Deinema 2017.

Methods	Study design: randomized, parallel‐group, controlled trial Study site(s): single center Number randomized (total and per group): 60 participants in total; 20 in each of the 3 intervention groups Unit of randomization (individual or eye): individual Exclusions after randomization: none Losses to follow‐up: 6 participants in total; 3 participants in the placebo group; 1 participant in the fish oil group; 2 participants in the krill oil group Unit of analysis (individual or eye): individual Reported power calculation? (Y/N): Y (90% power) Reported subgroup analysis? (Y/N): N
Participants	Baseline characteristics Country: Australia Age (mean ± SE, range): 39.4 ± 3.4 years in the fish oil group; 42.3 ± 3.8 years in the krill oil group; 46.2 ± 4.5 years in the placebo group Gender: 47% (n = 9) female in the fish oil group; 72% female (n = 13) in the krill oil group; 82% female (n = 14) in the placebo group Inclusion criteria: 1. ≥ 18 years of age 2. Provision of written informed consent to participate 3. For females of child‐bearing potential, a negative pregnancy test result at baseline 4. OSDI score ≥ 18 and < 65 5. Tear osmolarity ≥ 316 mOsmol/L in at least 1 eye 6. Distance best corrected visual acuity ≥ 20/40 Snellen equivalent in each eye 7. IOP ≤ 21 mmHg in both eyes 8. Ability to follow study instructions, with the intention of completing all required visits Exclusion criteria: 1. Any of the following general medical conditions: diabetes, bipolar disorder, atrial fibrillation, implanted defibrillator, familial adenomatous polyposis, systemic immunocompromise, bleeding disorders, or history of liver disease 2. A major change to diet or dietary supplement intake in the 3 months before enrollment 3. Consumption of essential fatty acid oral supplements in the 3 months before enrollment 4. Current consumption of any systemic anticoagulants 5. Known allergy or sensitivity to study supplements or any of their components (eg, any of fish, seafood, peanuts, nuts, oil, gelatin) 6. Females of child‐bearing potential who were planning a pregnancy over the course of the study, or currently pregnant or breastfeeding 7. Current enrollment in another interventional drug or device study or participation in such a study within 30 days of anticipated entry into this study 8. Anticipated contact lens wear during the study or contact lens wear in the month before enrollment 9. Any scheduled or planned ocular or systemic surgery or procedure during the study 10. Presence of severe dry eye at baseline, defined as (1) OSDI score > 65, and/or (2) corneal or conjunctival fluorescein staining of Grade 5 (Oxford scheme) in any zone of either eye 11. Start date of any systemic medication (including over‐the‐counter, herbal, prescription, or nutritional supplements) that may affect tear film or vision; less than 3 months before enrollment; or change in dosage anticipated during the study 12. Presence of any of active ocular infection or non‐keratoconjunctivitis sicca ocular inflammation, active ocular allergy, history of recurrent herpes keratitis, or active disease within 6 months of baseline; a corneal disorder or abnormality that affects corneal sensitivity or normal spreading of the tear film (except superficial punctate keratitis); severe blepharitis that in the judgement of the investigator may interfere with interpretation of study results 13. Occlusion of the lacrimal puncta with punctal plugs or cauterization in the 3 months before enrollment 14. History of ocular surgery/trauma that could affect corneal sensitivity and/or tear distribution within 6 months of enrollment 15. Use of any of the following topical medications in the 3 months before baseline: corticosteroids, non‐steroidal anti‐inflammatories, or cyclosporine 16. A medical or ocular condition or a personal situation that in the principal investigator's opinion may put the patient at significant risk, may confound study results, or may interfere significantly with participation in the study 17. Cultural or religious beliefs that exclude the consumption of certain or all animal products Equivalence of baseline characteristics? (Y/N): N (krill oil group had a significantly shorter NaFl TBUT at baseline compared with placebo and fish oil groups [P = 0.02], but this was not a primary outcome measure)
Interventions	Intervention #1 (treatment group #1): fish oil once daily (daily dose of 1000 mg EPA and 500 mg DHA) Intervention #2 (treatment group #2): krill oil once daily (daily dose of 945 mg EPA and 510 mg DHA) Intervention #3 (control group): 1500 mg of olive oil once daily Length of follow‐up: 90 days Notes: participants using topical lubricant eye drops at baseline (day 1) were allowed to continue to use these throughout the study. At each study visit, the investigator questioned participants with regard to how frequently they had used lubricant eye drops over the past month. There was no significant change in the frequency of eye drop utilization, compared with baseline, in any of the intervention arms over the study duration (data not shown) Participants were instructed to maintain their current dietary habits. Approximate dietary intake of omega‐3 essential fatty acids was determined by asking participants about their consumption of omega‐3‐rich foods over the preceding month. Participants were asked to quantify the approximate serving size (25, 50, 100, 150 g) and frequency of consumption of foods (including fish, oils, nuts, seeds, and spreads) containing greater than 1000 mg of combined EPA, DHA, docosapentaenoic acid, and alpha‐linolenic acid per 100 g edible portion (Australia New Zealand Food Authority, 2011; Nutrient Data Laboratory and Beltsville Human Nutrition Research Centre, 2011). At subsequent visits, participants were questioned about changes in their diet or medications and about compliance with taking the study supplements
Outcomes	Primary outcome(s): change from baseline in tear osmolarity (mOsmol/L) and OSDI score Secondary outcome(s): change from baseline in tear film stability (measured using fluorescein TBUT and non‐invasive TBUT); bulbar and limbal redness; ocular surface staining; tear production (Schirmer test with anesthesia); tear volume; anterior blepharitis; meibomian gland capping; basal tear levels of inflammatory cytokines Safety outcomes: change from baseline in best‐corrected visual acuity and intraocular pressure; and monitoring for adverse events Adverse events reported? (Y/N): Y Measurement time points (specify intervals at which outcomes were assessed): baseline, 30, 60, and 90 days Other issues with outcome assessment (eg, quality control for outcomes, if any): none
Notes	Study dates: October 29, 2014, to August 18, 2015 Funding source(s): "Rebecca L. Cooper Medical Foundation, Sydney, New South Wales, Australia (2015) and a University of Melbourne Early Career Researcher grant, Parkville, Victoria, Australia (2015)" Conflicts of interest: "L.E.D.: Grant from Rebecca L. Cooper Medical Foundation, Sydney, New South Wales, Australia (2015) and a University of Melbourne Early Career Researcher grant, Parkville, Victoria, Australia (2015). The sponsors had no role in the design or conduct of this research" Publication language: English Registered on clinical trials registry? (Y/N): Y ‐ ANZCTR (ACTRN12614001019695)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"An independent data manager generated a participant randomization sequence using a random number generator in Microsoft Excel (2007; Microsoft Corporation, Redmond, WA)"
Allocation concealment (selection bias)	Low risk	"This randomization schedule was provided to an independent compounding pharmacist (Dartnell’s Pharmacy, Victoria, Australia), who repackaged the study supplements into identical, opaque containers using the randomization schedule. Supplement containers were labeled with the appropriate participant randomization code (from 001 to 060). Eligible participants were sequentially enrolled by a masked research optometrist (L.A.D.), who dispensed the investigational product labeled with the appropriate code"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	"Participants were masked to treatment allocation, as achieved by all investigational products being dispensed in identical opaque containers." ; "All study personnel, including the principal investigator (L.E.D.), clinical outcome assessor (L.A.D.), laboratory outcome assessors (C.Y.W., D.C.J.), and co‐investigators (H.R.C., A.J.V.), were masked to participant allocation. Following completion of all participant visits, data were analyzed only with knowledge of the simple randomization code (i.e., group A, B, or C allocation). Full unmasking of treatment allocation by the independent data manager only occurred after statistical analyses were complete"
Blinding of outcome assessment (detection bias) All outcomes	Low risk	"All study personnel, including the principal investigator (L.E.D.), clinical outcome assessor (L.A.D.), laboratory outcome assessors (C.Y.W., D.C.J.), and co‐investigators (H.R.C., A.J.V.), were masked to participant allocation. Following completion of all participant visits, data were analyzed only with knowledge of the simple randomization code (i.e., group A, B, or C allocation). Full unmasking of treatment allocation by the independent data manager only occurred after statistical analyses were complete"
Incomplete outcome data (attrition bias) All outcomes	High risk	6 out of 60 (10%) participants who were randomized were not included in the final analysis
Selective reporting (reporting bias)	Low risk	Reported outcomes are consistent with prespecified outcomes in the clinical trial registry
Other bias	Low risk	No other apparent sources of bias