Mussels: the strong trials are about the extract

Almost everything solid that is known about mussels was tested as a capsule. A meta-analysis of 69 randomised trials covering 20 oral supplements for hand, hip or knee osteoarthritis found that only green-lipped mussel extract and undenatured type II collagen had clinically important effects on pain at medium term. The review grades its own evidence from very low to high, and that medium-term finding rests on few trials. A second review pooled five trials of the whole powder or the lipid extract and reported a moderate reduction in pain on a visual analogue scale, effect size -0.46 (95% CI -0.82 to -0.10, p = 0.01), an interval that nearly touches zero. None of that was mussels on a plate. For food the measurement is smaller and plainer. In a crossover feeding trial, 23 women with rheumatoid arthritis ate blue mussels five meals a week for 11 weeks. Those meals provided 0.18 g of EPA and 0.16 g of DHA a day averaged across the week, against 0.00 g of EPA in the meat period. Omega-3 content moves with water, season and feed, so read that as one worked example rather than a fixed figure.

The numbers, per 100 g

What a real portion looks like

• Blue mussels eaten at five meals a week in a controlled feeding trial delivered 0.18 g of EPA and 0.16 g of DHA a day, averaged across the week. — women with rheumatoid arthritis during the blue mussel period of a crossover feeding trial, five meals a week for 11 weeks (n = 23 completed) Evidence B source
  0.18 g EPA and 0.16 g DHA daily, against 0.00 g EPA and 0.01 g DHA in the meat period

What your body absorbs

• Eating blue mussels raised EPA and DHA in red blood cells, yet the oxylipins made from those fatty acids did not change with them. — women with rheumatoid arthritis in the mussel arm of a crossover trial (n = 23 completed) Evidence B source
  Significant rise in erythrocyte EPA, DHA and free EPA; no concurrent change in their oxylipins

What cooking and storage change

• Bivalve molluscs concentrate norovirus by filter feeding, and high viral loads survive both depuration and ordinary food preparation steps. — bivalve molluscs implicated in gastroenteritis outbreaks over two decades of published reports (n = two decades of outbreak and prevalence reports) Evidence C source
  Viral load persists after depuration and typical food preparation; no effect size is given

What it does to your health

• In a meta-analysis of 69 randomised trials covering 20 oral supplements for hand, hip or knee osteoarthritis, green-lipped mussel extract was one of only two supplements with clinically important effects on pain at medium term. — patients with hand, hip or knee osteoarthritis in randomised trials comparing oral supplements with placebo (n = 69 eligible studies, 20 supplements) Evidence A source
  Clinically important effect on pain at medium term for green-lipped mussel extract; no supplement kept a clinically important effect at long term
• Pooling five clinical trials of green-lipped mussel extract in osteoarthritis gave a moderate reduction in visual analogue scale pain, effect size -0.46 (95% CI -0.82 to -0.10). — osteoarthritis patients in nine clinical trials of whole or lipid green-lipped mussel extract, five of them pooled (n = 9 trials included, 5 pooled) Evidence B source
  Effect size -0.46 (95% CI -0.82 to -0.10), p = 0.01, on VAS pain
• In adults aged 40 to 75 with mild knee osteoarthritis, 1000 mg of green-lipped mussel powder a day for 12 weeks improved the WOMAC total score by 3.66 points more than placebo (95% CI 0.45 to 6.87). — adults aged 40 to 75 years with symptomatic radiographic knee osteoarthritis, Kellgren-Lawrence grade 1 or 2 (n = 100 randomised, 93 in the per-protocol set) Evidence B source
  Between-group difference in WOMAC total -3.66 (95% CI -6.87 to -0.45), p = 0.032, at week 12
• In 133 Chinese adults with type 2 diabetes, 1.6 g a day of blue mussel lipid extract for 60 days lowered fasting insulin (P = 0.005) and insulin resistance (P = 0.026), while fish oil and corn oil did not. — 133 Chinese adults with type 2 diabetes mellitus randomised to fish oil (n = 44), blue mussel lipid extract (n = 44) or corn oil (n = 45) for 60 days (n = 133) Evidence B source
  Fasting insulin P = 0.005 and HOMA-IR P = 0.026 fell in the mussel extract group only; no group difference in glycosylated hemoglobin
• In 20 untrained men, four weeks of Greenshell mussel powder before damaging eccentric exercise sped up recovery of isometric and concentric peak torque at 48 and 72 hours. — 20 untrained adult men aged 21 to 45 years in a double-blind placebo-controlled crossover study of bench-stepping exercise (n = 20) Evidence B source
  Isometric and concentric peak torque improved at 48 h and 72 h (p < 0.05); lower plasma creatine kinase at 72 h; treatment by time interaction for pain p = 0.007 and p = 0.018

Safety, limits and interactions

• 26 weeks of eating 1 kg of fish and mussels a week raised blood mercury by 5.1 ng/ml (95% CI 4.4 to 5.8) and arsenic by 7.1 ng/ml (95% CI 5.0 to 9.2) against controls. — 102 healthy non-smoking men and women aged 48 to 76 years, randomised to a high fish and mussel intake or to no intervention (n = 102 (51 intervention, 51 control)) Evidence B source
  Mercury +5.1 ng/ml (95% CI 4.4, 5.8); arsenic +7.1 ng/ml (95% CI 5.0, 9.2); lead +2.6 ng/ml (95% CI 0.0, 5.2)
• In the 1987 Prince Edward Island outbreak, 107 people met the case definition after eating cultivated mussels carrying domoic acid, 19 were hospitalised and 3 died. — people in Canada who ate cultivated mussels and developed gastrointestinal symptoms within 24 hours or neurologic symptoms within 48 hours (n = 107 cases out of more than 250 reports) Evidence C source
  Vomiting in 76 percent, abdominal cramps in 50 percent, loss of short-term memory in 25 percent; 19 hospitalised, 12 in intensive care, 3 deaths
• Two adult women had repeated bouts of severe vomiting and diarrhoea a few hours after eating mussels, with negative allergy testing, and were diagnosed with food protein-induced enterocolitis syndrome. — two adult women with recurrent reactions to mussels and negative allergy testing (n = 2) Evidence C source
  Severe vomiting and diarrhoea within a few hours of intake, recurring with each exposure
• A three-season survey of eleven commercial Tasmanian shellfish growing areas found Vibrio parahaemolyticus in every area in summer and autumn, with a prevalence of 8 to 100 percent. — eleven commercial oyster and mussel growing areas in Tasmania sampled at harvest between 2020 and 2022 (n = 11 growing areas) Evidence C source
  Prevalence 8-100 percent in summer and autumn, 17-33 percent in the five areas where it was found in winter and spring; levels usually below 10 MPN/g

What is still unknown

• The review that pooled 9 clinical trials of green-lipped mussel extract says the number and quality of those trials limit the answer, and that the best form of the extract is still unknown. — osteoarthritis patients across nine clinical trials of green-lipped mussel extract (n = 9 trials) Evidence B source
  No effect estimate; a statement about the state of the evidence

The bottom line

The newest trial shows how thin the ground still is. One trial randomised 100 adults aged 40 to 75 with symptomatic knee osteoarthritis of Kellgren-Lawrence grade 1 or 2, and analysed 93 of them per protocol. Taking 1000 mg of green-lipped mussel powder a day for 12 weeks improved the WOMAC total score by 3.66 points more than placebo (95% CI 0.45 to 6.87, p = 0.032). That result was significant only in the per-protocol set. In the intention-to-treat analysis it was a non-significant trend, and CRP and ESR did not move. Two other findings sit at the same scale and use the same kind of product. Among 133 Chinese adults with type 2 diabetes, 1.6 g a day of blue mussel lipid extract for 60 days lowered fasting insulin (P = 0.005) and the insulin resistance index (P = 0.026) within that group. The fish oil and corn oil groups showed no such change. Glycosylated haemoglobin did not differ between the arms, which keeps this a two-month signal on insulin. Twenty untrained men aged 21 to 45 took Greenshell mussel powder for four weeks before a bench-stepping protocol built to damage muscle. Isometric and concentric peak torque recovered better at 48 and 72 hours (p < 0.05), under a laboratory model of damage and again with powder rather than shellfish. The review that pooled the nine osteoarthritis trials says their number and quality limit the answer and that the best format of the extract is still unknown. No trial has tested eating mussels as food against osteoarthritis pain. Absorption has its own loose end. In the crossover trial above, erythrocyte EPA, DHA and free EPA rose significantly in the mussel group, and the oxylipins made from those fatty acids did not change with them, in 23 women with active rheumatoid arthritis who stayed on their medication. The cautions here are real and they are not the ones people expect. Two adult women had repeated bouts of severe vomiting and diarrhoea within a few hours of eating mussels, with allergy testing negative each time, because the reaction was not IgE-mediated and standard testing comes back normal. Two cases cannot say how common that is, and they describe one mechanism rather than shellfish allergy in general. Mussels feed by filtering water, and a review of two decades of outbreak and prevalence reports describes high norovirus loads persisting in bivalves after depuration and after typical food preparation steps, without naming a temperature or a time that would be enough. A survey of eleven commercial Tasmanian oyster and mussel growing areas, sampled at harvest between 2020 and 2022, detected Vibrio parahaemolyticus in every area across the three summer and autumn periods, at a prevalence of 8 to 100 percent. That describes the raw material at harvest rather than the plate. The rare risk is the serious one. In the 1987 Prince Edward Island outbreak, after people ate cultivated mussels carrying domoic acid, 19 patients were hospitalised and 12 needed intensive care for seizures, coma, heavy respiratory secretions or unstable blood pressure. That was one growing area in one year, found through voluntary reporting and before the harvest monitoring that came after it. Metals build up more quietly. A trial randomised 102 healthy non-smoking adults aged 48 to 76, half of them to eat 1 kg of fish and mussels a week for 26 weeks. Against the control group, blood mercury rose by 5.1 ng/ml (95% CI 4.4 to 5.8), arsenic by 7.1 ng/ml (95% CI 5.0 to 9.2) and lead by 2.6 ng/ml (95% CI 0.0 to 5.2). Fish and mussels were eaten together in that trial, so the metal load cannot be split between them, and the authors judged only lead to be above tolerable levels, from a baseline that was already high. We found no trial of mussels in pregnancy, in children, or against any medication.

Sources

1. musl-r5-09 · randomised crossover trial, methods section
   The blue mussels consumed every week provided 0.24 g saturated fat, 0.14 g monounsaturated fat, 0.47 g polyunsaturated fat, 0.18 g EPA, and 0.16 g DHA daily.
   Molecules, 2023 · checked 2026-09-24
2. musl-r5-10 · randomised crossover trial
   Despite significant changes in eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) and free EPA in erythrocytes in the mussel group, no concurrent changes in their oxylipins were seen.
   Molecules, 2023 · checked 2026-09-24
3. musl-r5-08 · review of observational studies
   The authors describe how their physiology of filter feeding can render them concentrated vehicles of norovirus contamination in polluted environments and how high viral loads persist in molluscs even after application of depuration practices and typical food preparation steps.
   Transbound Emerg Dis, 2020 · checked 2026-09-24
4. musl-r5-01 · systematic review and meta-analysis of RCTs
   Only green-lipped mussel extract and undenatured type II collagen had clinically important effects on pain at medium term.
   Br J Sports Med, 2018 · checked 2026-09-24
5. musl-r5-02 · systematic review of clinical trials
   Pooled results showed that GSM extracts (lipid extract or whole powder) provide moderate and clinically significant treatment effects on a visual analogue scale (VAS) pain score (effect size: - 0.46; 95% CI - 0.82 to - 0.10; p = 0.01).
   Inflammopharmacology, 2021 · checked 2026-09-24
6. musl-r5-04 · randomised controlled trial
   In the PP analysis, GLMP resulted in greater improvement in WOMAC total score at week 12 than placebo (between-group difference - 3.66; 95% CI - 6.87 to - 0.45; p = 0.032).
   BMC Complement Med Ther, 2026 · checked 2026-09-24
7. musl-r5-11 · randomised controlled trial
   The fasting serum concentration of insulin (P = 0.005) and the homeostasis model of insulin resistance (P = 0.026) were significantly decreased in the BMLE group, whereas no significant change was found in the FO or CO groups.
   J Sci Food Agric, 2023 · checked 2026-09-24
8. musl-r5-13 · randomised controlled crossover trial
   GSM powder promoted muscle function recovery, significantly improving (p < 0.05) isometric and concentric peak torque at 48 h and 72 h post exercise, respectively.
   Nutrients, 2023 · checked 2026-09-24
9. musl-r5-05 · randomised controlled trial
   The difference in mean change for intervention compared with control persons was 5.1 ng ml(-1) (95% confidence interval (CI) = 4.4, 5.8) for mercury, 7.1 ng ml(-1) (95% CI = 5.0, 9.2) for arsenic, and 2.6 ng ml(-1) (95% CI = 0.0, 5.2) for lead.
   Food Addit Contam Part A Chem Anal Control Expo Risk Assess, 2015 · checked 2026-09-24
10. musl-r5-06 · case series (outbreak investigation)
    Nineteen patients were hospitalized, of whom 12 required intensive care because of seizures, coma, profuse respiratory secretions, or unstable blood pressure.
    N Engl J Med, 1990 · checked 2026-09-24
11. musl-r5-07 · case series
    In this case report, we describe two adult women with recurrent episodes of severe vomiting and diarrhoea within a few hours of intake of mussels.
    Ugeskr Laeger, 2025 · checked 2026-09-24
12. musl-r5-12 · prevalence survey
    Vibrio parahaemolyticus was detected in all surveyed growing areas during the three summer/autumn sampling periods, with a prevalence of 8-100 %.
    Int J Food Microbiol, 2025 · checked 2026-09-24
13. musl-r5-03 · systematic review of clinical trials
    However, the current evidence is limited by the number and quality of studies, and further larger and high-quality studies are needed to confirm the effectiveness and to identify the optimal GSM format.
    Inflammopharmacology, 2021 · checked 2026-09-24