Factors Influencing Corneal Biomechanics in Diabetes Mellitus

Diabetes mellitus (DM) induces changes in corneal biomechanical properties. The influence of disease-specific factors was evaluated, and a novel DM index was created.

Eighty-one patients with DM and 75 healthy subjects were matched according to age, intraocular pressure, and central corneal thickness. Information on the disease was collected, and measurements with the Ocular Response Analyzer and the Corvis ST were taken. Results were compared between the groups, and the influence of disease-specific factors was evaluated. From dynamic corneal response parameters, a DM index was calculated.

In DM, corneal hysteresis was higher than in healthy subjects (10.5 ± 1.9 vs. 9.7 ± 1.9 mm Hg, P = 0.008). In addition, dynamic corneal response parameters showed significant differences. Among others, highest concavity (HC) (17.212 ± 0.444 vs. 16.632 ± 0.794 ms, P < 0.001) and A2 time (21.85 ± 0.459 vs. 21.674 ± 0.447 ms, P = 0.017) as well as A1 (0.108 ± 0.008 vs. 0.104 ± 0.011 mm, P = 0.019) and A2 deflection amplitudes (0.127 ± 0.014 vs. 0.119 ± 0.014 mm, P < 0.001) were increased in DM. In DM type 1, HC deformation amplitude (1.14 ± 0.19 vs. 1.095 ± 0.114 mm, P = 0.035) was higher than in type 2. The time of deflection amplitude max correlated with the severity of retinopathy (R = 0.254, P= 0.023). In case of diabetic maculopathy, A1 velocity (0.155 ± 0.018 vs. 0.144 ± 0.019 ms, P = 0.043) and A2 time (22.052 ± 0.395 vs. 21.79 ± 0.46 ms, P = 0.04) were increased. Deformation amplitude max (R = 0.297, P = 0.024), HC time (R = 0.26, P = 0.049), HC deformation amplitude (R = 0.297, P = 0.024), and A2 deformation amplitude (R = 0.276, P = 0.036) were associated to disease duration. The DM index revealed a sensitivity of 0.773 and a specificity of 0.808 (area under the curve of receiver operating characteristic = 0.833).

In DM, changes in corneal biomechanics were correlated with disease-specific factors. The DM index achieved reliable sensitivity and specificity values.