Predictive accuracy of the ABCD progression display among patients with keratoconus: A historic cohort analysis

To evaluate the accuracy of the ABCD Progression Display and the ABCD grading system in a population of adult patients with keratoconus.

A retrospective cohort analysis of all adult patients with keratoconus followed at the Shamir Medical Center between 2012 and 2017. A recommendation by the cornea specialist to undergo corneal crosslinking (CXL) was used as a surrogate of ectasia progression. The ABCD grading was not available to the treating physician and was computed post-hoc. Sensitivity and specificity of the ABCD Progression Display was calculated, and multivariate regression was used to estimate the risk to undergo CXL when the ABCD Progression Display indicated progression. The ABCD grading was compared between patients who required CXL to those who did not. A single eye of each patient was included. Sensitivity and specificity of the ABCD Progression Display were 82% and 73%, respectively. A multivariable model adjusted for possible confounders, found that ABCD Progression was associated with a 7-fold risk of undergoing CXL compared to a patient in whom progression was not recorded in the ABCD Progression Display (OR = 7.55; 95% CI = 3.82-14.93, p < 0.001).

293 eyes of 293 patients were analysed. Mean age at presentation was 26.92 ± 6.12 years. In 68 eyes, progression of keratoconus was recorded and CXL was performed (CXL-group).

The ABCD Progression Display demonstrated adequate sensitivity and specificity and high predictive capabilities of keratoconus progression. It can be effectively utilized as an initial screening test in adults with keratoconus.