Validation of a novel single-use esthesiometer to test corneal sensitivity

Abstract

Purpose : Assessment of corneal sensitivity (CS) allows to identify alterations of ocular surface sensory innervation associated with several conditions including neurotrophic keratopathy (NK) and dry eye disease (DED). Currently, Cochet-Bonnet (CB) esthesiometer represents the standard for measuring CS. A novel, single-use corneal esthesiometer (KeraSense^TM ) has been developed. The aim of this study is to evaluate the accuracy, reproducibility, and repeatability of the KeraSense^TM in comparison to CB.

Methods : 26 healthy subjects (60±13 years of age) and 41 patients (60±17 years of age) with NK (N=16), diabetes (N=14) and dry eye (N=11) were included. CS was assessed in all subjects with KeraSense^TM (Dompé Farmaceutici SPA, Milan, Italy) and CB. KeraSense^TM assesses CS by disposable devices of fixed nylon thread lengths: 15, 35 and 55 mm. Repeatability and reproducibility of KeraSense^TM were assessed by comparing results of 3 different measurements by the same masked investigator and by two different physicians, respectively. Sensibility, specificity, and accuracy for the diagnosis of NK have also been evaluated.

Results : CB evaluation showed that patients with NK had significant (P<0.001) decrease of CS (30±23 mm) when compared with healthy subjects (59±1 mm), and with patients with DM (57±5 mm) and dry eye (58±2 mm). KeraSense^TM showed repeatability and reproducibility, respectively with 100% agreement between different measurements and 99.6% concordance between different operators (Kappa index interrater agreement P<0.001). Corneal sensitivity values assessed by KeraSense^TM were significantly associated with CB values (ANOVA P<0.001). A 55mm KeraSense^TM value was adequate to exclude an NK diagnosis, while all patients with a KeraSense^TM value of 0 or 15mm had NK. KeraSense^TM response at 35mm showed good sensitivity and specificity to discriminate the presence of NK.

Conclusions : KeraSense^TM showed significant association with CB assessment and represents an expeditious and accurate device to assess CS without need of calibration, and being single-use, ensuring better safety also in the context of SARS-Cov2 outbreak.

This is a 2021 ARVO Annual Meeting abstract.

 

Stacked barplot of KeraSense^TM distribution by class. Chi-squared P value <0.001. Dashed lines represent thresholds for NK diagnosis.
KeraSense^TM values of 55mm were found in 100% of healthy subjects. Patients with NK did not respond to at 55mm, and showed values of 35mm or below.