A systematic review of the effect of omega-3 supplements on meibomian gland dysfunction

Identification of trials and consideration of data

This meta-analysis study has followed the standards and guidelines of Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). Randomized clinical trials (RCTs) that included nontreated groups versus treated groups selected by searching the MEDLINE and Clinicaltrials.gov. The following keywords were used in MEDLINE research: “Randomized Clinical Trials,” “Meibomian gland dysfunction,” “Omega-3” AND “Adults.” For Clinicaltrials.gov database: “Meibomian gland dysfunction,” “Omega-3,” “Adults” AND “Completed.” The selected studies applied no language or other restrictions. Database search was done to include only studies that had a nontreated group versus a treated group. RCTs that did not include placebo groups, and studies that have not being completed or terminated or are still under recruitment, or have not yet been officially approved, have been excluded.

For bias detection, in each eligible study a 7-point predefined quality control was used. The corresponding risk of bias was categorized as low (L), high (H), or unknown (U) for each quality item according to Higgins et al., 2011 & Liberati et al., 2009.^11,12 Unknown is used to judge insufficient information if the author does not report the required data. The complete outcome data were judged as “low risk” or “high risk” or “unknown.” The “low risk” is used when follow-up percentage of participants lost was lower than 5% and “high risk” when follow-up loss percentage was more than 20%. For other potential sources of bias, including the bias source, including the funding source reported in each protocol, the term “other bias” has been used.

The overall treatment effect was calculated and the study weight for each study was calculated. Due to the larger sample size of some of the studies, the study weight was calculated and the “true effect” for each study was shown. The larger sample size provides more information than the smaller sample size.

Statistical analyses

The odd ratios (OR) was calculated in each study protocol to distinct the comparison of placebo group to the omega-3 in MGD. OR values more than 1 mean that the investigated treatment has a positive effect in the MGD protection and treatment compared with placebo. The pooled ORs were calculated using a random effect model.

The equivalent Z test was done to test the null hypothesis for each pooled OR and if p < 0.05 shows that the comparison between the two groups is statistically significant which leads to null hypothesis rejection.

The study heterogeneity was assessed with the Cochran Q and I² statistics. For the heterogeneity qualitative interpretation, the I² values of at least 50% were considered to reflect considerable heterogeneity, while values of at least 75% indicated large heterogeneity, as per the Cochrane Handbook. Publication bias was evaluated using both the graphical funnel plot and the Egger’s statistical test. The Comprehensive Meta-Analysis Version 3 software was used for statistical analysis.

Selection and characteristics of the study

MEDLINE and Clinicaltrials.gov database searches resulted in six and one results, respectively. The CENTRAL Register of Controlled Trials database identified no additional RCTs. Duplicate studies were removed, titles and abstracts from the remaining seven studies were evaluated, one study was excluded, and six eligible studies were assessed for the systematic review, and four were assessed for meta-analysis study. Two articles were excluded from the meta-analysis study due to missing data of the primary outcome (TBUT in seconds) or the endpoint was 12 months instead of 90 days. Six studies were retained for qualitative eligibility,^13–18 and four studies were retained for quantitative eligibility primary outcome (TBUT)^13,14,17,18 and two studies were retained for secondary outcome (OSDI^13,17 & tear osmolarity^13,14). The flowchart presenting the selection of eligible studies is summarized in Figure 1. The Studies that met the inclusion criteria and the characteristics of the included studies are summarized in Table 1. All the enclosed studies included placebo group and approved doses of omega-3 treatment. The approved doses of omega-3 used that are included in the present analysis: Deinema and colleagues ¹³ (1000 mg/day eicosapentaenoic acid (EPA), 500 mg/day docosahexaenoic acid (DHA) for 90 days.), Epitropoulos and colleagues ¹⁴ (1680 mg of EPA/560 DHA daily for 12 weeks), Korb and colleagues ¹⁵ (combination treatment, 100 mg omega-3 for 3 months).

OPEN IN VIEWER

Figure 1.

Flowchart presenting the selection of eligible studies.

OPEN IN VIEWER

Table 1.

Characteristics of included studies.

Study ID	Study type	A	M	Intervention dosage	Pt group	Time frame	Age (years)	Actual enrollment	Primary outcome
Deinema and colleagues 13	I	R	D	Placebo (Olive oil 1500 mg/day) or Krill Oil (945 mg/day EPA + 510 mg/day DHA), Fish oil (1000 mg/day EPA +500 mg/day DHA)	Mild to moderate DED	90 days	⩾18	60	Mean change in tear osmolarity and DED symptoms (OSDI) score
Epitropoulos and colleagues 14	I	R	D	4 Soft gels total Treatment Group: 1680 mg of EPA /560 mg of DHA or control group = 3136 mg of linoleic acid daily	DED MGS STAGE 1 OR 2	12 weeks	⩾18	105	Tear osmolarity, TBUT, OSDI, Fluorescein corneal staining, Schirmer test
Korb and colleagues 15	I	R	D	Group A = Treatment Group lipid emulsion eye drops, omega-3 supplements, and lid hygiene with eyelid wipes Group B = Apply warm, wet compresses once daily, 8 minutes per day.	MGD	3 Months	⩾18	26	Meibomian gland functionality
Macsai 16	I	R	D	Group A: Oral omega-3 dietary supplementation: two 1000-mg capsules 3 times a day. Group B = Placebo (olive oil)	Blepharitis and simple obstructive MGD	12 Months	⩾18	38	TBUT, meibum score, and OSDI score
Malhotra and colleagues 17	I	R	D	Both groups received warm compresses, lid massage, and artificial tear substitutes. Treatment group: Oral supplements of 1.2 g volatile FAs per day	Moderate MGD	3 Months	⩾18	60	Measure contrast sensitivity under photopic and scotopic conditions
Olenik and colleagues 18	I	R	D	Group A: Cleaning the lid margins, artificial tears, a placebo oral agent. Group B Cleaning the lid margins with neutral baby shampoo and artificial tears, omega-3 FA oral supplementation	MGD and no tear stability	3 Months	23–85	61	TBUT, Schirmer I test, OSDI, MG Expression, Evaluation of lid margin inflammation, interpalpebral and corneal staining

Abbreviations: A, allocation; D, double (participant, investigator); DED, dry eye disease; FA, fatty acid; I, interventional (clinical trial); M, masking; MG, meibomian gland; MGD, meibomian gland dysfunction; OSDI: Ocular Surface Disease Index; P, prevention; PA, parallel assignment; Pt, patient; Q, quadruple (participant, care provided, investigator outcome assessor); R, randomized; T, treatment; TBUT, tear break up time.

Malhotra and colleagues ¹⁷ (1.2 g omega-3 fatty acids for 12 weeks), Macsai ¹⁶ (6 1000 mg capsules of flaxseed oil daily for 12 months). The studies duration varied from 12 weeks^{13–15,17,18} to 12 months. ¹⁶

Independent studies bias risk

The risk of bias of the included studies is summarized in Table 2. Random sequence generation was sufficient in all studies except for two.^17,18 Allocation concealment was adequately reported in three trials,^13,15,16 two trials were not reported in the article,^14,18 and one was high. ¹⁷ The blinding of participants and personnel was sufficient in two studies^13,16 except for four studies.^14,15,17,18 The blinding of the outcome assessment was sufficient in three studies except for three, one was unknown, and two were high risk. All of the study protocols reported a follow-up percentage loss between 5% and 20%.^{13,14,15,16,17,18} Selective bias and other bias were not detected in any of the six studies. The study protocols are supported financially, partly, or solely by the pharmaceutical companies that produce and market the drug under consideration in each study. Sources of funding were disclosed in all six study protocols to provide enough information to permit fair judgment.

OPEN IN VIEWER

Table 2.

Risk of bias summary. Review author’s judgments about each risk of bias item for each included study.

Study ID	Deinema et al., 2017	Epitropoulos et al., 2016	Korb et al., 2015	Macsai 2008	Malhotra et al., 2015	Olenik et al., 2013
Random Sequence Generation (Selection Bias)	L	L	L	L	H	U
Allocation Concealment (Selection Bias)	L	U	L	L	H	U
Blinding of Participants and Personnel (Performance Bias)	L	U	H	L	H	U
Blinding of Outcome Assessment (Detection Bias)	L	U	H	L	H	L
Incomplete Outcome Data (Attrition Bias)	L	L	L	L	L	L
Selective Reporting (Reporting Bias)	L	L	L	L	L	L
Other Bias	L	L	L	L	L	L

U=Unkown risk of bias, Insufficient information, L= Low risk of bias, H= High risk of bias, High= 75 % & Low=25 %

Overall final analyses

The systematic review of the selected six articles showed that omega-3 treatment solely or with combination treatment as artificial tears or vitamins improved from baseline in MGD patients. Combination treatment included lid hygiene, artificial tears with nonpreservatives, oral vitamin supplements, and 1000 mg of omega-3 fatty acids. The meta-analysis primary outcome was invasive NaFl-TBUT improvement from the baseline of the treatment group compared with control group. The secondary outcome was Schirmer I, Ocular Surface Disease Index (OSID), and tear osmolarity test.

Patients receiving omega-3 treated group had significantly better improvement outcome NaFl-TBUT compared with patients receiving placebo (OR = 8.72; 95% CI: 4.73, 16.09; p < 0.001; Figure 2). Heterogeneity was not statistically significant using the I² value of 18.60% and p = 0.383. In addition, no publication bias was detected using the funnel plot inspection (Supplemental S1-Figure 1) and the Egger’s statistical test (p = 0.15).

OPEN IN VIEWER

Figure 2.

Forrest plot depicting pooled analysis of the % of TBUT improvement of the treated group versus the control group.

The secondary outcome was Schirmer I test, OSID, and tear osmolarity. For the Schirmer test, Malhotra and colleagues, ¹⁷ Epitropoulos and colleagues, ¹⁴ and Deinema and colleagues ¹³ showed that the Schirmer I score improvement are not statistically significant in either group.

For OSID, patients receiving omega-3 had a significant improvement compared with patients receiving placebo, p < 0.005 (Figure 3). Only two studies included tear osmolarity, and both studies showed a statistically significant improvement in tear osmolarity in the omega-3 group compared with placebo.^13,14 These data suggest that omega-3 improves the signs and symptoms of MGD.

OPEN IN VIEWER

Figure 3.

Forrest plot depicting pooled analysis of the % of OSDI improvement of the treated group versus the control group.