Intereye Agreement in Dry-Eye Signs in the DREAM Study: Implications for Future Dry-Eye Trials

Abstract

Purpose: To investigate intereye agreement in dry-eye disease (DED) signs in the Dry Eye Assessment and Management study. Methods: Tear break-up time (TBUT), Schirmer test, conjunctival staining, corneal staining, meibomian gland dysfunction (MGD), and tear osmolarity were measured at baseline, 3, 6, and 12 months. Intereye agreement was assessed by intraclass correlation coefficient, weighted kappa, and percentage of participants with absolute intereye difference (AID) exceeding a clinically significant threshold (2 points for conjunctival staining and MGD, 2 seconds for TBUT, 3 points for corneal staining, 5 mm/5 minutes for Schirmer test, and 8 mOsms/L for osmolarity). The worse eye at each visit for each DED sign was determined as the eye with a sign worse than the contralateral eye by at least the clinically significant threshold. Results: DED signs had moderate-to-good intereye agreement with intraclass correlation coefficient ranging from 0.45 (tear osmolarity) to 0.81 (corneal staining and Schirmer test) and weighted kappa from 0.58 (plugging) to 0.69 (lid secretion). Percentage of participants exceeding threshold AID was 15% to 20% for conjunctival staining, 11% to 15% for TBUT, 17% to 21% for MGD, 13% to 18% for corneal staining, 21% to 23% for Schirmer test, and 44% to 47% for osmolarity. The eye with a worse DED sign ranged between 36% (TBUT) and 80% (osmolarity) of participants. Conclusions: Participants demonstrated moderate-to-good intereye agreement, yet a substantial portion showed clinically significant intereye differences in each sign. The worse eye was not the same eye in the majority during follow-up. These findings suggest considering signs from both eyes in future DED trials.