Abstract
Purpose :
Controversy exists regarding the utility of tear osmolarity (TO) as a marker for the clinical identification and management of dry eye disease (DED). In a secondary analysis of clinical trial data among patients with moderate to severe DED, we sought to determine whether TO levels are associated with the severity of symptoms and signs of DED, and whether changes in TO are associated with changes in signs and symptoms.
Methods :
Patients (N=405) in the Dry Eye Assessment and Management (DREAM) Study who enrolled at centers with a TearLab Osmolarity System had TO measurements taken of both eyes, before instillation of any eyedrops, at baseline, 6, and 12 months. At baseline, patients had an Ocular Surface Disease Index (OSDI) score of 23-80, and repeatable signs of DED (corneal and conjunctival staining, tear break-up time [TBUT], or Schirmer’s test) in at least one eye. The higher of the TO values from a patient’s two eyes was used for assessing the correlation (r) of TO with the OSDI scores, and the TO value for an eye was used for analyses with ocular signs from the same eye. Statistical analyses accommodated using both eyes.
Results :
Mean (SD) TO at baseline was 302.7 (16.3) in 794 eyes of 405 patients, and at month 12 TO had increased +0.7 (20.3) in 695 eyes of 357 patients. TO was not correlated with OSDI scores at any visit (absolute value of r (abs[r]) ≤0.04 at all times; all p≥0.46), and change in (Δ) TO was not correlated with ΔOSDI (abs[r] ≤0.03; p≥0.59). TO was positively correlated at each visit with corneal and conjunctival stainings (r= 0.12 to 0.23; all p≤0.003), and negatively correlated at most visits with TBUT and Schirmer’s test (r= -0.03 to -0.13; p<0.001 to 0.35). All correlations between change in TO and change in signs were small (abs[r] ≤0.09) and only the correlation between ΔTBUT and ΔTO at month 12 was significant, though in the unexpected direction (r=0.09, p=0.003). All correlation coefficients and p-values are shown in Tables 1 and 2.
Conclusions :
Within DREAM patients, TO and symptoms as measured by the OSDI were not associated at baseline, 6 or 12 months. TO was weakly associated with most signs at the different time points explaining 5% or less of the variability in signs. Changes in TO were not associated with changes in symptoms or signs.
This is a 2020 ARVO Annual Meeting abstract.