Chapter 203 – The Pupils
RANDY H. KARDON
• Pupillary disorders may be classified into two major categories—afferent and efferent.
• Afferent pupillary defects interfere with the input of light to the pupillomotor system by light blockage or deficits in any of the retinal layers, into the optic nerve, chiasm, optic tract, or midbrain pretectal area. All of these result in a symmetrical decrease in the contraction of both pupils to light given to the damaged eye, compared with light given to the other less damaged or normal eye.
• Efferent pupillary defects interfere with contraction or dilatation of the pupil due to damage in the midbrain, in the peripheral nerve that supplies the iris muscles, or in the iris muscles themselves, often leading to asymmetrical pupils (anisocoria).
• Relative afferent pupillary defects cause a reduction in pupil contraction when one eye is stimulated by light compared with when the opposite eye is stimulated by light.
• Efferent pupillary defects cause anisocoria, a difference in pupil size between the right and left eyes, the extent of which depends on the condition of lighting or near effort.
• Relative afferent pupillary defects may be associated with visual field or electroretinographic asymmetries between the two eyes. Asymmetrical differences in retinal appearance or optic nerve appearance may occur.
• Efferent pupillary defects may be associated with either damage to the parasympathetic or sympathetic nerves that supply the iris or direct damage to the iris sphincter or dilator muscles that results in immobility of the pupil.
In this chapter the pupil is discussed from a practical, clinical standpoint. The focus is on features of pupil examination that enable effective diagnosis and management of a variety of diseases of the afferent visual system and of diseases that affect pupil size. The chapter is divided into two main portions—one on the use of pupil examination to assess afferent visual input and the second on the diagnostic implications of abnormal integration of the efferent output to the pupils. Abnormal integration may result in pupils of unequal diameter (anisocoria), pupils that do not dilate well in darkness, or a light–near dissociation, in which pupil contraction to a near reflex greatly exceeds the pupil constriction to a light reflex.
RELATIVE AFFERENT PUPILLARY DEFECTS
In general, the most important clinical use of the pupil is in the assessment of afferent input from the retina, optic nerve, and subsequent anterior visual pathways (chiasm, optic tract, and midbrain pathways). Because the pupillary light reflex represents the sum of the entire neuronal input (photoreceptors, bipolar cells, ganglion cells, and axons of ganglion cells), damage anywhere along this portion of the visual pathway reduces the amplitude of pupil movement in response to a light stimulus.  Thus, the clinician can establish any asymmetrical damage between the two eyes by a simple comparison of how well the pupil contracts to a standard light shone into one eye compared with the same light shone into the other eye. Observation of pupil movement in response to alternating the light back and forth between the two eyes is the basis for the alternating light test, or “swinging flashlight” test, used to assess the relative afferent pupillary defect (RAPD).  
Another important aspect of pupil movement in response to light is that the pupillary light reflex summates the entire area of the visual field, with some increased weight given to the central 10°. Thus, in general terms, the pupillary light reflex is roughly proportional to the amount of working visual field. Damage to peripheral portions of the retina and visual field defects outside the central field reduce the amplitude of the pupillary light reflex. Such damage may not be established by other objective tests of visual function, such as the electroretinogram and visual evoked potential.
Standard flash electroretinogram findings are affected very little by focal retinal damage that produces a visual field defect. For example, a patient who has a disciform scar caused by aged-related macular degeneration or a branch artery occlusion gives a normal flash electroretinogram result. However, in such an example, the pupillary light reflex is reduced compared with that of the other eye, and an RAPD is obvious. In addition, optic nerve disease or damage to the retinal ganglion cells is not detected by standard flash electroretinography. Assessment of the pupillary light reflex readily enables the detection of such damage.
Because the visual evoked potential primarily samples the occipital pole or tip, which represents the central 5–10° of visual field, it is not affected to any great extent by peripheral visual field defects. In addition, cooperation of the patient is required to fixate on the center of a computer monitor while the visual stimulus is presented. A patient who chooses not to fixate properly or who has media opacities that reduce the clarity of the checkerboard pattern may produce an abnormal result, even if the retina and optic nerve function are normal (i.e., a false-positive test result). Similarly, peripheral visual field defects caused by glaucoma or anterior ischemic optic neuropathy may yield a normal visual evoked potential, or false-negative result, but the pupillary light reflex is reduced. 
Therefore, the pupillary light reflex is one of the few objective reflexes that can be used as a clinical test for the detection and
quantification of abnormalities of the retina, optic nerve, optic chiasm, or optic tract. Because the amount of RAPD is correlated, to a large extent, with the amount of asymmetry of visual field deficit between the two eyes, it also may be used to help substantiate abnormal results of perimetric testing    ; this often helps the clinician to determine whether a patient’s report of visual field defects is believable and trustworthy. The correlation between visual field asymmetry and RAPD also is a useful monitor of the course of disease for a worsening or improvement in function. RAPDs are, by definition, relative to the input of one eye compared with that of the other. Bilateral symmetrical damage does not produce RAPDs. Thus, a definite RAPD in one eye on the first visit but no RAPD on follow-up may represent improvement in the previously damaged eye or the development of damage in the previously better eye. Therefore, it is always important to remember that the RAPD is, indeed, relative to the other eye.
Estimation of the amount of RAPD in log units (asymmetry between the two eyes) provides an idea of how much visual field damage is present and whether it is consistent with the results of the visual field test. In addition, the amount of RAPD may indicate whether the cause of damage is consistent with the results of the pupil examination. For example, a patient affected by a small amount of macular degeneration in one eye and not the other is expected to have only a 0.3–log unit RAPD, but if that patient has a 1.0–log unit RAPD, then some other cause of visual loss is likely, such as a previous branch retinal artery occlusion or optic neuropathy ( Table 203-1 ).
In general, with unilateral visual loss, loss of the central 5° of the visual field results in an RAPD of approximately 0.3 log
TABLE 203-1 — COMMON DISEASES THAT PRODUCE RELATIVE AFFERENT PUPILLARY DEFECTS
Anterior chamber or vitreous (dense)
Density of hemorrhage
Anterior chamber (diffuse)
Density of hemorrhage
Preretinal (central vein occlusion or diabetic)
Preretinal location does not significantly reduce light
Diffusing media opacity
Cataract or corneal scar
0.0–0.3 in opposite eye
Dispersion of light produces increase in light input
Unilateral functional visual field loss
No real visual field loss
Central serous retinopathy or cystoid macular edema
Area of retina involved, depth of scotoma
Central or branch retinal vein occlusion
Area of visual field defect and degree of ischemia
Central or branch retinal artery occlusion
Area and location of retina involved
Area and location of detached retina (e.g., 0.6–0.9 log units for macula +0.3 log units for each quadrant)
Anterior ischemic optic neuropathy
Optic nerve head
Extent and location of visual field defect
Optic neuritis (acute)
Extent and location of visual field defect
Optic neuritis (recovered)
No visual field defect, residual relative afferent pupillary defect
Compressive optic neuropathy
Extent and location of visual field defect, other eye involvment
Asymmetry of visual field loss, unilateral central field involvement
Optic tract lesion
0.3–1.2 in the eye with temporal field loss
Incongruity of homonymous field defect, hemifield pupillomotor input asymmetry
Stimulus light size (no residual relative afferent papillary defect but definite pupil perimetry defects)
Midbrain tectal damage
Olivary pretectal area of pupil light input region of midbrain
Similar to optic tract lesions, but no visual field defect
The expected magnitude of defect is given as well.
units. Loss of the entire central area of field (10°) causes an RAPD of 0.6–0.9 log units. Each visual field quadrant outside of the macula is worth about 0.3 log units, but the temporal field loss seems to result in more loss of pupillary input compared with loss in the nasal field quadrants. The correlation between the relative afferent defect and the area and extent of visual field loss, however, is only approximate. Differences between the two may be important clues as to the cause and extent of damage to the anterior visual system.
Studies that used computerized pupillography to quantify the RAPD more precisely showed that some subjects who have normal visual fields and examination results can have a small (0.3–log unit) RAPD. 
The amount of pupillomotor input asymmetry (the RAPD) may be estimated roughly using the alternating light test (without any neutral density filters) and the subjective grades +1, +2, +3 or +4 for asymmetry of pupillary response. This subjective grading also may be categorized according to the amount of “pupil escape,” or dilatation of the pupils, as the light is alternated between the eyes. However, most subjective grading of RAPDs has serious limitations, such as some large-scale errors that arise from age variations in pupil size and pupil mobility. For example, a patient who has small pupils and small pupillary contractions to light may have a large RAPD, but this may appear deceptively small on the basis of small differences in pupil excursion observed as the light is alternated between the two eyes. However, the amount of neutral density filter needed to dim the better eye until the small contractions are equal represents substantial input damage. To estimate the size of RAPDs
without using filters is very much like the estimation of an ocular deviation “by Hirschberg” without a prism cover test. More accurate quantification of RAPDs is accomplished by determination of the log unit difference needed to “balance” the pupil reaction between the two eyes.  Photographic neutral density filters (49?mm, screw mount, 0.3 log, 0.6 log, and 0.9 log) often are available through local photography stores.
Measurement of the Relative Afferent Pupillary Defect
Measurement of RAPDs is the most important part of the pupil examination, because it may give the most valuable clinical information. The alternating light test for an afferent defect is based on the assumption that the irises are a matched pair—each has sphincter and dilator muscles of good shape and properly innervated—so that the light reactions can be compared. Therefore, it is important to first establish whether an anisocoria is present, which may indicate an efferent defect.
EVALUATION OF ANISOCORIA.
Pupillary inequality usually results from an iris innervation problem, so to evaluate anisocoria the iris sphincter and dilator muscles must be checked. In the office, the best way to decide whether the sphincter muscle or dilator muscle is weak is to compare the amount of anisocoria in darkness and in light, which can be carried out without any special equipment. The examiner must be able to change the lighting and still view the pupils. Of course, usually no anisocoria is found in darkness or in light, in which case the efferent arm of the light reflex arc is presumed intact, and the examiner proceeds to check for an afferent defect. When anisocoria is present, the examiner needs to establish whether it increases in darkness or in light. If one sphincter is weak, the investigator may still check for an afferent defect in the pupil that still works by comparison of its direct and consensual reactions. If no asymmetry of input is apparent, and hence no RAPD, this impression may be confirmed using the “tilt test” described below ( Fig. 203-1 ).
Anisocoria may influence the estimate of pupillary input asymmetry. Small pupils allow less light to pass and large pupils more. However, if neither pupil is less than 3.0?mm wide in light, then any anisocoria less than 2.0?mm difference may be disregarded—at least with respect to a false afferent defect induced by the pupillary inequality. Only very large anisocorias cause enough difference in retinal illumination between the two eyes to produce an apparent asymmetry of pupillomotor input.
ESTABLISHMENT OF RELATIVE AFFERENT PUPILLARY DEFECT.
To check for RAPD, the light is alternated from one eye to the other ( Fig. 203-2 ). If the light is too bright, the pupils do not redilate promptly, and very little pupil movement is seen as the light is alternated to the other eye. The problem may be solved by a direct reduction in stimulus intensity or if the light is moved 3–4 inches (8–10?cm) away from the eyes and alternated between them.
Figure 203-1 Checking for an RAPD using the “tilt test.” If there seems to be no input asymmetry, the “tilt test” can confirm this by inducing an RAPD of the same magnitude in each eye by holding a 0.3–log unit neutral density filter over one eye during the alternating light test and then repeating it with the filter over the other eye.
Observe the Illuminated Eye.
If the pupils react relatively weakly when one eye is stimulated and better when the other is stimulated, an afferent defect relative to the better eye (RAPD) has been identified.
Balance the Responses Using Filters.
To balance the response, a filter is held over the good eye and the alternating light test repeated. If the input asymmetry is still visible, the density of the filter over the good eye is increased until the amplitudes of the direct light reactions of the two eyes are balanced. To be certain of the measurement, the balance point may be overshot deliberately and then back titrated. When a dense filter is used, it may be necessary to look behind the filter to see the pupil ( Fig. 203-3 ).
“TILTING”—THE DUBIOUS RELATIVE AFFERENT PUPILLARY DEFECT.
If a very small asymmetry is suspected, such as a defect of less than 0.3 log units in the left eye, it may be just the result of noise in the system (e.g., “hippus”). An effort must made to confirm the asymmetry by “tilting” the RAPD to the right and to the left using a 0.3–log unit filter (see Fig. 203-1 ). If no RAPD exists, the examiner should be able to induce the same amount of input asymmetry by holding a 0.3–log unit filter over the right eye during the alternating light test and then repeating the test with the filter switched to the left eye. If a small RAPD is present, it will become more apparent when the filter is held over that eye.
Unfortunately the pupillary response to a repeated light stimulus is far from constant; it changes from moment to moment.  A common error is to judge the apparent asymmetry of the light reflex too quickly. It is important to alternate the light back and forth at least 3 times to obtain a mental average of any asymmetry. In this way, moment-to-moment fluctuations in the pupillary response are “averaged out.” Despite such efforts, the afferent asymmetry may still fluctuate slightly with time, even when carefully recorded using computerized pupillography.
Infants and small children may appear to have weak pupillary responses to light, largely because of the excitement
Figure 203-2 Demonstration of a large afferent defect in the right eye. This is best demonstrated when the light is alternated from eye to eye at a steady rate. The light is kept just below the visual axis and 1–2 inches (3–5?cm) from each eye. Each eye is illuminated for about 1 second and then the light switched quickly to the other eye; this allows comparison of the initial direct pupil contraction with light in each eye.
Figure 203-3 Balancing pupillary response using filters. A dark iris appears even darker behind the filter so it may be difficult to view the pupil, in which case it helps to peek behind the filter to obtain a better view of the iris.
and apprehension that inhibit the pupillary light reflex at the supranuclear level in the midbrain. Usually, after the light stimulus has been repeated several times, the light reaction begins to improve and “loosen up,” especially as the child becomes less anxious. A baby’s pupils are checked at about 3?ft (1?m) away with a direct ophthalmoscope. In a dark room, the brightest light and smallest spot are used, focus is on the red reflex, and the light is alternated from eye to eye. The baby usually is fascinated, and a filter sometimes can be placed in the beam to one eye.
ONLY ONE WORKING PUPIL.
When the input defect is in an eye that has an injured iris or a dilated pupil, the pupillary responses of the uninjured eye must be observed. The direct and consensual responses of the working pupil may be compared by alternation of the light from one eye to the other. While a measurement is made, the good eye is behind the filter and it may be hard too see the pupil. Sometimes it is necessary to use a side light on the healthy iris in order to see its consensual pupil reaction. Because this may corrupt the measurement, an infrared video system is used, if possible.
Instruments are available that give a more precise evaluation of the pupillary light reflex; these include infrared video recording equipment and a computerized interface to present controlled light stimuli and quantify the dynamics of pupil movement in response to each light stimulus.
Sometimes it is important to view the magnified movement of both pupils at once. Infrared videography ( Fig. 203-4 ) enables the examiner to see both pupils clearly in the dark, which is particularly helpful when difficult afferent pupillary defects need to be checked for (e.g., one pupil is fixed, or both irises are pigmented very darkly). Because melanin reflects infrared light, dark irises appear light and so the black pupils stand out in contrast and are viewed easily. Videography also is used to establish the dilatation lag of a Horner’s pupil, to catch the brief paradoxical constriction in patients who have some retinal abnormalities (found when lights are turned out), to transilluminate the iris in pigment dispersion syndrome, and in Adie’s syndrome.
Various computerized infrared-sensitive pupillometers are available commercially and can record precisely the dynamics of pupil movement in the light or in the dark; the results are analyzed by sophisticated software. Such systems provide quantitative information about the pupillary light reflex and, in the future, may help to automate the clinical determination of pupillary input deficits that result from retinal and optic nerve diseases.     In addition, information obtained using computerized pupillography provides evidence
Figure 203-4 The infrared video-pupillometer. The infrared sources (clusters of light-emitting diodes) are mounted in gooseneck lamps. The double base–out prisms bring the pupils close together on the screen, allowing increased magnification from the infrared videocamera and telephoto lens.
that a number of different types of visual stimuli can produce changes in the pupillary light reflex, related to color, form, movement, and acuity.      
An automated perimeter may be modified to record pupillary responses. A video camera is pointed at the pupil and the amplitude of each light reaction is measured and stored in the computer, which is helpful as an objective form of perimetry and to localize lesions in the pupillary pathways.  Pupil perimetry is also useful in cases of nonorganic, functional visual loss to show objectively that messages are, indeed, going normally into the brain from parts of the visual field in which the patient claims to see nothing.
EFFERENT PUPILLARY DEFECTS
As discussed above, a pupillary inequality usually indicates that one of the four iris muscles, or its innervation, is damaged ( Fig. 203-5 ). To establish which is the weaker muscle, it is useful to know how the anisocoria is influenced by light. An anisocoria always increases in the direction of action of the paretic iris muscle, just as an esotropia increases when gaze is in the direction of action of a weak lateral rectus muscle.
Pupillary Inequality That Increases in the Dark
In patients who have a pupillary inequality that increases in the dark ( Fig. 203-6 ), the problem is to differentiate Horner’s syndrome from a simple anisocoria (or physiological anisocoria)—in which the inequality also is greater in dim light. A simple anisocoria may vary from day to day, or even from hour to hour, and is visible in about one fifth of the normal population; it is not related to refractive error. Clinically, Horner’s syndrome is recognized by associated signs such as ptosis, “upside-down ptosis” of the lower lid and, in an acute case, conjunctival injection and lowered intraocular pressure.
Simple anisocoria is common (about 10% of normal subjects, examined in room light, have an anisocoria of 0.4?mm or more) and is not associated with disease. Simple anisocoria also, like Horner’s syndrome, decreases slightly in light, but it does not show a dilatation lag of the smaller pupil. It is believed that simple anisocoria most likely arises from asymmetrical inhibition at the Edinger–Westphal nucleus in the midbrain. Normally, during wakefulness some inhibition from the reticular activating formation keeps the pupils midsize or larger. During sleep, this inhibition fades and allows the neurons in the Edinger–Westphal nucleus to discharge, which results in miotic pupils. If, during wakefulness, the inhibition is greater to the right Edinger–Westphal nucleus than the left, the right pupil is larger, especially in dim light. When light is added or a near reflex is generated, this inhibition is overcome and the pupils become smaller and any asymmetrical inhibition diminishes. A reduction of the anisocoria results as the pupils become smaller.
PUPIL DILATATION RATE.
The characteristic “dilatation lag” of the pupil in Horner’s syndrome is seen easily in the office using a handheld light shone from below. The room lights are switched off and the smaller pupil examined for an apparent reluctance to dilate. Pupil dilatation is normally a combination of sphincter relaxation and dilator contraction, a combination that produces a prompt dilatation. The patient who has Horner’s syndrome has a weak dilator muscle in one iris and, as a result, that pupil dilates more slowly than the normal pupil. If the sympathetic lesion is complete, the affected pupil dilates only by sphincter relaxation. The resultant asymmetry of pupil dilatation produces an anisocoria that is largest 4–5?sec after the lights have been turned out—the process is much slower than is generally thought. At 10–20?sec after the lights have been put out, the anisocoria lessens as the sympathectomized pupil gradually catches up, a process referred to as dilatation lag. The test is a quick and simple way to differentiate Horner’s syndrome from simple anisocoria, and it does not require
Figure 203-5 Parasympathetic and sympathetic innervation of the iris muscles.
pupillary drug tests. It works well most of the time, particularly in young people who have mobile pupils, but if the dilatation lag is inconclusive, cocaine eye drops may be used to confirm the diagnosis of Horner’s syndrome.
DIAGNOSIS OF HORNER’S SYNDROME—COCAINE TESTING.
The action of cocaine is to block the reuptake of norepinephrine (noradrenaline) normally released from the nerve endings. If, because of an interruption in the sympathetic pathway, norepinephrine is not released, cocaine has no adrenergic effect. The affected pupil in a patient with Horner’s syndrome dilates less with cocaine than does the normal pupil, regardless of the location of the lesion. Cocaine drops are placed in both eyes and after 60 minutes the anisocoria has increased clearly, because the normal pupil has dilated more than the Horner’s pupil.
The author recommends cocaine hydrochloride 10% in both eyes (not more than two drops) to ensure that even the darkest iris receives a full mydriatic dose; corneal epithelial defects will not result from this dose. Cocaine 2%, 4%, and 5% also have been used in diagnostic tests for Horner’s syndrome and work well. Anisocoria is measured after 50–60 minutes have elapsed. If very little dilatation of the pupil occurs, even though an oculosympathetic defect is suspected, and the pupil did not dilate well in darkness even after 30 seconds before the cocaine test, then a false-positive cocaine test result must be considered. A false-positive result may occur if the iris is held in a miotic state, through either scarring or aberrant reinnervation of the iris sphincter. In such cases, the addition of a direct-acting sympathomimetic agent to both eyes (e.g., 2.5% phenylephrine) at the conclusion of a positive cocaine test should dilate the suspected eye easily and eliminate the cocaine-induced anisocoria almost immediately. For the reasons stated above, pseudo–Horner’s syndrome results in inadequate dilatation to direct-acting sympathomimetic agents. The likelihood of Horner’s syndrome increases steadily as the degree of pupillary inequality (measured after the instillation of cocaine) increases. Unlike in the hydroxyamphetamine test, calculation of the change in anisocoria from before to after cocaine application is unnecessary. If there is at least 0.8?mm of pupillary inequality after cocaine administration, Horner’s syndrome is highly likely.
LOCATION OF DAMAGE TO THE SYMPATHETIC PATHWAY.
Whether the damage to the sympathetic pathway is in the postganglionic neuron is a question of considerable clinical importance, because many postganglionic defects are caused by benign, vascular headache syndromes or carotid dissections, and a preganglionic lesion sometimes results from the spread of a malignant neoplasm.
Hydroxyamphetamine eyedrops help to localize the lesion in Horner’s syndrome. The clinician needs to know where the lesion is to direct the radiographic workup, for example, to the internal carotid artery rather than to the pulmonary apex. Horner’s syndrome sometimes manifests so characteristically that further efforts to localize the lesion are not needed, as with patients who have cluster headaches.
Hydroxyamphetamine releases norepinephrine from storage in the sympathetic nerve endings. When the lesion is postganglionic, the third order nerve is dead and no norepinephrine stores are available for release at the iris. When the lesion is complete, the pupil does not dilate at all. However, the dying neurons and their stores of norepinephrine may last for almost 1 week from the onset of damage. Therefore, a hydroxyamphetamine test administered within 1 week of a postganglionic lesion may give a false preganglionic localization if some of the norepinephrine
Figure 203-6 Diagnosis of pupillary abnormalities in which anisocoria increases in dim light. If the anisocoria is greatest in dim light and diminishes in bright light, then the pupillary inequality is either physiological (a simple anisocoria) or arises from the loss of sympathetic innervation to the dilator muscle (Horner’s syndrome). A few other conditions need to be considered, but this chart is concerned only with acute damage to a single intraocular muscle or its innervation. (Adapted from Thompson HS, Kardon RH: Clinical importance of pupillary inequality, Focal Points: Clinical Modules for Ophthalmologists, vol 10, no 10, December, 1992, American Academy of Ophthalmology.)
stores remain. When Horner’s syndrome is caused by preganglionic or central lesions, the pupils dilate normally, because the postganglionic third order neuron and its stores of norepinephrine, although disconnected, are still intact; when the lesion is in the preganglionic neuron, the involved pupil often becomes larger than the normal pupil after hydroxyamphetamine administration, apparently because of “decentralization supersensitivity.”
INTERPRETATION OF THE HYDROXYAMPHETAMINE TEST.
The test is simple—the pupil diameters are measured before and 40–60 minutes after hydroxyamphetamine drops have been placed in both eyes. The change in anisocoria in room light is noted. If the affected pupil—the smaller one—dilates less than the normal pupil, an increase in anisocoria occurs, and the lesion is in the postganglionic neuron. If the smaller pupil now dilates so much that it becomes the larger pupil, the lesion is preganglionic and the postganglionic neuron is intact. The examiner must wait at least 2 days after cocaine has been used before the administration of hydroxyamphetamine; cocaine seems to block its effectiveness.
In about one half of ambulatory patients with Horner’s syndrome, the location of the lesion is identified satisfactorily by the nature and location of the injury or disease. The other half of these patients offer no clues as to the location of the damage— a pharmacological localization of the lesion in these patients can be most helpful.
The author has attempted to apply the results of hydroxyamphetamine mydriasis in those patients with a known lesion location to those in whom the lesion location is unknown. It appears that postganglionic lesions (along the carotid artery) can be separated from the nonpostganglionic lesions (in the brainstem, spinal cord, upper lung, and lower neck) with a degree of certainty that varies with the amount of anisocoria induced when the drops are placed in both eyes.
CONGENITAL HORNER’S SYNDROME.
When a child is observed to have a unilateral ptosis and miosis, the first question is
Figure 203-7 Horner’s syndrome clearly acquired in infancy must be evaluated for neuroblastoma, a treatable tumor. This baby, with a right ptosis and miosis, developed a flush during cycloplegia that made the vasomotor abnormality very clear—the Horner’s side remained pale. The baby had no sign of Horner’s syndrome during her first 8 months, but at 16 months Horner’s syndrome is obvious (ptosis, miosis, and upside-down ptosis). Because the syndrome was acquired, a chest radiograph was ordered; it showed a mass in the pulmonary apex. Magnetic resonance imaging confirmed the lesion. Surgery showed it to be a neuroblastoma.
to ascertain whether Horner’s syndrome is present. The ptosis of Horner’s syndrome is moderate, never complete. Sometimes the elevation of the lower lid is helpful. A child who has congenital Horner’s syndrome and naturally curly hair has, on the affected side of the head, hair that seems limp and lank. The shape of the hair follicles apparently depends on intact sympathetic innervation, as does the iris pigment. A child who has blond, straight hair and very pale, blue eyes does not have any visible hair straightness or iris heterochromia.
A weaker solution of cocaine (two drops of cocaine 2% in each eye) is used in children. The most telling symptom is the hemifacial flush (blanch on the affected side) that occurs with nursing or crying. Generally, the affected side is pale. In an air conditioned office, it may be hard to decide whether decreased sweating on the affected side is present. A cycloplegic refraction sometimes produces an atropinic flush everywhere except on the affected face and forehead and, thus, provides additional evidence toward diagnosis because of lack of sympathetic innervation to the skin vasculature.
In infants, hydroxyamphetamine drops do not help to localize the lesion, because orthograde transsynaptic dysgenesis takes place at the superior cervical ganglion after early interruption of the preganglionic oculosympathetic neuron. Fewer postganglionic neurons result, even though no direct postganglionic injury has occurred, which produces weak mydriasis and ambiguous results in children. A patient with Horner’s syndrome clearly acquired in infancy must be evaluated for neuroblastoma—a treatable tumor ( Fig. 203-7 ).
Pupillary Inequality That Increases With Light
For a patient who has pupillary inequality that increases with light, several problems must be addressed ( Fig. 203-8 ).
SLIT-LAMP EXAMINATION OF THE IRIS.
Trauma to the globe usually results in a torn sphincter and an iris border that transilluminates at the slit lamp. The pupil often is not round and other evidence of ocular injury may be present. Naturally, such a pupil does not constrict well to light. The residual reaction often is segmental in a traumatic iridoplegia. An atrophic sphincter caused by previous herpes zoster iritis also may reveal transillumination defects, as seen with the slit lamp, that arise from previous ischemic insults to the iris. If, however, the iris looks normal, further investigation is required, as outlined below.
RESIDUAL LIGHT REACTION.
If no residual light reaction is present, the possibility of pharmacological mydriasis must be explored. However, a completely blocked light reaction sometimes may occur when the sphincter is denervated by either a preganglionic lesion (third cranial nerve palsy) or a postganglionic lesion (acute, complete tonic pupil), in acute angle closure (iris ischemia), or with an intraocular iron foreign body (iron mydriasis). If the dilated pupil still has some response to light, the dilatation may result from partial denervation of the sphincter, incomplete atropinization, or adrenergic mydriasis. When the light reaction is poor because the dilator muscle is in spasm (as a consequence of adrenergic mydriatics such as phenylephrine), then the pupil is very large, the conjunctiva is blanched, and the lid is retracted. In such cases, any decrease in the amplitude of accommodation is minor and is the result of spherical aberration and a shallow depth of field—both optical results of the dilated pupil, or from the small inhibitory effect of sympathetic receptor activation or accommodation.
SEGMENTAL PARALYSIS OF THE IRIS SPHINCTER.
When some residual light reaction occurs, the iris sphincter is examined for sector palsy using the slit lamp. When the dilator is in a drug-induced adrenergic spasm or when the cholinergic receptors in the iris sphincter are blocked by an atropine-like drug, the entire sphincter muscle (all 360°) is less effective. This does not happen when postganglionic parasympathetic nerve fibers have been interrupted. In patients with Adie’s syndrome, all pupils that have a residual light reaction (about 90%) show segmental contractions of the sphincter (so-called vermiform movements). Thus, a pupil that has a weak light reaction and no segmental palsy usually indicates a drug-induced mydriasis, but signs of a third cranial nerve paresis (preganglionic parasympathetic nerve) must also be sought.
PUPILLARY SUPERSENSITIVITY TO CHOLINERGIC DRUGS.
If weak pilocarpine (about 0.1%) or weak methacholine (2.5%) is applied to both eyes (with both corneas healthy and untouched), and the affected (dilated) pupil constricts more than the normal pupil to become the smaller pupil, that iris sphincter is denervated. It seems likely that with a postganglionic denervation (ciliary ganglion to the eye), the sphincter will show a little more supersensitivity than in the preganglionic case (third cranial nerve palsy); however, the differences are not great. Cholinergic supersensitivity of the iris sphincter is considered now to be only a weak sign of Adie’s syndrome. As the iris sphincter is reinnervated by cholinergic accommodative fibers and becomes smaller over time, supersensitivity can be lost.
Ptosis or diplopia must be re-evaluated, because it is very rare for an ambulatory patient to have an isolated sphincter palsy from damage to the intracranial third nerve. If the normal pupil constricts a little and the dilated pupil not at all, the mydriasis may result from a local dose of an anticholinergic drug such as atropine. A stronger concentration of pilocarpine is needed to establish this.
PUPILLARY RESPONSE TO A MIOTIC DOSE OF PILOCARPINE.
If, on application of pilocarpine 1% in each eye, the affected pupil reacts little or not at all and the unaffected pupil constricts normally, the pupil was not dilated because of innervation problems but because of a problem in the sphincter muscle, itself. Non-neuronal causes of mydriasis are:
• Anticholinergic mydriasis (e.g., scopolamine [hyoscine], cyclopentolate, atropine)
• Traumatic iridoplegia (sphincter rupture, pigment dispersion, angle recession)
• Angle-closure glaucoma (ischemia of the iris sphincter)
• Fixed pupil after anterior segment surgery
• Bound down iris (synechia) after iritis
The cause for complete loss of function of the iris muscles after anterior segment surgery is unknown. Sometimes an excessive rise in intraocular pressure during or after surgery can cause ischemic damage to the iris sphincter.
Tonic Pupil of Adie’s Syndrome
Young adults (more women than men) may discover that one pupil is large or that they cannot focus up close with one eye. Slit-lamp examination usually shows segmental denervation of the iris sphincter. Within the first week, supersensitivity to cholinergic substances may be demonstrated. After about 2 months, nerve regrowth is active and fibers originally bound for
Figure 203-8 Diagnosis of pupillary abnormalities in which anisocoria increases in bright light. Initial pupillary inequality greater in bright light than in the dark indicates that the sphincter of the large pupil is weak or that a parasympathetic lesion is present on that side.
the ciliary muscle (they outnumber the sphincter fibers by 30:1) start to arrive (aberrantly) at the iris sphincter, which produces the characteristic light–near dissociation of Adie’s syndrome. Eventually, the affected pupil becomes the smaller of the two pupils, especially in dim light, because of the aberrant reinnervation by accommodative fibers (“little old Adie’s pupil”). The segmental palsy of the iris sphincter is seen particularly well using infrared video recording of transillumination of the iris.
Fixed, Dilated Pupil
When a pupil is dilated by an atropinic medication, the resultant condition can be differentiated with confidence from an innervational palsy by its tendency to resist the miotic effects of cholinergic drops such as pilocarpine. Pilocarpine 1% is a sufficient miotic dose for any eye, but a sphincter with all its cholinergic receptors blocked by atropine or tropicamide does not constrict with pilocarpine 1%. If the anticholinergic drug starts to wear off such that a small light reaction begins to return, pilocarpine 1% may only cause minimal constriction compared with the normal pupil.
Third Cranial Nerve Palsy
An old clinical rule of thumb states that if the pupillary light reaction is spared, the third cranial nerve palsy probably does not result from compression or injury, but more likely from small-vessel disease such as might be seen in diabetes. The rule still applies,
but a very small number of pupil-sparing third cranial nerve palsies arise from midbrain infarcts.
Aberrant Regeneration of the Third Cranial Nerve
The third cranial nerve carries instructions to several different muscles, so when the nerve is injured and the fibers regrow, they may grow into the wrong place. This most commonly occurs when the glial scaffolding, which normally segregates nerve bundles, is disrupted by trauma or external compression by a tumor. For example, the eye may inappropriately turn in when the patient is trying to look down, or the pupil may inappropriately constrict with depression, adduction, or supraduction of the globe.
LIGHT–NEAR DISSOCIATION: EVALUATION OF THE NEAR RESPONSE
The pupillary contractile response to a near effort is observed as a standard part of the pupil evaluation. If the light reaction seems a little weak, the examiner must check whether the pupils constrict better to near than to light. If they do, this is called light–near dissociation, the causes of which are summarized in Table 203-2 .
How to Test for a Pupillary Near Response
The pupillary near reaction usually is weak when tested in dim light or in the dark. The patient needs a clear view of an object before it can be brought into focus. The examiner must not induce a near response at the slit lamp unless magnification is required to view segmental contractions of the iris sphincter; the room usually is dark and too much equipment is close to the patient’s face in this situation.
The near response is examined in moderate room light, such that the patient’s pupils are midsize and the near object is clearly visible. The patient is given an accommodative target to view—something with fine detail on it. Sometimes a better response is obtained if some other sensory input is added to the stimulus, such as a ticking watch or clicking fingernails; or something proprioceptive for example, the patient’s own thumbnail can be brought into the patient’s view (perhaps, in the case of a child, with a little face drawn on it).
Convergence indicates how hard the patient is trying. The near response, although it may be triggered by blurred or disparate imagery, has a large volitional component, and the patient may need encouragement. If, for some reason, the patient has not made a near effort recently (for example, because stereopsis
TABLE 203-2 — CAUSES OF LIGHT–NEAR DISSOCIATION OF THE PUPIL
Severe loss of afferent light input to both eyes
Anterior visual pathway (retina, optic nerves, chiasm)
Damage to the retina or optic nerve pathways
Loss of pretectal light input to Edinger–Westphal nucleus
Tectum of the midbrain
Infectious (Argyl Robertson pupils) or compression (pinealoma)
Aberrant reinnervation of sphincter by accommodative neurons
Third cranial nerve aberrant reinnervation
Aberrant reinnervation of sphincter by accommodative neurons or medial rectus neurons
is not achieved at near), then a few practice runs may be needed. Often, on the third or fourth try, a good near response is obtained. Sometimes the near effort must be generated for 5 to 10 seconds to see a good near reaction of the pupil. A lack of near response usually indicates that the patient (or the doctor) is not trying hard enough, or not enough time is given at the peak of convergence for the pupil to constrict maximally.
A patient who is completely blind and has no pupillary reaction to light sometimes provides a good near response when asked to “cross your eyes like you did when you were a child.” If a patient cannot produce a near response, the lid closure reflex is tried: the patient faces the examiner with eyes squeezed shut while the examiner tries, with both hands, to hold one of the patient’s eyes open. This often produces a surprisingly strong near response, and is called the eye closure pupil reaction.
Recognition of Light–Near Dissociation
Sometimes it is difficult to establish whether the near response is clearly greater than the light reaction. In such cases, when the examiner faces the patient with pocket light in hand, three levels of light are used:
• Darkness, with a light shining tangential on the pupils from below
• Room light
• Room light with an additional bright light shone in the eyes
With the patient looking in the distance, the bright light is shone in the eye for 1–2 seconds 3 or 4 times, which indicates how small the pupils may become using a light stimulus only.
TABLE 203-3 — CAUSES OF POOR PUPIL DILATATION IN DARKNESS
Past inflammation or surgical trauma
Posterior iris surface or sphincter
Scarring or synechiae of the iris because of past iritis
Prostaglandin release causes sphincter spasm
Adie’s syndrome tonic pupil
Third nerve aberrant reinnervation
Aberrant regeneration of iris sphincter by accommodative or extraocular motor neurons that are not inhibited in darkness
Unilateral episodic spasm of miosis
Postganglionic parasympathetic neuron
Uninhibited episodic activation of postganglionic neurons
Congenital miosis (bilateral)
Loss of inhibition at midbrain from reticularactivating formation
Lymphoma, inflammation, infection
Periaqueductal gray matter
Interruption of inhibitory fibers to the Edinger–Westphal nucleus
Reticular activating formation, midbrain
Narcotics, general anesthetics
Old age (bilateral miosis)
Reticular activating formation, midbrain
Loss of inhibition at midbrain from reticular activating formation
Sympathetic neuron interruption
The near response must not be judged by the addition of a near stimulus to a bright light stimulus, which almost always produces an apparent light–near dissociation, because the near stimulus inevitably adds something to the light stimulus. A real light–near dissociation is present only if the near response (tested in moderate light) exceeds the best constriction that bright light can produce.
POOR PUPIL DILATATION
When one or both pupils stay small and miotic, even in darkness, a number of factors may be responsible ( Table 203-3 ). To better understand the different mechanisms possible it is important to understand the normal process in darkness that allows the pupil to dilate. When a light stimulus is terminated, two mechanisms cause the pupil to dilate. The greater part of pupil dilatation arises from inhibition to the Edinger–Westphal nucleus in the midbrain, which reduces the firing of the preganglionic parasympathetic neurons in the Edinger–Westphal nucleus and results in relaxation of the iris sphincter. Within a few seconds, sympathetic nerve firing increases, which augments the pupil dilatation by active contraction of the dilator muscle. The combined inhibition of the iris sphincter and stimulation of the iris dilator is a carefully integrated neuronal reflex. Therefore, inability of the pupil to dilate in darkness may occur because of a sympathetic nerve palsy, but also from mechanical limitations of the pupil (scarring), pharmacological miosis, aberrant reinnervation of cholinergic neurons to the iris sphincter that are not normally inhibited in darkness (accommodative or extraocular motor neurons), or inhibitory input signal not received by the Edinger–Westphal nucleus.
1. Lowenstein O, Kawabata H, Loewenfeld I. The pupil as indicator of retinal activity. Am J Ophthalmol. 1964;57:569–96.
2. Loewenfeld IE. The pupil: anatomy, physiology, and clinical applications. Ames, Iowa: Iowa State University Press; Detroit: Wayne State University Press; 1993.
3. Levatin P. Pupillary escape in disease of the retina or optic nerve. Arch Ophthalmol. 1959;62:768–79.
4. Thompson HS, Corbett JJ, Cox TA. How to measure the relative afferent pupillary defect. Surv Ophthalmol. 1981;26:39–42.
5. Thompson HS, Corbett JJ. Asymmetry of pupillomotor input. Eye. 1991;5:36–9.
6. Cox TA, Thompson HS, Hayreh SS, Snyder JE. Visual evoked potential and pupillary signs. A comparison in optic nerve disease. Arch Ophthalmol. 1982;100:1603–7.
7. Thompson HS, Montague P, Cox TA, Corbett JJ. The relationship between visual acuity, pupillary defect, and visual field loss. Am J Ophthalmol. 1982;93:681–8.
8. Brown RH, Zillis JD, Lynch MG, Sanborn GE. The afferent pupillary defect in asymmetric glaucoma. Arch Ophthalmol. 1987;105:1540–3.
9. Johnson LN, Hill RA, Bartholomew MJ. Correlation of afferent pupillary defect with visual field loss on automated perimetry. Ophthalmol. 1988;95:1649–55.
10. Kardon RH, Haupert C, Thompson HS. The relationship between static perimetry and the relative afferent pupillary defect. Am J Ophthalmol. 1993;115:351–6.
11. Kawasaki A, Moore P, Kardon RH. Variability of the relative afferent pupillary defect. Am J Ophthalmol. 1995;120:622–33.
12. Kawasaki A, Moore P, Kardon RH. Long-term fluctuation of relative afferent pupillary defect in subjects with normal visual function. Am J Ophthalmol. 1996;122: 875–82.
13. Bell RA, Waggoner PM, Boyd WM, et al. Clinical grading of relative afferent pupillary defects. Arch Ophthalmol. 1993;111:938–42.
14. Haynes Wl, Alward WLM, McKinney K, et al. Quantitation of iris transillumination defects in eyes of patients with pigmentary glaucoma. J Glaucoma. 1994;3: 1106–13.
15. Fison PN, Garlick DJ, Smith SE. Assessment of unilateral afferent pupillary defects by pupillography. Br J Ophthalmol. 1979;63:195–9.
16. Cox TA. Pupillography of a relative afferent pupillary defect. Am J Ophthalmol. 1986;101:320–4.
17. Cox TA. Pupillographic characteristics of simulated relative afferent pupillary defects. Invest Ophthalmol Vis Sci. 1989;30:1127–31.
18. Young RSL, Han B, Wu P. Transient and sustained components of the pupillary responses evoked by luminance and color. Vision Res. 1993;33(4):437–46.
19. Young RSL, Kennish J. Transient and sustained components of the pupil response evoked by achromatic spatial patterns. Vision Res. 1993;33(16):2239–52.
20. Barbur JL, Harlow AJ, Sahraie A. Pupillary responses to stimulus structure, colour, and movement. Ophthalmic Physiol Opt. 1992;12:137–41.
21. Slooter JH, van Noren D. Visual acuity measured with pupil responses to checkerboard stimuli. Invest Ophthalmol Vis Sci. 1980;19:105–8.
22. Ukai K. Spatial pattern as a stimulus to the pupillary system. J Opt Soc Am A. 1985;2:1094–100.
23. Barbur JL, Thomson WD. Pupil response as an objective measure of visual acuity. Ophthalmic Physiol Opt. 1987;7:425–9.
24. Cocker KD, Moseley MJ. Visual acuity and the pupil grating response. Clin Vision Sci. 1992;7:143–6.
25. Kardon RH, Kirkali PA, Thompson HS. Automated pupil perimetry. Pupil field mapping in patients and normal subjects. Ophthalmology. 1991;98:485–96.
26. Kardon RH. Pupil perimetry. Curr Opin Ophthalmol. 1992;3:565–70.
27. 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: 384–7.
28. Cremer SA, Thompson HS, Digre KB, Kardon RH. Hydroxyamphetamine mydriasis in Horner’s syndrome. Am J Ophthalmol. 1990;110:71–6.
29. Weinstein JM, Zweifel TJ, Thompson HS. Congenital Horner’s syndrome. Arch Ophthalmol. 1980;98:1074–8.
30. Thompson HS, Newsome DA, Loewenfeld IE. The fixed dilated pupil: sudden iridoplegia or mydriatic drops? A simple diagnostic test. Arch Ophthalmol. 1971; 86:21–7.
31. Kardon RH, Corbett JJ, Thompson HS. Segmental denervation and reinnervation of the iris sphincter as shown by infrared videographic transillumination. Ophthalmology. 1998;105:313–21.