Effect of new classification on ICRS outcomes
My colleagues and I were keen to identify if grading patients according to the new classification system had a noticeable
impact on the predictability of ICRS outcomes. Among the database group, 563 eyes were implanted with ICRS and of these 304
had undergone implantation under similar pre-, intra-and postoperative circumstances. Of note, 51% were implanted with Intacs
(Addition technology, Sunnyvale, California, USA), 44% had Keraring (Mediphacos, Belo Horizonte, Minas Gerais, Brazil) and
5% had Myorings (Dioptex, Linz, Austria).
We found some evidence of correlation between keratoconus classification and visual outcomes. For example, patients classified
as having grade I and grade IV plus keratoconus had the lowest ICRS failure rates, when failure was defined as a loss of one
or more lines in BCDVA or uncorrected distance visual acuity (UCDVA). However, we also noticed a lack of congruity and some
conflicting results. Patients with grade I keratoconus gained one or more lines in UCDVA post-ICRS implantation more often
than any other keratoconus class. But these patients gained one line in BCDVA less commonly than patients in all other keratoconus
Interestingly, among a best case group of 114 patients (with baseline parameters that reflected best suitability for ICRS
implantation and who had their ICRS implanted using femtosecond laser) post-implantation failure rates were the same as among
all other patients. This observation, considered along with the conflicting outcomes seen when patients were classified according
to conventional keratoconus parameters, indicated that ICRS outcomes are difficult to predict using current nomograms. In
light of this, my colleagues and I developed our own predictive model using refractive, corneal aberrometric data (CSO, Florence,
Italy) and corneal biomechanical data obtained with the ocular response analyser (ORA, Reichert, New York, USA).
Equation 1: The function developed for predicting postoperative BCDVA.
The model (Equation 1) was based on multiple regression analysis and we found that it provided a much more accurate method
of predicting postoperative BCDVA in keratoconus patients implanted with an ICRS than existing strategies. The efficacy of
this model arises from the manner in which it takes into account not just the presence of factors, such as primary coma, higher
order aberration and cylindrical refractive power, but also the way these interact with each other in eyes with specific biomechanical
Improved classification of keratoconus is a clear strategy by which eye specialists and patients can better understand the
condition. It is clear that although existing nomograms and disease parameters have already provided some guidance on how
to approach the treatment of keratoconus patients, these require more work if predictability of ICRS outcomes is to improve.
As manufacturers continue to tackle this by developing new ICRS models, there is work to be done from the patient selection
point of view. Even the very best device cannot work to its optimum if used in the wrong way or for the wrong task. In the
same light, the efficacy of ICRS is heavily dependent on the eyes in which they are implanted. The best way to ensure that
patients implanted with these segments achieve the results they expect is for more focus to be placed on the development of
models that better predict outcomes preoperatively.
1. A. Vega-Estrada et al., Am. J. Ophthalmol., 3 Dec 2012, doi: 10.1016/j.ajo.2012.08.020. [Epub ahead of print].
2. D.P. Piñero and J.L. Alio, Clin. Exp. Ophthalmol., 2010; 38(2):154–167.