Corneal thickness indices discriminate between keratoconus and contact lens-induced corneal thinning.

PURPOSE

To evaluate two indices generated from measurements obtained from the Orbscan Corneal Topography System (CTS; Orbscan, Inc., Salt Lake City, UT) to distinguish contact lens-induced corneal thinning from keratoconus. The corneal thickness index (CTI) was used to compare central and peripheral corneal thicknesses. The discriminant function 1 (DF1) was used to evaluate corneal thickness and central keratometry measurements.

DESIGN

Case-control study.

PARTICIPANTS

Fifty-four patients with keratoconus or suspected keratoconus, 75 contact lens wearers, and 67 normal controls.

METHODS

In the initial model-building study, the central and peripheral corneal thickness and central keratometry (K) readings were evaluated retrospectively in 1 eye each of 23 patients with clinically defined keratoconus, 31 contact lens wearers, and 43 normal patients with the Orbscan CTS. Two methods (corneal thickness index [CTI] and discriminant function 1 [DF1]) were evaluated for their efficacy in classifying these conditions. The CTI was designed as the ratio between mean peripheral corneal thickness and the central corneal thickness, and the formula for the DF1 was as follows: DF1 = 0.044 x central thickness - 0.030 x nasal thickness + 9.210 x CTI - 0.157 x Max K - 8.9. In the subsequent validation study, the results were verified in a separate group of keratoconus patients (n = 23), keratoconus suspects (n = 8), contact lens wearers (n = 44), and normal patients (n = 24).

MAIN OUTCOME MEASURES

Corneal thickness index (CTI) and discriminant function 1 (DF1).

RESULTS

The corneal thickness in patients with keratoconus and contact lens wearers was significantly thinner than that of normal eyes in the central and eight peripheral measured sites (P < 0.001). The corneal thickness in contact lens wearers was significantly greater than in keratoconus patients in the inferotemporal (P = 0.013), inferior (P = 0.003), and central (P < 0.001) sites and was borderline different in the superior site (P = 0.07). In the model-building study, the CTI in keratoconus patients (1.28 +/- 0.15) was significantly greater than in contact lens-wearing (1.10 +/- 0.03; P < 0.001) and normal eyes (1.09 +/- 0.04; P < 0.001). The CTI was not significantly different between normal and contact lens-wearing eyes (P = 0.68). A CTI value of 1.16 or more showed a sensitivity of 91% and specificity of 99% in differentiating keratoconus from contact lens-wearing and normal eyes. The DF1 value was significantly lower in keratoconic eyes than in contact lens-wearing and normal eyes. A DF1 value of -0.6 showed a 96% sensitivity and 99% specificity in differentiating keratoconus from contact lens-wearing and normal eyes. Similar results were obtained in the validation study.

CONCLUSIONS

Corneal thickness indices generated from the Orbscan CTS appear to be sensitive and specific for diagnosing keratoconus. These indices may prove to be clinically useful parameters for distinguishing keratoconus from contact lens-induced corneal thinning.