e30 Volume 13 Number 1 / February 2009

study examines the relationship between rate of vascular changeand plus disease diagnosis.

Methods: During ROP examinations of 74 eyes from 37 infants,wide-angle images were taken bilaterally at 31-33 and 35-37 weeks'postmenstrual age (PMA). The semiautomated Retinal Image multi-Scale Analysis system was used to measure parameters for all arte-rioles and venules: integrated curvature (IC), diameter, and tortuosityindex (TI). A reference standard diagnosis (plus vs not plus) was de-fined for each eye by consensus of five experts at 35-37 weeks PMA.Weekly rate of change in parameters was compared in eyes withplus versus not plus disease. Receiver operating characteristicarea under the curve (AUC) was calculated for plus disease detec-tion based on: 1) weekly rates of parameter change between 31-33and 35-37 weeks' PMA, and 2) parameter values at 35-37 weeks only.

Results: Weekly rates of change in all venular parameters weresignificantly different in eyes with plus versus not plus disease, par-ticularly for TI (p\0.0004) and diameter (p 5 0.02). Using weekly rateof change, we found that AUC for plus disease detection was great-est for venular TI (0.818) and diameter (0.702). Using the 35-37 weekPMA image only, we found that AUC was greatest for venular IC(0.952) and diameter (0.760).

Conclusions: Rate of change in venular, but not arteriolar, param-eters is correlated with plus disease development. This did notappear to contribute information beyond analysis of an image at35-37 weeks PMA only.

118 The cost effectiveness of lay vision screening. JoannahVaughan, David Wheeler

Introduction: The purpose of this institutional review board-ap-proved study is to test the theory that a lay vision screening programcan be an effective method for identifying preschool children whoneed eye care. An extensive follow-up program was created tofind out if the children referred by the screening actually went infor a complete eye examination. The study also tried to determinethe major obstacles for children who were referred but did not goin for a complete eye examination.

Methods: Volunteers were shown a training video. Staff assistedvolunteers with the set up of the screening area prior to the screen-ing. Children 3-5 years of age were give a distance test and a stere-opsis screens. Children who failed were referred for a complete eyeexam. A follow-up program determined if they received an eye exam,or why they did not. The chart notes were studied from each exam.

Results: A total of 2,017 children consented to be in the study.Eleven percent of the children screened were referred for a completeeye examination. Eighty-four percent of the children passed; 5% ofthe children could not be screened; and 212 were referred.

Discussion: The program is a cost effective method to detect chil-dren with eye disease.

Conclusions: There is a huge cost differential between giving 212children a complete examination versus the cost of giving 2,017 chil-dren a complete eye examination.

119 Measurement of retinal vascular dilation using ROPtool.David K. Wallace, Sharon F. Freedman, Zheen Zhao

Background: Plus disease is the primary criterion for laser treat-ment in retinopathy of prematurity, but its assessment is subjective.ROPtool is a computer program that measures retinal arteriolartortuosity. Our aim was to assess the accuracy of ROPtool's newlydeveloped measurement of retinal vascular width (dilation).

Methods: ROPtool was used to measure the width of 154 bloodvessels in 20 high-quality RetCam images from 20 premature infants.

The accuracy of the ROPtool was determined by comparing resultsto the mean grades of 2 authors who scored retinal vascular dilationusing a 10-point scale.

Results: There was very good correlation (r 5 0.80) between ROP-tool's measurement of retinal vascular dilation and author judgment.Areas under receiver operating characteristic (ROC) curves for iden-tification of dilation sufficient for plus disease and for pre-plus dis-ease were 0.93 and 0.90, respectively. At an optimal point on theROC curve, ROPtool's sensitivity for diagnosing dilation sufficientfor plus disease was 89% (24/27), and its specificity was 83% (106/127).

Conclusions: In addition to measuring tortuosity, ROPtool nowaccurately calculates retinal vascular width in high-quality RetCamimages. Application of this technology has the potential to removesubjectivity from the diagnosis of plus disease.

120 Progression of with-the-rule astigmatism in children withinfantile nystagmus syndrome. Jingyun Wang, Lauren M. Wyatt,Joost Felius, David R. Stager Jr., David R. Stager Sr., Eileen E Birch

Introduction: Our previous cross-sectional study showed a highprevalence of with-the-rule (WTR) astigmatism in infants and chil-dren with infantile nystagmus syndrome (INS). This study investi-gated the longitudinal development of WTR astigmatism in childrenwith idiopathic INS and INS associated with albinism.

Methods: Two groups of children with INS (Idiopathic: n 5 22; Al-binism: n 5 27) were included. Ages ranged from 1.3 to 202 monthsold. They were refracted at $3 visits, starting before 48 months ofage and separated typically by $6 months. Cycloplegic refractionswere converted into power vector components: M (spherical equiv-alent), J0 (positive J0 indicates WTR astigmatism) and J45 (obliqueastigmatism). The slope of the best linear fit to each child's J0data as a function of age characterized the rate of change of astig-matism in that individual.

Results: Our previous results demonstrated that the mean J0 of in-fants with INS is more positive than normal infants before 6 months ofage. We now report that the mean J0 increases with age in both INSgroups, indicating progressive increase in WTR astigmatism. Themean slopes of the best fit lines are 0.011 � 0.004D per month inthe idiopathic group and 0.010 � 0.003D per month in the albinismgroup, indicating that J0 changes similarly in both groups.

Conclusions: Although children with INS have more WTR astigma-tism than normal children during infancy, the increase in WTR astig-matism with age supports the hypothesis that the presence ofnystagmus is associated with progressive increase of astigmatismin these patients.

121 Retinal hemorrhages defining the probability of inflicted braininjury: A systematic review. Patrick Watts, Nicole Bowden, DanielFarewell, Sabine Maguire, Alison Kemp

Purpose: To assess the value of the presence of retinal hemor-rhages as one indicator of inflicted brain injury (iBI).

Methods: A systematic review of all relevant studies identifiedfrom an all language literature search of 10 databases between1970 and 2008 using standardized critical appraisal methodologywas performed. This study is part of a larger review looking ata number of clinical features that may be associated with iBI. Com-parative studies of iBI and non inflicted brain injuries (niBI) in chil-dren (0-18 years) where there was a clear history of the etiology ofthe injury were included. Brain injury was defined as intracranialhemorrhage (ICH) visible on neuro-imaging. A mulitilevel logistic re-gression was performed to determine the positive predictive values

Journal of AAPOS

Volume 13 Number 1 / February 2009 e31

(PPV) with 95% confidence interval, of retinal haemorrhages (RH)for iBI.

Results: Of 320 studies reviewed, 14 were included, representing779 children with iBI from a total of 1,655. The positive predictivevalue of retinal hemorrhages alone as an indicator of iBI was PPV71%, OR 3.504, p 5 0.03. The only other single feature that wasmore significant was apnoea (PPV %, OR 17.06, p \ 0.001), but therewas less data to support this finding.

Conclusions: This study is the first systematic review generatinga multilevel logistic regression to determine the PPV of retinalhemorrhages for identifying iBI. Clearly, they are confirmed asa strong indicator that an ICH in a young child with associatedRH has a high probability of being the consequence of an inflictedinjury.

122 Statewide preschool vision screening in Oregon. David T.Wheeler, Joannah Vaughan, Allison Summers

Introduction: There is great interest in preschool vision screen-ing as a means of identifying children at risk for amblyopia ina timely and accurate fashion but no standard protocol hasemerged. We developed a method similar to that recommendedby Prevent Blindness America and partnered with Head Startsthroughout Oregon.

Methods: Volunteers were shown an instructional video and weretrained on site. Staff assisted volunteers with the set up of thescreening area. Children aged 3-5 were screened using Lea symbolsor HOTV letters and the Random Dot E. Children who failed were re-ferred for a complete eye examination. A follow-up program deter-mined if they received an eye exam, or why they did not. Chartnotes were requested for each child who had an eye exam. The studyalso tried to determine the major obstacles for children who were re-ferred but did not follow through with a complete exam.

Results: The study included 2,017 children. Of these, 84% passed,11% were referred, and 5% could not be screened. Eye exam datawere obtained for 212 of the 225 children who were referred. Therewas a strong correlation between screening and examination re-sults. The main reason for incomplete follow-up was poor accessto medical care.

Discussion: The Oregon program is an accurate and cost effectivemethod to detect children with eye disease.

Conclusions: A lay vision screening program measuring monocu-lar distance acuity and stereopsis can be an effective method foridentifying preschool children who need eye care.

123 A pilot study evaluating use of VisionQuest to perform visionscreening. M. Edward Wilson, Rupal H. Trivedi, M. MillicentPeterseim, Kali B. Cole, Ronald G. W. Teed

Objective: To compare vision screening results using automatedVisionQuest (VQ) software on an off-the-shelf laptop computer tothat of 1) an ophthalmic technician using an electronic visual acuity(EVA) tester and 2) professional eye examination results by a pediat-ric ophthalmologist for clinically significant vision disorders in chil-dren.

Methods: In this pilot study, 72 children were evaluated with theuse of VQ (patch and goggles), EVA and examination by pediatricophthalmologist. The order of presentation of the vision screeningassessments was assigned randomly. VQ registers pass/fail resultsfor threshold VA testing at 20/32 and stereopsis of 300 arcsec. A sim-ilar threshold was used for comparative analysis.

Journal of AAPOS

Results: Age: 11.4 (SD 2.2) years.

%

VQ usingpatch and

EVA

VQ using patchand Professional

examination

VQ usinggogglesand EVA

VQ usinggoggles andProfessionalexamination

False positive 7 8 15 17False negative 3 4 3 4Sensitivity 92 88 92 88Specificity 89 87 77 74Predictive

value 182 79 68 65

Predictivevalue �

95 93 95 92

Discussion: VQ using a patch achieved the best results. The lowernumber of false negatives compared to false positives is encourag-ing. The main advantage of VQ is ease of use, allowing a lay person touse it without compromising the outcome. The VQ software used inthe study has now undergone several upgrades, which may help im-prove outcomes.

Conclusions: VQ has potential for use as a vision screener.

124 Bilateral medial rectus muscle recession results in childrenwith developmental delay compared to normally developedchildren: More follow-up. Tamara Wygnanski-Jaffe, AbrahamSpierer

Introduction: Children with developmental delay have a high rateof esotropia. Nevertheless, strabismus surgery for neurologically im-paired children is less predictable. In a former report we comparedbimedial rectus muscle recession (BMR) results for esotropia in chil-dren with developmental delay to the results in normal children andreported a high rate of undercorrection in this group. We presenlycontinued the follow-up and report our results.

Methods: A retrospective analysis of children that underwentBMR surgery for esotropia during a 16-year period was undertaken.Surgical results of children with developmental delay were com-pared with those of normal children.

Results: At the last follow-up visit only 56% of the developmentallydelayed children achieved surgical success. Among failures, 86%were undercorrected initially and 42% remained undercorrecetedafter additioanl surgical procedures; only one patient developed con-secutive exotropia, during the first period of follow-up, neverthelesswith longer follow-up 4 patients became exotropic, and 50% of thepatients required an additional surgical intervention.

Discussion: A greater rate of surgical failure was found in devel-opmentally delayed children who received a smaller recession of themedial rectus muscles when compared with the developmentallynormal children receiving the standard amount of recession. Themain reason for surgical failure in the developmentally delayed groupwas undercorrection and remained so after additional follow-up.

Conclusions: Strabismus surgery in children with developmentaldelay is less predicatable. We need to better delineate the idealamount of surgery in this special group of children.

125 Down syndrome and surgical outcomes after strabismussurgery. Claudia Yahalom, Hadas Mechoulam, Evelin Cohen, IreneAnteby

Introduction: Children with Down Syndrome have a high preva-lence of strabismus. The purpose of our study was to describe thestrabismus characteristics and to evaluate the response to strabis-mus surgery in this population.