Meibomian Gland Contrast Sensitivity and Specificity in the Diagnosis of Lipid-deficient Dry Eye: A Pilot Study

Abstract

Significance: Lipid deficiency due to meibomian gland (MG) dysfunction is believed to account for the vast majority of patients with dry eye compared with aqueous deficiency. Clinicians commonly evaluate MG length to determine a disease, but our research with isotretinoin users suggests that MG contrast is also an important characteristic to consider.

Purpose: This study aimed to determine the sensitivity and specificity of MG contrast for the diagnosis of lipid-deficient dry eye (LDDE).

Methods: This case-control study used demographic data, Standard Patient Evaluation of Eye Dryness (SPEED) scores, average tear lipid layer thickness (TLLT), fluorescein tear breakup time (FTBUT), upper eyelid meibography images, and meibum quality and quantity scores for individuals with LDDE (SPEED score ≥10 and TLLT ≤35 interferometric color units) and normal individuals (SPEED ≤2 and TLLT ≥80 interferometric color units).

Results: Thirty-one eyes of 22 controls (mean ± SD age, 22.7 ± 5.5 years) and 13 eyes of 12 cases (mean ± SD age, 43.9 ± 17.2 years) were included. Normalized MG contrast was significantly correlated with FTBUT (r = 0.35, P = .02), percent MG atrophy (r = -0.50, P < .001), and SPEED scores (r = -0.49, P < .001). The receiver operating characteristic curve for LDDE diagnosis classifiers MG contrast, MG atrophy, and meibum quantity score had areas under the curve of 0.83, 0.64, and 0.73, respectively. Meibomian gland contrast cutoff at 28.3 intensity units yielded optimal correct classification of subjects (84.1%; sensitivity, 0.69; specificity, 0.90). Cases had shorter FTBUT (P < .001), worse meibum quality (P = .02) and quantity (P = .02) scores, and lower MG contrast (P < .001) compared with controls. Subjects with low MG contrast (≤28.3) had 14.9 higher odds of having LDDE (95% confidence interval, 2.84 to 78.4) compared with subjects with high MG contrast (>28.3).

Conclusions: Meibomian gland contrast correlates well with clinical parameters and symptoms, shows good sensitivity and excellent specificity for diagnosing LDDE, and can be a useful diagnostic parameter for monitoring MG changes due to age, disease, or intervention.

Conflict of interest statement

Conflict of Interest Disclosure: TNY is currently an employee for a for-profit company; however, the study was designed and the manuscript was written when she was a researcher under a National Institutes of Health training grant. MCL reports no financial conflict of interest. Each of the authors had full access to the study data and takes full responsibility for their presentation in this article.

Copyright © 2021 American Academy of Optometry.

Figures

Figure 1.

Receiver Operating Characteristic (ROC) curves for meibomian gland contrast, normalized meibomian gland contrast, percent meibomian gland atrophy area, and meibum quantity scores in the diagnosis of lipid-deficient dry eye.

Figure 2.

Meibography images for a control subject (leftmost) and cases. Despite the presence of meibomian gland atrophy in the control subject, tear lipid layer (TLL) was thick provided good tear film stability (fluorescein tear breakup time (FTBUT)) and no dryness symptoms (Standard Patient Evaluation of Eye Dryness (SPEED) score). Cases on the right have minimal atrophy but have thin TLL and severe symptoms. All cases had meibomian gland (MG) contrast less than or equal to 28.3 intensity units, the cut-off for optimal sensitivity and specificity for the diagnosis of lipid deficient dry eye in this study.