Assessing Agreement of Clinical Estimation of Vertical Optic Disc Cupping against Retcam Assessment in the Premature Neonate and Prospective Observation for Changes

Abstract

Aim: To assess agreement between binocular assessment of optic disc cupping and objective evaluation with RetCam imaging in premature infants. To look for difference in changes observed in disc cupping between different birthweight groups and effect of retinopathy of prematurity, intraventricular hemorrhage and sepsis. Methods: A cohort of premature infants were examined prospectively. At Timepoint I (31 – 36 weeks), the vertical optic disc cupping(vCDR) was assessed clinically through a 28D lens. Two RetCam fundus images centred on the optic discs were taken. A blinded assessor measured the vCDR and the average was calculated. This was repeated at Timepoint 2(37– 40 weeks). Exclusion criteria included IOP >21mmHg, abnormality in cornea, enlarged cornea size (>10.0mm) and abnormal eye. Consistency between the 2 methods was assessed with intraclass correlation coefficient (ICC). Bland-Altmann plots were used for agreement analysis. Results: In total, 126 eyes (n= 63 infants) were examined, 504 images captured and assessed. The average vCDR by clinician was 0.30(SD:0.1) and by RetCam 0.32(SD:0.1). The ICC showed moderate correlation however, Bland-Altman plot showed a wide 95% limits of agreement (-0.24, 0.26). Paired t-test did not show significant change in vCDR between the time-points for infants in different birthweight groups but trend of change were identical for both methods. Retinopathy of prematurity, intraventricular hemorrhage and sepsis were not associated with changes in cupping in this cohort. Conclusion: Optic disc cupping with clinical estimation is not interchangeable with RetCam values but showed identical trend of change. Mean optic disc vertical cupping in this population was 0.3 at term.