The Effect of Repeated Measures on the Evaluation of Non-Invasive Tear Break Up Time (NITBUT)

Abstract

Purpose : Videokeratographs are used to obtain objective measures of NITBUT. Multiple measures are averaged with the goal of reducing measurement variability, but repeated measures may destabilize the tear film and increase measurement variability. The effect of averaging NITBUT for subjects with mild to moderate dry eye disease undergoing low-level red light therapy (LLLT) is examined.

Methods : Thirty participants (mean ±SD age: 31.1 ±9.5 years, range 24-58 years, male = 14, female = 16) with mild to moderate dry eye disease completed the study. Three NITBUT measurements of the right eye were taken with the Oculus Keratograph 5M in a temperature- and humidity-controlled room. Subjects were instructed to take as much time as needed between measurements for ocular comfort, with a minimum 1 minute required break between measures. Within-subjects standard deviation was compared for 2 and 3 repeated NITBUT measurements, and Bland Altman plots of within-subject standard deviation and single vs. average NITBUT time differences examined.

Results : Average within-subject standard deviation for 1^st NITBUT pre-LLLT was 1.65 seconds for 2 repeated measures and 2.18 seconds for 3 repeated measures, and 1.60 seconds for 2 measures and 1.78 seconds for 3 measures of Average NITBUT. Bland Altman analysis showed a mean difference of 0.53 (95% Limits of Agreement -1.59 to 2.65) seconds in standard deviation between 3 and 2 repeated measures for 1^st NITBUT, and a mean difference of 0.18 seconds (95% Limits of Agreement -1.75 to 2.11) in standard deviation between 3 and 2 repeated measures for Average NITBUT. Average within-subject standard deviation after three 15-minute sessions of LLLT was 2.66 seconds for 2 measures and 3.73 seconds for 3 measures of 1^st NITBUT, and 2.06 seconds for 2 measures and 3.01 seconds for 3 measures of Average NITBUT. Bland Altman analysis showed a mean difference of 1.07 seconds (95% Limits of Agreement -2.89 to 5.04) in standard deviation between 3 and 2 repeated measures for 1^st NITBUT, and a mean difference of 0.95 seconds (95% Limits of Agreement -2.59 to 4.49) in standard deviation between 3 and 2 repeated measures.

Conclusions : Compared to 2 repeated measures, averaging three NITBUT measures did not reduce variability of measures for 1^st and Average NITBUT prior to or after LLLT. The number of measures and timing between acquisitions can impact the evaluation of NITBUT.

This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.