[Detection of keratoconus using wavefront analysis]

To determine the value of wavefront analysis in the detection of keratoconus.

A prospective study was conducted from May 2008 to May 2010. Four groups were formed: patients with a "healthy" cornea (group 0; n=48), patients with keratoconus "suspect" (group 1; n=43), patients with keratoconus "forme fruste" (group 2; n=31), and patients with "beginning" keratoconus (group 3; n=47). Total and corneal aberrations were analysed. The Fisher test and Student t-test were used to compare the different groups. Receiver operating characteristic (ROC) curves were plotted when aberration differentiated groups 0 and 1.

Total and corneal coma (Z 3), the corneal trefoil, (Z 3) and the corneal secondary astigmatism (Z 4) differentiated groups 0 and 1. Sensitivities and specificities of total and corneal coma, corneal trefoil, and corneal secondary astigmatism were, respectively, 67.4%, 56.3% (Youden=0.237), 60.5%, 72.9% (Youden=0.334), 83.7%, 39.6% (Youden=0.233), and 65.1%, 58.3% (Youden=0.234). Total, corneal coma, and corneal trefoil differentiated all severity groups. The third high-order aberrations differentiated groups 1 and 2.

High-order aberrations are good indicators for grading keratoconus. Total high-order aberrations are less discriminating than corneal high-order aberrations. Corneal coma (Z 3) is the most discriminating aberration to differentiate the healthy cornea from the suspect cornea. Nevertheless, it does not optimally detect keratoconus suspect (Youden=0.334 and sensitivity=60.5%). The statistical analysis suggests that keratoconus suspect and forme fruste are different entities.

Wavefront analysis can be used for the detection of keratoconus suspect, forme fruste, and beginning keratoconus but must be associated with videotopography and pachymetry during consultation of refractive surgery.