Medical records of canine patients that had undergone TF examination after a diagnosis of OSD at the Ophthalmology Referrals Unit of the Turin Veterinary Centre, Italy from 2016-2019 were reviewed. Informed consent form was obtained from the owners for the use of the patient data. Diagnosis of OSD was made after a complete ophthalmic examination and was based on clinical signs of loss of ocular surface homeostasis caused by anatomical and/or functional alterations of one or more components of the lacrimal functional unit including: the ocular surface (cornea and conjunctiva) with sensory nerve endings, the afferent and efferent innervation to stimulate tear secretion, the lacrimal gland, the meibomian gland, the conjunctival goblet cells and the eyelids.
Information from each case included gender, breed, skull conformation (divided into brachycephalic and non-brachycephalic skull), age, eye examined (left or right), concurrent ocular diseases, meibography and interferometry.
All dogs in this study had undergone a complete and bilateral ophthalmic examination by a board-certified veterinary ophthalmologist. The examination included assessments of the palpebral reflex, menace response, pupillary light and dazzle reflexes, Schirmer tear test 1 (STT-1) (Eickemeyer®; Tuttlingen, Germany), slit-lamp biomicroscopy (Keeler®; PSL Classic, Windsor, Berkshire, UK), fluorescein staining (Optitech® eyecare; An-Vision, Hennigsdorf, Germany), applanation tonometry (Tono-Pen Vet®; Reichert Technologies, Depew, NY, USA), and indirect ophthalmoscopy (Omega® 2000; Heine, Herrsching, Germany).
Patients that were on any kind of OSD treatment met inclusion criteria only if treatment had been suspended ≥2 days before performing the TF examination. Records of dogs with orbital or nictitating membrane disorders, with intraocular diseases (i.e. uveitis, cataract, glaucoma, retinal detachment), or with a recent history of surgery (less than one month) were excluded from the study.
A hand-held ocular surface analyzer (OSA-VET®, SBM Sistemi, Torino, Italy), equipped with infrared and white led lights for meibography and interferometry, respectively, was used to perform non-contact infrared meibography of the upper lid and interferometry (Fig. 2).
Meibography was mainly directed at the diagnosis of MGD, and only 2 categories were considered: MGD affected or unaffected. A diagnosis of MGD was achieved when terminal duct obstruction was present along over 75% of the eyelid margins and/or when meibomian glands dilatation, shortening, atrophy or dropout affected more than 50% of the eyelid . (Fig. 3).
Meibomian gland expression has been used in some cases to evaluate the color and consistency of the expressed meibum or to confirm ductal occlusion.
Interferometry was used to assess the TF-LL patterns, each one reflecting a specific thickness of the TF-LL. The patterns were classified according to a grading scale recommended by the instrument manufacturer, with categories adopted for humans [8, 9, 29], and adapted for veterinary use, according to the available literature [10-12]. A five-interval scale was used as follows (Fig.4): Grade 0 included cases of almost complete absence of the aqueous phase, with lipid-contaminated mucus over the surface of the corneal epithelium; this grade was added since, without a liquid component present on the cornea it was impossible to evaluate the lipids, sometimes scattered over the ocular surface in static colored islets. Grade 1 (15-30 nm), when faintly visible homogeneous meshwork pattern was present; grade 2 (31-60 nm), when a more compact meshwork pattern with grey waves was observed; grade 3 (61-100 nm), when a meshwork with waves and interference fringes with some colors was noted and grade 4 (>100 nm), when waves with many colors were present.
STT-1 was used to assess tear production. The continuous variable was subdivided into three categories: < 10 mm/min, 10-15 mm/min and >15 mm/min.
In order to avoid excessive inter-observer differences in the assignment of ranks the interferometry was done by the same experienced operator.
The data were retrospectively collected and reviewed, and elaborated for statistical analysis.
Normality of the underlying distribution of the continuous variables was assessed using the Shapiro-Wilk test.
Continuous variables with normal distribution were presented as mean and standard deviation (SD), and those with non-normal distribution were presented as median (range), while absolute frequencies and percentage were used for categorical variables.
The MGD frequency distribution was evaluated within different categories (skull conformation, gender, eye and STT-1) using the Chi-squared test.
Independent variables (gender, age, skull conformation) were evaluated as risk factors for developing MGD (dependent variable) in a logistic regression model and reported as odd ratios (ORs) with their 95% confidence (95% CI) intervals. ORs were adjusted by gender and age. The skull conformation was considered as a categorical variable and brachycephaly was used as the reference level to which nonbrachycephaly was compared. Association between the presence of MGD and interferometry grades was analyzed by ORs.
Probability values of <0.05 were considered significant.
Statistical analysis was performed with R software (version 3.4.3. R project, Auckland, New Zealand) and MedCalc software (version 19.0.6 Bvba, Ostend, Belgium).