The effects of omega-3 PUFA on muscle and whole-body protein synthesis (Therdyothin 2025)
TL;DR
This systematic review and meta-analysis (PROSPERO CRD42022366986) pooled human trials testing omega-3 polyunsaturated fatty acid (PUFA) supplementation against muscle protein synthesis (MPS) and whole-body protein synthesis rates. The critical detail for interpreting the headline result is the estimand: the main MPS pool (k=6, SMD 0.03, 95% CI −0.35 to 0.40, I²=30%, P=0.89) predominantly measured basal (unstimulated) fractional synthetic rate (FSR) — the authors state that “basal unstimulated rates of MPS, as opposed to MPS rates measured after anabolic stimulation, were used to calculate the pooled effect in 50% of included studies,” and that in the one study reporting both, basal FSR was selected for the pooled effect. This is the same estimand on which Smith et al. 2011’s own randomized trial found nothing (that trial’s positive result was in clamp-stimulated, not basal, FSR) — so the null main-pool result does not refute the Smith trial or the broader sensitization hypothesis. Two separate stimulated-condition pools existed: acute-exercise-stimulated (k=3, SMD 0.41, 95% CI −0.16 to 0.98, I²=0%, P=0.16) and anabolic-stimulus-during-measurement (k=4, SMD 0.52, 95% CI −0.64 to 1.67, I²=77%, P=0.28). Both were directionally positive but non-significant and imprecise (small k, wide CIs). The authors conclude they “did not detect any increase in MPS in response to anabolic stimulation with n-3 PUFA ingestion,” while also stating “although n-3 PUFA ingestion might not augment basal rates of MPS, it is suggested that n-3 PUFA ingestion increases the response of MPS to anabolic stimuli” and calling for adequately powered trials that measure basal and stimulated FSR together — i.e., they treat the sensitization hypothesis as unconfirmed and not reliably replicated, not refuted. Separately, pooled across only 3 comparisons, omega-3 PUFA did increase whole-body protein synthesis (SMD 0.51, 95% CI 0.12–0.90, I²=0%, P=0.01).
Design
| Field | Value |
|---|---|
| Search databases | PubMed, Web of Science, Cochrane Library, Scopus |
| Search window | Database inception to December 2022 |
| Studies screened | 302 |
| Studies included | 8 |
| Statistical model | Random-effects inverse-variance model; standardized mean differences (SMD) with 95% CI |
| Risk-of-bias tool | Cochrane RoB2 |
| Registration | PROSPERO CRD42022366986 |
| Primary outcome | Muscle protein synthesis (fractional synthetic rate) |
| Secondary outcome | Whole-body protein synthesis rate |
| Planned subgroup analyses | Age, omega-3 PUFA dose, supplementation duration, method used to measure fractional synthetic rate |
Of the 8 included studies, only a subset contributed to each pooled outcome (k=6 for muscle protein synthesis, k=3 for whole-body protein synthesis) — the review’s own included-study count (8) is larger than either individual outcome pool, reflecting that not every trial measured both outcomes. Only the Smith et al. 2011 AJCN randomized trial 1 is among the 8 included studies; the companion, uncontrolled 2011 Clinical Science paper 2 is cited in the review’s reference list but was not included in the meta-analysis.
Key results
| Outcome | Pooled effect | k (comparisons) | I² | P | Interpretation |
|---|---|---|---|---|---|
| Muscle protein synthesis (MPS), main pool — predominantly basal/unstimulated FSR | SMD 0.03 (95% CI −0.35 to 0.40) | 6 | 30% | 0.89 | No detectable pooled effect on basal muscle-specific protein synthesis. Basal FSR was used to calculate the pooled effect in 50% of the 6 included studies (and was the preferred value when a study reported both basal and stimulated FSR) — this pool does not directly test the anabolic-stimulus-sensitization hypothesis. |
| MPS, acute-exercise-stimulated subset | SMD 0.41 (95% CI −0.16 to 0.98) | 3 | 0% | 0.16 | Directionally positive but not statistically significant; small k and a CI crossing zero. |
| MPS, anabolic-stimulus-during-measurement subset | SMD 0.52 (95% CI −0.64 to 1.67) | 4 | 77% | 0.28 | Directionally positive but not statistically significant; wide CI and substantial heterogeneity (I²=77%). |
| Whole-body protein synthesis | SMD 0.51 (95% CI 0.12–0.90) | 3 | 0% | 0.01 | Statistically significant increase, but rests on only 3 comparisons — a very small pool that should not be over-read as a robust or generalizable effect. |
| MPS subgroup: age | No effect detected | subset of 6 (k up to 5) | — | — | Exploratory subgroup on a small pool; not a robust independent replication of the null. |
| MPS subgroup: omega-3 PUFA dose | No effect detected | subset of 6 (k=4 for >3 g/d, k=2 for ≤3 g/d) | — | — | Each dose stratum has only 2–4 studies; wide, imprecise intervals. |
| MPS subgroup: supplementation duration | No effect detected | subset of 6 (k=3 for <8 wk, k=3 for ≥8 wk) | — | — | Each duration stratum has only 3 studies. |
| MPS subgroup: FSR measurement method | No effect detected | subset of 6 (k=4 deuterium oxide, k=2 phenylalanine infusion) | — | — | Each method stratum has only 2–4 studies. |
The muscle-specific vs whole-body dissociation is a structural finding worth noting, but the muscle-specific null itself must be read through the basal-vs-stimulated estimand distinction above: an intervention that does not raise basal muscle protein synthesis can still plausibly sensitize the muscle’s response to an anabolic stimulus (meal, exercise) — which is exactly what the underpowered stimulated-state pools (k=3, k=4) are directionally, if not significantly, consistent with.
Why this reframes the muscle-protein-synthesis narrative
A widely repeated claim in the sarcopenia/omega-3 literature is that fish-oil-derived omega-3 PUFA sensitizes muscle protein synthesis to an anabolic stimulus (a meal, exercise), particularly in older adults — a claim that traces substantially to Smith et al. 2011’s randomized older-adult trial 1, which found no effect on basal FSR but a large increase in clamp-stimulated FSR. An uncontrolled companion study in younger adults 2 is often cited alongside it, but it was not among the 8 studies Therdyothin et al. included in this meta-analysis, so it does not contribute to any of the pooled estimates discussed here. The same group’s 2015 trial is often cited alongside them too, but it measured muscle volume and strength rather than protein synthesis and so is not part of this meta-analysis’s muscle-protein-synthesis pool either 3.
The key methodological fact for interpreting Therdyothin et al.’s null MPS result is that the main pooled analysis (k=6) predominantly used basal, not stimulated, FSR — the authors state basal FSR was used to calculate the pooled effect in 50% of included studies, and was the preferred value (over stimulated FSR) whenever a single study reported both. This means the headline null (SMD 0.03, P=0.89) is largely answering the same question Smith 2011 itself answered in the negative — that omega-3 does not raise basal muscle protein synthesis — and does not directly test the sensitization/stimulated-state claim that is the more clinically relevant part of the original hypothesis. The two separate stimulated-condition pools (k=3 acute-exercise-stimulated, SMD 0.41, 95% CI −0.16 to 0.98, P=0.16; k=4 anabolic-stimulus-during-measurement, SMD 0.52, 95% CI −0.64 to 1.67, P=0.28) were both directionally positive but non-significant, with wide confidence intervals reflecting very small k. The MPS subgroup analyses (age, dose, duration, FSR method) each drew on only 2–5 of the 6 MPS studies and should be read as exploratory, not as robust independent replications of the null.
The correct summary is therefore: the sensitization hypothesis is unconfirmed and has not been reliably replicated at pooled scale — it has not been refuted. The authors themselves state they “did not detect any increase in MPS in response to anabolic stimulation with n-3 PUFA ingestion” in their pooled analysis, while also concluding “although n-3 PUFA ingestion might not augment basal rates of MPS, it is suggested that n-3 PUFA ingestion increases the response of MPS to anabolic stimuli,” and call for future trials that measure basal and stimulated FSR within the same design. This is consistent with — not contradictory to — the original Smith 2011 trial’s own basal/clamp dissociation; the stimulated-state pooled estimates are directionally aligned with it but underpowered to confirm it.
The whole-body protein synthesis finding is a genuine positive result but must not be conflated with the muscle-specific question: whole-body protein turnover integrates hepatic, visceral, and other tissue compartments, not just skeletal muscle. A pooled result from only 3 comparisons is itself fragile — a single additional null or negative trial could shift the pooled estimate’s confidence interval to cross zero. The null basal-MPS result, the non-significant stimulated-state pools, and the positive whole-body result should all be read as provisional, small-pool findings, not settled science. gap/needs-replication
Aging relevance
| Dimension | Status | Notes |
|---|---|---|
| Human evidence? | yes | All included studies were human trials; this is a meta-analysis of primary human data, not model-organism extrapolation. |
| Aging/older-adult population directly tested? | partial | Subgroup analysis by age was performed (k up to 5 of the 6 MPS studies) and found no effect modification, but this is an exploratory, small-k stratification, not a dedicated older-adult-only pooled estimate. |
| Muscle-specific sarcopenia-relevant outcome tested? | partial, and negative for the basal estimand | MPS/fractional synthetic rate is the most directly muscle-relevant outcome for sarcopenia, but the main pooled null (SMD 0.03, P=0.89) predominantly reflects basal FSR; the stimulated-state pools (k=3, k=4) — arguably the more sarcopenia-relevant estimand, since anabolic-stimulus responsiveness is what declines with age — were directionally positive but non-significant, not negative. |
| Replicated across subgroups? | no — subgroups are exploratory, not independent replications | The null basal-MPS finding held across small (k=2-5) age/dose/duration/method strata, but these are subsets of the same 6-study pool sharing most of its studies, not independent replications. |
Limitations
- Only 8 studies met inclusion criteria after screening 302 — this is a small evidence base for a meta-analysis, and the muscle protein synthesis pool (k=6), its stimulated-state subsets (k=3, k=4), and the whole-body pool (k=3) are each individually very small. Absence of a significant pooled effect on stimulated-state MPS is not proof of no effect — it may reflect being underpowered to detect a real effect of the size reported in Smith 2011.
- The main MPS pool is a mixed and predominantly basal-FSR estimand (basal FSR used in 50% of the 6 studies, and preferred over stimulated FSR when a study reported both) — this is a significant interpretive caveat that changes what the headline null does and does not tell us about the anabolic-stimulus-sensitization hypothesis (see § Why this reframes the muscle-protein-synthesis narrative above).
- The whole-body protein synthesis result (SMD 0.51, P=0.01) rests on only 3 comparisons with I²=0% (low heterogeneity, but heterogeneity estimates from 3 studies are themselves imprecise) — this finding should be treated as preliminary pending independent replication, not as an established effect. gap/needs-replication
- Search cutoff was December 2022; trials published after this date (including any updates to the Smith-trial line of evidence or newer RCTs) are not captured. gap/long-term-unknown
- The authors themselves call for future trials that measure basal and stimulated FSR within the same design and are adequately powered to detect a stimulated-state effect — the current stimulated-state pools (k=3, k=4) are directionally positive but too imprecise to confirm or rule out the sensitization hypothesis. gap/needs-replication
- Publication bias / small-study-effects assessment (e.g., funnel plot, Egger’s test) was not identified as reported for these small pools; with only 3-6 comparisons per outcome such tests would have low power in any case.
Cross-references
- omega-3-fatty-acids — parent compound family page
- protein-synthesis — the underlying process this meta-analysis interrogates
- sarcopenia — the aging phenotype for which muscle protein synthesis is a candidate mechanistic lever
- smith-2011-omega3-muscle-protein-synthesis-older-adults — small early trial whose clamp-stimulated finding motivates the sensitization hypothesis this meta-analysis’s basal-FSR-predominant null does not directly test
- smith-2015-omega3-muscle-mass-function-older-adults — companion/follow-up trial from the same group
- santo-andre-2023-n3pufa-muscle-meta-analysis — related prior meta-analytic synthesis of omega-3 and muscle outcomes
Footnotes
Cited by wiki pages
- omega-3-fatty-acids (pending propagation)
Footnotes
-
smith-2011-omega3-muscle-protein-synthesis-older-adults · small early trial reporting no effect of omega-3 PUFA on basal muscle protein synthesis but a large increase in clamp-stimulated muscle protein synthesis in older adults; the only Smith 2011 study among this meta-analysis’s 8 included studies ↩ ↩2
-
doi:10.1042/CS20100597 · PMID 21501117 · Smith GI et al. · Clinical Science 2011;121(6):267–278 · n=9 · single-arm pre/post, no control group · p<0.05 · model: healthy adults aged 25–45 · uncontrolled companion to the randomized older-adult trial; cited in Therdyothin et al.’s reference list but not among the review’s 8 included studies ↩ ↩2
-
smith-2015-omega3-muscle-mass-function-older-adults · n=60 randomized / 44 completed · randomized controlled trial · model: healthy adults aged 60–85 · 6 months · measured thigh muscle volume, grip strength, and 1-RM strength — not protein synthesis, so it does not contribute to this meta-analysis’s muscle-protein-synthesis pool ↩