Primarily, to evaluate the repeatability and reproducibility of the new noncontact esthesiometer (NCE) in healthy subjects. Secondarily, the corneal sensitivity threshold measurements of the NCE were compared with those of the Cochet-Bonnet esthesiometer (CBE).

Assessment reliability study.

Two examiners measured bilateral corneal sensitivity thresholds by NCE and CBE. Triple NCE measurements were performed per eye at all 5 levels, whereas single CBE measurements were performed per eye. When appropriate, NCE (mBar) and CBE (mm) measurements were converted to millinewtons (mN) for direct comparison between devices. The NCE measurement variations were calculated with the intra-class correlation coefficients (ICC) for the primary objective. For the secondary objective, Bland-Altman plots with 95% limits of agreement (95%-LoA) and Spearman's rank correlation coefficients were used to evaluate the level of agreement and linear relationship between the corneal sensitivity thresholds obtained with both devices. Generalized estimating equation models were used to account for the inter-eye correlation of the same study participant.

Fifty subjects (100 eyes) aged 29 years (median) were included. There were no statistically significant differences in the pressure measured by both observers at each NCE level (all P > .05). There was a high intra-observer and inter-observer repeatability of the NCE measurements (ICC > 0.90) and a strong linear correlation (rho > 0.7) at each NCE level. The difference between the mean corneal sensitivity thresholds (mN) measured with the NCE (0.052 ± 0.021 mN) and CBE (0.046 ± 0.005 mN) was statistically significant ( P = .001). Bland-Altman analysis revealed a differential bias of +0.00578 mN (95%-LoA -0.03677 – +0.04833 mN) between threshold measurements.

The NCE provides reliable user-independent corneal sensitivity measurements. The NCE and CBE do not yield similar values. Thus, both devices cannot be used interchangeably.