Pretutorial
    Assessment

Tutorial
Introduction
Anatomy, Physiology, and Pharmacology of the Pupillary Light and Near Reflexes
Anatomy and Physiology of the Sympathetic System
Pharmacology of the Sympathetic Nervous System
Clinical Abnormalities of Pupillary Function
Abnormalities of the Afferent Pupillary Pathway
Abnormalities of the Efferent Pupillary Pathway
Abnormalities of the Sympathetic Pathway (Horner Syndrome)
References

Slides

The Pupil

Joel M. Weinstein, MD

Introduction

The size of the pupillary aperture is controlled by two opposing smooth muscles, the dilator and the sphincter. The dilator muscle originates at the iris root and lies within the cytoplasm of the anterior portion of the anterior layer of the iris pigment epithelium. This radially oriented muscle terminates approximately 2 mm from the pupillary margin. Shortening of the radially oriented dilator muscle widens the pupillary aperture. The iris sphincter muscle occupies the area within 2 mm to 3 mm of the pupillary margin. It lies superficially in the iris stroma, running in a circumferential fashion around the pupillary aperture. Contraction of the circular sphincter muscle in a "coil-spring" fashion serves to narrow the pupillary aperture.

The sphincter muscle changes its tonus in response to two types of physiologic stimuli - changes in the level of ambient retinal illumination (the light reflex) and changes in viewing distance as part of the near triad (the near reflex), which consists of accommodation, convergence, and miosis. The tonus of the dilator muscle, on the other hand, increases in response to activity of the sympathetic nervous system. This may occur in response to increased outflow from the central sympathetic center in the hypothalamus and increased levels of circulating catecholamines from the adrenal glands.

In this tutorial, the details of the neuroanatomy and pharmacology of the sympathetic and parasympathetic pathways to the pupil will be reviewed. In addition, the clinical disorders that can alter pupillary size and reactivity by affecting each step of the anatomic/pharmacologic chain will be discussed. The key to the correct diagnosis of pupillary disorders is a two-step process: (1) localizing the neuroanatomic site of the lesion using appropriate clinical observations and pharmacologic tools, and (2) determining which disease is the most likely problem at the neuroanatomic site in question.

Anatomy, Physiology, and Pharmacology of the Pupillary Light and Near Reflexes

The anatomy of the pupillary light reflex is illustrated in Slide 1. The pupillary light reflex begins in the retina with transduction of light energy by photoreceptors. The photoreceptors synapse with bipolar cells, which communicate with each other via horizontal and amacrine cells. The bipolar cells then synapse with ganglion cells. Axons of ganglion cells that subserve the pupillary reflex, similar to ganglion cells that subserve vision, undergo a hemidecussation in the optic chiasm (Slide 1). In the optic tract, some ganglion cell axons appear to bifurcate, projecting both to the lateral geniculate body (subserving vision) and to the pretectum (subserving the pupillary light reflex).¹ These axons leave the posterior optic tract, enter the pretectal area of the midbrain, and synapse in the pretectal olivary and sublentiform nuclei.² Each pretectal nucleus is an analogue of the lateral geniculate body in that it receives light input from the opposite hemifield, just as each lateral geniculate nucleus subserves vision from the opposite hemifield. The pretectal nuclei distribute input equally between the two Edinger-Westphal nuclei. The result of this distribution is that unequal input arising from lesions of one optic nerve or one optic tract does not result in anisocoria, that is, the afferent input, regardless of its source, is equally distributed to both third nerves. The Edinger-Westphal nucleus is the autonomic subnucleus of the third nerve and gives rise to pupillomotor and ciliary muscle innervation. The preganglionic fibers to the pupil follow the third nerve into the orbit and synapse in the ciliary ganglion. Postganglionic fibers follow the short posterior ciliary nerves and run in the suprachoroidal space to innervate the pupillary sphincter. Further details of the pathway from the third nerve to the pupil are discussed below.

SLIDE 1

The neuroanatomic pathway for the near reflex is less well characterized. It is believed to involve a closed feedback loop, which begins with estimation of target distance, by visual association areas within the occipital and posterior parietal lobes. This information is relayed to one or more near reflex centers somewhere in the rostral midbrain that instruct the third nerve nuclei to increase or decrease the tonus of the intrinsic and extrinsic muscles of the eye that regulate accommodation, convergence, and miosis.³ Continuous feedback to the visual cortex serves to complete the closed loop and fine-tune the reflex.

The cholinergic innervation of the iris sphincter consists of three components - a parasympathetic motor nerve ending; a myonerual junction consisting of pre- and postsynaptic components with postsynaptic muscarinic receptors; and iris smooth muscle (sphincter). Action potentials are initiated via release of acetylcholine, which is stored in vesicles of all presynaptic cholinergic nerve terminals. The contents of these vesicles enter the synaptic cleft, causing stimulation of the postsynaptic cholinergic receptors. Termination of a parasympathetic action potential is accomplished when acetylcholine is hydrolyzed by the enzyme acetylcholinesterase.

Pharmacologic agents that affect the sphincter muscle can be divided into several categories: direct-acting muscarinic agonists, such as acetylcholine and pilocarpine; indirect-acting muscarinic agonists (cholinesterase inhibitors); and muscarinic antagonists (e.g., atropine, homatropine). The properties of each class of drugs follow.

Direct-Acting Muscarinic Agonists
Acetylcholine is not used topically but can be a useful adjunct to intraocular surgery. When acetylcholine in a concentration of 1:1,000 is instilled intracamerally, prompt miosis to approximately 2 mm results. The miotic action lasts approximately 10 minutes.

Pilocarpine acts directly on muscarinic receptors in the pupillary sphincter muscle. Pilocarpine is invaluable in the treatment of glaucoma, and from the standpoint of pupillary disorders, it is the drug of choice for the diagnosis of tonic pupils. The systemic adverse effects of topical pilocarpine are related to stimulation of other systemic muscarinic receptors. These effects include lacrimation, salivation, sweating, vomiting, and diarrhea. These side effects are uncommon in the doses used either in pupillary drug testing or in the treatment of glaucoma.

Methacholine bromide is similar to acetylcholine in its pharmacologic activity. This drug has been used topically in the diagnosis of Adie tonic pupil and in the diagnosis of familial dysautonomia. Methacholine bromide is not generally available for ophthalmic use and has been demonstrated to be less effective than dilute pilocarpine in the diagnosis of tonic pupils. For this reason, methacholine bromide is rarely used for the diagnosis of pupillary disorders.

Indirect-Acting Muscarinic Agonists
The most commonly encountered indirect acting muscarinic agents are diisopropyl fluorophosphates (DFP), Echothiophate (phospholine), and neostigmine. These indirect agonists produce ocular effects that are similar to those of direct-acting agents, although their mechanism of action is different. These medications prevent the degradation of acetylcholine by the enzyme acetylcholinesterase, thereby producing miosis, ciliary muscle contraction, and ocular hypotension. These agents have been useful in the treatment of glaucoma, accommodative esotropia with a high AC/A ratio, and louse blepharitis.

Muscarinic Antagonists
Muscarinic antagonists compete with acetylcholine for muscarinic receptor sites on the sphincter muscle. Muscarinic antagonists are usually subdivided into naturally occurring agents, such as atropine or scopolamine, and synthetic agents, such as cyclopentolate, homatropine, and tropicamide. Of these agents, atropine is the most potent and long lasting cycloplegic, requiring an average of 12 days for full recovery. The cycloplegic effect of cyclopentolate is superior to both tropicamide and homatropine and has the advantage over atropine of being short acting. Because the antimuscarinic activity of these medications is competitive, it can be reversed, at least in part, by muscarinic agonists.

Anatomy and Physiology of the Sympathetic System

The iris dilator muscle is controlled by the sympathetic nervous system. The tonus of the dilator muscle is, for all practical purposes, independent of the light and near stimuli. The tonus of the dilator muscle changes in response to the level of circulating catecholamines and discharge from the sympathetic nervous system. The sympathetic pathway to the eye is illustrated in Slide 2. The first neuron of this three-neuron chain begins in the posterolateral hypothalamus on the same side as the dilator muscle. The first-order, or "central," neuron descends through the brain stem and synapses in the intermediolateral gray matter of the lower cervical and upper thoracic spinal cord (the ciliospinal center of Budge).

SLIDE 2

The second-order, or "preganglionic," neuron is situated in the intermediolateral gray and gives rise to an axon that leaves the spinal cord with the ventral roots between C8 and T2. These fibers pass over the apex of the lung and ascend in the cervical sympathetic plexus, passing through the stellate (inferior cervical) ganglion and finally synapsing in the superior cervical ganglion at the level of the angle of the jaw.

The third order, or "postganglionic," neuron originates in the superior cervical ganglion and follows the neural plexus along the internal carotid artery. Within the cavernous sinus, the sympathetic fibers pass close to the sixth nerve (perhaps joining it briefly in some patients) and enter the orbit along with the nasociliary branch of V_I. Fibers to the dilator muscle enter the globe with the long ciliary nerves. Postganglionic sympathetic fibers regulating facial sweating and vasoconstriction also originate in the superior cervical ganglion but follow the external carotid artery and its branches to the face, except for a small patch on the forehead, which in some patients may be supplied by fibers that follow the internal carotid. This anatomic variation may have clinical significance in localizing the site of lesion producing Horner syndrome (see below).

Pharmacology of the Sympathetic Nervous System

The iris dilator muscle contains alpha-adrenergic receptors and responds most strongly to adrenergic agents with primarily alpha-2 agonist properties. Among these, phenylephrine has been found to have the most widespread clinical use. Clinical studies have demonstrated that the 2.5% solution is almost as effective as the 10% solution and produces fewer cardiovascular adverse effects. In premature neonates, in whom a single drop of the 2.5% solution may produce serious cardiovascular adverse effects, the 1% solution is commonly used and has been shown to be equally as effective.

Sympathetic action potentials are terminated by reuptake of approximately 98% of released norepinephrine back into the presynaptic nerve terminal. When the entire three-neuron sympathetic pathway to the pupil is intact, the resting tonus of the dilator muscle is maintained by a continuous balance between release and reuptake of norepinephrine. Cocaine exerts its agonist effect by preventing reuptake of norepinephrine, allowing accumulation of the neurotransmitter and continuous stimulation of the postsynaptic receptors. Cocaine will dilate the pupil subnormally if a lesion at any point in the sympathetic pathway interrupts the tonic release of norepinephrine, preventing accumulation of neurotransmitter. All sympathetically denervated pupils, regardless of the site of lesion, show a relatively poor dilation to cocaine compared with the fellow eye. One drop of 10% cocaine is placed in each eye and repeated in about 5 minutes. The results are interpreted at 45 minutes. More than 0.8 mm of posttest anisocoria, regardless of pretest pupillary size, is considered a positive result (mean odds ratio of 1050:1 that Horner syndrome is present, lower 95% confidence limit = 37:1).⁴

Hydroxyamphetamine dilates the pupil by releasing norepinephrine from intact postganglionic neurons. A lesion of the sympathetic pathway affecting the postganglionic neuron will impair pupillary dilation by hydroxyamphetamine. Lesions of the central or preganglionic neuron, on the other hand, have little or no effect on the activity of topically applied hydroxyamphetamine. Hydroxyamphetamine therefore differentiates between postganglionic lesions, on the one hand, which dilate poorly to hydroxyamphetamine, and central or preganglionic lesions, on the other hand, which dilated normally to hydroxyamphetamine. On the basis of clinical data and pharmacologic testing in 54 patients with Horner syndrome, Cremer and colleagues⁵ concluded that after instillation of 1% hydroxyamphetamine, a difference in dilation of 1 mm indicates an 85% probability of a postganglionic lesion; a difference of 1.5 mm would correspond to a 96% probability of a postganglionic lesion.

No pharmacologic test is available that can differentiate a central from a preganglionic lesion. This must be done on the basis of other clinical data (i.e., the presence of hypothalamic, brain stem, or spinal cord symptoms in the case of a central lesion versus chest or neck signs and symptoms in a preganglionic lesion). This clinical differentiation is usually not difficult.

The hydroxyamphetamine and cocaine tests should not be performed on the same day, because cocaine may inhibit the uptake of hydroxyamphetamine by postganglionic nerve endings, yielding spurious results. Although some degree of denervation supersensitivity to dilute adrenergic agents is found in patients with postganglionic lesions, preganglionic agents can also display this property. The degree of overlap may preclude differentiation, and super-sensitivity testing has been largely abandoned in favor of the hydroxyamphetamine test.

Clinical Abnormalities of Pupillary Function

Simple Anisocoria
Approximately 15% of the normal population has clinically detectable anisocoria.⁶ The difference in pupillary diameter is usually in the range of 0.3 mm to 0.7 mm. Rarely is the difference more than 1 mm. The anisocoria in these normal patients is either equal in light and dark or is slightly greater in darkness. The anisocoria is never greater in light than dark. The designation "simple anisocoria" has been applied to the pupils in these normal patients. Although the anisocoria, once present, is usually unchanged on future examinations, the degree of anisocoria may vary in some patients, and the size of the pupils may even reverse, with the originally larger pupil becoming the smaller of the two.

Differences that help distinguish between simple anisocoria and Horner syndrome include lack of ptosis; no change in anisocoria between light and dark (although many patients with simple anisocoria have more anisocoria in darkness, as is seen in Horner syndrome); no "dilation lag" compared with the larger pupil (can be detected by flash Polaroid or digital photos); and a normal response to cocaine. Many patients referred for evaluation of Horner syndrome actually have physiologic anisocoria and a non-neurologic cause of ptosis, such as chronic blepharitis or age-related levator dehiscence. The cocaine test is useful in convincing referring physicians that the ptosis and anisocoria are non-neurogenic. Old photographs, such as a driver's license or other identification card, can also be useful.

Abnormalities of the Afferent Pupillary Pathway

A relative afferent pupillary defect (RAPD) is an extremely sensitive indicator of optic nerve disease and may be found in the presence of normal or minimally impaired visual acuity. For all practical purposes, afferent pupillary defects are not found in eyes with cataract, refractive error, or mild macular disease. When present, this sign indicates impaired function of the retina or optic nerve of one eye relative to the other eye (Table 1).

SLIDE 3

A proper technique for RAPD testing is crucial⁷ (Slide 3). Testing should be performed in a semi-dark room with the patient fixating at a distance. A bright light is used and should be alternated from one eye to the other every 3 to 4 seconds. The normal response to the swinging flashlight consists of a brief constriction of the stimulated pupil, followed by redilation, whereupon the pupils reach a stable diameter after some oscillation (hippus). When an afferent defect is present, both pupils are larger when the defective eye is stimulated and smaller when the good eye is stimulated. Only one working iris sphincter is required to perform the test, which can be carried out in the presence of unilateral posterior synechiae, third nerve palsy, corneal opacities, or hyphema (Slide 4). Only the reaction of the working pupil must be observed by using a dim sidelight while each eye is alternately illuminated with the swinging flashlight. The afferent pupil defect is on the side which, when stimulated, results in dilation of the observed pupil.

Considerable controversy has surrounded the question of optimal brightness of the stimulus used for the swinging flashlight test. The answer to this question appears to be that brighter is not necessarily better. Rizzo and colleagues⁸ have demonstrated that a very bright light may mask some RAPDs by creating a maximal response in the defective pupil that almost equals the normal response. In such cases, a midrange light intensity may be more effective in bringing out mild asymmetry. Therefore, it is most likely worthwhile to use a light with a variable rheostat (e.g., a muscle light or a Finoff transilluminator). When a small RAPD is suspected and is not adequately demonstrated with a bright light, a lower setting on the rheostat can be tried.

SLIDE 4

A fully developed, "barn-door" RAPD is characterized by obvious dilation of the affected pupil when the light is swung from the normal to the defective eye. But which component of the pupillary response is most useful for detecting smaller afferent defects? The possibilities include amplitude of initial constriction, amount of redilation, final size after dilation, and minimum size. Any one of these may be observed clinically during the swinging flashlight test, but keeping track of all four simultaneously in both eyes is impossible. According to Cox, "although clinicians have been looking at pupils for years, no one has shown what component of the pupillary response is most useful for detecting small afferent pupillary defects."⁹ To answer this question, Cox simulated RAPDs in normal subjects by using a neutral-density filter over one eye. Cox concluded that "to detect a relative afferent pupillary defect, the clinician should look for a difference in amplitude of the initial constriction of the two pupils during the alternating light test, and a pupil defect should be diagnosed when the consensual response is greater than the direct response." In other words, when a small RAPD is suspected, the clinician should focus attention on the suspected defective eye, which can be illuminated by a dim sidelight. As the bright light is alternated between the two eyes, the responses of the defective eye to direct and consensual illumination are compared. If the initial constriction to consensual stimulation exceeds that for direct stimulation, an RAPD is present. If the initial constriction is not well seen (i.e., too small), a brighter light can be used or the light may be alternated more slowly.

An RAPD can be quantified by "partial occlusion" of the good eye with a neutral density filter (Slide 5). The filter absorbs some of the incident light and reduces the response of the good eye. The swinging flashlight test is performed in the usual manner and the filter is increased or decreased until the RAPD is eliminated. This measurement can then be compared with later measurements as an index of recovery or progression of optic nerve (or retinal) disease. Thompson and colleagues have provided discussions of the fine points and possible pitfalls of quantitative RAPD testing.^7,10

SLIDE 5

It is important to remember that unilateral optic nerve disease does not result in anisocoria. This is because there is summation in the midbrain of all input from both optic nerves, followed by distribution of equal output to each oculomotor nucleus ( Slide 1). It is also important to keep in mind that the afferent pupillary defect is a relative test. It is a comparison of the pupillomotor input between the two optic nerves. Asymmetric optic nerve disease (e.g., asymmetric chiasmal compression) will give rise to an afferent pupil defect. Lesions of the optic tract may also cause afferent pupil defects. These lesions tend to produce incongruous (i.e., asymmetric) homonymous hemianopias with an afferent pupil defect on the side of greater field loss. When a tract lesion produces a complete homonymous hemianopia, the field loss in the ipsilateral eye (i.e., the eye with a temporal field defect) is greater than the field loss in the eye with the nasal defect. This asymmetry is often sufficient to produce a small but easily detectable afferent pupillary defect. Symmetric optic nerve disease may mask an afferent pupil defect, leading to the incorrect conclusion that both optic nerves are normal. For example, in patients with acute optic neuritis associated with multiple sclerosis, a relative afferent pupil defect may not be found, or may be smaller than expected, due to subclinical optic neuropathy on the asymptomatic side.

Although afferent pupil defects are usually considered to be characteristic of optic nerve disease, this sign may also be present in retinal lesions including detachment, age-related macular degeneration, histoplasmosis, and retinal vein occlusion. Although the mechanism is not clear, afferent pupil defects may be seen in some patients with amblyopia.¹¹ The severity of the RAPD is not related to the degree of amblyopia. Although opacities of the ocular media, such as cataracts and corneal opacities, filter a significant fraction of incoming light, they do not produce an RAPD.

Abnormalities of the Efferent Pupillary Pathway

Whereas afferent defects are characterized by a pupil that responds better to consensual than to direct stimulation, efferent defects are characterized by pupils that respond poorly to both direct and consensual stimulation. Abnormalities of the efferent pathway may be caused by lesions anywhere from the midbrain to the pupillary sphincter muscle. Before an etiologic diagnosis is possible, the clinician must localize the lesion by clinical, radiologic, and pharmacologic methods. The classification scheme outlined below is useful for that purpose.

Dorsal Midbrain Lesions Producing Light-Near Dissociation
Lesions of the dorsal midbrain often produce pupillary light-near dissociation, with impairment of the light reflex and relative sparing of the near response. Compression of the dorsal midbrain can selectively damage the dorsally located light reflex fibers in and around the posterior commissure, while leaving the more ventrally located near reflex fibers intact (Slide 6). The most frequent cause is a tumor of the pineal gland. Other causes include hydrocephalus and metastatic tumors in an around the periaqueductal area. The pupils in this syndrome are usually large, may be slightly unequal, and are often oval in shape. Light-near dissociation is only one component of the dorsal midbrain syndrome (Parinaud syndrome). Other components of the fully developed dorsal midbrain syndrome include a supranuclear upgaze palsy, impaired convergence, convergence-retraction nystagmus on attempted upgaze, and pseudoabducens paresis due to inappropriate convergence on attempted lateral gaze.

SLIDE 6

Perhaps the best known, albeit rare, form of light-near dissociation is the Argyll Robertson pupil. Argyll Robertson pupils are small, often irregular in shape and dilate poorly to all mydriatics. The condition is frequently bilateral, and the majority of patients have neurosyphilis. Good vision in both eyes is a prerequisite for this diagnosis, since optic neuropathy must be ruled out as a cause for impairment of the pupillary light reflex. It should be remembered that tonic pupils resulting from lesions of the ciliary ganglion and/or the short posterior ciliary nerves (see below) also exhibit light-near dissociation.

Lesions of the Third Nerve
Efferent pupil defects of neurogenic origin may occur as part of a third-nerve palsy or as an isolated internal ophthalmoplegia. The pupillary fibers of the third nerve arise from the Edinger-Westphal subnucleus, a midline structure that innervates both pupils (Slide 1). Nuclear lesions affecting the third nerve, therefore, cause bilateral (sometimes asymmetric) pupillary involvement if the pupils are involved at all. As shown in Slide 1, the fascicular portion of the third nerve (the portion within the midbrain) passes through the red nucleus and the corticospinal tracts, then emerges in the interpeduncular cistern. Midbrain lesions of various causes may result in a nuclear or fascicular third nerve plays with associated brain stem signs, including contralateral tremor due to involvement of the red nucleus (Benedikt syndrome) and contralateral hemiparesis due to involvement of the corticospinal tracts (Weber syndrome). Third nerve palsies caused by fascicular lesions often involve the pupil, although partial sparing of one or more extraocular muscles is not uncommon. Neuroradiologic evidence, however, suggests that midbrain ischemia is a frequent cause of pupil-sparing third-nerve palsy in diabetic patients.¹² These findings contradict earlier autopsy studies demonstrating infarction of the subarachnoid or intracavernous portion of the nerve.

The pupillary fibers occupy the dorsomedial portion of the third nerve, where their superficial location renders them vulnerable to external compression. The most common cause of third nerve compression is an aneurysm of the posterior communicating artery at its junction with the internal carotid artery (intracranial pressure aneurysm). Aneurysms at this location usually point downward and medially, immediately contacting with the pupillary fibers of the third nerve. Pupil sparing, therefore, is uncommon in third-nerve palsy resulting from ICPC aneurysms.

The issues involved in managing patients with pupil-sparing third-nerve palsy have been reviewed in detail elsewhere.^13,14 A complete review of this problem is beyond the scope of this chapter. The critical question, of course, is whether or not the palsy represents an ischemic process, with an excellent prognosis for recovery, or whether it represents a compressive lesion, possibly an ICPC aneurysm, which may be immediately life-threatening. Although the problem is still controversial and the management of each patient must be individualized, the following guidelines are helpful.

First, the majority (more than 90%) of patients with ischemic third-nerve palsies are diabetic. The diagnosis of ischemic third-nerve palsy should be viewed with caution in patients without diabetes mellitus. Approximately 10% of patients with nondiabetic ischemic third-nerve palsy are usually elderly and/or have a well-documented history of long-standing hypertension or arteriosclerotic cardiovascular disease, often with prior cardiovascular complications.

Second, pupil-sparing ischemic third-nerve palsies involve all of the extraocular muscles. Therefore, the term "pupil sparing" should not be applied to partial third-nerve palsies that spare one or more extraocular muscles. Partial third nerve palsies of this type should be considered nonischemic (but not necessarily aneurysmal) until proven otherwise.

Third, third-nerve palsy due to migraine is basically a disease of children and this diagnosis should be regarded with suspicion in anyone older than 20. Finally, high-resolution magnetic resonance imaging and magnetic resonance angiography are not adequate substitutes for cerebral angiography which is, to date, the only test capable, when performed adequately, of ruling out an aneurysm as the cause of third-nerve palsy. In summary, patients who are not diabetic or do not have well-established cardiovascular disease, as well as those with other atypical features mentioned above, should usually undergo neuroradiologic studies, including arteriography, to rule out an aneurysm or other compressive lesion. However, future advances in magnetic resonance angiography and other noninvasive techniques may greatly simplify the management of these patients, perhaps obviating standard angiography, over the next few years.

Aberrant regeneration of the third nerve may affect the eyelid (pseudo-Graefe sign) or the pupil and is almost always a sign of chronic compression.¹⁵ Aberrant regeneration involving the pupil is manifested by constriction of the affected pupil when the eye moves up, down, or medially, activating one of the muscles innervated by the third nerve. In patients with aberrant regeneration, axons destined for one of the extraocular muscles become "miswired," making their way to the ciliary ganglion instead. This miswiring, and hence the aberrant sphincter response, is often limited to one or more segments of the pupillary sphincter. Therefore, different areas of the sphincter may constrict, depending on the direction of intended gaze. Aberrant regeneration is characteristic of compressive lesions and does not occur with ischemic third-nerve palsy. Patients with slowly evolving compressive third nerve lesions may have only minimal paresis and little or no diplopia, due to a balance between injury and regrowth of axons. In such cases, the presence of aberrant regeneration provides a clue to the existence of a long-standing compressive third-nerve palsy.

Mass lesions within the cavernous sinus typically produce pupil-involving third nerve palsies. These include intracavernous carotid aneurysms, sphenoid meningiomas, lymphomas, pituitary adenomas, and metastases. Lesions within the cavernous sinus may involve both the third nerve and the sympathic fibers to the pupillary dilator muscle, producing a third-nerve palsy with a nondilated pupil which is nevertheless fixed to light. This should not be confused with a pupil-sparing third-nerve palsy. Simultaneous involvement of the sympathetic innervation to the pupil, which can be confirmed by cocaine testing, localizes the lesion to the cavernous sinus.

A potentially confusing situation may occur in diabetic patients with peripheral autonomic neuropathy. A third-nerve palsy may occur in the setting of a poorly reactive pupil caused by diabetic autonomic neuropathy affecting the innervation of the pupillary sphincter (see Tonic Pupils section). These patients usually have bilateral pupillary involvement and the other features of tonic pupils, including segmental palsy, light near dissociation, and a tonic near response, facilitating the diagnosis. However, testing for denervation supersensitivity with dilute pilocarpine will not differentiate pupillary involvement from third-nerve palsy (preganglionic lesion from an autonomic neuropathy postganglionic lesion).¹⁶

Tonic Pupils
The third-nerve palsy passes through the superior orbital fissure and into the orbit, where the pupillary fibers synapse in the ciliary ganglion. The ciliary ganglion may be affected by benign or malignant orbital tumors, orbital infection, or autoimmune orbital inflammation (orbital pseudotumor). As shown in Slide 1, the motor root to the ciliary ganglion arises from the inferior division of the third nerve as it courses along the lateral border of the inferior rectus muscle on its ways to innervate the inferior oblique. Trauma to the orbital floor, including entrapment of the inferior rectus muscle, may produce traction on the motor root, causing pupillary paresis. Lesions of the ciliary ganglion and/or short posterior ciliary nerves (of any etiology) produce a characteristic pupillary abnormality, the tonic pupil, characterized by segmental palsy of the iris sphincter (Slide 7); a tonic response to both light and near; light-near dissociation (i.e., the light reflex is usually affected to a greater extent than the near reflex); and denervation supersensitivity to dilute cholinergic agents.¹⁷

SLIDE 7

The most common cause of isolated internal ophthalmoplegia is Adie tonic pupil syndrome.¹⁸ The typical patient is a woman, 20 to 40 years of age, complaining of unilateral blurred vision at near, or in whom an asymptomatic anisocoria has been noted. In the earliest stage, the involved pupil is dilated and reacts poorly to light. The near response tends to be slow and tonic but is usually better than the light response. A "tonic" near response is one in which the pupils remains constricted long after the patient has discontinued accommodative effort (i.e., there is a delayed and prolonged relaxation of the sphincter after an accommodative response). For reasons that are unclear, deep tendon reflexes are diminished in most patients. The hyporeflexia is usually diffuse but may spare one or more reflexes and may be asymmetric.

The sphincter palsy in tonic pupils is almost always segmental (i.e., the response is preserved in some segments and impaired in others). This is most easily observed with slit lamp magnification as "bunching up" of the iris stroma and pupillary margin in segments with intact innervation and spreading of the stroma in flaccid segments. The radial strands of the iris stroma are tugged toward areas of working sphincter. Similar to the near response, the light response is usually tonic. This segmental tonic response has caused some observers to characterize the movement of the iris as "vermiform." Another common feature of Adie pupils is depigmentation of the iris collarette at the pupillary margin.

Denervation supersensitivity to dilute pilocarpine is present in 80% to 90% of patients with Adie syndrome,¹⁸ athough the diagnosis can often be made on the basis of clinical criteria alone. Pilocarpine 1/8% is the drug of choice; 1 drop is instilled in each eye and repeated 5 minutes later. The test should not be performed within 24 hours of other procedures that may affect corneal permeability, such as applanation or instillation of anesthetic drops. Pupil diameter is reevaluated at 45 minutes. A change in anisocoria of 1 mm or more compared with pretest measurement is considered a positive result. When bilateral tonic pupils are suspected, a comparative test with pilocarpine 1/8% would not be useful. Instead, pilocarpine 1/16% is instilled bilaterally (twice at 5 minutes apart). Normal pupils respond little, if at all, to this dilute solution; constriction greater than 1 mm of either pupil is considered highly probable for a tonic pupil.

The characteristics of an Adie pupil change with time.18 Approximately 50% of patients will recover almost full accommodation within 2 years. With time, the Adie pupil decreases in size and may actually become smaller than the normal pupil, although the tonic near response, the poor light response, the segmental nature of the palsy, and the denervation supersensitivity all persist. Palsy of previously normal segments often occurs. Involvement of the second eye is common and occurs at the rate of approximately 5% per year, so that after 10 years a patient who began with a unilateral Adie pupil has about a 50% chance of bilateral involvement.

Although the cause of Adie syndrome is unknown, the site of pathology is almost certainly the ciliary ganglion and/or the short posterior ciliary nerves. This view is supported by both histologic and physiologic evidence. Many patients with Adie syndrome have mildly impaired corneal sensation, presumably due to involvement of trigeminal branches to the short posterior ciliary nerves that pass through the ciliary ganglion without synapsing. The presence of denervation supersensitivity also supports a postganglionic lesion, although preganglionic lesions due to third nerve palsy may produce supersensitivity¹⁶. The light-near dissociation that occurs in Adie syndrome is believed to be due to aberrant regeneration in a mixed nerve. Near reflex fibers, primarily those destined for the ciliary muscle, outnumber light reflex fibers by approximately 30 to 1. When axonal (or ciliary ganglion) damage occurs, followed by regeneration, these odds overwhelmingly favor appropriate re-innervation of the ciliary muscle by an accommodative axon. If re-innervation of the iris sphincter occurs, the odds are 30 to 1 in favor of re-innervation by an accommodative axon rather than by an axon driven by the light reflex. It is believed that this phenomenon is the basis for the recovery of accommodation and the pupillary light-near dissociation.

Pharmacologic Paresis: The Fixed, Dilated Pupil
Pharmacologic sphincter paresis, inadvertent or intentional, is common in neuro-ophthalmologic practice. This may occur as part of a behavioral disorder or when pharmacists, nurses, or other medical personnel inadvertently handle cycloplegic drops or other anticholinergic substances. In this setting, the ophthalmologist is faced with the differential diagnosis of a pupil that is fixed to both light and near. The differential diagnosis of isolated mydriasis includes all of the causes of sphincter paresis, beginning in the midbrain and ending at the pupillary sphincter. These causes include:

• dorsal midbrain compression
• early third-nerve palsy due to a midbrain (nuclear or fascicular) or extra-axial lesion
• a lesion involving the ciliary ganglion or short posterior ciliary nerves
• pharmacologic sphincter paresis
• damage to the pupillary sphincter muscle from either inflammation or trauma

Myopathic injuries can usually be ruled out with certainty by inspection of the iris and anterior segment during slit lamp examination. Dorsal midbrain lesions virtually always cause bilateral pupillary abnormalities. Pupillary light-near dissociation is usually present, at least in the early stages. Other signs of dorsal midbrain compression include upgaze paresis with convergence-retraction nystagmus on attempted upgaze and pseudoabducens paresis. Compression of the third nerve rarely produces isolated pupillary paresis without extraocular muscle abnormalities. Even in the acute phase of transtentorial herniation, extraocular muscle abnormalities follow within hours of the pupillary abnormality, if not sooner. In addition, patients with this problem are neurologically ill. The differential diagnosis of isolated pupillary mydriasis is usually tonic pupil or pharmacologic paresis. The pharmacologic differentiation should proceed as outlined below.

Pilocarpine 1/8% should be used first to diagnose a tonic pupil. A differential response between the two pupils of more than 1 mm after installation of pilocarpine 1/8% is highly suggestive of denervation supersensitivity. This will be positive in at least 85% of patients with this disorder. If supersensitivity is present, no further testing is needed. If supersensitivity is not present, then 1% pilocarpine should be instilled bilaterally. If both pupils react equally to this agent, pharmacologic paresis, sphincter damage, and sphincter inflammation are ruled out. "Equal" reaction means the same change in pupillary size in each eye, for example, a change from 7 mm to 4 mm in one eye and from 6 mm to 3 mm in the other eye. If neither pupil constricts more than 1 mm to pilocarpine 1%, pilocarpine 4% is used and a differential response is again looked for. More than 1 mm difference in the response to 4% pilocarpine is considered indicative of pharmacologic paresis.

Abnormalities of the Sympathetic Pathway (Horner Syndrome)

The syndrome of ptosis, miosis, and facial anhidrosis was first attributed to a lesion of the sympathetic pathway by Friedrich Horner in 1858. Other inconsistent features, including elevation of the lower lid ("upside-down ptosis"), ocular hypotony, and conjunctival hyperemia in the acute stage were subsequently noted. Although the complete triad of ptosis, miosis, and facial anhidrosis is diagnostic, lesions distal to the superior cervical ganglion usually spare the sweating and vasoconstrictor fibers, leaving only ptosis and miosis. In addition, sweating can be difficult to assess in a dry, air-conditioned office or hospital. Many patients referred for evaluation of Horner syndrome actually have pseudo-Horner syndrome, with miosis due to physiologic anisocoria and ptosis due to a non-neurologic cause, such as blepharochalasis, age-related levator dehiscence, or old trauma.¹⁹ Pharmacologic testing should therefore be used to confirm the presence of Horner syndrome and to localize the lesion. These patients are often seen in the setting of evaluation for headache or other neurologic complaints in which the presence of Horner syndrome would significantly affect diagnosis and management. Pharmacologic confirmation of the diagnosis is particularly important in these patients.

When observed in room light, the anisocoria in Horner syndrome may be as little as 1 mm or even slightly less. Placing the patient in darkness accentuates the anisocoria by turning off the sphincter muscles and directly comparing the strength of the two dilator muscles. Anisocoria that does not increase in darkness is not likely to be due to Horner syndrome. In addition to increased anisocoria in darkness, the Horner pupil dilates more slowly than normal, often requiring 10 to 15 seconds to dilate completely.

Lesions of the sympathetic pathway to the eye can be classified anatomically as central, preganglionic, or postganglionic (Slide 2). Anatomic localization by clinical and pharmacologic criteria is useful in diagnosing the underlying cause of the lesion. Central lesions causing Horner syndrome originate in the hypothalamus, brain stem, or the cervical or upper thoracic spinal cord. Perhaps the most characteristic syndrome involving the central sympathetic pathway is Wallenberg lateral medullary syndrome. The syndrome includes ipsilateral Horner syndrome, facial hemianesthesia, paresis of ninth to twelth cranial nerves (one or more), and contralateral body hemianesthesia. It is caused by infarction of the lateral medulla, including the sympathetic pathway, the lower cranial nerve nuclei, the spinal tract and/or nucleus of the trigeminal nerve (ipsilateral facial hemianesthesia), and the ascending spinothalamic tract (contralateral body hemianesthesia). Lateral medullary infarction is usually caused by thrombosis of the posterior inferior cerebellar artery (PICA) or its parent vertebral artery. As noted earlier, pupils with central Horner syndrome fail to dilate to cocaine but dilate normally to hydroxyamphetamine.²⁰

Preganglionic Horner syndrome originates from lesions in the neck or chest. The best known cause is a tumor at the apex of the lung (Pancoast syndrome). The majority of patients with tumors in this location have symptoms of compression of the brachial plexus, especially pain in the hand or arm. All patients with pharmacologic evidence of preganglionic Horner syndrome require detailed imaging studies (magnetic resonance imaging or computerized tomography) of the chest and neck to rule out neoplasia. As noted earlier, pupils in patients with preganglionic and central lesions dilate normally to hydroxyamphetamine, and clinical criteria must be used to differentiate the two. However, this differentiation can usually be made reliably.

Postganglionic Horner syndrome originates from a lesion along the course of the carotid artery in the neck or at the base of the skull, or from a lesion within the cavernous sinus or at the orbital apex. Dissecting aneurysms of the carotid artery usually occur in the subcranial portion of the vessel and are characterized by the acute onset of pain and Horner syndrome, the latter occurring in approximately 20% of patients.²¹ The pain may be referred to the neck, jaw, pharynx, ear, cheek, or almost any craniofacial location. Dysgeusia, or a sensation of unpleasant taste, is sometimes present. Although many of these dissections are nonprogressive and self-healing, microembolization may occur and many patients will require anticoagulation. The diagnosis can often be made by the noninvasive technique of magnetic resonance angiography.

In 1924, Raeder described two syndromes to which his name was subsequently attached. The first type of patient described by Raeder (subsequently referred to as Raeder type 1) presented with Horner syndrome, chronic facial pain (often in the periorbital region), and trigeminal hypesthesia. These patients had tumors or space-occupying lesions at the base of the skull, involving the trigeminal ganglion or its branches and the sympathetic plexus on the carotid artery. The mass was often located near the fossa containing the trigeminal ganglion (Meckel cave) and adjacent to the carotid artery carrying the sympathetic plexus. The second type of patient described by Raeder (Raeder type 2) had what we would now recognize as the cluster headache syndrome, that is, intermittent periorbital headaches clustering in time and accompanied during one or more episodes by Horner syndrome, which may become permanent. The critical differentiating factors between the two types of patients are the presence of trigeminal hypesthesia and the more or less continuous characteristic of the pain associated with paratrigeminal masses. The eponyms Raeder type 1 and type 2 serve no useful purpose. Some patients with tumors at the base of the skull are unfortunately labeled by clinicians as "Raeder syndrome," with the inference that their problem is benign. It seems more logical to discard the term "Raeder syndrome" altogether and to remember that postganglionic Horner syndrome may be caused by a benign vascular headache syndrome or by lesions at the base of the skull that are frequently malignant. Although the absence of trigeminal hypesthesia is reassuring, most patients with postganglionic Horner syndrome should probably undergo detailed magnetic resonance imaging of the base of the skull.

Congenital Horner syndrome, or Horner syndrome occurring in the first 2 years of life, is often marked by hypochromia of the affected iris.²² The reason for this is not clear, but electron microscopic studies have demonstrated adrenergic nerve terminals associated with iris melanocytes. The trophic influence of the sympathetic nervous system is presumably required for melanocytic growth, which normally occurs in the iris stroma during early childhood. It is also important to note that testing with hydroxyamphetamine may not be reliable in early childhood. Preganglionic lesions (e.g., chest tumors) in children frequently produce a Horner syndrome that behaves pharmacologically as though the lesion were postganglionic, that is, the pupil fails to dilate to hydroxyamphetamine.²² This phenomenon has been attributed to transsynaptic degeneration of the postganglionic neuron after early preganglionic lesions, a phenomenon that has been abundantly documented in experimental animals. An acquired Horner syndrome in the first few years of life is an ominous occurrence and is most frequently caused by neuroblastoma involving the sympathetic chain in the chest or neck.²³ These tumors may be difficult to detect, and the assistance of a pediatric oncologist should be sought in children with acquired Horner syndrome. Horner syndrome that is present at birth, on the other hand, is almost always benign and a cause is rarely found. Cases of intrauterine or infantile neuroblastoma associated with congenital Horner syndrome have been described, however, and therefore a pediatric evaluation is advisable in these patients.

References

1. Hada J, Hayashi Y. Retinal X-afferents bifurcate to lateral geniculaste X-cells and to the pretectum or superior colliculus in cats. Brain Res. 1990;515:149-154.



2. Pierson R, Carpenter MB. Anatomical analysis of pupillary reflex pathways in the rhesus monkey. J Comp Neurol. 1974;158:121-144.



3. Mays LE, Gamlin PD, et al. Neuronal circuitry controlling the near response. Curr Opin Neurobiol. 1995;5(6):763-768.



4. Kardon RH, Denison CE, Brown CK, Thompson HS. Critical evaluation of the cocaine test in the diagnosis of Horner's syndrome. Arch Ophthalmol. 1990;108(3):834-387.



5. Cremer SA, Thompson HS, Digre KB, Kardon RH. Hydroxyamphetamine mydriasis in Horner's syndrome. Am J Ophthalmol. 1990;110(1):66-70.



6. Loewenfeld IE. Simple, central anisocoria: A common condition, seldom recognized. Trans Am Acad Ophthalmol Otolaryngol. 1977;83:OP832-839.



7. Thompson HS, Corbett JJ, Cox TA. How to measure the relative afferent pupillary defect. Surv Ophthalmol. 1981;26:39-42.



8. Rizzo JF. Pupillodynamic study of the swinging flashlight test: Intensity of stimulating light variably affects components of the pupil response. Abstract presented at the North American Neuro-Ophthalmology Society Meeting, February 24-28, 1991, Park City, UT.



9. Cox TA. Pupillographic chracteristics of simulated relative afferent pupillary defects. Invest Ophthalmol Vis Sci. 1989;30:1127-1131.



10. Thompson HS, Corbett JJ. Swinging flashlight test: Letter to the editor. Neurology. 1989;39:154-157.



11. Portnoy JZ, Thompson HS, Lennarson L, et al. Pupillary defects in amblyopia. Am J Ophthalmol. 1983;96:609-614.



12. Breen LA, Hopf HC, Farris BK, Gutmann L. Pupil-sparing oculomotor nerve palsy due to midbrain infarction. Arch Neurol. 1991;48:105-106.



13. Jacobson DM. Pupil involvement in patients with diabetes-associated oculomotor nerve palsy. Arch Ophthalmol. 1998;116(6):723-727.



14. Trobe JD. Managing oculomotor nerve palsy. Arch Ophthalmol. 1998;116(6):798.



15. Czarnecki JC, Thompson HS. The iris sphincter in aberrant regeneration of the third nerve. Arch Ophthalmol. 1978;96:1606-1610.



16. Jacobson DM. Pupillary responses to dilute pilocarpine in third nerve disorders. Neurology. 1990;40:804-808.



17. Thompson HS. A classification of tonic pupils. In: Thompson HS, Daroff RB, Firsen L, et al, eds. Topics in Neuro-Ophthalmology. Baltimore, MD: Williams & Wilkins; 1979:95-96.



18. Thompson HS. Adie's syndrome: Some new observations. Trans Am Ophthalmol Soc. 1977;75:587-626.



19. Thompson BM, Corbett JJ, Kline LB, Thompson HS. Pseudo-Horner's Syndrome. Arch Neurol. 1982;39:108-111.



20. Maloney WF, Young BR, Moyer NJ. Evaluation of the causes and accuracy of pharmocologic localization in Horner's syndrome. Am J Ophthalmol. 1980;90:394-402.



21. Kline LB, Vitek JJ, Raymon BC. Painful Horner's syndrome due to spontaneous carotid artery dissection. Ophthalmology. 1987;94:226-230.



22. Weinstein JM, Zweifel TJ, Thompson HS. Conenital Horner's syndrome. Arch Ophthalmol. 1980;98:1074-1078.



23. Woodruff G, Buncic JR, Morin JD. Horner syndrome in children. J Pediatr Ophthalmol Strabismus. 1988;25:41-44.