Purchase this article with an account.
Marta Jiménez-García, Elke O. Kreps, Sorcha Ní Dhubhghaill, Carina Koppen, Jos J Rozema; How to define keratoconus progression?. Invest. Ophthalmol. Vis. Sci. 2022;63(7):3082 – F0554.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Although corneal crosslinking (CXL) was a major breakthrough in keratoconus (KC) management, there has been debate about its efficiency. Cochrane reviews have only found limited, low quality evidence and there is no consensus on which KC cases need CXL, nor on how to establish progression. This is complicated further by noise in the measurements. Some progression criteria may better reflect our clinical knowledge of KC evolution, such as the % of progressive KC based on clinical changes, hence, this study aimed to verify the performance of diverse criteria.
The retrospective REDCAKE study included data from 743 KC patients measured longitudinally with Pentacam (Oculus, DE). Habitual progression criteria based on (combinations of) the maximum keratometry (KMAX), astigmatism (AF), and minimum pachymetry (PMIN), or based on ABCD Progression Display were analyzed. For each criterion and cut-off, we calculated the eyes flagged progressive at some point (RPROG), the individual consistency CIND (% of examinations after detection of progression that would still be considered progressive), and the population consistency CPOP (% of eyes with CIND > 66%). Finally, we studied more monotonic and consistent variables, such as the front steep keratometry (K2F), the keratometry in a 3 mm area around KMAX (KZONAL3mm), and the mean radius of the back surface (RmB).
Use of a single criterion (e.g. △KMAX >1D) led to rather high values of RPROG, unless the cut-off exceeded the measurement noise. When two criteria were required, (KMAX AND AF) led to worse CPOP and higher variability than (KMAX AND PMIN); alternative criteria (KZONAL3mm AND RmB) and (K2F AND RmB) obtained the best CPOP and the lowest variability (Fig 1; all p < 0.0001). ABC, as defined by its authors, obtained a very high RPROG of 74.2%. Using wider 95% confidence intervals (95CI) and requiring two of ABC over the 95CI reduced RPROG to a more realistic 27.9%.
Clinical observation of contact lens parameters, visual acuity, incidence of scarring or corneal transplants suggest 20–35% of KC cases are progressive. This clinical RPROG value should be considered when defining KC progression to avoid overtreatment. By using combinations of alternative variables and wider 95CI for ABC, the RPROG can be brought closer to clinical observations; furthermore, these approaches obtained better longitudinal consistency than current definitions.
This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.
Fig 1. Progression surfaces for different criteria
This PDF is available to Subscribers Only