In this issue of the Journal of Neuro-Ophthalmology, Almog et al (1) present the results of their review of 52 adults referred for outpatient or inpatient consultation to a neuro-ophthalmologist for evaluation of Horner syndrome.
They found that in two-thirds of patients, the cause of the Horner syndrome will already be known at the time of the first neuro-ophthalmic consultation (usually surgery or trauma to the head, neck, or chest, dorsolateral medullary stroke, or carotid dissection). Among one-sixth of the patients in whom the cause of Horner syndrome is not yet known, there will be clinical clues to allow localization of the lesion, such as acute neck or face pain (cervical region), arm pain or weakness (brachial plexus or paraspinal region), or sixth cranial nerve palsy (cavernous sinus); in those patients, targeted imaging will usually find the lesion. In the remaining one-sixth of patients, there will be no localizing clues for the Horner syndrome; in those patients, nontargeted imaging of the head, neck, and chest will rarely find a responsible lesion (only 1 case of thyroid carcinoma).
This “real world” study, together with previous studies, provides valuable guidance in the evaluation of Horner syndrome in adults and suggests the following approach:
Step 1. Confirm that there really is a Horner syndrome. The clinical features-ptosis and miosis-are calling cards, but they may be exceedingly subtle or transient (2). (The report of anhydrosis or absent facial flushing is helpful but rarely elicited.) Because there are other causes of miosis and ptosis (3), topical pharmacologic testing should be used-and results may be positive even if signs are equivocal or completely absent (2)! Cocaine has been the traditional agent, but it is a weak pupil dilator and, as a controlled substance, often is not readily available. It has been supplanted by topical 0.5% apraclonidine (4) (except in children younger than 1 year of age, in whom it may cause serious acute dysautonomia [5,6]), although wider experience will be needed to establish how reliable it is, especially in an acute Horner syndrome (7,8).
The traditional use of pharmacologic agents such as topical hydroxyamphetamine to assist in localization is a waste of time. These agents are difficult to obtain and do not provide information reliable enough to allow targeted imaging (9).
Step 2. Determine whether there has been previous accidental or surgical trauma to the neck, upper spine, or chest that will explain the Horner syndrome. (Include carotid endarterectomy/stenting and epidural anesthesia among legitimate causes [10,11].) If so, no further diagnostic work-up is necessary. Move to…
Step 3. Determine whether there are localizing clinical features for the Horner syndrome. For example, ataxia and nystagmus would suggest a medullary lesion. Arm pain/weakness/numbness or myelopathic features would direct attention to the lung apex, brachial plexus, and cervical spine. Acute neck or face pain would direct attention to the cervical carotid artery. (The oculosympathetic fibers crawling up the outside of the cervical carotid artery are exquisitely vulnerable to compression, trauma, dissection, and inflammation, including arteritis.) Beware of attributing persistent Horner syndrome to trigeminal autonomic syndromes such as cluster headache; carotid artery dissection can perfectly mimic these syndromes (12). Ear pain or hearing loss would direct attention to the temporal bone and carotid canal (13). Ipsilateral sixth cranial nerve palsy would direct attention to the cavernous sinus. Perform targeted (selective) imaging on the basis of this kind of information. If there are no localizing features, then the Horner syndrome is considered “isolated,” and you move to…
Step 4. Perform nontargeted (nonselective) imaging of the upper chest and neck as far up as the skull base to encompass the second (preganglionic) segment and the extracranial part of the third (postganglionic) segment of the oculosympathetic pathway. If the Horner syndrome is truly “isolated,” imaging need not extend above the skull base, as a cavernous sinus or orbit lesion would be extremely unlikely to be the cause. Anticipate that the yield of a causative lesion will be low (many such cases remain unsolved) but, as Almog et al (1) point out, even if 1 tumor is found, the gesture may be life-saving.
What kind of imaging should be performed? Because carotid artery dissection is such a common cause (even in the absence of pain), a vascular study must be included. To rule out nonvascular masses of the neck or chest, soft-tissue imaging must also be there. Although MRI/MRA has often been recommended in this setting (11,14), my neuroradiologic colleagues advise CT/CT angiography (CTA). CT provides adequate resolution of soft tissue masses, and CTA provides good pictures of the carotid artery lumen (15). Narrowing of the lumen is likely to be the most important predictor of stroke. MRI/MRA does have the advantage of showing not only the vascular lumen but also its wall. When the patient lies perfectly still, the wall hematoma can be beautifully visualized on fat-saturated T1 axial MRI. But MRI is often difficult to obtain promptly, relatively expensive, and readily degraded by patient motion.
How urgent is imaging? It is not urgent unless the patient reports recent ipsilateral neck trauma, neck/face pain, ipsilateral transient monocular visual loss, or contralateral transient weakness or numbness, which suggest acute cervical carotid dissection. Within the first 2 weeks after onset of Horner syndrome, there is a substantial risk of hemispheric (middle cerebral artery distribution) stroke (16), which may be attenuated by aspirin or anticoagulation treatment (although there are no rigorous studies to prove that point). If the Horner syndrome has been present for more than 4 weeks, the threat of stroke is much less, but imaging should not be unduly delayed.
In children with clinical features of Horner syndrome, topical cocaine should be used in preference to apraclonidine to confirm the diagnosis if the child is younger than 1 year old (5). Neither heterochromia nor a history of birth trauma entirely excludes the possibility of a causative mass in the chest or neck, usually a neuroblastoma, although the chance of negative results is high (17,18). Urine catecholamine metabolite studies, which have traditionally been performed in the investigation of neuroblastoma, are not sufficiently sensitive to that diagnosis (19). MRI (rather than CT) of the neck and chest must be performed. There may be an additional yield from I-123 metaiodobenzylguanidine (MIBG) scintigraphy, which can “light up” small tumor foci beyond the detection of MRI (20).
1. Almog Y
, Gepstein R, Kesler A. Diagnostic value of imaging in Horner syndrome in adults. J Neuroophthalmol
2. Leira EC
, Bendixen BH, Kardon RH, et al. Brief, transient Horner's syndrome can be a hallmark of carotid artery dissection. Neurology
3. Thompson BM
, Corbett JJ, Kline LB, et al. Pseudo-Horner's syndrome. Arch Neurol
4. Koc F
, Kavuncu S, Kansu T, et al. The sensitivity and specificity of 0.5% apraclonidine in the diagnosis of oculosympathetic paresis. Br J Ophthalmology
5. Mughal M
, Longmuir R. Current pharmacologic testing for Horner syndrome. Curr Neurol Neurosci Rep
6. Lebas M
, Seror J, Debroucker T. Positive apraclonidine test 36 hours after acute onset of Horner syndrome in dorsolateral pontomedullary stroke. J Neuroophthalmol
7. Dewan MA
, Harrison AR, Lee MS. False-negative apraclonidine testing in acute Horner syndrome. Can J Ophthalmol
8. Kardon R
. Are we ready to replace cocaine with apraclonidine in the pharmacologic diagnosis of Horner syndrome? J Neuroophthalmol
9. Maloney WF
, Younge BR, Moyer NJ. Evaluation of the causes and accuracy of pharmacologic localization in Horner's syndrome. Am J Ophthalmol
10. Lee AG
, Kawasaki A. Horner syndrome [MedLink Neurology Web site]. Available at: www.medlink.com
. Accessed December 26, 2009.
11. Kedar S
, Biousse V, Newman NJ. Horner's syndrome [MedLink Neurology Web site]. www.utdol.com
. Accessed December 26, 2009.
12. Rigamonti A
, Iurlaro S, Reganati P, et al. Cluster headache and internal carotid artery dissection: two cases and review of the literature. Headache
13. Spector RH
. Postganglionic Horner syndrome in three patients with coincident middle ear infection. J Neuroophthalmol
14. Reede DL
, Garcon E, Smoker WR, et al. Horner's syndrome: clinical radiographic evaluation. Neuroimaging Clin N Am
15. Walton KA
, Buono LM. Horner syndrome. Curr Opin Ophthalmol
16. Biousse V
, D'Anglejan-Chatillon J, Touboul PJ, et al. Time course of symptoms in extracranial carotid artery dissection. A series of 80 patients. Stroke
17. George ND
, Gonzalez G, Hoyt CS. Does Horner's syndrome in infancy require investigation? Br J Ophthalmol
18. Jeffery AR
, Ellis FJ, Repka MX, et al. Pediatric Horner syndrome. J AAPOS
19. Mahoney NR
, Liu GT, Menacker SJ, et al. Pediatric Horner syndrome: etiologies and roles of imaging and urine studies to detect neuroblastoma and other responsible mass lesions. Am J Ophthalmol
20. Papaioannou G
, McHugh K. Neuroblastoma in childhood: review and radiological findings. Cancer Imaging