Abstract
Background
Head pain is a cardinal feature of primary headache disorders (PHDs) and is often accompanied by autonomic and vasomotor symptoms and/or signs. Spontaneous extracranial hemorrhagic phenomena (SEHP), including epistaxis, ecchymosis, and hematohidrosis (a disorder of bleeding through sweat glands), are poorly characterized features of PHDs.
Aim
To critically appraise the association between SEHP and PHDs by systematically reviewing and pooling all reports of SEHP associated with headaches.
Methods
Advanced searches using the PubMed/MEDLINE, Web of Science, Cochrane Library, Google Scholar, and ResearchGate databases were carried out for clinical studies by combining the terms “headache AND ecchymosis”, “headache AND epistaxis”, and “headache AND hematohidrosis” spanning all medical literature prior to October 10, 2015. Preferred Reporting Items for Systematic Reviews and Meta-analyses and Meta-analysis of Observational Studies in Epidemiology guidelines were applied.
Results
A total of 105 cases of SEHP associated with PHDs (83% migraine and 17% trigeminal autonomic cephalgias) were identified (median age 27 years, male to female ratio 1:2.3); 63% had epistaxis, 33% ecchymosis, and 4% hematohidrosis. Eighty-three percent of studies applied the International Classification of Headache Disorders diagnostic criteria. Eighty percent of the reported headaches were episodic and 20% were chronic. Twenty-four percent of studies reported recurrent episodes of SEHP.
Conclusions
Our results suggest that SEHP may be rare features of PHDs. Future studies would benefit from the systematic characterization of these phenomena.
Keywords
Introduction
Head pain is commonly the chief symptom reported by patients; epidemiological studies estimate a lifetime prevalence of head pain of 64% and a one-year prevalence of 46% (1,2). Head pain is a cardinal feature of primary headache disorders (PHDs), including migraine, tension-type headaches, and trigeminal autonomic cephalgias (TACs) (3). The pain may be associated with autonomic and vasomotor features, such as lacrimation and conjunctival injection (3). These cranial autonomic symptoms (CAS) are thought to result from activation of the trigemino-autonomic reflex in which nociceptive stimuli trigger parasympathetic ganglia, resulting in parasympathetic overflow to the head and orofacial vasculature (4). The frequency of CAS in PHDs ranges from 14 to 100% (5–9). Up to 49% of patients with migraine report CAS (9,10). CAS are a definitive feature in patients with TACs, although the occurrence of specific CAS is variable.
Spontaneous extracranial hemorrhagic phenomena (SEHP) as a feature of PHDs are poorly characterized and rarely recognized (11). This bleeding can take the form of blood extravasation through nasal mucosa (termed epistaxis (12)), through subcutaneous tissue (termed ecchymosis when the lesions are \gt1 cm (13,14)), and through sweat glands in hematohidrosis (15). To date, no study has systematically evaluated the occurrence of SEHP in the setting of PHDs.
We conducted a systematic review and pooled analysis of all published case studies of SEHP in association with PHDs described prior to October 10, 2015. We characterized and critically appraised SEHP as a feature of PHDs and hypothesized that SEHP, specifically epistaxis, ecchymosis, and hematohidrosis, are features of PHDs and should be included among the autonomic and vasomotor changes seen in TACs and migraine.
Methods
The following seven strategies were used to capture our topic of interest.
A PubMed/MEDLINE search was carried out for Clinical Studies Categories on the PubMed Clinical Queries tool combining the terms “headache AND ecchymosis/epistaxis/hematohidrosis”. The Clinical Studies Category was selected for “Therapy”, “Etiology”, “Diagnosis”, “Prognosis”, “Clinical Prediction Guides”, and the scope of the search was made specific to “Broad” to enable sensitivity and specificity search values of 99 and 70% (16), respectively. A PubMed/MEDLINE search was carried out for Clinical Studies Categories and Systematic Reviews on the PubMed Clinical Queries Tool combining the terms “Headache[Title] AND Ecchymosis[Title]”, “Headache[Title] AND Epistaxis[Title]”, “Headache[Title] AND Hematohidrosis[Title]”. The Clinical Studies Category was selected for “Therapy” and the scope of search was made specific to “Narrow” to enable sensitivity and specificity search values of 93 and 97% (16), respectively. PubMed/MEDLINE search was carried out without using the Clinical Queries Tool. Search terms utilized were “Headache[Title/Abstract] AND Ecchymosis/Epistaxis/Hematohidrosis[Title/Abstract]”. An Advanced PubMed/MEDLINE search was used by implementing auto-suggested Medical Subject Headings (MeSH Terms) and the Boolean logic operator “AND” as “(headache[MeSH Terms]) AND ecchymosis/epistaxis/hematohidrosis[MeSH Terms]”. The Cochrane Library Advanced Search strategy was used by adding the search line “Search All Text” and using the search terms “headache AND ecchymosis/epistaxis/hematohidrosis”. A Web of Science Advanced Search was carried out using the field tag “TS” for topic, the Boolean operator “AND” and parentheses to create our query as “TS=(headache AND ecchymosis), (headache AND epistaxis), (headache AND hematohidrosis)” on Indexes= CPCI (Conference Proceedings Citation Index), Science Citation Index Expanded (SCI-EXPANDED), Timespan=All years. The results included all languages and all document types. An unpublished studies and relevant reference search was exhaustively conducted using Google Scholar and ResearchGate. This enabled us to capture unpublished studies from the gray literature (e.g. conference abstracts or research letters) with the goal of avoiding publication bias.
All published studies up to October 10, 2015 were covered. Two authors (AMP and YWW) reviewed each article and, where disagreement occurred, discussion and consensus were achieved with input from the third author (RPC). The methods were in accordance with the Meta-analysis of Observational Studies in Epidemiology (MOOSE) guidelines (17) and the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) flow diagram (18) (see Figure 1).
PRISMA (18) flowchart used for selection of reviewed studies.
Inclusion and exclusion criteria
Following identification, screening, and eligibility, the inclusion criteria were set to clinical studies that focused on headache and SEHP. The studies excluded at screening consisted of clinical studies of hemorrhage of no relevance to PHDs; these included traumatic forms of hemorrhage and hemorrhage resulting from other bleeding disorders such as hereditary hemorrhagic telangiectasias (19). Other forms of hemorrhage, including intraocular hemorrhage and menstrual-related hemorrhagic phenomena such as menorrhagia, were also excluded. The paper by Hedges (20) describing a personal communication with H. Naquin detailing over 100 patients with hemorrhagic phenomena occurring in association with migraine, including subconjunctival, lid, and retinal hemorrhage as well as epistaxis, was excluded from this review due to the lack of pertinent clinical data detailing the type of hemorrhage or linking the occurrence of these hemorrhages temporally with PHDs. The epistaxis–migraine cohort within the paper by Ikonomoff et al. (21) was included in the total number of patients, but was excluded from the detailed analysis as no further pertinent clinical data was available.
Data extraction and pooled analysis
The data extracted were: first author, age, gender, headache description, headache onset to onset of ecchymosis/epistaxis/hematohidrosis, headache frequency at presentation, hours of headache prior to ecchymosis/epistaxis/hematohidrosis formation, sites of ecchymosis/epistaxis/hematohidrosis, time to ecchymosis/epistaxis/hematohidrosis recovery, recurrence of ecchymosis/epistaxis/hematohidrosis, medical history, physical examination, medications, International Classification of Headache Disorders (ICHD) diagnosis, family history, treatments, laboratory investigations, imaging, biopsy, duration of follow-up, and prognosis were extracted from each case study. Considering the method’s robustness to outlier data, the median and interquartile range (IQR) were selected to compare age distribution. Data were pooled to obtain a summary from the different but related studies. Percentages were used to describe the proportions of results.
Results
A total of 34 papers (17 papers on ecchymoses (21–37), 14 papers on epistaxis (12,19,20,21,38–47), and three papers on hematohidrosis (15,48,49)) were included within the final systematic review for qualitative and quantitative analysis (Figure 1). Thirty-two papers were in English, one paper was in German (22), and one was in French (21). The articles in German and French were translated into English.
Epidemiology
Little is known about the true incidence of hemorrhagic phenomena in PHDs because no population-based epidemiologic study is available. Case reports and case series have been reported, providing a crude estimate of its incidence. The low number of reported cases suggests that it is an uncommonly reported feature of PHDs. A total of 105 case studies were identified over the past 120 years: 58 cases of epistaxis, 41 cases of ecchymosis, five cases of hematohidrosis, and one case of simultaneous ecchymosis, epistaxis, and subconjunctival hemorrhage in association with PHDs.
Journals of publication
The journals publishing reports of cases of SEHP in relation to PHDs spanned a wide range of specialties, suggesting that these patients present to different specialists. Thirty-one percent of case reports were published in journals of headache medicine, 22% in neurology and psychiatry, 16% in internal medicine, and 7% in ophthalmology, with the remaining 24% including journals of pain medicine, dermatology, neuro-ophthalomology, pharmacology, hematology, pediatrics, and otolaryngology.
Demographics
The median age of presentation was 27 years (IQR 21–47): this was 26 years in the epistaxis, 46 years in the ecchymosis, and 26 years in the hematohidrosis cohorts. The overall male to female ratio was 1:2.3. This was 1:1.4 in the epistaxis, 1:3.2 in the ecchymosis, and 0:5 in the hematohidrosis cohorts.
Primary headache types
Among the cases reported after the initial publication of the ICHD, 88% applied the ICHD criteria; 7% of case reports were published before publication of the ICHD. A total of 83% of all case reports applied the ICHD diagnostic criteria. The PHDs reported were migraine (81%), TACs (18%), and a description of “tension-type” (not using ICHD criteria) headache in one case report. Epistaxis was seen in 67% of patients with migraine and 13% of patients with TACs. Ecchymosis was seen in 87% of patients with TACs and 33% of patients with migraine (Figure 2). The headache types were not classified in the hematohidrosis cohort, with the exception of one patient with “tension-type” headache. Among the patients with migraine, 50% had migraine without aura, 23% had migraine with aura, 1% had abdominal migraine, and 12% of studies did not classify the type of migraine. Among the patients with TAC, 80% had cluster headache, while 12.5% had chronic paroxysmal headache, and the remaining 6% had hemicrania continua. The frequency of all types of headaches was episodic in 80% of patients and chronic in 20% of patients.
Number of reported cases of SEHP among specific primary headache disorders based on patient age.
Patient characteristics
Information regarding preexisting medical disorders and prescribed drugs was provided in 47% of case (33% of patients with TACs and 48% of patients with migraine). Among those providing prior medical history, 10% reported hypertension (6% among patients with migraine, 40% among patients with TACs), three patients reported chronic dermatological disorders (discoid lupus erythematoses, psoriasis, and sarcoidosis), none reported coagulation disorders, and none reported cardiac disease. Other past medical histories reported included dysmenorrhea, depression, insomnia, hypothyroidism, hyperlipidemia, hyperthyroidism, and diabetes.
Among the patients specifying the prescribed drugs taken, the use of non-steroidal anti-inflammatory drugs (NSAIDs) was reported by 22% of patients with TACs and 5% of patients with migraine; the use of triptans was reported among 6% of patients with migraine, whereas no patient with TAC reported triptan use. Antiepileptic drugs (specifically topiramate) were taken by 1% of patients with TACs and 3% of patients with migraine. Other drugs taken included acetazolamide (3%), acetaminophen (8%), flunarizine (8%), intravenous corticosteroids (3%), propranolol (8%), and hydrocodone (3%). Among the ecchymosis cohort, 18% reported the use of NSAIDs, 18% triptan, and 10% antiepileptic drugs. Among the epistaxis cohort, 3% reported the use of NSAIDs, none reported the use of triptan, and 3% reported the use of antiepileptic drugs. Among the hematohidrosis cohort, none reported the use of NSAIDs, triptan, nor antiepileptic drugs; 50% of patients with hematohidrosis were taking β blockers (propranolol). Overall, 28% of patients were taking some kind of prescribed drugs at the time of presentation.
Hemorrhage characteristics
The median time from first headache to initial hemorrhage was 1.5 years (range 0–3 years). The median number of hours between the start of pain and the SEHP was one hour in patients with epistaxis, 21 hours in patients with ecchymosis, and 0.5 hours in patients with hematohidrosis (Figure 3(a)). Thirty-two percent of all patients with epistaxis and PHD reported the relief of their headache following epistaxis.
(A) Time from onset of pain to SEHP based on type of hemorrhage. The time from headache onset to SEHP ranged from simultaneous with headache onset to up to one week after headache onset. Of the four cases of trigeminal autonomic cephalgias reporting time to ecchymosis, all reported onset of ecchymosis within minutes of headache onset. (b) Sites of ecchymoses. Among ecchymosis–migraine cohort, bleeding was ipsilateral to the site of pain in 67% and bilateral in 33% of reported cases. Among the trigeminal autonomic cephalgias cohort, ecchymosis was ipsilateral to the site of pain in 100% of reported cases. Other sites included periauricular, glabella, buttocks, and legs. (c) Laterality of SEHP. Among the epistaxis–migraine cohort, 53% had a nose bleed ispilateral to the site of head pain and 47% had bilateral epistaxis. In two reported cases of epistaxis and trigeminal autonomic cephalgia, bleeding was ipsilateral to the site of head pain. In the hematohidrosis cohort, the distribution of bleeding included bilateral nasal mucocutaneous, bilateral hemolarcira, forehead, bilateral hands, nails, umbilicus, and wrists and legs.
Severity of pain with SEHP
Among the cases in which headache intensity was reported, 80% of the ecchymosis cohort and 100% of the hematohidrosis cohort described severe pain during the episodes of headache. During headaches without ecchymosis, their head pain was described as mild (22,50).
Site of SEHP
Sites for spontaneous ecchymosis included bilateral periorbital (11%), temporal (11%), conjunctival (8%), buccal (8%), preauricular (5.5%), palpebral (2.8%), glabellar (2.8%), and suborbital (2.8%) (Figure 3(b)). The distribution of bleeding in the hematohidrosis cohort included bilateral nasal mucocutanoeus (5.6%), bilateral hemolacria (lacrimal ducts) (11%), forehead (22%), bilateral hands (5.6%), nails (5.6%), umblicus (17%), ear canal (17%), nasal bridge (5.6%), neck (5.6%), and wrists and leg (5.6%). Bleeding in the epistaxis cohort was either from one or both nares.
Laterality of SEHP
Among patients with migraine, the distribution of ecchymosis was ipsilateral to the side of head pain in 67% and bilateral in 33%. Among patients with TACs, the ecchymosis was always ipsilateral (Figure 3(c)). Among the epistaxis–migraine cohort, 53% had nose bleeding ipsilateral to the site of head pain, whereas this was bilateral in the remaining 47% (Figure 3(c)). In the two reported cases of TACs with epistaxis, bleeding was ipsilateral to the site of head pain. Any laterality of hemorrhagic phenomena was not specified in the hematohidrosis cohort.
Associated features
Vomiting in association with SEHP was reported in 13% of the reviewed case reports of ecchymosis, in 2.5% of case reports of epistaxis, and none in case reports of hematohidrosis. Of the patients who had CAS occurring in association with SEHP, the presentations included lacrimation (43%), ptosis (14%), eye redness (14%), facial flushing (14%), and nasal congestion (14%).
Ancillary studies
To further elucidate the underlying diagnosis, various imaging modalities were explored and revealed incidental findings such as widespread hyper-intensities on T2-weighted MRI. A detailed description of imaging is available in the Supplementary Material.
Biopsy
A skin biopsy of the ecchymotic lesions was obtained in four patients (24,25,30,35). Biopsy results in three of the four case reports were consistent with ecchymoses, with extensive red blood cell extravasation in the absence of vascular damage (24,25,30,35). No underlying vasculitic or amyloid change was noted (24,25,30,35). In one 69-year-old man with cluster headaches who developed an ecchymotic lesion over his left eyelid, the biopsy results showed telangiectatic blood vessels in the papillary dermis, some with perivascular infiltrates of lymphocytes and plasma cells (35). These findings were consistent with a capillary malformation or port wine stain (PWS) (35). A skin biopsy was obtained in one patient with hematohidrosis and was normal (15). No skin biopsy was obtained from the epistaxis cohort.
Treatment and prognosis
Successful drug treatments.
Among the migraine cohort, successful treatments included paracetamol, domperidone, bromide, phenazon, propranolol, sumatriptan, and flunazarine. Among the trigeminal autonomic cephalgias cohort, successful treatments included topamax, hydrocodone, sumatriptan and oxygen.
NSAIDs: non-steroidal anti-inflammatory drugs.
Discussion
The mechanism by which SEHP occurs in PHDs is unknown. With spontaneous hemorrhagic phenomena occurring commonly in the general population (up to 60% of patients develop epistaxis in their lifetime (51–53), although estimates of the prevalence of spontaneous ecchymosis and hematohidrosis are not available) and PHDs occurring in 11% of the population for patients with migraine (54) and 1% for patients with TACs (55), the simultaneous or near-simultaneous occurrence of SEHP and PHDs may be one of chance alone. However, it is also possible that SEHP is a rare feature of PHDs with a shared pathophysiology. The relative infrequency of SEHPs associated with PHDs compared with other features of PHDs such as photophobia and CAS suggests that patients with SEHP may also have a specific anatomical, biochemical, or neurogenic susceptibility to developing hemorrhagic phenomena.
Multiple hypotheses have been proposed to account for the pathogenesis of SEHP in PHDs. We propose that the formation of SEHP in PHDs occurs by a similar mechanism to CAS through activation of the trigemino-autonomic reflex and sterile neurogenic inflammation (4,10,56). The observed pattern of SEHP in PHDs lends further support to the hypothesis that SEHP occurs by a mechanism similar to the formation of CAS in PHDs. Cranial autonomic features tend to be bilateral in migraine, although they can be unilateral, unlike in TACs, in which these features are almost invariably unilateral and ipsilateral to the painful side (5,8,9). Our study corroborates these known characteristics of the CAS; all SEHP among patients with TACs were reported as unilateral, whereas among patients with migraine, the SEHP were either unilateral or bilateral.
With regards to specific hemorrhagic phenomena, epistaxis may result from activation of the trigeminovascular system and subsequent vasodilation of local nasal mucosal vessels, specifically Kiesselbach’s plexus (42). Similarly, perioribtal and facial ecchymosis may result from local vascular processes which precipitate the extravasation of blood into adjacent subcutaneous tissue. The case report of Kakisaka et al. (28) of a young girl with abdominal migraine who developed ecchymoses remote from her visceral pain further highlighted the role of PHDs in producing hemorrhagic phenomena. In a rodent model for visceral pain, insults to pelvic organs resulted in the extravasation of blood into skin regions innervated by the affected neurons (57). We hypothesize that the underlying pathophysiologic processes occurring in PHDs may, through the activation of nerves and the release of neuropeptides and neurotransmitters, facilitate hemorrhage remote from the site of pain.
Intraocular hemorrhage and the formation of a PWS were identified as possible complications of SEHP associated with PHDs. Intraocular hemorrhage have been reported in relation to PHDs and may share underlying pathophysiologic mechanisms with other forms of extracranial hemorrhage (20,68). Acquired forms of PWS are thought to result from decreased autonomic innervation to the lesion site leading to neuronal injury and loss and the subsequent formation of ectatic papillary dermal capillaries and postcapillary venules (35). Vagefi et al. (35) proposed that repeated disturbances of the cranial autonomic nervous system, as occurs in cluster headaches, predisposed the patient in their case report to the development of a PWS. If this is truly a complication in patients with PHDs and SEHP, it highlights the importance of managing headaches to minimize the risk of complications.
Differential diagnoses
Differential diagnosis for spontaneous extracranial hemorrhagic phenomena associated with headaches.
NSAIDs: non-steroidal anti-inflammatory drugs; SSRIs: selective serotonin reuptake inhibitors.
Vomiting can precipitate SEHP by transmitting increased intracranial pressure to adjacent spaces (including the periorbital region and nasal mucosa). In the case study by Al-Sardar et al. (23), the occurrence of ecchymosis was attributed to vomiting; however, the patient’s history of migraine occurring in correlation with each episode of ecchymosis renders a true determination of the etiology of ecchymosis, either related to valsalva maneuver as opposed to a feature of PHDs, difficult to determine. As PHDs are commonly associated with vomiting, the occurrence of SEHP as a result of this valsalva maneuver is an important diagnostic consideration. However, if the SEHP is induced by vomiting, the distribution of bleeding would be expected to be bilateral without a clear predilection for any particular side. In this systematic review, bilateral ecchymoses were only seen in 33% of patients with migraine, making a pathophysiologic mechanism beyond increased intracranial pressure from valsalva likely to be responsible for SEHP formation.
A careful review of drugs should be conducted by the clinician to determine if any of these predispose patients to the development of SEHP. Hematological adverse events have been reported with a number of prescribed drugs. The most commonly cited drugs causing bruising include anticoagulant drugs (59), antiplatelet drugs (60), and glucocorticoids (61). Antidepressants, specifically selective serotonin reuptake inhibitors (62), antibiotics, and migraine-preventive drugs (most commonly topiramate (63,64)) and chemotherapeutic drugs, although less recognized, can also precipitate SEHP.
Limitations
This systematic review included only 105 case reports, limiting the generalization of the results. The frequency of SEHP associated with PHDs may not be accurately reflected in published work as a result of ascertainment bias, with clinicians not routinely assessing for SEHP. In addition, the lack of a standardized approach to the case reports makes conclusions about a true correlation between SEHP and PHDs, the associated symptoms, appropriate work-up, and prognostication difficult. Although 50% of patients reportedly experienced the spontaneous resolution of SEHP, the duration of follow-up was variable, ranging from three weeks to 27 years, therefore an assessment of the recurrence and duration of SEHP is limited.
Conclusions
Based on our systematic review and pooled analysis of all case reports of SEHP in PHDs, we concluded that SEHP may occur as a feature of PHDs. The clinician should carefully assess for secondary causes of headache and hemorrhage given the consequences of failing to identify more serious diagnoses. Utilizing clinical data in conjunction with laboratory and imaging studies, secondary headaches can reasonably be excluded, making SEHP as a feature of PHDs a relevant and important diagnostic consideration.
Footnotes
Clinical implications
Spontaneous extracranial hemorrhagic phenomena, including epistaxis, ecchymosis, and hematohidorsis, may be features of primary headache disorders.
As a feature of primary headache disorders, hemorrhagic phenomena tend to occur in the setting of severe head pain, are strictly unilateral and ipsilateral to the pain in trigeminal autonomic cephalgias, and can occur unilaterally or bilaterally in migraine.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Acknowledgments
AMP and YWW were responsible for the conception, design, data acquisition, analysis, and interpretation of data, drafting the article, and revising it critically for important intellectual content. AMP was responsible for revising the draft article critically for important intellectual content. RCP was responsible for the conception, design, and interpretation of data, and revising the draft article critically for important intellectual content.