Detection of subclinical keratoconus through non-contact tonometry and the use of discriminant biomechanical functions

• Published in: Journal of biomechanics 2016-02, Vol.49 (3), p.353-363
• Main Authors: Peña-García, Pablo, Peris-Martínez, Cristina, Abbouda, Alessandro, Ruiz-Moreno, José M
• Format: Article
• Language: English
• Subjects: Adolescent; Adult; Aged; Area Under Curve; Biomechanics; Collagen; Confidence intervals; Cornea - pathology; Cornea - physiopathology; Corneal biomechanics; Correlation; Discriminant Analysis; Discriminant functions; Eyes; Female; Glaucoma; Humans; Intraocular Pressure; Keratoconus - diagnosis; Male; Mathematical analysis; Middle Aged; Patients; Physical Medicine and Rehabilitation; Receiving; Rest; ROC Curve; Subclinical keratoconus; Tonometry; Tonometry, Ocular; Topography; Variables; Young Adult
• ISSN: 0021-9290; 1873-2380
• DOI: 10.1016/j.jbiomech.2015.12.031

Abstract The purpose of the present study was to develop a discriminant function departing from the biomechanical parameters provided by a non-contact tonometer (Corvis-ST, Oculus Optikgeräte, Wetzlar, Germany) to distinguish subclinical keratoconus from normal eyes. 212 eyes (120 patients) were divided in two groups: 184 healthy eyes of 92 patients aged 32.99±7.85 (21–73 years) and 28 eyes of 28 patients aged 37.79±14.21 (17–75 years) with subclinical keratoconus. The main outcome measures were age, sex, intraocular pressure (IOP), corneal central thickness (CCT) and other specific biomechanical parameters provided by the tonometer. Correlations between all biomechanical parameters and the rest of variables were evaluated. The biomechanical measures were corrected in IOP and CCT (since these variable are not directly related with the corneal structure and biomechanical behavior) to warrant an accurate comparison between both types of eyes. Two discriminant functions were created from the set of corrected variables. The best discriminant function created depended on three parameters: maximum Deformation Amplitude (corrected in IOP and CCT), First Applanation time (corrected in CCT) and CCT. Statistically significant differences were found between groups for this function ( p =2·10−10 ; Mann–Withney test). The area under the Receiving Operating Characteristic was 0.893±0.028 (95% confidence interval 0.838–0.949). Sensitivity and specificity were 85.7% and 82.07% respectively. These results show that the use of biomechanical parameters provided by non-contact tonometry, previous normalization, combined with the theory of discriminant functions is a useful tool for the detection of subclinical keratoconus.