NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Hodge W, Barnes D, Schachter HM, et al. Effects of Omega-3 Fatty Acids on Eye Health. Rockville (MD): Agency for Healthcare Research and Quality (US); 2005 Jul. (Evidence Reports/Technology Assessments, No. 117.)

  • This publication is provided for historical reference only and the information may be out of date.

This publication is provided for historical reference only and the information may be out of date.

Cover of Effects of Omega-3 Fatty Acids on Eye Health

Effects of Omega-3 Fatty Acids on Eye Health.

Show details



Only 16 studies, described in 16 published journal reports, were deemed relevant for inclusion in this systematic review. These studies addressed nine of the 23 questions that we posed in this evidence synthesis. This includes a focus on the possible adverse effects of the intake of omega-3 fatty acids. A brief overview of the studies is now presented, and it is organized in terms of which eye disease/visual impairments constitute the most important public health concerns, particularly in developed countries.

The eye disease/visual impairment likely constituting the most important public health concern is ARMD, and it was investigated by the largest number of studies. Eight studies examined either its primary79–84 or secondary prevention.60, 76 The ARMD studies involved various research designs, including one RCT76 and various observational designs.60, 79–84 Therefore, half of the included studies examined the possible protective influence of the intake of omega-3 fatty acids on ARMD.

The next most important health concern is vascular disease of the retina in patients with diabetes. However, only one (quasi-experimental) study investigated the secondary prevention of events related to diabetic retinopathy.75 Three studies examined the questions of primary85, 86 or secondary prevention78 with respect to cataract. These primary and secondary prevention studies employed observational85, 86 and quasi-experimental designs,78 respectively. The possible impact of the intake of omega-3 fatty acids on RP was examined in four studies,72, 73, 77 with a report published by Hoffman et al. describing two of these investigations.72 Study designs included one RCT73 and three quasi-experimental designs.72, 77 No study was found which investigated outcomes related to occlusions of the retinal veins or arteries.

For each question, in turn, we now present a synthesis of the key findings. This includes a critical appraisal of the individual studies from which the results were drawn. Attention is paid to the numbers, size, quality and applicability (i.e., to relevant North American populations) of studies in trying to ascertain larger patterns of result. The broader implications of these findings, including potential future research, are highlighted. We begin with the cross-cutting issue of safety.

Evidence Synthesis and Appraisal

Only one interventional study collected adverse event data from study participants. When they carefully observed the possible links between the characteristics of their intervention (i.e., source of exposure, amount of omega-3 fatty acid content, numbers of capsule, etc.), Hoffman et al. reported only minor, invariably transient, effects primarily of a gastrointestinal variety.73 None produced a discontinuation. This benign safety profile is consistent with the profiles observed both in our review of the impact of omega-3 fatty acids on asthma71 and our recently completed synthesis examining the impact of omega-3 fatty acids on mental health. Yet, the very low doses of both the omega-3 fatty acids (400 mg/day DHA) and the oil in which it was contained could account for Hoffman et al.'s results.73 Moreover, when the adverse event data in Hoffman et al.'s placebo group are juxtaposed with those safety data obtained from participants receiving the omega-3 fatty acids, one possible interpretation is that the minor adverse effects were more closely linked to the daily consumption of oil than to the omega-3 fatty acid contents in the oil.73 The placebo exposure contained some omega-6 fatty acid content. The ability of purified forms of omega-3 fatty acid exposures (e.g., ethyl esterifications) to contribute to maintaining blinding in experimental studies, due in large part to the minimization of the exposure's taste and odor, could not be evaluated because there were virtually no studies with which to assess this possibility.

There are sufficient between and within study conflicts (e.g., results of univariate vs multivariate analyses) in the results to preclude drawing any inference that is conclusive with respect to the value of the intake of omega-3 fatty acids to prevent ARMD (i.e., primary prevention). If it can be assumed that the study designs likely best suited to address this question should be both controlled and prospective, none of the included studies would qualify.

The only prospective study, which was conducted by Cho et al., included a large sample and appropriately conducted multivariate analysis, and controlled for key confounders.79 These investigators observed that the consumption of canned tuna fish or more than four fish servings per week each played a protective role against ARMD.79 Subgroup analyses purportedly revealed similar results (no data reported) when ARMD cases were classified by subtype (i.e., wet ARMD, and early and dry ARMD with geographic atrophy), and for advanced ARMD. However, Cho et al.'s results also indicated that several types of oily fish well known to have high concentrations of DHA and EPA (i.e., sardines, mackerel) failed to show a similar, protective effect. These discordant observations will require an explanation before anything conclusive can be asserted based on this study alone.

Moreover, Cho et al.'s study design did not a priori employ a separate, unexposed cohort as a control to establish a comparison with individuals consuming omega-3 fatty acids. Finally, while their dietary intake assessment was quite comprehensive, an assessment of the possible misclassification of non cases (i.e., no ARMD) was not performed. This is likely a result of the exposure-outcome question failing to be the primary focus of this study. That said, this observational study's sample exhibited strong applicability to the North American population.

The remaining relevant studies cannot resolve the divergent primary prevention results described by Cho et al., however. The two controlled studies employed a case-control design, and focused on whether the intake of omega-3 fatty acids exclusively protected against advanced forms of ARMD.81, 82 Seddon et al.'s study conducted appropriate multivariate analyses, which controlled for key confounders. Then, either with or without stratification for LA intake, Seddon et al. failed to find statistically significant associations between advanced ARMD and any intensity of fish intake, or combined EPA and DHA intake.82 At best, tests for trend suggested a possible protective value of fish consumption for advanced ARMD in this sample population primarily consisting of white Americans. When LA intake was then controlled for in Seddon et al.'s study, the results of a test for trend suggested effective protection against advanced ARMD only when LA consumption was low.82 The possible meaning of this finding is discussed below. Results of Ouchi et al.'s very small case-control study (n=21) were produced by univariate analysis, involved a population whose results offer little potential for generalizability to the North American population, and revealed no association between fish intake and advanced ARMD.81

Heuberger et al.'s single population cross-sectional study, with multiple covariates entered appropriately into multivariate analysis, revealed that fish intake frequency was not associated with early or late ARMD.84 Smith et al.'s multivariate analysis of data from a single population cross-sectional study also provided no evidence of a significant association between fish intake and either early or late ARMD.83 However, the type of study design employed by these two studies cannot be considered ideal to address the question of prevention. Causality cannot be inferred without some degree of temporal separation between exposure and outcome. This is why prospective designs are preferred. Mares-Perlman et al.'s retrospective population-based cohort study, while potentially affected by recall bias, likewise showed no protective role of seafood consumption in early or late forms of ARMD.80

These six included studies varied in their definitions of the exposure, clinical outcome and/or confounders, which together with their conflicting results, makes it impossible to draw a definitive conclusion regarding the potential of the intake of omega-3 fatty acids to prevent the onset of either early or late ARMD. And, neither study quality nor applicability can serve to clarify the inconclusive picture revealed by these six studies. The meaning of Ouchi et al.'s biomarker results is unclear.

The nature of the RCT design and the “cocktail-like” exposure employed by Scorolli et al. made it impossible to examine the specific impact of omega-3 fatty acids on slowing the progression of ARMD.76 A small sample size, the uncommonness and dubious clinical relevance of the visual recovery outcome, low study quality, and little or no applicability to the North American population suggest that there are, at present, no data with which to meaningfully address this research question.

Seddon et al.'s single prospective cohort study found that fish intake did not affect the progression to advanced ARMD overall, or in a high LA consumption group, but did protect against the progression to advanced ARMD in the low (below median consumption) LA consumption group.60 This parallels what was observed exclusively via a significant test for trend in the Seddon et al. study described earlier with reference to its investigation of the influence of the intake of omega-3 fatty acids on preventing the onset of advanced ARMD.82 The implications of this observation are discussed below. Both Seddon et al. studies exhibited the highest absolute study quality scores observed in the entire systematic review.60, 82 Nevertheless, the results from neither study can be used as yet to provide a conclusive answer to their respective research questions. The results from the Seddon et al. study addressing the progression to advanced ARMD require replication.60

The four studies examining whether the intake of omega-3 fatty acids slows the progression of RP do not provide a conclusive answer to this question.72, 73, 77 Hoffman et al.'s good quality RCT constituted the most rigorous test and revealed conflicting results.73 That said, rod and cone functional loss showed effect modification by age, with rod loss significantly reduced in the prepuberty group supplemented with DHA compared with placebo, and cone loss significantly reduced in the post-puberty group supplemented with DHA compared with placebo. The observation that certain analyses failed to reveal statistically significant between-group differences could be explained by this having been an underpowered trial.73

By virtue of its research design, which did not permit the isolation of the specific impact of omega-3 fatty acids on slowing the progression of RP, results from Dagnelie et al.'s internet-based comparative before-after study cannot be used to meaningfully address this question.77 In Hoffman et al.'s two very small noncomparative before-after studies of short duration, ERG results did not reveal statistically significant changes following supplementation.72 Thus, until Hoffman et al.'s RCT73 is replicated with a much larger sample size, little that is conclusive can be said about the potential value of the intake of omega-3 fatty acids in slowing the progression of RP.

Sorokin et al.'s noncomparative before-after study received a low study quality score and failed to resolve the questions of whether the intake of omega-3 fatty acids can slow the progression of either proliferative retinopathy or clinically significant macular edema in patients with diabetic retinopathy.75 This study did not constitute the best test of either of these possibilities, however. The most relevant clinical outcome by North American standards entailed fundus assessments, yet few details were reported. Covariates were not measured, and the univariate analysis of the data was flawed. Thus, the results of this study are inconclusive with respect to these two possible benefits of the intake of omega-3 fatty acids in diabetic retinopathy.

Although both the Arnarsson et al.85 and Cumming et al.86 studies are well known population-based risk factor studies, in neither of them was the association between the intake of foods or oils containing omega-3 fatty acids and age-related cataract prevalence the primary question. That said, no statistically significant associations were observed. Cross-sectional designs constitute very limited tests of this question.

Suzuki et al.'s noncomparative before-after study did not assess cataract status as its clinical outcome, preferring instead to examine visual acuity.78 Thus, with improvements in visual acuity unlikely to have been produced by reduced cataract formation, this study does not directly address the question of whether the intake of omega-3 fatty acids can slow the rate of progression of age-related cataracts.

Overall, we identified very few relevant studies per research question, which is the first important observation stemming from this review. Second, where multiple studies were found to address a given question, small sample sizes and either specific flaws or notable variability in the research designs, as well as in the definitions of the study populations, exposures/interventions, clinical outcomes, or covariates/confounders employed to investigate it, made it impossible or inappropriate to consider conducting a quantitative synthesis (i.e., meta-analysis) or a systematic appraisal of the role played by key covariates/confounders (e.g., reliable relationships between clinical and biomarker effects in interventional studies). Moreover, if it can be assumed that the best tests of these questions concerning prevention would involve study designs that are both prospective and controlled, while including large samples, other than the two RCTs,73, 76 none of the studies would qualify. Finally, many of the included studies received low total quality scores and exhibited limited or no applicability to the North American population. The implications of these observations are considered next.

Clinical Implications

The intake of omega-3 fatty acids in the present review's collection of interventional studies was not associated with moderate or severe adverse events. Supplementation was well-tolerated, with some mild and transient, mostly gastrointestinal events occurring occasionally. This finding may misrepresent what actually occurred in these studies if safety data were under-reported.

Nothing conclusive can be asserted with respect to the possible eye health benefits accruing to the intake of omega-3 fatty acids. At the same time, nothing at all can be said with respect to the potential eye health-related value of various sources (e.g., marine, plant), types (e.g., DHA, EPA, ALA) or doses/serving sizes of (foods containing) omega-3 fatty acids. Likewise, nothing can be said about for whom (e.g., subpopulations) these exposures/interventions might (not) be beneficial. Much more research is required.

Research Implications and Directions

ARMD is a significant public health problem. Treatment of dry ARMD with antioxidant vitamins provides only a small benefit, while treatment of advanced ARMD with PDT may provide greater statistical, than clinical, evidence of benefit. These observations, seen in light of an ageing North American population, suggest that ARMD may soon become an even greater public health challenge. That said, the question of its progression to advanced forms is likely the question in the field of ARMD. The development of early ARMD is of lesser clinical importance and may be an inevitable part of ageing. Advanced forms, on the other hand, cause considerable morbidity. This includes greater loss of vision and a decreased quality of life. Progression to advanced forms can carry a similar burden.

The possible worsening of this public health problem associated with an ageing population likely justifies additional research, and with a special focus on the secondary prevention of advanced forms of ARMD. The ideal study would likely enroll patients at high risk for development of advanced ARMD. Incident cases of advanced forms of ARMD would serve as the primary outcome (e.g., geographic atrophy or choroidal neovascular net formation). An adequately powered RCT could be conducted to test the placebo-controlled efficacy of omega-3 fatty acid supplementation. Indeed, there exists a biologically plausible relationship between omega-3 fatty acid intake and ARMD progression, given there are high levels of omega-3 fatty acids, especially DHA, in the retina.60

To exercise control over the type of omega-3 fatty acids and its dosing, and thereby maximize the interpretability of results, specific capsulized doses of purified DHA should likely be employed as the active intervention. Servings of food, or even oils poured from bottles, can lead to situations whereby within- and between-subject variability in daily on-study intake of omega-3 fatty acids can confound study results.71 Thus, proxies for omega-3 fatty acids (e.g., fish intake) should likely be avoided. Given the relative safety of omega-3 fatty acid supplementation, a high dose exposure might be wise, to avoid a false negative trial result.

The purification of DHA would minimize active contents' odor and taste, which in turn should help maintain experimental blinding. Given that it does not contain PUFA content, liquid paraffin could be used as the placebo content. At this point in time, however, it is unknown which dose(s) might prove efficacious, and so pilot work beforehand in uncontrolled studies is indicated. Moreover, additional preliminary work could recommend that a blend of DHA and EPA be used instead of DHA alone. The final trial design could accommodate multiple levels of dosing or combinations of omega-3 fatty acid type. Also, pilot research could help define the minimum duration of supplementation required to notice a (lack of) benefit. Ideally, it would be helpful if followups lasted on the order of years, given that progression to key milestones (e.g., advanced ARMD) likely does not occur rapidly. Finally, it is possible that endocannabinoids and lipoxins, which have been linked to LC PUFAs, will prove more biologically important than the omega-3 fatty acids themselves. Pilot research is indicated.

The two ARMD studies conducted by Seddon and colleagues have suggested that any trial should likely control, either experimentally or analytically, for the intake of omega-6 fatty acids.60, 82 They noted eye health benefits associated with omega-3 fatty acid intake only in those participants consuming lower levels of LA. This observation provides tentative support for the hypothesis that omega-3 fatty acid supplementation is more likely to have a clinical impact if omega-6 fatty acid intake is minimized. Omega-3 and omega-6 fatty acids stand in competitive relationship for enzymes (e.g., desaturases) and for positioning in cell membranes, and if there is large intake of omega-6 fatty acids, omega-3 supplementation may fail to (enter cell membranes from which to) make a clinical difference.71 Seddon and colleagues' observations suggest to us that, at minimum, the measurement of LA intake could help identify those individuals and eye diseases/visual impairments for whom omega-3 fatty acid supplementation might prove clinically beneficial.60, 82

In effect, consideration should be given to accounting for the on-study omega-6/omega-3 fatty acid intake ratio from both supplementation and the background diet in making sense of study outcomes. The intervention could be even more complex: e.g., actively decreasing the intake of omega-6 fatty acids, while omega-3 fatty acid intake is increased. This may be important given that certain investigators have suggested that various types of disease are linked to higher omega-6/omega-3 fatty acid intake ratios;46–58 and, above a certain value, this intake ratio could preclude consumed omega-3 fatty acids from becoming incorporated within the human biosystem (i.e., cell membranes), and in turn contributing to (ameliorating) eye health. Trial investigators might consider accounting for prestudy/baseline omega-6/omega-3 fatty acid intake levels as well. The reader is reminded that populations in different geographic locations and countries vary in terms of their omega-6/omega-3 fatty acid intake ratio,46–58 and often in no small part because of their intake of marine sources of omega-3 fatty acids.71 It is with this understanding that the present review's indices of applicability were defined, with the North American population as the reference point.

Confounders also requiring consideration include, but are not restricted to, exposure to sunlight, smoker status, and alcohol consumption. The latter two variables can influence both health status and essential fatty acid status.59, 60 No evidence reviewed here indicated the presence of reliable empirical relationships between (e.g., blood lipid) biomarker content and ARMD, or any other eye health outcomes.

RP is a relatively rare condition, but its public health importance is notable because it affects young and very young individuals. The impact can be observed as the loss of many potential years of vision. Another measure of RP's public health significance is the relatively large number of funding organizations and support groups whose central focus is this condition.

However, clinical research with regards to RP is complicated by a few observations. While RP is a rare disorder, its forms are quite heterogeneous. That said, an intervention found to be efficacious for XLRP may produce minimal or no benefit in autosomal dominant RP. Furthermore, even a single genetic classification such as XLRP refers to an heterogeneous set of genetic entities.

It is our view that, at this time, more observational research should be taken to begin to identify the specific links between exposures to omega-3 fatty acids and the onset as well as progression of the different RP subtypes. If prospective cohort studies were undertaken that used dietary intake data obtained from existing RP registries, reasonable, RCT-testable hypotheses could be generated. These prospective studies could also permit the identification of promising exposures and intervention lengths for each of the RP subtypes. It is possible that supplementation would be needed before a critical time period of programmed cell death. The reviewed study of XLRP by Hoffman et al. is a laudable effort, whose replication should include a larger sample size.73

As one of the most important morbidities associated with diabetes mellitus, diabetic retinopathy is the most important ophthalmic public health problem among adults of working age. There is face value to expecting that omega-3 fatty acids might be helpful in diabetic retinopathy, given that lipid exudation plays such an important role in background and preproliferative retinopathy. Given the extremely flawed nature of the single study identified by our review, it is our view that research with this focus has essentially not begun.75 Therefore, prospective observational designs (e.g., cohort study) likely constitute the best option with which to begin studying the value of omega-3 fatty acids in diabetic retinopathy. The key outcome would be clinically significant macular edema, as it has been measured in all previous ETDRS studies. Important covariates to assess include control of diabetes via HgA1c, blood pressure, and smoker status.

Cataracts are the most important cause of blindness worldwide, while also constituting one of the most costly burdens on Western health care systems. Unfortunately, little is understood about its primary prevention. A nutritional contribution to cataract formation is nevertheless conceivable given its higher prevalence in underprivileged populations.

The sparse evidence identified by our review85, 86 suggests that some prospective observational studies be conducted first. The two cross-sectional studies could shed little or no light on possible causal links between exposures and outcomes. A prospective cohort study, which entails clear documentation of the intake of omega-3 fatty acids and regular lens assessments, might be the ideal starting point. Stratification by type of cataract would be essential to account for this source of clinical heterogeneity. Also important would be to control for possible etiologic factors such as smoking, sunlight exposure, and medication use.

As it was observed with respect to ARMD, cataract progression, and not its formation, has the largest public health impact. A mild nuclear sclerotic cataract, which reduces visual acuity by one or two lines, is usually well-tolerated by patients. However, progression to advanced forms of cataract significantly affects one's ability to drive and quality of life. A prospective cohort study, which identifies the impact of the intake of omega-3 fatty acids on the progression of cataracts from mild to severe, would be an ideal study focus.

Given that studies were not found to have examined the relationship between the intake of omega-3 fatty acids and either retinal vein or retinal artery occlusions, preliminary observational work is likely indicated. Finally, since several questions of pertinence to this systematic review could be addressed within a longterm (e.g., 10-year) prospective cohort design, one such study could simultaneously investigate many of the outcomes that our review sought to examine (e.g., cataract, diabetic retinopathy, ARMD, and perhaps even progression of some forms of RP).

If future research is going to produce data that are unequivocally applicable to North Americans, it will likely need to enroll either North American populations or populations exhibiting a high omega-6/omega-3 fatty acid intake ratio similar to what has been observed in the diet of North Americans. It is our view that the dietary omega-6/omega-3 fatty acid intake ratio may eventually be seen to play an important role in the prevention and treatment of various forms of eye disease/visual impairment.

Limitations of the Review

While there are some limitations characterizing the present systematic review, almost none could likely be considered a serious impediment to the interpretation of the evidence we identified and synthesized. Overall, we found too few studies investigating a given question, and employing an appropriate research design of sufficient size and sound methodology, to have these limitations alter the view that, at present, we cannot conclude anything definitive about the eye disease/visual impairment-specific or overarching roles of omega-3 fatty acids in eye health. As well, the possible roles played by likely covariates and confounders could not be evaluated as planned. We were limited in what we could observe because of the paucity of relevant studies per question, and because many studies did not specifically investigate the influence of these variables. Or, if they did, the definition of these factors varied to some extent.

As stated in Chapter 2, in light of the relatively limited details often provided in reports about the ways in which lipid samples were extracted, stored and analyzed, we could only readily identify situations where investigators described inappropriate methods. It is unclear how this state of affairs might have influenced the very few observations we gleaned from the evidence base concerning the role of omega-3 fatty acids in eye health.

Time constraints made it impossible to complete dual-assessor appraisals of the quality (i.e., internal validity) or applicability of studies. One experienced quality assessor conducted these evaluations. At the same time, we conducted these quality assessments using items we either modified from existing instruments, or which we had to develop outright because no similar tools existed (e.g., noncomparative before-after studies). A design-specific, total quality score was then generated for each study, from which a single summary value was derived (i.e., A, B, C). This simplification permitted the entry of these values into summary matrices. However, the design-specific cutpoints used to assign these values were established without any validational basis, and so their value is likely extremely limited. The applicability indices, while continuing the work we did when we systematically reviewed the evidence for the health effects of omega-3 fatty acids on asthma,71 also did not receive any validational support. Nevertheless, given the limited number of studies addressing a specific question, and using a design whose data could meaningfully elucidate it, it is unlikely that these shortcomings could have had a meaningful impact on the limited “take home messages” highlighted by our review. Formal statistical assessments of the impacts of study quality or applicability on study outcomes could not be conducted. Finally, with meta-analysis deemed impossible or inappropriate, an examination of the possible presence and impact of publication bias could not be conducted.


Based on the studies identified by this review, it is apparent that clinical research has only scratched the surface with respect to understanding the possible utility of the intake of omega-3 fatty acids as a primary or secondary prevention in eye health. Moreover, seen from the point of view of clinical research's typical, linear arc—which moves from basic science to observational research to RCTs, and culminating in the systematic review/meta-analysis of the observations obtained by these primary studies—there is a paucity of solid observational research with which to construct an experimental framework affording the meaningful conduct of RCTs. For example, there is little understanding of the exact sources, types and doses of omega-3 fatty acids, or even the possible duration of their use, which might usefully serve as definitions of a prevention-centered “intervention” for any of the eye diseases/visual impairments examined in our review. Moreover, a single study reporting adverse event data likely does not permit laying to rest all possible concerns regarding the short- or long-term safety of such an intervention.

It is therefore our view that much more research will need to be conducted before anything conclusive can be asserted with respect to the effects of omega-3 fatty acids on eye health. It is also our understanding that sorting out the possible benefits of the intake of omega-3 fatty acids in eye health might profit from taking into consideration the impact of the concurrent intake of omega-6 fatty acids and, by definition, the omega-6/omega-3 fatty acid intake ratio. Finally, any notable causal or correlational relationships observed between the omega-6/omega-3 fatty acid intake ratio and the development or progression of eye disease/visual impairment may then be “explained” by future studies, which focus on observing patterns of omega-6/omega-3 fatty acid content in peripheral, or even brain, biomarkers.

PubReader format: click here to try


  • PubReader
  • Print View
  • Cite this Page

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...