reported olfactory disturbances.
Finally, 839 (75%) patients (494 females) were eligible for the assessment of OH. Their age ranged from 4 to 18.3 years (mean = 11.4 years). Mean headache history was 2.12 years and mean headache frequency was 13 per month. Headache diagnoses on the basis of ICHD-2 included migraine (n = 537), tension type headaches (n = 117) and other primary headache types (n = 47). The remaining 138 patients had not yet specified headaches. The migraine category included 230 migraine with aura (MA) and 307 migraines without aura (MO).
Clinical data and characteristics of olfactory hallucinations.
F: female; M: male; OH: olfactory hallucination; C: Caucasian; min: minutes; n: nausea; v: vomiting; pt: photophobia; phn: phonophobia; va: visual aura; sa: somatosensory aura; ma: motor aura; C: chronic; E: episodic; MA: migraine with aura; MO: migraine without aura.
Using the ICHD-2, all patients with OH had migraine (15 MA, 6 MO). Eight patients had chronic MA; seven had episodic MA; four had chronic MO; and two patients had episodic MO. In our patients, typical migraine aura as defined by the ICHD-2 included visual aura (n = 10); somatosensory aura (n = 2); motor aura (n = 1); and a combination of visual and somatosensory aura (n = 2). In 12 out of the 15 patients, OH occurred simultaneously with migraine aura, whereas in the remaining three patients, OH preceded aura.
Follow-up for a mean period of 10.8 months (range = 8–18 months) showed headache that disappeared or improved considerably in 15 patients, and was static in the remaining six patients. Patients with static or improved headaches continued to experience OH during headache attacks.
Discussion
In this study we examined data regarding OH in 21 patients. Migraine was the only associated headache type among these patients. Migraine with aura was the major linked headache as 15/21 (71.5%) patients with OH experienced aura symptoms according to the ICHD-3. We found prevalence of OH to be 3.9% among migraineurs and 6.5% among those with MA. We were unable to discover OH amongst non-migraineurs. The patients in our study described in detail the perception of abnormal smells without corresponding source of odour and without warning. Such hallucination appeared spontaneously and was associated completely with attacks of migraine. None of the patients perceived unusual smell in between headache attacks. In our patients, OH was not accompanied by any other forms of hallucination, such as tactile, gustatory or bodily, or other forms of olfactory disturbances, such as osmophobia, although it may be difficult to distinguish OH and osmophobia. None of our patients attributed the smells to body odour or perfumes. The perceived smells were characteristic, recognisable and usually unpleasant. The majority of our patients described graphically and consistently the smell of food items, fish tanks, burnt plastic, petrol, cigarettes and chlorine. Food items included rotten cheese, smelly fish, canned beef, cat food, burned sweets and vinegar. In a few cases, flowery smells and that of rotten medicine were reported. Odour triggering headaches were reported by 16% of children with migraine (21). In our patients, attacks of migraine were triggered by other precipitants that were unrelated to smell. Triggers such as stress at school, cheese, noise and inadequate fluid intake were identified in 25% of patients with OH.
So far, to the best of our knowledge, there are 40 reported cases of OH in patients with primary headaches. Therefore, the entire cohort of patients, including our cases, comprises a total of 61 cases. Others reported OH in 14 patients with primary headache (4). Although age at onset of OH was not mentioned for eight of their patients, none of their patients experienced OH during childhood or adolescence. Recently, Demarquay et al. reported recurrent unpleasant OH during migraine attacks in a 33-year-old male (5). Review of the literature also revealed another 25 cases of OH and primary headaches (6–19). In these patients, age at onset of OH was not mentioned in eight patients. Of the remaining 17 patients, there were only three with OH from an earlier age, one with OH from the age of 12 years, one from the age of 17 years and one who developed OH during childhood.
Olfactory hallucination is usually an acquired condition of central origin. It can be a symptom of other conditions (1–3). None of our patients suffered from systemic illnesses, head injury, brain tumour, cerebral aneurysm or stroke that could cause OH. In addition, our patients carried on normal daily life activities and their guardians, staff at school and the referral clinicians expressed no concerns regarding psychiatric illnesses among our patients. Sinusitis was excluded clinically and on imaging procedures. Additionally, none of these patients were on any form of medication that is known to cause olfactory disturbance. OH occurs in association with uncinate seizures in temporal lobe epilepsy but rarely as an isolated epileptic seizure. It is generally known that uncinate seizures are often accompanied by odd feelings, sensation of familiarity and peculiar taste. In this condition, perception of smell is characterised by strange unpleasant and ill-defined odour and is difficult for patients to describe. None of our patients reported such findings.
The olfactory system lacks the exactness found in other sensory systems such as vision and skin, and its topography is disguised in its first order projection. Major neurons and structures, from which fibres project to cells within the olfactory bulb, originate from many parts of the brain including locus ceruleus and raphi nuclei. It is also known that major afferent connections project from the olfactory bulb via the amygdala and insular cortex to the brainstem regions to execute autonomic functions. The olfactory tubercle, entorhinal cortex, amygdala, hippocampus and hypothalamus are associated structures to the brainstem, olfaction and limbic systems. It is also known that olfactory receptors and the trigeminal afferents are located more closely in the nasal mucosa (22). A recent fMRI study showed that olfactory stimuli increased activity in the olfactory system and the rostral part of the pons (23). A number of sites of interaction between the olfactory and the trigeminal systems have been reported (24). The overlapping of the neuro-anatomy for the olfactory system and the areas involved in the pathogenesis of migraine provide a scientific hint to consider migraine as one of the causes of OH. In our study, the close temporal relationship between OH and headache, headache associated symptoms (nausea, vomiting, photophobia and phonophobia) and typical aura symptoms support such a hypothesis and provide further clinical evidence of the interaction between the trigeminal and the olfactory systems.
There are strong data pointing to cortical spreading depression as the basic mechanism of migraine aura. This phenomenon can be implicated in migraine attacks without clinically perceived aura (25). Hence, one can argue that cortical spreading depression might be involved in the mechanism of OH in our patients (cases 9, 12, 15, 16, 20) with MO. The question regarding this hypothesis remains open for debate and certainly requires further well-designed studies.
In our view, experience of OH during migraine attacks should not be considered in the same manner as the presence of accompanying symptoms of migraine such as nausea, vomiting, photophobia, phonophobia or osmophobia. In our study, we are positive that OH was frequently concurrent with the onset of migraine attacks, occurred on more than two separate attacks of migraine and was linked with other associated symptoms. In these patients, OH lasted for more than 5 minutes and less than an hour, and was then followed by spontaneous full recovery. The most reasonable hypothesis to explain our findings is to consider OH as similar to other forms of migraine aura. One may argue that OH was uncommonly reported among patients with migraine in our study (6.5% migraine with aura). However, a large nosographic population based study of patients with migraine with aura reported motor aura in only 6% of patients (26). Another study showed that the occurrence of aura without visual aura was rare (27).
Strengths and limitations
Our study has a number of strengths. To the best of our knowledge, this is the first study to systematically address OH in a paediatric headache population. It includes a large number of healthy subjects with diagnoses based on the ICHD-3. Additionally, we had strictly defined case eligibility. Patients reported OH during headache attacks on initial assessment and during the study follow-up period. Also, to make sure certain inclusion and exclusion criteria were met, all patients were evaluated by one investigator who has a special interest in childhood headaches.
There are a few potential limitations of this study that need to be taken into account when interpreting the data. We accept that it could not be easy, specifically for children, to distinguish between different types of olfactory disturbance. It is also hard for children and adolescents to tell phantosmia apart from other forms of olfactory dysfunction such as osmophobia, and it can be difficult for children to be certain about smell as a trigger of migraine. Additionally, we could not be certain about the gradual progression and development of OH among our patients as they could be affected by recall bias. Another weak point of this study is the small sample of patients with phantosmia. For this reason we cannot comment on the significant correlation of OH and other accompanied symptoms. Despite this, we found a conclusive association of OH with migraine attacks.
In conclusion, findings derived from the present study reveal that OH occurs, although uncommonly, in patients with migraine. It is imperative to be aware of this in order to avoid misdiagnosis of other aetiologies of OH. Currently olfactory hallucination is not a recognized migraine aura and is omitted from the latest ICHD-3-beta classification. Our case series show that olfactory hallucination is not an uncommon aura in migraine in children and should be included at least in the appendix aura criteria in the final ICHD-3. There is a need for further studies to look at OH as an olfactory aura among migraineurs and examine its pathophysiological mechanism.
Clinical implication
Olfactory hallucinations are not uncommon in children with primary headache and should be considered as migraine aura.
Footnotes
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Conflict of interest
None declared.