To determine whether rates of strabismus and associated visuomotor deficits differed among children with different severities of periventricular leukomalacia (PVL).
Retrospective, case-control study.
Brain magnetic resonance images (MRI) obtained from 98 children aged ≥2 years were analyzed using a standardized scoring system: 67 of 98 had PVL (mean GA 31 weeks) and 31 of 98 did not have PVL (mean GA 29 weeks). Severity of PVL was scored as degree of damage to the posterior optic radiations and the splenium of the corpus callosum on MRI. Ophthalmologic examination data were collated to assess the prevalence of visuomotor deficits and the relationship to PVL severity (grades 1-3, mild to severe).
Infantile strabismus was documented in 61% of children with mild, 74% with moderate, and 88% with severe PVL (esotropia: exotropia ratio 3.5:1). Associated ocular motor deficits also increased systematically with PVL severity: latent (“fusion maldevelopment”) nystagmus (20%, 47%, and 40%, respectively), dissociated vertical deviation (13%, 28%, and 30%), and nasotemporal pursuit/optokinetic nystagmus asymmetry (23%, 38%, and 54%). Additionally, the prevalence of retrograde optic neuropathy increased with PVL severity (5%, 26%, and 38%). The prevalence of each of these signs was substantially lower in children who had no PVL.
Children who suffer PVL are likely to develop the deficits of the infantile strabismus complex. The deficits tend to increase systematically as a function of PVL severity. These findings provide evidence that infantile strabismus is linked to perinatal damage to cerebral vergence and gaze pathways.
P eriventricular leukomalacia (PVL) is the most common form of brain damage in premature infants. , The term (from Greek is leukos = white + malakia = soft) describes enlargement of the lateral ventricles and their posterior horns resulting from loss of white matter. , The injury is characterized by focal necrosis of neurons and diffuse loss of myelin within the optic radiations, which travel along the outer wall of the lateral ventricles. , White matter tracts communicating information between the right and left visual cortex also course through the splenium of the corpus callosum.
Eye alignment is dependent upon the sensorimotor circuits of the optic radiations: the geniculostriate visual pathway. Monocular axons from the right and left eyes project from the lateral geniculate nucleus to layer 4C of striate visual cortex or area V1. The monocular neurons of layer 4C in turn conjoin to make binocular neurons of layer 4B. Non-correspondence of monocular inputs generates a binocular disparity signal in V1. The disparity signals feed to extra-striate cortical regions and in turn to brainstem motoneurons to drive vergence eye movements, which in normal brains restore binocular correspondence. ,
Periventricular leukomalacia damages geniculostriate visual axons and their myelin sheaths, impairing signal conduction. Degradation of conduction through these axons would be expected to degrade the fidelity of binocular disparity-driven vergence and conjugate gaze signals. Maldevelopment of disparity vergence is hypothesized to be a major cause of early onset strabismus. ,
To quantify the association between PVL and strabismus, this study collated the clinical findings of children born prematurely who had brain magnetic resonance imaging (MRI) and ocular motor examinations. The first question that was asked was whether rates of strabismus differed between groups of premature infants who had or did not have neuroimaging evidence of PVL. The second question that was asked was whether rates of strabismus and other visuomotor deficits differed among infants with different severities of PVL.
The medical records were retrospectively reviewed of 98 children examined in the Ophthalmology Department at St Louis Children’s Hospital who had a history of premature birth and a brain MRI performed. The review collated the presence of strabismus and associated ocular motor signs, as well as any documented optic neuropathy or stage 3 retinopathy of prematurity (ROP3). Information regarding neurodevelopmental disorders was also extracted. Brain MRI data were then analyzed for the presence or absence of PVL (see below). The MRI examiners were masked to the presence or absence of any clinical signs. On the basis of the MRI results, the study group was divided into two cohorts: those with PVL (67 children) versus those without PVL (31 children). The study protocol complied with the Association for Research in Vision and Ophthalmology resolution on the use of human subjects in research and was approved by the Institutional Review Board of the Washington University School of Medicine. Health Insurance Portability and Accountability Act requirements were also met.
INCLUSION AND EXCLUSION CRITERIA
Inclusion criteria were: (1) birth at gestational age ≤36 weeks (the standard definition of prematurity); (2) an MRI performed at St Louis Children’s hospital beyond 2 years of age (to avoid difficulty in scoring due to incomplete myelination); (3) a minimum of 2 comprehensive ophthalmological examinations, including orthoptic measurements after age 6 months; and (4) a quantitative measure of visual acuity (listed further in Methods below). Exclusion criteria were: (1) any history of progressive neurologic or progressive cerebral white matter abnormality (eg, mitochondrial or other metabolic or genetic disease); (2) paralytic or restrictive strabismus (eg, cranial nerve palsy or Duane co-contraction syndrome); (3) funduscopic abnormality of the retina or optic nerve not attributable to prematurity and/or PVL (eg, coloboma, chorioretinal atrophy, albinism, or other retinal maldevelopment/dystrophy); and (4) any anterior segment or adnexal eye abnormality (eg, developmental cataract, glaucoma, corneal opacity, or blepharoptosis).
A history of neonatal intraventricular hemorrhage (IVH) or neonatal hydrocephalus was not included in the database. The focus was on the presence of PVL after 2 years of age. It is reasonable to assume that some infants in this study had a history of IVH. About 3% of preterm infants with a history of no IVH or mild-moderate IVH have PVL, while 17% with a history of severe IVH have PVL and a slightly higher percentage of hydrocephalus.
Ophthalmic and orthoptic assessments were conducted, including: age-appropriate and cognitive level-appropriate testing of best corrected distance visual acuity (CDVA) in each eye; pupillary examination (for anisocoria, iridoplegia, or afferent defects); sensorimotor examination of eye alignment/eye movement/binocular fusion; cycloplegic refraction with manual and, when feasible, automated retinoscopy; slit lamp biomicroscope evaluation of the anterior segment; measurement of intraocular pressure by applanation (Tonopen XL); indirect ophthalmoscopy with a 15-diopter (D) lens for high-magnification assessment of optic disc/foveal anatomy; and a 22-D lens for assessment of the peripheral fundus. Optic neuropathy was defined as: segmental/sectoral or diffuse pallor; enlarged optic cup; or reduced optic disc diameter (ie, retrograde optic neuropathy). Visual acuity (best corrected) was monocularly quantified for each eye using optotype (ETDRS letter, HOTV, or Allen-type figures) testing when feasible. Alternatively, spatial-sweep visually evoked potentials or forced preferential looking methods were employed (Teller grating acuity or Cardiff Acuity Test cards). Visual acuity at last examination was reported. Age of onset was extracted from medical records or parental history. Collated ocular motor measures were those deemed most accurate across all office visits; the type of strabismus recorded was that present before any strabismus surgery. Ocular motor signs of the infantile strabismus complex were identified as follows:
Latent (“fusion maldevelopment” ) nystagmus
Measured under conditions of monocular viewing while fixating a stationary 1° or 2° target, nasalward slow-phase drifts of eye position with respect to the fixating eye, interrupted by temporalward saccadic fast phases; the direction of the nystagmus inverts 180° instantaneously with a change of fixating eye, also usually evident, but more subtle, when viewing binocularly (manifest latent nystagmus); and the slow-phase waveform is linear or of decreasing velocity on eye movement tracings.
Dissociated vertical deviation
Upward deviation of an eye without a corresponding hypotropia of the fellow eye; occurrs spontaneously or evoked by cover test; the upwardly deviating eye is also excyclotorted; and occurs when vertical vergence is activated to damp the minor cyclotorsional and vertical components of horizontal latent nystagmus. ,
Pursuit/optokinetic nystagmus asymmetry
Measured under conditions of monocular viewing with each eye alternately elicited by a 1° spot or 2° physical target; for pursuit, sustained horizontal smooth eye-tracking for nasalward target motion, but subnormal tracking for temporalward motion, evident as the eye-lagging target position and substitution of catch-up saccades; for optokinetic nystagmus (OKN) elicited by a 90° by 90° large field projected image of vertically oriented 80% contrast 5° stripes, a reduced number of sustained slow phases, and nystagmus fast phases for temporalward stripe motion; the direction of the asymmetry inverts 180° instantaneously with a change of fixating eye; and eye movement tracings reveal consistently lower smooth eye velocities for temporalward motion.
Descriptions of the testing and grading of these visuomotor and sensory signs have been described in previous reports.
BRAIN MRI EXAMINATION AND SCORING
All of the children had a Siemens Magnetom 1.5 Tesla brain MRI performed. The exam consisted of axial and coronal T1, T2, and fluid attenuated inversion recovery images of the entire brain. In the majority of patients, a 3D technique was employed (1 mm slice thickness) and in the remainder a 2D technique (2-4 mm slice thickness). The MRI scans were scored by two independent examiners and pooled to obtain a consensus score. The grading algorithm ( Table 1 ) scored PVL as: (1) white matter signal abnormality over the course of the posterior optic radiations (ie, peritrigonal radiation and occipital horn); (2) white matter loss; (3) ventricular dilation; and (4) thinning of the splenium of the corpus callosum. Figure 1 shows representative brain MRI and example scores for each of the 4 gradings. For measures 1 to 3, the right and left sides of each child’s brain were scored independently and the worst score was used for that child; right and left scores were symmetrical in all but 4 of 67 children (6%). The PVL global score was defined as the sum of the 1 to 4 scores. The global score ranged from 0 (no white matter injury) to 11 (most severe white matter injury). Using these global scores, 3 groups of severity of PVL were defined: mild = scores 1 to 3; moderate: scores 4 to 7; and severe scores 8 to 11.
|White Matter |
|0 = normal signal |
1 = ≤25% white matter affected
2 = 25%-75% affected
3 = >75% affected
|White Matter Loss ||0 = normal |
1 = <33% mild reduction of white matter
2 = 33%-66% moderate reduction
3 = severe reduction: gray matter abuts
|Splenium Injury||0 = normal |
1 = mild thinning
2 = moderate thinning
3 = splenium not visible
|Ventricular Dilation |
At Occipital Horn
|0 = no dilation |
1 = moderate dilation
2 = severe dilation
Study children were compared for each of the following clinical variables: birth gestational age; birth weight; racial category; age at MRI; age at most recent eye examination; CDVA; spherical equivalent refractive error; developmental delay; cerebral palsy; and cerebral palsy Gross Motor Function Classification Scale severity score. , , Visual acuities were converted to logMAR for calculation of geometric means. The presence or absence of each of the clinical signs of the infantile strabismus complex (or constellation) were also tabulated: strabismus; latent nystagmus; nasotemporal pursuit/OKN asymmetry; and disassociated vertical deviation (DVD). Apart from ocular motor signs, the presence of optic neuropathy and ROP3 was also compiled for each child. The presence of these signs was compared between children with and without PVL. For children with PVL, comparison was also conducted for global PVL score versus ocular motor sign, optic neuropathy, or ROP3. The linear regression of gestational age and PVL score conformed to standard statistical methods. , Means for parametric data were compared with one-tailed or two-tailed paired t tests. Ordinal and dichotomous variables were compared with the Fisher exact test. Significance was defined as P < .05.
The birth weight and age, racial distribution, age at examinations, visual measurements, and neurodevelopmental disorders in the children with and without PVL are listed in Table 2 . The mean gestational age at birth for the children with PVL (31.0±3.1 weeks) was comparable with that of the children without PVL (28.9±3.6 weeks). Within the PVL group there was a non-significant correlation ( r = 0.22) between gestational age and PVL score, reinforcing that factors more important than gestational age contribute to PVL. There were no significant differences between the PVL and no-PVL groups when comparing: birth weight; racial distribution; age at MRI; age at most recent eye examination; and spherical equivalent refractive error.
|Characteristic||PVL||No PVL||P Value|
|Number of Children||67||31||–|
|Birth Gestational Age (weeks)||31.0±3.1||28.9±3.6||.81|
|Birth Weight (grams)||1885±1006||1101± 35||.06|
|Sex (% female)||51%||55%||.80|
|Other or Mixed Race||6.3%||3.8%||.08|
|Age at MRI (years)||7.1±4.4||6.7±3.2||.54|
|Age Recent Eye Exam (years)||8.6±2.8||7.6±3.9||.44|
|CDVA Recent Eye Exam (logMAR; Snellen)||0.51±0.38; |
|SEQRE (D)||+ 0.60±0.9||+ 0.51±4.6||.15|
|Cerebral Palsy GMFCS Score||2.2±0.9||1.7±0.5||.41|