Evaluation of tear film interference patterns and measures of tear break-up time

Academic Article


  • Purpose. The purpose of this study was to compare the agreement within and between examiners when evaluating real-time and digital photographs of TearScope Plus tear film interference patterns. We also evaluated between-examiner agreement of noninvasive digital tear break-up time (NITBUT) using the TearScope, the relation between NITBUT and tear film interference patterns, and the relation between NITBUT and tear break-up time (TBUT). Methods. Forty nondry eye patients were seen for digital imaging of the tear film using the TearScope Plus mounted on a slit-lamp with the EyeCap Ophthalmic Image Capture System. Two independent examiners first assessed the tear interference pattern in real-time and then captured the patterns and NITBUT digitally for a period of approximately 60 s. Each image series was later graded for interference pattern and NITBUT by the two independent, masked examiners. Results. Within-examiner comparisons of real-time and digital tear pattern photograph grading showed moderate to substantial agreement for experienced examiners (kappa for examiner 1 = 0.76 and kappa for examiner 2 = 0.55). Between-examiner agreement when grading tear patterns was also substantial in real-time (kappa = 0.72) and moderate when using digital photography (kappa = 0.59). The between-examiner mean NITBUT was 11.2 ± 6.8 s and the 95% limits of agreement between examiners were -19.2 to +9.2 s. The mean TBUT was 7.6 ± 10.4 s, and it was on average 3.7 ± 12.0 s shorter than NITBUT (p = 0.06). Thicker tear film interference patterns graded photographically were associated with longer NITBUTs for both examiners (p = 0.001). Conclusions. There is moderate to substantial within- and between-examiner agreement when comparing real-time and digital tear interference patterns photographs when using the TearScope. Although there is considerable between-examiner variability with the NITBUT test, the thicker lipid layer tear patterns tend to be associated with longer NITBUTs.
  • Published In

    Digital Object Identifier (doi)

    Author List

  • Nichols JJ; Nichols KK; Puent B; Saracino M; Mitchell GL
  • Start Page

  • 363
  • End Page

  • 369
  • Volume

  • 79
  • Issue

  • 6