Cortical Visual Impairment in Congenital

Cytomegalovirus Infection
Haoxing Douglas Jin, MD; Gail J. Demmler-Harrison, MD; Jerry Miller, MS, PhD; Jane C. Edmond, MD;
David K. Coats, MD; Evelyn A. Paysse, MD; Amit R. Bhatt, MD; Kimberly G. Yen, MD;
Joseph T. Klingen, MD; Paul Steinkuller, MD; for The Congenital CMV Longitudinal Group

ABSTRACT Conclusions: CVI may result from symptomatic con-

Purpose: To describe the presentation, evolution, and
long-term outcome of cortical visual impairment (CVI)
in patients with symptomatic congenital cytomegalo-
virus (CMV) infection, and to identify risk factors for the
development of CVI in patients with symptomatic con-
genital CMV.
genital CMV infection. The relationship of CVI and its risk
factors in patients with CMV suggests the potential to
predict the development of CVI through predictive mod-
eling in future research. Early screening of CVI in children
born with symptomatic congenital CMV can facilitate
educational, social, and developmental interventions.

Methods: Retrospective subanalysis of a long-term [J Pediatr Ophthalmol Strabismus. 2019;56(3):194-202.]

prospective cohort study with data gathered from 1982
to 2013.
Results: Eleven of 77 (14.3%) patients with symptomatic
CMV, 0 of 109 with asymptomatic CMV, and 51 control
patients had CVI. Overall, patients with symptomatic
CMV had worse vision than patients with asymptomatic
CMV, who in turn had worse vision than control patients.
Microcephaly, intracranial calcification, dilatation of ven-
tricles, encephalomalacia, seizure at birth, optic atrophy,
chorioretinitis/retinal scars, strabismus, and neonatal
onset of sensorineural hearing loss were risk factors as-
sociated with CVI.
INTRODUCTION
Cortical visual impairment (CVI) is the most
common cause of visual impairment in children in
the developed world.1 It is characterized by visual
impairment or impaired functionality in performing
vision-guided tasks due to damage to the central ner-
vous system (CNS) that does not involve the ocular
structures.2,3 It is most commonly caused by perina-
tal or postnatal injuries to the developing brain by in-
sults such as brain malformation, hypoxia/ischemia,
prematurity, trauma, infection, ventriculo-peritoneal

From the Departments of Pediatric Ophthalmology (JCE, DKC, EAP, ARB, KGY, PS) and Pediatrics, Section of Infectious Disease (HDJ, GJD-H, The
Congenital CMV Longitudinal Study Group), Baylor College of Medicine, Houston, Texas; Texas Children’s Hospital, Houston, Texas (HDJ, GJD-H, JM,
JCE, DKC, EAP, ARB, KGY, JTK, PS, The Congenital CMV Longitudinal Group); and the Mitchel and Shannon Wong Eye Institute, Dell Medical School
at the University of Texas–Austin, Austin, Texas (JCE).
Submitted: October 25, 2018; Accepted: February 14, 2019
Supported in part by the CMV Research Fund Donors at Baylor College of Medicine; the Woman’s Hospital of Texas Research Foundation; the Office
of Research Resources and the General Clinical Research Center for Children at Texas Children’s Hospital and Baylor College of Medicine (NIH 5M0I
RR00188-33); the Mental Retardation Research Center at Baylor College of Medicine (NIH-CHHD 5 P30 HD24064P); Research to Prevent Blindness,
Inc., New York, NY; the Deafness Foundation, Houston, TX; the Vale Ashe Foundation, Houston, TX; the Maddie’s Mission Foundation, Katy, TX; the
Naymola Charitable Foundation, Beaumont, TX; the American Pediatric Society-Society for Pediatric Research Summer Student Research Program (NIH-
CHHD); and Merck & Co. and the Centers for Disease Control and Prevention (Cooperative Agreement FOA IP 10-006).
The authors have no financial or proprietary interest in the materials presented herein.
The authors thank The Congenital CMV Longitudinal Group (members listed in Table A, available in the online version of this article) and the study
patients and their families for their lifetime of dedication and support for this study.
Correspondence: Gail J. Demmler-Harrison, MD, Texas Children’s Hospital, Feigin Center Suite 1150, 1102 Bates Street, Houston, TX 77030. E-mail:
gdemmler@bcm.edu
doi:10.3928/01913913-20190311-01

194 Copyright © SLACK Incorporated

shunt blockage, drugs, and certain neurological dis-
eases.1,3-5 Patients diagnosed as having CVI often
experience severe central vision loss, despite having
otherwise normal eye examinations.6 Reports on in-
trauterine infection as potential causes of CVI are
scarce.7
The World Health Organization estimates 39
million people are blind and 246 million are visu-
ally impaired worldwide.8 Children younger than
14 years make up 3.6% (1.42 million) of the to-
tal blind and 7.7% (18.9 million) of the visually
impaired populations.8 In the United States, CVI,
retinopathy of prematurity (ROP), and optic nerve
hypoplasia/atrophy have been found to be the top
three causes of blindness and visual impairment in
children.9-12 It is estimated that 23.6% of all child-
hood blindness is caused by CVI.12
Cytomegalovirus (CMV) is the most common
congenital viral infection in the United States, af-
fecting approximately 0.5% to 1% (approximately
40,000) of all live births annually.13 It is transmit-
ted transplacentally from the pregnant mother to
the unborn child via primary infection or viral
reactivation. Approximately 10% to 15% of new-
borns with congenital CMV infection will have one
or more signs (eg, jaundice, hepatosplenomegaly,
low birth weight, microcephaly, seizure, purpura,
or chorioretinitis) at birth, and thus are classified
as having symptomatic infection, with the remain-
ing newborns having no signs or symptoms at birth
and classified as asymptomatic.13 Congenital CMV
infection is known to cause ophthalmologic, au-
diologic, and other neurodevelopmental sequelae
in patients.14-16 Most of the publications on vi-
sion and congenital CMV infection focus on the
ocular sequelae, such as chorioretinitis, optic nerve
atrophy, and strabismus. There has been a limited
investigation on CVI and congenital CMV infec-
tion.14,15,17 Although CVI is known to occur among
patients with symptomatic congenital CMV, there
is no method to accurately predict which patients
will develop CVI at birth. The presentation, evolu-
tion, risk factors, and long-term outcome of CVI in
congenital CMV infection have not been fully de-
scribed. Because each child with CVI may present
differently, management should be fitted to each
child’s unique functional status; furthermore, early
detection of CVI in patients with congenital CMV
allows timely intervention to maximize functional
vision over time.
Journal of Pediatric Ophthalmology & Strabismus • Vol. 56, No. 3, 2019
PATIENTS AND METHODS
The purpose of this study was to describe the
presentation, evolution, and long-term outcome of
CVI in patients with symptomatic congenital CMV
infection, and to identify risk factors for the devel-
opment of CVI in patients with symptomatic con-
genital CMV.
The Houston Congenital CMV Longitudinal
Study is a prospective, multidisciplinary cohort study
conducted since 1982 with a total of 237 children
enrolled, of which 77 were categorized as symptom-
atic at birth, 109 were asymptomatic at birth, and
51 were uninfected controls.18 The current study
is a post hoc retrospective analysis of The Houston
Congenital CMV Longitudinal Study. This study
was approved by the institutional review board of
Baylor College of Medicine and conformed to the
requirements of the Health Insurance Portability and
Accountability Act of 1996. Informed consent was
obtained from human participants.
Patients in the study cohort received serial
age-appropriate ophthalmologic examinations at
follow-up study examinations. Best corrected visual
acuity (BCVA) was classified as normal, moderate
impairment, or severe impairment. Normal BCVA
was defined as 20/40 or better on optotype visual
acuity testing in each eye for patients who were able
to cooperate, or fixing on and following a near ob-
ject for nonverbal and preverbal children. Moderate
visual impairment was defined as BCVA of 20/200
or better, BCVA of 20/40 or worse, or the presence
of fixation on a near object but no following. Se-
vere impairment was defined as BCVA of 20/200 or
worse, no demonstrable fix or follow behavior, or no
reaction to light.
CVI is a severe form of vision loss that is primar-
ily caused by damage to the posterior visual path-
ways of the central nervous system, with reduced
BCVA as a result of a disease process that does not
primarily involve the ocular structures.3,19
The diagnosis of CVI was made clinically in
study patients when they had an otherwise normal
eye examination, yet showed abnormal visual re-
sponses, including difficulty to fix and follow pur-
posefully, or when the examination findings, such
as small peripheral retinal lesions, could not explain
the degree of visual impairment. CVI was further
confirmed by visual evoked responses in one of the
patients. Neurological imaging studies were also re-
viewed when available.
195
 TABLE 1
Demographic Data of Patients in the Current Study (N = 77)a
Symptomatic CMV Symptomatic CMV
Characteristic With CVI (n = 11) Without CVI (n = 66) P
Male 6 (54.5%) 30 (45.5%) .576
African American 2 (18.2%) 10 (15.2%) .678
Asian 0 (0%) 2 (3.0%) 1.000
White 9 (81.8%) 54 (81.8%) .640
Hispanic 2 (18.2%) 17 (25.8%)
.722
Non-Hispanic 9 (81.8%) 49 (74.2%)
Mother’s age (y), mean ± SD 24.3 ± 8.1 23.7 ± 6.6 .795
GA (weeks), mean ± SD 36.9 ± 2.1 37.5 ± 2.2 .417
Birth weight (g), mean ± SD 2,073.0 ± 532.6 2,448.6 ± 637.8 .069
Cesarean section birth 3 (27.3%) 30 (45.5%)
.214
Vaginal birth 8 (72.7%) 36 (54.5%)
Age at most recent examination (y), mean ± SD (range) 11.0 ± 6.5 (2 to 21.1) 12.2 ± 8.0 (2.4 to 27.3) .647
Avg. no. of eye examinations, mean ± SD (range) 10.3 ± 6.4 (2 to 21) 7.1 ± 4.9 (0 to 19) .058
Ganciclovir treatment 5 (45.5%) 13 (19.7%) .116
CMV = cytomegalovirus; CVI = cortical visual impairment; SD = standard deviation; GA = gestational age
a
Values are expressed as number of patients (percent) unless otherwise stated.

A retrospective review of clinical records identi-
fied 11 patients diagnosed as having CVI. Patient
records were analyzed and each patient’s presenta-
tion, clinical course, clinical correlations, and long-
term outcome were described. Demographics, birth
characteristics, and ocular findings are described in
patients with CVI and compared to patients with
non-symptomatic CVI in the study cohort.
Statistical analysis employed chi-square or Fish-
er’s exact tests when comparing categorical variables
and t tests when comparing continuous variables.
Risk factors were assessed using regression models to
determine their association with the future develop-
ment of CVI. Various characteristics found at birth
and before the CVI diagnosis were entered into uni-
variate logistic regressions to assess the strength and
significance of these risk factors for the development
of CVI. Multivariate logistic regression models were
developed using some of these variables in combina-
tion to predict CVI in individual cases. The logistic
regression procedure normally predicts the probabil-
ity of each patient having the outcome of interest
(CVI) from 0 to 1, with 0.5 as the default (ie, if
the prediction is 0.5 or higher, then the outcome
is predicted to occur; if the prediction is below 0.5,
then it is predicted to not occur). In our models, we
adjusted the cutoff level to identify more possible
cases of CVI, albeit at the cost of identifying more
false-positives results along with them. By using the
predicted versus the actual number of CVI cases
identified in each multivariate model, the sensitiv-
ity, specificity, false-positive rate, false-negative rate,
positive and negative predictive values, and overall
performance of each model were calculated.
RESULTS
There were 11 patients who developed CVI,
and all were in the group who were symptomatic
at birth (11 of 77 [14.3%]). Comparisons were
made between these 11 patients and the remaining
66 symptomatic patients who did not develop CVI.
Demographic and clinical information are sum-
marized in Table 1. There were no significant dif-
ferences between gender, race/ethnicity, gestational
age, and ganciclovir treatment status. The average
birth weight was lower in patients with CVI than
in patients without, but it did not reach statistical
significance (P = .069). The mean number of eye
examinations was 10.3 (range: 2 to 21) for the 11
patients with CVI compared with 7.1 (range: 0 to
19) for the patients with symptomatic congenital
CMV (P = .058) (Table 1). The mean age at CVI

196 Copyright © SLACK Incorporated

 TABLE 2
Ocular and Visual Findings (N = 77)a
Symptomatic CMV Symptomatic CMV
Characteristic With CVI (n = 11) Without CVI (n = 66) P
Percent (%) of total patients 14% 86%
Age at CVI diagnosis (y), mean ± SD (range) 3.3 ± 3.0 (11 to 9.9) –
Vision at most recent examination
Normal 0 (0%) 51 (77.3%) < .001
Moderate impairment 4 (36.4%) 2 (3.0%) .003
Severe impairment 7 (63.6%) 3 (4.5%) < .001
Not evaluated 0 (0%) 10 (15.2%) .341
Optic atrophy 4 (36.4%) 5 (7.6%) .020
Unilateral 1 (25%) 1 (20%)
Bilateral 2 (50%) 2 (40%)
Laterality not recorded 1 (25%) 2 (40%)
Chorioretinitis or retinal scars 6 (54.5%) 13 (19.7%) .022
Unilateral 2 (33.3%) 7 (53.8%)
Bilateral 2 (33.3%) 5 (38.5%)
Not noted 2 (33.3%) 1 (7.7%)
Retinal detachment 0 (0%) 2 (3.0%) 1.000
Strabismus 7 (63.6%) 11 (16.7%) .002
Esotropia 0 (0%) 3 (27.3%) .245
Exotropia 7 (100%) 8 (72.7%)
Amblyopia 0 (0%) 3 (4.5%) 1.000
Anterior segment abnormality 1 (9.1%) 4 (6.1%) .548
Nystagmus 5 (45.5%) 6 (9.1%) .007
Astigmatism 5 (45.5%) 17 (25.8%) .227
Cupping 2 (18.2%) 3 (4.5%) .146
Vitreous hemorrhage 0 (0%) 7 (10.6%) .584
Cataract 1 (9.1%) 1 (1.5%) .267
Ocular abnormalities, mean ± SD b
4.1 ± 1.7 1.0 ± 1.4 < .001
CMV = cytomegalovirus; CVI = cortical visual impairment; SD = standard deviation; Avg = average
a
Values are expressed as number (percent) unless otherwise stated.
b
Average number of ocular abnormalities per patient.

diagnosis was 3.3 years (range: 11 months to 9.9
years) (Table 2).
Table 2 shows the significance of the differences
in rates of visual impairment between patients with
and without symptomatic CVI, with 36.4% versus
3% having moderate impairment (P = .003) and
63.6% versus 4.5% having severe impairment (P <
.001). Seventy-seven percent of the patients without
symptomatic CVI had normal vision (P < .001). Ocu-
lar abnormalities were more common among patients
with CVI than those without, including optic atro-
phy (36.4% vs 7.6%, P = .020), chorioretinitis/retinal
scars (54.5% vs 19.7%, P = .022), strabismus (63.6%
vs 16.7%, P =.002), and nystagmus (45.5% vs 9.1%,
P = .007) (Table 2). No significant changes were ob-
served in visual acuity, including fix and follow behav-
ior over the study period in patients with CVI.
Nonophthalmologic sequelae were also ob-
served in patients with symptomatic CVI. Senso-
rineural hearing loss (SNHL) occurred in all 11
(100%) patients with CVI and in 46 (69.7%) pa-
tients without symptomatic CVI (P = .057). There

Journal of Pediatric Ophthalmology & Strabismus • Vol. 56, No. 3, 2019 197
 TABLE 3
Nonophthalmologic Sequelae and Comorbidities (N = 77)a
Symptomatic CMV Symptomatic CMV
Characteristic With CVI (n = 11) Without CVI (n = 66) P
SNHL 11 (100%) 46 (69.7%) .057
Unilateral 1 (9.1%) 10 (21.7%) .672
Bilateral 10 (90.9%) 36 (78.3%) –
Neonatal onsetb 7 (63.6%) 20 (43.5%) .229
Onset after 28 days 4 (36.4%) 26 (56.5%) –
Congenital onset c
8 (72.7) 25 (54.3%) .326
Delayed onsetd 3 (27.3%) 21 (45.7%) –
Age at SNHL diagnosis (y), mean ± SD (range) 0.67 ± 1.78 (birth to 6 y) 1.14 ± 2.58 (birth to 11 y) .566
Hearing aid 7 (63.6%) 30 (45.5%) .264
Cochlear implant 0 (0%) 15 (22.7%) .109
SGA 6 (54.5%) 22 (33.3%) .194
Prematurity 4 (36.4%) 20 (30.3%) .732
Petechiae 9 (81.8%) 46 (69.7%) .334
Anemia 0 (0%) 3 (4.5%) .626
Splenomegaly 5 (45.5%) 34 (51.5%) .481
Hepatomegaly 4 (36.4%) 40 (60.6%) .120
Jaundice 3 (27.3%) 27 (40.9%) .306
Bilirubin > 3 mg/dL 2 (18.2%) 26 (39.4%) .193
Platelets < 75/mm 3
5 (45.5%) 45 (68.2%) .080
ALT > 100 IU 1 (9.1%) 11 (16.7%) .676
Neonatal pneumonia 0 (0%) 2 (3.0%) .733
Microcephaly ­e
9 (81.8%) 16 (25.4%) .001
Seizure at birth 3 (27.3%) 2 (3.0%) .019
Neurological abnormality 6 (54.5%) 17 (25.85%) .061
Abnormal head CT results 11 (100.0%) 52 (78.8%) .194
Intracranial calcifications 10 (90.9%) 33 (50.0%) .019
Dilatation of ventricles 9 (81.8%) 28 (42.4%) .020
Gray matter abnormality 1 (9.1%) 2 (0.9%) .133
White matter abnormality 6 (54.1%) 34 (15.0%) .004
Cerebral immaturity 3 (27.3%) 20 (8.8%) .079
Encephalomalacia 3 (27.3%) 4 (1.8%) .002
CMV = cytomegalovirus; CVI = cortical visual impairment; SNHL = sensorineural hearing loss; SD = standard deviation; SGA = small for gestational
age; ALT = alanine aminotransferase; CT = computed tomography
a
Values are expressed as number (percent) unless stated otherwise.
b
Onset within 28 days of birth.
c
Onset within 3 months of birth.
d
Onset after 3 months.
e
Olsen 3rd percentile.

were no differences in the age of diagnosis or the atic CVI (P = .229). In 52 symptomatic patients,
laterality for the hearing loss. SNHL occurred in the SNHL was severe enough to require hearing devices
neonatal period in 7 (63.6%) of 11 patients with (37 hearing aids and 15 cochlear implants). Seven of
CVI versus 20 (30.3%) patients without symptom- 11 patients with CVI required hearing aids but no

198 Copyright © SLACK Incorporated

 TABLE 4
Logistic Regressions of CVI Risk Factors and Predictive Modeling
Variable Univariate Logistic
No. Regressions Variable Description OR LCL UCL P Cutoff
1 Microcephaly (Olsen 3rd percentile) 12.97 2.54 66.10 .002 0.5
2 Seizures at birth 12.00 1.73 83.03 .012 0.5
3 Optic atrophy before CVI 2.71 0.46 16.12 .273 0.5
4 Chorioretinitis/retinal scar before CVI 2.33 0.59 9.17 .226 0.5
5 Strabismus before CVI 6.00 1.55 23.19 .009 0.5
6 Intracranial calcification at birth 10.00 1.21 82.61 .033 0.5
7 Nystagmus before CVI 1.00 0.11 9.21 1.000 0.5
8 Neonatal hearing loss 4.03 1.06 15.30 .041 0.5
9 Severe vision impairment before CVI > 1,500 0.00 ∞ a
.999 0.5
10 Encephalomalacia at birth 5.81 1.10 30.82 .039 0.5
11 Leukomalacia at birth 1.55 0.16 15.32 .708 0.5
CVI = cortical visual impairment; OR = odds ratio; LCL = lower confidence limit; UCL = upper confidence limit; cutoff = logistic regression link function
value (from 0 to 1) used to predict patients with CVI;
a
The symbol denotes infiinity because odd ratios may be between zero and infinity.

patients underwent cochlear implantation. Among DISCUSSION

the other birth characteristics for symptomatic pa-
tients, there was high statistical significance for mi-
crocephaly (Olsen 3rd percentile)20 (P = .001), white
matter abnormality (P = .004), seizure at birth (P =
.019), intracranial calcifications (P = .019), enceph-
alomalacia (P = .002), and dilatation of ventricles
(P = .020), which all occurred more frequently in
patients with CVI than without (Table 3).
Five predictive models based on 11 variables
were created using logistic regression to predict
which symptomatic patients would have developed
CVI (Tables 4-5). Model 5 showed the variables
of microcephaly, seizure at birth, optic atrophy,
chorioretinitis/retinal scars, strabismus, intracranial
calcifications, neonatal hearing loss, and severe vi-
sual impairment were predictors of CVI. Further-
more, model 5 had the most desirable parameters of
any predictive model. This model correctly identi-
fied 10 of 11 cases of CVI that occurred, giving the
lowest false-positive rate (23.1%), high sensitivity
(90.9%) and specificity (95.5%), and high negative
predictive value (98.4%) for CVI (Table 5). Model
3 also had good overall performance characters. If
not all variables were known, then model 1, which
used presence of microcephaly at birth as the only
parameter, performed well as a predictive risk factor
for CVI with high sensitivity and reasonably good
specificity (Table 5).
It is widely known from the literature that con-
genital CMV infection can cause long-lasting visual
sequelae such as optic atrophy and chorioretini-
tis.13-18 The 11 patients with CVI in our study were
symptomatic at birth by having one or more clinical
symptoms of congenital CMV infection. In clinical
settings, 10% to 15% of patients who are congeni-
tally infected with CMV show symptoms at birth,21
whereas the remaining 85% to 90% are asymptom-
atic at birth. Despite few reports of CVI in congeni-
tal CMV infection in the literature, our findings
support that CVI cases are observed frequently in
symptomatic congenital CMV infection.14,15,17 In
comparison, the 109 children born with asymptom-
atic congenital CMV infection and the 51 uninfect-
ed controls did not develop CVI during the study
period of observation.
Prematurity is a known major risk factor for
CVI. Optic radiations are supplied with blood from
the germinal matrix, and in premature infants, the
germinal matrix is a watershed area of the brain that
is prone to hypoxic, ischemic, and hemorrhagic in-
sults,1 thus causing CVI. In our study, there were no
significant differences in prematurity or gestational
ages in the group of patients with and without symp-
tomatic CVI (Table 3). This finding suggests that,
although symptomatic congenital CMV can be asso-
ciated with premature birth, something in addition

Journal of Pediatric Ophthalmology & Strabismus • Vol. 56, No. 3, 2019 199
200
TABLE 5
CVI Predictive Performance Characteristics of Selected Models (N = 77)
Correctly Classified
False- False-
Positive Negative Patients With Patients Without Overall
Model Variablea Sensitivityb Specificityc Rated Ratee PV+f PV-g Cutoff CVI (n = 11) CVI (n = 66) Performanceh
Model 1 1 81.8% 74.2% 65.4% 3.9% 34.6% 96.1% 0.2 9 49 75.3%
Model 2 1, 2, and 3 81.8% 72.7% 66.7% 4.0% 33.3% 96.0% 0.2 9 48 74.0%
Model 3 1, 2, 3, 5, 81.8% 95.5% 25.0% 3.1% 75.0% 96.9% 0.3 9 63 93.5%
6, 8, and 9
Model 4 1, 2, 3, 5, 90.9% 87.9% 44.4% 1.7% 55.6% 98.3% 0.2 10 58 88.3%
6, 8, and 9
Model 5 1, 2, 3, 4, 5, 90.9% 95.5% 23.1% 1.6% 76.9% 98.4% 0.2 10 63 94.8%
6, 8, and 9
CVI = cortical visual impairment; PV+ = predictive value of a positive test; PV- = predictive value of a negative test; cutoff = logistic regression link function value (from 0 to 1) used to predict patients with CVI
a
1 = microcephaly; 2 = seizures at birth; 3 = optic atrophy before CVI; 4 = chorioretinitis/retinal scar before CVI; 5 = strabismus before CVI; 6 = intracranial calicification at birth; 7 = nystagmus at birth; 8 = neonatal
hearing loss; 9 = severe vision impairment before CVI; 10 = encephalomalacia at birth; 11 = leukomalacia at birth.
b
Sensitivity is the proportion of patients who developed CVI and were predicted by the model to develop CVI. The probability of correctly identifying a patient with CVI in the screened population.
c
Specificity is the proportion of patients who did not develop CVI and were predicted by the model to not develop CVI. The probability of correctly identifying a patient without CVI in the screened population.
d
False-positive rate is the rate of patients who were predicted by the model to develop CVI but did not develop CVI.
e
False-negative rate is the rate of patients who were predicted by the model to not develop CVI and developed CVI.
f
The predictive value of a positive test.
g
The predictive value of a negative test.
h
Overall performance is the percentage of correctly classified patients with CVI and the percentage of correctly classified patients without CVI.

our study.

earlier in utero than those without CVI.

acquiring congenital CMV infection is
ocular comorbidities include optic nerve

Copyright © SLACK Incorporated

ocular, and neurodevelopmental sequel-
over time. Nevertheless, it is important
ity or functional fix and follow behavior
enced measurable changes in visual acu-
symptomatic congenital CMV experi-
dence that our patients with CVI with
vision4,24; however, there is no clear evi-
ma or hypoxic-ischemic injuries tend to
suspect these patients acquired infection
times more ocular findings compared to
nosed as having CVI had on average four

have gradual and partial recovery of their

Children with CVI resulting from trau-
patients without CVI. Therefore, we
ity.23 In the current study, patients diag-
connected to the level of a child’s disabil-
atrophy, chorioretinitis/retinal scars, stra-
in severe visual impairment in patients
or improved vision in the patients with

with symptomatic CMV. Common

symptomatic congenital CMV infec-
and ganciclovir treatment was not as-
atic neonates.22 However, Kimberlin
developmental outcomes of symptom-
can improve the audiologic and neuro-
nous ganciclovir and oral valganciclovir
terized.21 Reports have shown intrave-
infection remain incompletely charac-
infection, yet the precise neurodevelop-
the setting of symptomatic congenital
ity, or encephalomalacia) observed in
microcephaly, white matter abnormal-
dominance of CNS abnormalities (eg,
CNS cell lines is evident from the pre-
The high tropism of CMV for infecting
to prematurity must be contributing to

bismus, and nystagmus. The timing of

mental pathway and cellular targets of
viral interference with neurogenesis.
the development of CVI, such as direct

our study also had multiple other visual,

We found that CVI often resulted
tion and neurologic involvement in
sociated with fewer occurences of CVI
et al.22 did not report visual outcomes,

to be aware that the patients with CVI in

ae from congenital CMV infection, which could have
masked improvement in CVI.
In addition to ocular and visual comorbidities,
symptomatic congenital CMV infection causes se-
vere neurodevelopmental and sensorineural sequelae.
Sensorineural hearing loss, microcephaly, intracranial
calcification, dilatation of ventricles, encephalomala-
cia (leukomalacia), and seizures at birth were all as-
sociated with the presence of CVI. Taking these risk
factors into account, a predictive model could help
to screen patients with congenital CMV who are at
risk for CVI, which can aid clinicians, parents, and
educators to take preemptive steps toward early inter-
vention and rehabilitation.
There were some limitations with predictive
modeling. First, our current predictive modeling
assumed the presence of the predictors to have oc-
curred before and not after the diagnosis of CVI,
but in actual practice CVI was diagnosed concur-
rently with certain model variables at the same
evaluation visit for some patients. The mean age at
CVI diagnosis was 3.3 years (range: 11 months to 9
years); therefore, although not guaranteed, it would
be reasonable to assume that the predictors had oc-
curred prior to CVI diagnosis. Second, our sample
size was small. Developing models from larger or
pooled populations of symptomatic patients would
be desirable. Finally, although birth characteristics
such as microcephaly would likely be known for
other patients with symptomatic CMV, we assessed
other characteristics (ie, intracranial calcification)
that might not be identified in patients not included
in the study. This could limit the applicability of a
predictive model that relies on knowing the presence
or absence of detailed visual and non-visual findings
from serial examinations or neuroimaging. Despite
these limitations, logistic regression-based predic-
tive models have potential applicability in clinical
medicine. Evidence-based predictive models could
become useful precision medicine tools for patients
with symptomatic congenital CMV, who are known
to be at risk for CVI.
To help children with CVI, ophthalmologists
first have to correctly identify this condition. Al-
though logistic regression-based predictive model-
ing is not a diagnostic tool, it could be developed
into a screening instrument or clinical decision
support tool to help identify cases that may require
heightened medical surveillance. A thorough clini-
cal examination of the visual system should follow,
Journal of Pediatric Ophthalmology & Strabismus • Vol. 56, No. 3, 2019
and is often sufficient to diagnose CVI. However, in
equivocal cases, visual evoked potentials and brain
imaging can offer additional confirmation.24
CVI can affect many aspects of vision, includ-
ing visual acuity, visual fields, color vision, contrast
sensitivity, perception of movement, ability to see
details in complex visual scenes, visual memory for-
mation, and recognizing the significance of facial
expression.25,26 When diagnosing CVI, clinicians
should further specify which aspects of visual func-
tions are affected in the child. Examples of func-
tional vision testing include visually guided motor
response, attentive gaze, visual recognition of ob-
jects, pictures, and faces, preferences for moving
versus static stimuli, preferences for familiar versus
novel stimuli, and suppression of visual responses
by auditory or tactual simulation.25,26 In 2006,
Roman-Lantz developed the CVI Range, which is
a reliable functional vision assessment tool for chil-
dren with CVI.6,27 Detailed descriptions of visual
function impairment help target specific rehabilita-
tion planning, the intervention of CVI should be
tailored to each child’s unique functional status, and
management goals should focus on early interven-
tion and collaboration with parents, educators, and
other health care providers to maximize functional
vision over time. Future larger-scale studies focus-
ing on predictive models, functional vision progress,
rehabilitation methods, and educational accommo-
dation for children with congenital CMV and CVI
are needed.
CONCLUSION
CVI may be suspected when one or more more
classic CMV symptoms present at birth. Microceph-
aly, intracranial calcification, dilatation of ventricles,
encephalomalacia, seizure at birth, optic atrophy,
chorioretinitis/retinal scars, strabismus, and neonatal
onset of sensorineural hearing loss are risk factors for
CVI. The relationship between CVI and its risk fac-
tors suggests it is possible to predict the development
of CVI through predictive modeling in future re-
search. Early screening for CVI in children born with
symptomatic congenital CMV can help with early
educational, social, and developmental intervention.
REFERENCES
1. Carden SM, Good WV. Cortical visual impairment. In: Wright
KW, Spiegel PH, Thompson LS, eds. Handbook of Pediatric Neuro-
Ophthalmology, 1st ed. New York: Springer-Verlag; 2006:247-254.
2. Lehman SS. Cortical visual impairment in children: identification,
201
 evaluation and diagnosis. Curr Opin Ophthalmol. 2012;23:384-
387.
3. Good WV, Jan JE, DeSa L, Barkovich AJ, Groenveld M, Hoyt
CS. Cortical visual impairment in children. Surv Ophthalmol.
1994;38:351-364.
4. Handa S, Saffari SE, Borchert M. Factors associated with lack of
vision improvement in children with cortical visual impairment. J
Neuroophthalmol. 2018;38:429-433.
5. Dutton GN, Jacobson LK. Cerebral visual impairment in chil-
dren. Semin Neonatol. 2001;6:477-485.
6. Roman-Lantz C. Cortical Visual Impairment: An Approach to As-
sessment and Intervention. New York: AFB Press, American Foun-
dation for the Blind; 2007.
7. Boot FH, Pel JJ, van der Steen J, Evenhuis HM. Cerebral visual
impairment: which perceptive visual dysfunctions can be expected
in children with brain damage? A systematic review. Res Dev Dis-
abil. 2010;31:1149-1159.
8. Pascolini D, Mariotti SP. Global estimates of visual impairment:
2010. Br J Ophthalmol. 2012;96:614-618.
9. Steinkuller PG, Du L, Gilbert C, Foster A, Collins ML, Coats
DK. Childhood blindness. J AAPOS. 1999;3:26-32.
10. Kong L, Fry M, Al-Samarraie M, Gilbert C, Steinkuller PG. An
update on progress and the changing epidemiology of causes of
childhood blindness worldwide. J AAPOS. 2012;16:501-507.
11. Swaminathan M. Cortical visual impairment in children: a new
challenge for the future? Oman J Ophthalmol. 2011;4:1-2.
12. Hatton DD, Schwietz E, Boyer B, Rychwalski P. Babies count: the
national registry for children with visual impairments, birth to 3
years. J AAPOS. 2007;11:351-355.
13. Demmler GJ. Infectious Diseases Society of America and Centers for
Disease Control.Summary of a workshop on surveillance for con-
genital cytomegalovirus disease. Rev Infect Dis. 1991;13:315-329.
14. Jin H, Demmler-Harrison GJ, Coats DK, et al. Long-term visual
and ocular sequelae in patients with congenital cytomegalovirus
infection. Pediatr Infect Dis J. 2017;36:877-882.
15. Coats DK, Demmler GJ, Paysse EA, Du LT, Libby C. Ophthal-
mologic findings in children with congenital cytomegalovirus in-
fection. J AAPOS. 2000;4:110-116.
16. Lanzieri TM, Chung W, Flores M, et al. Hearing loss in children
202
with asymptomatic congenital cytomegalovirus infection. Pediat-
rics. 2017;139:e20162610. doi:10.1542/peds.2016-2610.
17. Ghekiere S, Allegaert K, Cossey V, Van Ranst M, Cassiman C,
Casteels I. Ophthalmological findings in congenital cytomegalo-
virus infection: when to screen, when to treat? J Pediatr Ophthal-
mol Strabismus. 2012;49:274-282.
18. Noyola DE, Demmler GJ, Williamson WD, et al. Cytomegalo-
virus urinary excretion and long term outcome in children with
congenital cytomegalovirus infection. Congenital CMV Longitu-
dinal Study Group. Pediatr Infect Dis J. 2000;19:505-510.
19. Chhablani PP, Kekunnaya R. Neuro-ophthalmic manifestations
of prematurity. Indian J Ophthalmol. 2014;62:992-995.
20. Olsen IE, Groveman SA, Lawson ML, Clark RH, Zemel BS. New
intrauterine growth curves based on United States data. Pediatrics.
2010;125:e214-e224.
21. Cheeran MC, Lokensgard JR, Schleiss MR. Neuropathogenesis
of congenital cytomegalovirus infection: disease mechanisms and
prospects for intervention. Clin Microbiol Rev. 2009;22:99-126.
22. Kimberlin DW, Jester PM, Sanchez PJ, et al. Valganciclovir for
symptomatic congenital cytomegalovirus disease. N Engl J Med.
2015;372:933-943.
23. Bosnjak VM, Dakovíc I, Duranovic V, Lujic L, Krakar G, Marn
B. Malformations of cortical development in children with con-
genital cytomegalovirus infection: a study of nine children with
proven congenital cytomegalovirus infection. Coll Antropol.
2011;35 Suppl 1:229-234.
24. Good WV, Jan JE, Burden SK, Skoczenski A, Candy R. Recent
advances in cortical visual impairment. Dev Med Child Neurol.
2001;43:56-60.
25. AFB Press, American Foundation for the Blind. In: Dennison
E, Luck AH, eds. Proceedings of the Summit on Cerebral/Cortical
Visual Impairment: Educational, Family, and Medical Perspectives:
April 30, 2005. New York: AFB Press; 2006.
26. Hyvärinen L. Considerations in evaluation and treatment of the
child with low vision. Am J Occup Ther. 1995;49:891-897.
27. Newcomb S. The reliability of the CVI range: a functional vision
assessment for children with cortical visual impairment. Journal of
Visual Impairment & Blindness. 2010;104:637-647.
Copyright © SLACK Incorporated
 TABLE A
The Congenital CMV Longitudinal Study Group Members
Amit R. Bhatt, MD Sara Reis, RN
Peggy Blum, AuD Ann Reynolds, MD
Frank Brown, MD Judith Rozelle, MS
Francis Catlin, MD Sherry Sellers Vinson, MD
Alison C. Caviness, MD, PhD, MPH O’Brien Smith, PhD
David K. Coats, MD Paul Steinkuller, MD
Jane C. Edmond MD Marie Turcich, MS
Daniel Franklin, MD John Voigt, MD
Jewel Greer Bethann Walmus
Carol Griesser, RN Daniel Williamson, MD
Mohamed A. Hussein, MD Kimberly G. Yen, MD
Isabella Iovino, PhD Martha D. Yow, MD
Allison Istas, MPH Texas Children’s Hospital Audiology
Mary K. Kelinske, OD Shahzad Ahmed, MD
Antone Laurente, PhD Hanna Baer, MD
Thomas Littman, PhD Gail J. Demmler-Harrison, MD
Jerry Miller, PhD Marily Flores, MS
Mary Murphy, MS Cindy Gandaria
Christopher Nelson, MD Haoxing Douglas Jin, MD
Daniel Noyola, MD Joseph T. Klingen, MD
Evelyn A. Paysse, MD Jill Williams, MA
Alan Percy, MD
Reproduced with permission of copyright owner. Further reproduction
prohibited without permission.