PURPOSE: To assess clinical outcomes and tear cytokine levels in patients with moderate and severe meibomian gland dysfunction (MGD) after treatment with oral minocycline and artificial tears versus artificial tears only. DESIGN: Prospective, randomized clinical trial. METHODS: Sixty eyes of 60 patients with stage 3 or 4 meibomian gland dysfunction were enrolled. We evaluated the tear film break-up time, Schirmer test results, corneal and conjunctival fluorescein staining results, biomicroscopic examination results of lid margins and meibomian glands, and tear cytokine levels before and after 1 month and 2 months of oral minocycline and artificial tears (group 1) or artificial tears only (group 2). Tear samples were collected and analyzed using a BD Cytometric Bead Array (BD Bioscience, San Jose, California, USA) for detection of interleukin (IL)-1β, IL-6, IL-7, IL-8, IL-12p70, IL-17α, interferon-γ, tumor necrosis factor-α, and monocyte chemotactic protein-1. The Wilcoxon signed-rank test, Mann-Whitney U test, generalized linear model, and linear mixed model were performed. RESULTS: Patients in group 1 showed statistically significant improvement in all clinical signs and symptoms after 1 month and 2 months of treatment. Patients of group 1 showed more significant improvement compared with those in group 2. Patients in group 1 also showed statistically significant reductions in IL-6, IL-1β, IL-17α, tumor necrosis factor-α, and IL-12p70 after 2 months of treatment. CONCLUSIONS: Oral minocycline can provide clinical benefits in treating moderate and severe meibomian gland dysfunction by reducing inflammatory cytokine levels.