Damage to the pre-geniculate anterior visual pathway, including the retina and optic nerve, impairs the rate and amplitude of the direct and consensual pupillary light reaction when the affected eye is stimulated. When the opposite eye is stimulated, the direct and consensual response will not be similarly affected unless it too is injured. The degree to which pupillary light reaction is impaired is directly proportional to the degree of damage to the visual neural elements. For example, with complete unilateral optic nerve damage no direct or consensual pupillary response is observed when the injured side is stimulated by light, a sign referred to as an amaurotic pupil. Because of the bilateral projections of the afferent pupillomotor input to the Edinger-Westphal nuclei, the pupils are equal in size in a patient with an amaurotic pupil.
When less-than-complete injury to the afferent visual system occurs, the resulting impairment of the pupillary light reaction is referred to as a relative afferent pupillary defect. The direct pupillary light reaction is characterized by a reduced rate of initial constriction, reduced maximal amplitude of contraction, and early or excessive escape relative to the direct light reaction of the opposite pupil (47, 48). At the beginning of the 20th century, the English ophthalmologist Robert Marcus Gunn (49) described a pupillary sign associated with retrobulbar neuritis whereby the pupil of the affected eye demonstrated secondary dilation when it was continuously stimulated by direct light for several seconds, a phenomenon now referred to as pupillary escape. Clinicians often generalize his original description of this test and refer to the abnormal pupillary reaction associated with disease of the afferent visual system as a Marcus Gunn pupil. I prefer the term "relative afferent pupillary defect", because pupillary escape, by itself, is an insensitive sign of optic nerve disease (50, 51) and because this term more aptly describes the pupillary signs associated with disorders of the afferent visual pathway.
The most sensitive way to detect a relative afferent pupillary defect at the bedside is to compare the direct reactions of both pupils by alternating the light stimulus back and forth between the two eyes serially, a test referred to as the alternating (or swinging) light test (of Levatin) (52). First, ask patients to stare at a distant fixation target to prevent their pupils from intermittently constricting as a result of the near reaction during the test. Then, use a bright hand-held light and stimulate each eye back and forth at a regular interval while observing the reactions of the directly stimulated pupil. Use dim ambient illumination and a bright stimulating light, such as a hand-held Finhoff transilluminator, to facilitate large amplitude and brisk pupillary reactions. Finding the best combination of ambient illumination and stimulating light intensity may require some initial trial-and-error test runs. For example, if the room light is too bright, then the pupils will be small and the amount that they constrict when stimulated will be too tiny to allow you to perceive much difference between the two sides. If the light stimulus is too dim, then the evoked pupillary reactions may be too small, even if the starting sizes of the pupils are appropriate. But, the intensity of the light stimulus, by itself, probably does not influence the size of an existing relative afferent pupillary defect (53).
As you alternately stimulate each pupil, note the rate of contraction, the absolute degree of contraction and, of least importance and hardest to judge, the latency and degree of escape of the stimulated pupil. Any asymmetry of these pupil reactions indicates a relative difference in the pupillary light reflex, a relative afferent pupillary defect. In some cases, injury to the afferent visual system is so extensive that the pupil on the affected side appears to dilate when the light is shifted from the normal eye to the affected eye (Fig. 9). This occurs because of the physiologic principle that the size of the stimulated pupil is proportional to the degree that light energy is conducted through the afferent visual system. The size of this pupil (its consensual response) when the opposite normal eye is stimulated is small. When the light is shifted to stimulate the affected pupil, not as much energy is conducted through the injured afferent visual system, so that the size of this pupil now must become relatively larger to reach its new steady state stimulated size (its direct response) based on impaired conduction.
Fig. 9 Right-sided relative afferent pupillary defect in a man with optic nerve glioma. When the unaffected left eye is stimulated by light, both pupils constrict (top). When the light is then swung over to the affected right eye, both pupils dilate (bottom). This indicates that pupillomotor conduction through the right optic neve is markedly impaired.
Thompson and Corbett (54) have outlined the preferred technique of performing the swinging flashlight test. The examiner should take great care to hold the light before each eye for the same brief period so that the retina of one eye, especially one that has a suspicious pupillary reaction, does not become bleached to light and develop an afferent defect on that basis alone. A judgment regarding whether any asymmetry between the light reactions of the two pupils exists should be made within the first five cycles to reduce the chances of inducing asymmetry of retinal bleach between the two eyes. If one cannot make a judgment at that point, stop the test and allow several seconds to pass so that each retina can recover before starting another set of alternating light sequences. Shining the light into both eyes for a few seconds will also ensure that both retinas are equally adapted before starting the test again.
The rate of alternating the light between the two eyes may help to isolate certain aspects of the light reaction that assist one in identifying an afferent pupillary defect. Stimulating each eye for one second is more sensitive for detecting asymmetry of the initial rate and amplitude of constriction of the pupils and is generally the preferred technique (54). Stimulating each eye for 3 seconds allows one to detect asymmetry in the appearance and amplitude of pupillary escape, an endpoint generally considered to be less sensitive and specific for identifying an afferent pupillary defect than observing characteristics of the initial pupillary light reaction. In some patients, using both time intervals may be useful for assessing the presence of a relative afferent pupillary defect. Regardless of which time interval is used for light stimulation, it is important to quickly swing the light from one to the other eye to avoid a long dark interval. When the dark interval is too long, the light reactions become too variable (55).(Video 1 )
Video 1. Right-sided relative afferent pupillary defect in a patient with ischemic optic neuropathy.
A clinician can test for an afferent pupillary defect even if one iris does not react normally due to an ocular disorder (e.g., synechiae), Adie's pupil, or pupil-involving third cranial nerve palsy. Because the direct and consensual efferent pupillary light responses are generally equal, the examiner can use the pupil of the eye with a normal iris as an indicator of afferent conduction when either eye is stimulated (Fig. 10).
Fig. 10 Assessment of an afferent pupillary defect when only one iris is functional. In this example, a right-sided parasellar tumor is compressing both the optic and oculomotor nerves, causing an optic neuropathy and a pupil-involving third crainial nerve palsy. The pupil on the affected side has both an afferent as well as an efferent defect. The room is darkened and the examiner uses a dim light to illuminate the unaffected left pupil from below. The right-sided afferent defect is detected by observing that the left pupil dilates when the right eye is stimulated (top), and then constricts when the left eye is stimulated (bottom), during the swinging flashlight test.
The size of the relative afferent pupillary defect is generally proportional to the amount of damage to the retina or optic nerve axons, reflected in the amount of associated visual field loss. In clinical practice, the amount of afferent pupillary defect correlates best with the degree of visual field impairment and is poorly correlated with other measures of visual sensory function, such as visual acuity (22-24). For example, consider two patients. The first has a central retinal artery occlusion associated with a ciliary retinal artery that preserves foveal circulation so that their visual acuity is reduced only to 20/30. They have marked loss of visual field outside of their preserved central vision. The second has anterior ischemic optic neuropathy associated with visual acuity of 20/400 and a small but deep central scotoma. The first patient will have a much larger relative afferent pupillary defect than the second patient, despite better visual acuity.
Inasmuch as an afferent pupillary defect occurs only when there is imbalance of pupillomotor input between the two eyes, this sign is not present in patients with disorders causing symmetric bilateral injury to the afferent visual system, such frequently occurs in retinitis pigmentosa or papilledema. If a disorder causes bilateral, but asymmetric, injury to the afferent visual pathway, then an afferent pupillary defect is found in the most affected eye. For example, parasellar tumors that compress the optic chiasm frequently cause asymmetric visual field loss and commonly produce an afferent pupillary defect despite the presence of bilateral visual field loss.
In addition to ophthalmoscopic observations, a relative afferent pupillary defect is the only other objective sign of anterior visual pathway dysfunction available to the neurologist evaluating a patient complaining of visual loss. The ability to quantify this defect makes assessment of an afferent pupillary defect a powerful objective tool that can be used to follow the course of diseases affecting the retina, optic nerve, and chiasm. Knowing the size of an afferent pupillary defect, and comparing it with other measures of the visual sensory examination, will often reveal additional information regarding the site and nature of the disturbance responsible for a patient's visual loss. For example, if a patient with sudden visual loss is found to have a small (e.g., 0.3 log) relative afferent pupillary defect and no ophthalmoscopic abnormality of the optic disc, then the examiner should direct particular attention to that patient's macula as the responsible location for the visual loss. As another example, consider a patient with neurologically isolated homonymous hemianopia and a small afferent pupillary defect (e.g., 0.6 log). The presence of this defect almost certainly guarantees that the responsible lesion resides in the optic tract, not the occipital lobe.
Quantification of an afferent pupillary defect can be determined using neutral density filters that can be obtained through most photographic equipment suppliers (Fig. 11 ). When an impaired pupillary reaction is found in one eye relative to the other during the alternating light test, a small strength neutral density filter is then placed in front of the unaffected eye to reduce the intensity of light input into that eye when stimulated by light (56). The alternating light test is then repeated. If the afferent defect is not neutralized, successively larger filters are placed before the unaffected eye until the defect of the affected eye is balanced by the induced defect of the unaffected eye, at which point no difference in the pupillary light reaction between the eyes may be identified. If the density of filters required to balance the defect is so large that it becomes difficult to view the pupil of the unaffected eye behind them, then simply watch the movements of the affected pupil, which is not covered, behind the filters to determine the balance point. Use a dim light held below eye level or off to one side to illuminate the pupil so its movements can be easily observed. The size of the defect is determined by adding the amount of neutral density filters required (in log units) to balance the pupils, with an additional 0.1 log added whenever two filters are stacked together.(Video 1 )
Fig. 11 Hand-held equipment used to measure a relative afferent pupillary defect and to record pupil sizes. Four neutral density filters (0.3, 0.6, 0.9, 1.2 log units) are conveniently carried in a soft cloth carrying pouch. A bright light source (a Finhoff model illuminator is shown here) is ideal for stimulating the direct pupillary response. A pupil gauge is necessary to quantify pupil diameters when evaluating anisocoria.
Like most biologic systems, the pupillary light reflex contains a certain amount of so-called noise, which contributes to variability. Because of this inherent noise, the size of a relative afferent pupillary defect in a patient with a static afferent visual deficit may vary by as much as roughly 0.3 log from examination to examination (55, 57). Use of computerized infrared pupillography, although currently impractical for most office settings, provides a method for further refining the reliability of measuring a relative afferent pupillary defect.
One can qualitatively grade the relative afferent pupillary defect using a predetermined and standardized set of criteria based on certain characteristics of the pupillary light reaction. For example, the system advocated by Bell and colleagues (58) incorporates the following levels of abnormal pupillary defects identified during the alternating light test: grade 1, weak initial constriction and greater redilation; grade 2, initial stall before constriction and greater redilation; grade 3, immediate dilation; grade 4, immediate dilation with a secondary constriction following prolonged illumination of the good eye for 6 seconds; grade 5, immediate dilation with no secondary constriction, or no response to light, following prolonged illumination of the good eye for 6 seconds.
Although I am a strong advocate of quantifying afferent pupillary defects, I acknowledge that the use of a standardized grading scheme is a marked improvement over making no judgment of the size of the afferent pupillary defect, or simply eyeballing the pupillary responses and subjectively recording it as, for example, "a 2+ relative afferent pupillary defect". When properly performed, a standardized grading scheme defines afferent pupillary defects that roughly correlate with their size when quantified using neutral density filters (58). The so-called eyeball technique of estimating the size of an afferent pupillary defect is inherently imprecise because the examiner is usually comparing only the amplitude of constriction of the two pupils. For example, a large afferent pupillary defect in a patient with small pupils will be underestimated, and a small afferent pupillary defect in a patient with large pupils may be overestimated.
One advantage of measuring a relative afferent pupillary defect with neutral density filters is that a numeric value can be assigned to this sign. This value can then be compared with so-called rules that have been established through experience regarding the expected size of a relative afferent pupillary defect in certain disorders affecting the afferent visual pathway.
There should be no significant asymmetry of the pupillary light reflex in normal subjects. Although a small (i.e., 0.3 log or less) relative afferent pupillary defect can be found in some normal subjects using computerized infrared pupillography, this should be considered a laboratory phenomenon and not a rule that should be applied in clinical practice (57). Finding an afferent pupillary defect demands further testing of visual function.
No relative afferent pupillary defect is seen in a patient whose visual loss is due to hysteria or who is malingering. If this pupillary sign is detected in a patient originally suspected of having functional visual loss, then the examiner must consider that an organic process is either responsible for, or contributing to, the patient's presentation.
If one looks carefully, a small defect usually measuring no more than 0.5 log will be identified in roughly one half of patients with amblyopia (59, 60). The size of the defect does not correlate with visual acuity. For all practical purposes, however, when an obvious relative afferent pupillary defect is identified in a patient with visual loss, with or without a history of amblyopia, a neurologist should generally assume that some disorder other than amblyopia is responsible.
A moderate-size afferent defect occurs in the eye opposite a completely patched or occluded eye (61). The retina under cover becomes dark adapted relative to the light-adapted retina of the unoccluded eye. A relative afferent pupillary defect is produced in the open eye because the light-adapted retina is relatively less sensitive to direct light stimulation than the dark-adapted retina. The size of this defect can be as large as 1.5 log if an eye is fully occluded from any light perception (62). This is an important clinical rule to remember when examining a patient whose eye is occluded due to ptosis or occluded with a patch for ocular reasons. This pupillomotor defect resolves about 10 minutes after removing the occlusion.
Most lesions of the retina confined to the macula do not produce an obvious relative afferent pupillary defect unless the visual acuity if markedly impaired (63). Maculopathies are generally not associated with afferent pupillary defects larger than 0.6 log. Thus, if one measures a large defect in a patient with a macular lesion, the examiner should consider an additional process (e.g., optic neuropathy) responsible for the patient's visual loss. Central serous chorioretinopathy, a common maculopathy in young adults whose clinical features may mimic those of optic neuritis, is often associated with a very small afferent pupillary defect (i.e., 0.4 log or less) during its acute stage that will disappear as the maculopathy resolves (64). With retinal detachment, expect a 0.3 log defect for each quadrant of affected retina, and an additional 0.6 log defect if the macula is detached (65). A central retinal artery occlusion produces a huge afferent defect, typically greater than 3.0 log, because of the extensive amount of nerve fiber layer infarction in this condition. In cases of bilateral retinopathies--such as retinitis pigmentosa, cone dysfunction syndromes, or retinal toxicity--an afferent pupillary defect generally not found (66).
Nearly all acute or progressive unilateral optic neuropathies are associated with an afferent pupillary defect, the size of which roughly correlates with the degree of impaired visual field, not visual acuity (22-24). Detecting a change in the measured afferent pupillary defect of a patient with a lesion compressing her or his optic nerve is typically more sensitive than detecting a change in the size of the lesion using neuroimaging. In the case of optic neuritis, the presence of an afferent pupillary defect is at least as sensitive as detecting prolonged visual evoked potential latency response in the affected eye (67). Furthermore, an afferent pupillary defect usually persists even after the vision has recovered, just as does the abnormal visual evoked potential.
Tumors or aneurysms that compress the optic chiasm frequently produce differences in the degree of temporal visual field loss in both eyes. Accordingly, an afferent pupillary defect is often found in the eye with the more extensive loss of visual field, which may not necessarily be the eye with the worst visual acuity (68). Looking for an afferent pupillary defect, or a change in the measurement of a preexisting defect, is a clinically useful way to identify progression of tumor injury of the anterior visual pathway and additional optic nerve compression when following patients with mass lesions. Loss of visual sensory function and progression of measured afferent pupil defect will frequently identify progressive loss of function before a neuroimaging study demonstrates growth of the underlying tumor when the anterior visual pathway is compressed.
As a result of the slightly greater number of crossed relative to uncrossed fibers that are present in each optic tract, complete (and sometimes incomplete) lesions of the optic tract are often associated with an afferent pupillary defect in the eye with the temporal visual field defect (i.e., the eye contralateral to the side of the lesion in which the nasal crossing fibers originate). The size of this defect is 0.3 to 0.6 log (69). This so-called clinical rule can be used to predict the presence of lesions in the optic tract, not the occipital lobe, in a patient with a complete homonymous hemianopia who has no other localizing neurologic symptoms and signs. Knowledge of a potential optic tract lesion is important because they are often tumors or other mass lesions that may require further focused neuroimaging.
Two other pupillary signs associated with optic tract disorders should be mentioned, although they are apocryphal and primarily of historical, not clinical, interest. Early in the 20th century, the German ophthalmologist Carl Behr noted that the pupil ipsilateral to the side of a homonymous hemianopia in a patient with an optic tract lesion tended to be a bit larger than its fellow pupil. The so-called Behr pupil has not been uniformly observed in modern series of carefully examined patients with optic tract lesions (70-72). The pupillary hemiakinesia reaction of Wernicke, described by the German neurologist Karl Wernicke in the late 1800s, refers to the relative unreactivity of either pupil when a light stimulus is projected into the side of the retina that projects the axons into the damaged tract, and the preserved light reaction when the opposite, intact, hemiretina is stimulated. This sign is not clinically useful because light scatters throughout the retina regardless of which hemiretina one tries to stimulate using most hand-held light sources (71).
A lesion involving the pretectal nucleus or the brachium of the superior colliculus can interrupt the pupillomotor afferent fibers after they have left the optic tract, resulting in a small (i.e., 0.6 log or less) contralateral afferent pupillary defect in the absence of any visual field impairment (Fig. 12) (73-77). Unilateral, or asymmetric bilateral, contraction anisocorias may rarely be a sign of a midbrain lesion, although they are usually too small to be appreciated at the bedside (28, 78).
Fig. 12 Magnetic resonance image of an enhancing bladder metastasis involving the tectum of the midbrain of a 56-year-old man who developed double vision resulting from skew deviation and divergence insufficiency. He also had a left-sided relative afferent pupillary defect measuring 1.4 log units but showed normal visual acuity, color vision, and visual fields when tested using automated perimetry.
Given that afferent pupillomotor fibers leave the optic tract before the lateral geniculate nucleus, no afferent pupillary defect occurs with lesions of the lateral geniculate nucleus or the retrogeniculate pathway. Some disorders that affect the lateral geniculate nucleus (e.g., anterior choroidal artery infarction) or optic radiations (e.g., infiltrative tumor), however, may also involve the adjacent optic tract so that a relative afferent pupillary may be seen contralateral to the side of the lesion. Simultaneous involvement of the optic tract is probably also the explanation for why a small relative afferent pupillary defect is identified in some patients with homonymous hemianopia due to a retrogeniculate lesion (79). In some patients with congenital lesions involving the optic radiations or the occipital lobe, a contralateral relative afferent pupillary defect is observed, reflecting retrograde transsynaptic degeneration of the optic tract fibers (80).