Marine-derived fatty acid therapy for stroke

Review question

What are the effects of marine-derived omega-3 fatty acids for stroke after a short (up to three months) and a longer (more than three months) follow-up?


The term stroke refers to a group of diseases of blood vessels in the brain. Stroke can be caused by either bleeding or blockage in these vessels that leads to a loss of function of brain cells. Transient ischaemic attack (TIA), also called a 'mini-stroke', is a temporary disruption of blood supply to the brain. Stroke is a disabling disease that usually requires prolonged specialised care and currently there are few treatment options for stroke patients. Omega-3 fatty acids (eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA)), present in oily fish, have important functions in the brain. In animal research, they appear to protect the brain cells after stroke, particularly if they are administered very early. However, the effects of EPA and DHA as a treatment for stroke in humans are unclear.

Study characteristics

We identified 30 studies that included participants with stroke or TIA and we found relevant information in nine of them (3339 participants in total). Three had a short follow-up (up to three months) and six had a longer follow-up. Three studies compared marine-derived omega-3 fatty acids to normal care and the remainder used a placebo (dummy). Not all of the studies assessed all outcomes.

Key results

The effects of marine-derived omega-3 fatty acids on stroke recovery are unclear. Only two very small studies reported it, one in each follow-up category, without finding meaningful differences. One study in the short follow-up group found less improvement in mood with marine-derived omega-3 fatty acids but the evidence was of low certainty. The effect of marine-derived omega-3 fatty acids on deaths due to blood vessel disease, recurrence of stroke, adverse events, and quality of life after having a stroke or TIA is unclear in both follow-up categories, due to the small number of studies that have assessed them.

Certainty of the evidence

In both the short and longer follow-up studies the certainty of the evidence was very low or low. The frequency of other type of stroke and quality of life were not reported in the longer follow-up.

Date of the evidence

This review updates one carried out originally in 2019. It is now up to date to 31 May 2021.

Authors' conclusions: 

We are very uncertain of the effect of marine-derived n-3 PUFAs therapy on functional outcomes and dependence after stroke as there is insufficient high-certainty evidence. More well-designed RCTs are needed, specifically in acute stroke, to determine the efficacy and safety of the intervention.

Studies assessing functional outcome might consider starting the intervention as early as possible after the event, as well as using standardised, clinically relevant measures for functional outcomes, such as the modified Rankin Scale. Optimal doses remain to be determined; delivery forms (type of lipid carriers) and mode of administration (ingestion or injection) also need further consideration.

Read the full abstract...

Currently, with stroke burden increasing, there is a need to explore therapeutic options that ameliorate the acute insult. There is substantial evidence of a neuroprotective effect of marine-derived n-3 polyunsaturated fatty acids (PUFAs) in animal models of stroke, leading to a better functional outcome.


To assess the effects of administration of marine-derived n-3 PUFAs on functional outcomes and dependence in people with stroke.

Search strategy: 

We searched the Cochrane Stroke Trials Register (last searched 31 May 2021), the Cochrane Central Register of Controlled Trials (CENTRAL; 2021, Issue 5), MEDLINE Ovid (from 1948 to 31 May 2021), Embase Ovid (from 1980 to 31 May 2021), CINAHL EBSCO (Cumulative Index to Nursing and Allied Health Literature; from 1982 to 31 May 2021), Science Citation Index Expanded ‒ Web of Science (SCI-EXPANDED), Conference Proceedings Citation Index-Science – Web of Science (CPCI-S), and BIOSIS Citation Index. We also searched ongoing trial registers, reference lists, relevant systematic reviews, and used the Science Citation Index Reference Search.

Selection criteria: 

We included randomised controlled trials (RCTs) comparing marine-derived n-3 PUFAs to placebo or open control (no placebo) in people with a history of stroke or transient ischaemic attack (TIA), or both.

Data collection and analysis: 

At least two review authors independently selected trials for inclusion, extracted data, assessed risk of bias, and used the GRADE approach to assess the certainty of the body of evidence. We contacted study authors for clarification and additional information on stroke/TIA participants. We conducted random-effects meta-analysis or narrative synthesis, as appropriate. The primary outcome was efficacy (functional outcome) assessed using a validated scale, for example, the Glasgow Outcome Scale Extended (GOSE) dichotomised into poor or good clinical outcome, the Barthel Index (higher score is better; scale from 0 to 100), or the Rivermead Mobility Index (higher score is better; scale from 0 to 15). Our secondary outcomes were vascular-related death, recurrent events, incidence of other type of stroke, adverse events, quality of life, and mood.

Main results: 

We included 30 RCTs; nine of them provided outcome data (3339 participants). Only one study included participants in the acute phase of stroke (haemorrhagic). Doses of marine-derived n-3 PUFAs ranged from 400 mg/day to 3300 mg/day. Risk of bias was generally low or unclear in most trials, with a higher risk of bias in smaller studies. We assessed results separately for short (up to three months) and longer (more than three months) follow-up studies.

Short follow-up (up to three months)

Functional outcome was reported in only one pilot study as poor clinical outcome assessed with the GOSE (risk ratio (RR) 0.78, 95% confidence interval (CI) 0.36 to 1.68, P = 0.52; 40 participants; very low-certainty evidence). Mood (assessed with the GHQ-30, lower score better) was reported by only one study and favoured control (mean difference (MD) 1.41, 95% CI 0.07 to 2.75, P = 0.04; 102 participants; low-certainty evidence).

We found no evidence of an effect of the intervention for the remainder of the secondary outcomes: vascular-related death (two studies, not pooled due to differences in population, RR 0.33, 95% CI 0.01 to 8.00, P = 0.50, and RR 0.33, 95% CI 0.01 to 7.72, P = 0.49; 142 participants; low-certainty evidence); recurrent events (RR 0.41, 95% CI 0.02 to 8.84, P = 0.57; 18 participants; very low-certainty evidence); incidence of other type of stroke (two studies, not pooled due to different type of index stroke, RR 6.11, 95% CI 0.33 to 111.71, P = 0.22, and RR 0.63, 95% CI 0.25 to 1.58, P = 0.32; 58 participants; very low-certainty evidence); and quality of life (physical component, MD −2.31, 95% CI −4.81 to 0.19, P = 0.07, and mental component, MD −2.16, 95% CI −5.91 to 1.59, P = 0.26; 1 study; 102 participants; low-certainty evidence).

Adverse events were reported by two studies (57 participants; very low-certainty evidence), one trial reporting extracranial haemorrhage (RR 0.25, 95% CI 0.04 to 1.73, P = 0.16) and the other one reporting bleeding complications (RR 0.32, 95% CI 0.01 to 7.35, P = 0.47).

Longer follow-up (more than three months)

One small trial assessed functional outcome with both the Barthel Index for activities of daily living (MD 7.09, 95% CI −5.16 to 19.34, P = 0.26), and the Rivermead Mobility Index for mobility (MD 1.30, 95% CI −1.31 to 3.91, P = 0.33) (52 participants; very low-certainty evidence). We carried out meta-analysis for vascular-related death (RR 1.02, 95% CI 0.78 to 1.35, P = 0.86; 5 studies; 2237 participants; low-certainty evidence) and fatal recurrent events (RR 0.69, 95% CI 0.31 to 1.55, P = 0.37; 3 studies; 1819 participants; low-certainty evidence).

We found no evidence of an effect of the intervention for mood (MD 1.00, 95% CI −2.07 to 4.07, P = 0.61; 1 study; 14 participants; low-certainty evidence). Incidence of other type of stroke and quality of life were not reported.

Adverse events (all combined) were reported by only one study (RR 0.94, 95% CI 0.56 to 1.58, P = 0.82; 1455 participants; low-certainty evidence).