Chapter 1

Introduction

Visual unresponsiveness, in an otherwise healthy infant, is an alarming finding. Parents are staring down the cannon at a lifetime of potential blindness in their baby and are understandably anxious and inquisitive about the cause, severity, and prognosis of the condition. Depending on the underlying cause, the visual outcome may range from normal vision to complete blindness. The importance of establishing an accurate diagnosis in this setting is obvious. This chapter emphasizes the congenital visual disorders in infancy but includes some discussion of other visual system disorders that may manifest later in childhood.

Decreased vision in infancy is generally due to developmental malformations or acquired lesions of the eyes, anterior visual pathways, or posterior visual pathways. Some causes involving ocular structures are readily identifiable on careful eye examination (e.g., cataracts, corneal opacities, refractive errors). Most congenital retinal dystrophies (e.g., Leber congenital amaurosis [LCA], congenital stationary night blindness [CSNB], achromatopsia) lack conspicuous ophthalmoscopic signs in early infancy (although arteriolar constriction is usually detectable on direct ophthalmoscopy)231 and therefore necessitate electroretinography (ERG) to establish the diagnosis.252 Because rod and cone waveforms mature slowly over the first year of life,183,232,610 we generally wait until 1 year of age to perform ERG. However, some specialized electrophysiology laboratories can derive meaningful informationfromstudiesperformedearlierininfancy.Neurological visual impairment (e.g., cortical visual impairment [CVI]) can also be suspected clinically but requires neuroimaging to confirm.

Mentally retarded or autistic children may appear visually unresponsive despite intact visual pathways. However, physically or mentally disabled children may also have occult ophthalmologic disorders that are difficult to diagnose because of their disability.173 Compared with children who are only blind, blind children who have additional neurologic handicaps show a more marked motor delay in postural and locomotor development and developmental skills such as

reach on sound.148 The diagnosis of disorders causing visual disability in infants and children depends first and foremost on a pertinent clinical history and a thorough examination. The information thus obtained should enable a clinician with a thorough grasp of the various clinical entities that may cause an infant to act blind to formulate a list of differential diagnoses. The correct diagnosis may then be reached using a thoughtful diagnostic paradigm to work up such patients (Fig. 1.1).

Important clues to the cause of blindness in an infant may be derived from various aspects of the ophthalmologic evaluation. Infantile nystagmus is generally absent in children with cortical visual loss but is a common feature in those with congenital ocular or anterior visual pathway disorders. However, it also manifests infrequently in premature children with subcortical injury to the optic radiations (periventricular leukomalacia).286 As discussed in Chap. 8, many children with infantile nystagmus have underlying visual sensory disorders,187,211 even when the eyes appear structurally normal.608 It should be emphasized that the clinical appearance and the electro-oculographic waveforms of infantile nystagmus are identical whether or not a sensory deficit is detectable.

If damage to ocular or anterior visual pathway structures occurs postnatally, the nystagmus is said to appear about 1 month after visual loss and develop only when the visual loss occurs prior to 2 years of age.93 The term “congenital nystagmus” is a misnomer because the nystagmus usually develops between 8 and 12 weeks of age.290 Thus, during the first 2 months of life, the absence of nystagmus (in infants who will subsequently acquire it) eliminates an important diagnostic clue in differentiating an anterior visual pathway disorder from lesions of the posterior pathway. This distinction becomes especially important when dealing with ocular conditions that show minimal ophthalmoscopic signs in early infancy (e.g., LCA) or at any age (e.g., achromatopsia, CSNB).

Infants with nystagmus due to anterior visual pathway abnormalities typically have certain directions of gaze in which the nystagmus is less intense and the vision is better (null points or null zones). Such children may hold their heads at eccentric angles when fixating an object. The presence

of such a preferential head posture usually implies the presence of fixation and functional vision. Unlike head nodding observed in patients with spasmus nutans, the head shaking seen in some patients with infantile nystagmus and poor vision presumably does not prolong foveation time and would not, therefore, be expected to improve vision. However, this is a controversial issue, with some authorities maintaining that the head shaking is a learned, voluntary neurovisual adaptation to improve vision.294 This is supported by the observation that the head shaking is noted during intense visual fixation.294

Infantile nystagmus due to sensory visual dysfunction should be distinguished from “roving” or “drifting” eye movements, the latter implying worse visual function. Roving eye movements are often seen in affixational patients with ocular or anterior visual pathway lesions whose vision is worse than 20/400. They consist of slow, aimless drifting of the eyes back and forth, usually horizontally.290 Jan et al290 likened fixation to an “anchor” without which the eyes “rove” back and forth. Nystagmus may be seen in some patients with roving eye movements when an object is held close enough to allow some fixation, or it may replace roving altogether in those whose vision improves.290

Jan et al290 observed that some characteristics of infantile nystagmus due to anterior visual pathway abnormalities correspond with the age of the onset of visual loss and level

of vision. Thus, nystagmus associated with extremely poor vision and/or vision loss before 6 months of age showed slow velocity and large amplitudes. Roving eye movements may represent one extreme in this continuum. In such infants, it is important to ask about disruption of normal sleep patterns as the absence of light entrainment of circadian rhythms may respond to oral melatonin.

To summarize, nystagmus in an apparently blind infant is a valuable clinical marker for anterior visual pathway disease. Patients with bilateral disorders of the eye or anterior visual pathways may display roving eye movements if the vision is extremely poor with absent fixation; horizontal nystagmus, if the vision is less than 20/70 in the better eye, but fixation is present; or neither, if the vision is better than 20/70. The 20/70 cutoff is somewhat arbitrary, and variations on this are common. It is not unusual to see patients with CSNB, albinism, or blue-cone monochromatism, with visual acuity as good as 20/40, who display nystagmus. It is therefore probably an oversimplification to suggest that the nystagmus is a result of the visual deficit in such patients. The finding of ocular movement abnormalities, including nystagmus, in obligate carriers of blue-cone monochromatism who had visual acuity of 20/20 or better suggests that the nystagmus is intrinsic to the disease and can appear independent of the visual impairment.214 Theoretically, the two traits (the cone disorder and the associated nystagmus)

may be inherited through linked genes rather than a single gene.

The pupillary examination may provide valuable clues to the diagnosis in this setting to the extent that it can be reliably performed in small, uncooperative infants. Infants with blindness due to congenital retinal disorders show sluggishly reactive pupils, whereas the pupillary light reaction is usually spared in patients with pure CVI. A “paradoxical” pupillary response (initial constriction of the pupil to darkness) is classically present in CSNB (an isolated rod dystrophy) and congenital achromatopsia (an isolated cone dystrophy).39,175,470 It is not entirely specific to these disorders, however, and has been occasionally in a variety of developmental optic nerve disorders and even normal eyes.181 Paradoxical pupillary responses are often difficult to detect in infants but become more apparent during the first few years of life. As discussed in Chap. 8, the existence of melano- psin-containing retinal ganglion cells probably explains the paradoxical pupillary phenomenon.

Certain congenital retinal disorders are characteristically associated with high refractive errors: high hyperopia in some types of LCA24,631 and high myopia in patients with CSNB and other congenital retinal dystrophies.346 Children with albinism may have high hyperopia or high myopia. These associations are not constant but are sufficiently frequent to warrant consideration of retinal disorders in a blind infant with nystagmus.

The funduscopic appearance of the infant eye differs sufficiently from that of the adult eye in ways that may cause a diagnostic problem for the clinician. The optic discs of young infants often appear pale even when undue pressure on the globe is avoided while opening the eyelids. In equivocal cases, the presence of asymmetric disc appearance or peripapillary nerve fiber layer dropout may serve to corroborate the impression of genuine disc pallor. The fundus of young infants often has a pale, speckled appearance that may be difficult to distinguish from an abnormal fundus with ophthalmoscopy. Foveal hypoplasia is one of the more difficult causes of decreased vision to diagnose. Although a common feature of ocular albinism and aniridia, it occasionally occurs as an isolated familial disorder.446

Some congenital retinal disorders are associated with various degrees of photophobia (intolerance to light). Most notably, children with congenital achromatopsia display an extreme aversion to light. Marked photophobia may also be seen in achromatopsia, cone–rod dystrophy, and LCA. For reasons that are poorly understood, the extreme photophobia that characterizes achromatopsia may not be apparent early in infancy. Children with optic nerve hypoplasia and dominant optic atrophy are often mildly photophobic. Photophobia and glare arising from corneal or lenticular opacities can be readily classified as such on ocular examination. Even when other ocular disorders (e.g., media opacities, iritis, albinism, aniridia) are excluded, photophobia is not invariably of retinal origin. Photophobia has been reported in a patient with no

light perception.14 It seems plausible that, in patients with deficient photoreceptors, the newly discovered melanopsincontaining retinal ganglion cells may play a role in producing photophobia.57,237 A variety of neurological disorders are associated with photophobia, including meningitis, subarachnoid hemorrhage, migraine, trigeminal neuralgia, thalamic infarct, head injuries, and tumors compressing the anterior visual pathways.126,352,567

CVI can also be associated with photophobia (approximately one-third of patients in the study by Jan et al)293, presumably a result of associated thalamic damage or the cortical lesion itself. The photophobia in most patients is mild, with a tendency to resolve or diminish along with visual and other symptomatic improvement. Jan et al293 did not find a close relationship between the photophobia and the severity of the visual loss or the peripheral field defects. It should be noted that some patients with CVI show a compulsive tendency to gaze at room lights, especially fluorescent lights, or other bright objects, including the sun (light gazing).295 Photophobia and light gazing in patients with CVI are not mutually exclusive, with many children, paradoxically, exhibiting both.293 Jan et al293 feel that “light-gazing is such a compulsive behavior that even the presence of photophobia is not a deterrent.” The presence of mild photophobia cannot, therefore, be used to distinguish a primary retinal disease from cortical visual loss. However, more severe photophobia is highly suggestive of congenital retinal dystrophy.

Some children with very poor vision habitually press their eyes with a finger or a fist. This “oculodigital sign” appears to be specific for bilateral congenital or very early-onset blindness due to retinal disease (most often in LCA but, rarely, in severe cases of retinopathy of prematurity [ROP]).291 It does not occur in children with only one blind eye, irrespective of cause, and is not seen in children with cortical blindness, media opacities, or optic nerve disease. Children who engage in frequent eye poking often exhibit sunken eyes because of orbital fat atrophy. This probably explains the observation that children with CVI do not show deep-set orbits.296 Jan et al291 speculated that eye pressing stimulates the visual cortex by mechanically triggering ganglion cell action potentials (phosphenes). However, we believe that this compulsion will ultimately prove to be a phantom limb syndrome. Eye pressing should be distinguished from eye rubbing and eye poking.292 For example, children who are sleepy tend to rub their eyes, and those with blinding ocular disorders tend to press their eyes, whereas severely mentally disabled children with self-injurious behavior may poke their eyes or even rub their corneas, sometimes with disastrous results.292 We have seen vigorous eye poking in children with Down syndrome leading to dislocation of the crystalline lens. In addition to eye pressing, children with retinal blindness may wave their fingers between their eyes and a bright light (finger waving). Finger waving may also be seen in photoconvulsive epilepsy and autism.

In children with congenital retinal blindness, functional magnetic resonance imaging (MRI) has shown the functioning visual cortex to be capable of supporting novel nonvisual functions through mechanisms of crossmodal plasticity.443 In these children, the occipital cortex participates in semantic processing of spoken language, with greater activation in the left occipital lobe, possibly relating to its spatial proximity to other left hemispheric language areas.469 The visual cortex can be activated by Braille reading in blind people.504 Patients say that they can localize hearing in space, so that their ears become like eyes. Because of cortical plasticity, these patients effectively acquire the ability to “see with their ears.”480 Some children with bilaterally poor vision display a phenomenon termed “overlooking” (Fig. 1.2). Instead of looking at the object of regard directly, affected children look above the object. Initially attributed to relative preservation of the inferior visual field in patients with certain retinal disorders,559 it was later reported to be not disease-specific but rather to represent a sign of bilateral central scotomas (and vision of 20/200 or worse) in children from a variety of causes.203

Nevertheless, most patients who display this sign are found to have congenital retinal disorders. Overlooking may initially be mistaken for lack of either cooperation or comprehension on the part of the patient. Indeed, some children and adults with central scotomata can be trained to “overlook” in order to establish a preferred retinal locus for reading by eccentric fixation.442 Cruysberg et al100 have found that children with overlooking frequently display dystonic posturing, in which the eyes and head are maintained in an upward position. These authors have attributed the overlooking in some of the children with neuronal ceroid lipofuscinosis to tonic upgaze (probably associated with brain stem pathology) rather than greater sensitivity of the lower retina. We believe that bilateral congenital visual loss directly inhibits the cerebellar flocculi which normally exert downward innervation via the vestibulo-ocular pathways. In the absence of macular function, the resulting tonic upgaze is not corrected by fixation.

Blind patients at various ages may experience formed and/or unformed hallucinations.348 In the absence of epilepsy, these have been considered to represent a release

Fig. 1.2  Overlooking. Eight-year-old boy habitually views objects by looking above them. (a) Patient attempting to view target attached to camera’s lens. (b and c) Fundus pictures depict arteriolar narrowing and

mottling of retinal pigment epithelium. He was diagnosed as having rod–cone dystrophy

phenomenon akin to the Charles Bonnet syndrome.519 Release phenomena refer to neurological disorders resulting from maladaptive activity of disinhibited neurons following damage to their source of inhibition.

The fixation pattern of children with congenital visual defects follows a general pattern based on the extent of the vision loss: Children with vision better than 20/200 follow mostly with their eyes, those with 20/200 follow with both their eyes and head, and those with severe vision loss follow mostly with their head.290 Those with severe, congenital loss of vision have difficulty producing willful saccades into any suggested direction, with upward gaze being most markedly limited.290 The suggestion that this limitation in upgaze results because the superior fields are relatively unimportant (and thus infrequently utilized)290 may be valid although this dictum would appear to be more applicable to adults than to children, in whom (because of short stature) much of the world resides in the superior fields.

A preference to view objects of regard at very close range is seen occasionally in normal children, reflecting a transient behavioral pattern. Children with poor vision often do so consistently to produce linear magnification (by shortening the focal length), to damp an existing nystagmus with convergence to improve vision or, in the case of uncorrected aphakic children, to induce a miotic response to increase depth of focus and create a pinhole effect.208 In the hyperopic aphakic child, the effects of linear magnification are more offset than those of increased image blur.

Assessment of vision in children with severe visual deficits requires the utilization of specialized techniques or specific adaptations of techniques commonly used in children.201 The usual qualitative subjective methods of assessing the ability of the child to fix and follow, the steadiness of fixation, and the ability to maintain fixation are not helpful because most such children are affixational. It is practically more relevant to obtain a measure of overall visual function than to simply attempt measurement of distance visual acuity, which is difficult or impossible with severe visual impairment. Snellen acuity charts and similar tests are often useless. Parental accounts of the child’s visual behavior are usually good starting points. Patiently playing with affected children using a variety of toys of different colors and sizes provides useful information. Can the child recognize various objects in the environment, navigate effectively, and interact visually with other people? How does the child respond to the examiner’s face, larger objects, or movement of the parent nearby? How does the child react to penlight or to flickering of room lights on and off? A widening of the palpebral fissures when room lights are extinguished indicates the presence of at least light perception.

A measure of vision can be achieved with the dynamic vestibulo-ocular reflex. This is evoked by holding the infant face-to-face with the examiner at about arm’s length and

spinning the infant around. The infant develops nystagmus as the eyes move counter to the direction of rotation and, at the limit of their excursion, make a quick movement in the opposite direction before the cycle is repeated. When the spinning stops, the inertia of the endolymph in the semicircular canal evokes nystagmus in the opposite direction that is dampened within 5 s in a child with good vision; the nystagmus lasts longer in a blind infant because of poor fixation. Visual function may also be qualitatively evaluated with the optokinetic reflex. When the visual field moves with respect to the eyes, as with a rotating optokinetic drum, the eyes track the moving field to the limit of their excursion and then make a recovery saccade in the opposite direction and so on, producing optokinetic nystagmus. Totally blind infants cannot generate optokinetic nystagmus. It has been estimated that visually impaired patients with horizontal nystagmus who are able to generate an optokinetic nystagmus to a vertically rotating drum should be able to achieve a significant measure of visual independence (i.e., they probably will not be required to be in a school for the blind).

Generally, the vision of infants and children with various neurological diseases (e.g., cerebral palsy, mental retardation) is more difficult to evaluate irrespective of its level.188 Numerous studies have demonstrated a higher prevalence of subnormal acuity in children with cerebral palsy than in agematched controls.188,390 Hertz and Rosenberg247 found that the more severe the physical and neurological disability in children with cerebral palsy, the worse the visual performance on the acuity cards. They reasoned that poorer visual performance may reflect a combination of genuine poor visual function and difficulty in test administration and interpretation in this group of patients. The poorer vision in more severely affected children with cerebral palsy is not surprising, given the diffuse nature of the encephalopathic process. The vision of infants and children with cerebral palsy and mental retardation, with or without severe visual impairment, may be quantitatively evaluated with preferential looking techniques (e.g., Teller acuity cards).130,166,247 These techniques are also useful to document the visual improvement that occurs in some of these patients, especially those with CVI or delayed visual maturation (DVM).218 However, the results of Teller acuity card testing and similar tests of grating acuity should be interpreted with caution.334 Hoyt261 suggested that Teller acuity card methods have low sensitivity in detecting significant visual dysfunction during infancy. Infants who score entirely within the normal range with Teller acuity cards may later be found to have significantly reduced acuity with recognition visual acuity testing (e.g., Snellen acuity).

Visual function in visually impaired children is more accurately evaluated with a combination of Teller acuity cards and a battery of other visual function behavioral tests than by Teller cards alone.132 Citing these and other potential pitfalls, Kushner337 cautioned against using the results of