Smith 2015 — omega-3 PUFA, muscle mass and function in healthy older adults
Smith GI, Julliand S, Reeds DN, Sinacore DR, Klein S, Mittendorfer B · American Journal of Clinical Nutrition 102(1):115–122 · 2015 · DOI: 10.3945/ajcn.114.105833 · PMID 25994567 · PMC4480667
NCT01308957 (ClinicalTrials.gov).
TL;DR
A 6-month, double-blind, randomized, placebo(corn oil)-controlled trial in 60 healthy adults aged 60–85 tested whether 3.36 g/day EPA+DHA (fish oil) slows age-related muscle decline. Among the 44 participants who completed the trial (73% of the n-3 arm, 75% of control — a ~27% overall dropout), the n-3 group showed three statistically significant increases relative to control: MRI-measured thigh muscle volume (+3.6%), handgrip strength (+2.3 kg), and one-repetition-maximum (1-RM) strength (+4.0%); isokinetic muscle power did not reach significance (+5.6%, p=0.075). The muscle-volume and 1-RM-strength estimates carry meaningful uncertainty: their 95% confidence-interval lower bounds sit close to zero (0.2% and 0.8% respectively), and the completer sample is small (n=44) with a substantial dropout rate that adds further uncertainty; handgrip strength’s CI (0.8 to 3.7 kg) sits further from zero and is comparatively less borderline. This trial’s positive findings diverge from the subsequent pooled evidence base: multiple later meta-analyses find no effect of omega-3 supplementation on lean/muscle mass in older adults on pooled average, with at most a very small strength effect 1 2. gap/contradictory-evidence
Background and rationale
The authors’ own prior mechanistic work (the same group’s earlier trial, smith-2011-omega3-muscle-protein-synthesis-older-adults) had shown that fish oil-derived n-3 PUFA increases the rate of muscle protein synthesis in older adults, motivating a longer-duration functional trial to test whether this translated into measurable gains in muscle mass and strength — a candidate intervention against age-related sarcopenia.
Study design
| Feature | Detail |
|---|---|
| Design | Double-blind, randomized, placebo-controlled |
| Randomization ratio | 2:1 (n-3 PUFA : corn oil control) |
| Participants | Healthy men and women aged 60–85 years; no clinically significant physical impairments at baseline |
| n (randomized) | 60 total — n-3 group n=40, control n=20 3 |
| Dosing | n-3 group: 4 pills/day, 1.86 g EPA + 1.50 g DHA = 3.36 g/day total EPA+DHA; control: 4 identical corn oil pills/day 3 |
| Duration | 6 months |
| Registration | NCT01308957 |
| Primary outcomes | MRI-measured thigh muscle volume (1.5-T MRI, 22 consecutive 8-mm bilateral T1-weighted axial images); handgrip strength; 1-RM muscle strength; isokinetic leg power |
| Funding | NIH DK 56341, DK 20579, UL1 TR000448; study drugs gifted by GlaxoSmithKline plc 3 |
Completion
| Arm | Randomized | Completed | Completion rate |
|---|---|---|---|
| n-3 PUFA | 40 | 29 | 73% |
| Control (corn oil) | 20 | 15 | 75% |
| Total | 60 | 44 | ~73% |
Analysis was completers-only (n=44); this ~27% dropout is a material limitation the authors themselves flag. Compliance among completers was high (93.6% n-3, 91.8% control), and red-blood-cell n-3 PUFA content rose ~135% in the treatment group, confirming biochemical adherence 3.
Results
| Outcome | Effect (n-3 vs. control) | 95% CI | P |
|---|---|---|---|
| Thigh muscle volume | +3.6% | 0.2% to 7.0% | <0.05 |
| Handgrip strength | +2.3 kg | 0.8 to 3.7 kg | <0.01 |
| 1-RM muscle strength | +4.0% | 0.8% to 7.3% | <0.05 |
| Isokinetic muscle power | +5.6% | −0.6% to 11.7% | 0.075 — not significant |
All values from the PMC open-access full text 3. Secondary outcomes: body weight and total body fat were unchanged in both groups; intermuscular fat fell −4.0 cm³ in the n-3 group but this was not statistically significant. Blood pressure, lipid panel, liver enzymes, and glucose tolerance showed no adverse changes. Adverse events were mild and mostly gastrointestinal (fishy aftertaste 34% of n-3 group, diarrhea 11.4%, constipation 5.7%, indigestion 8.6%), with similar overall GI symptom rates between arms (31–33%) — no serious treatment-related adverse events 3.
Why the muscle-volume and 1-RM results carry substantial uncertainty
Three primary outcomes reached statistical significance: thigh muscle volume, handgrip strength, and 1-RM strength. Of these, two have confidence-interval lower bounds sitting very close to zero: thigh muscle volume’s lower bound is 0.2 percentage points, and 1-RM strength’s lower bound is 0.8 percentage points — both close to the boundary of significance. Handgrip strength’s lower bound (0.8 kg on a 0.8–3.7 kg CI) sits comparatively further from the null. With a completer sample of only 44 participants split 29:15, all three estimates carry meaningful sampling uncertainty, and the ~27% completer attrition adds further uncertainty about whether the completer sample is representative of all randomized participants. Isokinetic power — arguably the most functionally relevant strength measure, and the one with the widest CI — did not reach significance (p=0.075) and must not be reported as a positive finding.
This trial’s findings relative to later pooled evidence
Smith 2015’s positive muscle-mass finding stands in contrast to the subsequent meta-analytic evidence base on omega-3 and muscle outcomes in older adults:
- Santo André et al. 2023 (Advances in Nutrition 14(1):115–127) pooled RCT evidence and found no effect of omega-3 supplementation on muscle mass (SMD 0.07, 95% CI −0.02 to 0.17) or muscle function, with only a very small effect on muscle strength (SMD 0.12, 95% CI 0.006 to 0.24) 1.
- Cornish et al. 2022 (Nutrients 14(11):2221) similarly found no effect on lean tissue mass (SMD 0.09, 95% CI −0.10 to 0.28) 2.
Taken together, the pooled evidence suggests Smith 2015’s muscle-volume result — while internally well-conducted (randomized, double-blind, biochemically confirmed adherence) — was not supported by the pooled average across broader syntheses once more and larger trials were aggregated. A meta-analysis’s pooled average is not a replication trial: it does not test whether this specific trial’s result is reproducible, only whether the aggregate of available trials (of varying dose, duration, and population) centers on a null. Plausible contributors to the discrepancy between this trial and the pooled average include: this trial’s comparatively small completer sample and wide confidence intervals; possible chance findings given multiple outcome comparisons (a limitation the authors themselves note); and heterogeneity in dose, duration, and population across the broader trial literature that the meta-analyses pool over. gap/contradictory-evidence gap/needs-replication
See santo-andre-2023-n3pufa-muscle-meta-analysis for the meta-analytic synthesis.
| Dimension | Status |
|---|---|
| Pathway conserved in humans? | not applicable — human RCT |
| Phenotype conserved in humans? | not applicable — human RCT |
| Replicated in humans? | contradicted — later meta-analyses (Santo André 2023, Cornish 2022) find no muscle-mass effect and at most a marginal strength effect |
Author-identified limitations
- The fishy aftertaste experienced by the n-3 group (34%) may have partially compromised blinding.
- Dietary intake was not controlled or monitored during the trial.
- Multiple outcome measures were tested, raising the risk of false-positive findings.
- Analysis was completers-only (n=44 of 60 randomized), not intention-to-treat.
- The cohort was restricted to healthy older adults without clinically significant physical impairment — the authors explicitly note they “could not determine whether n-3 PUFAs have therapeutic effects” in sarcopenic or frail populations.
Cross-references
| Page | Relationship |
|---|---|
| omega-3-fatty-acids | Intervention tested; family-level compound page |
| sarcopenia | Clinical phenotype this trial’s outcomes bear on |
| smith-2011-omega3-muscle-protein-synthesis-older-adults | Same research group’s mechanistic predecessor (muscle protein synthesis rate) motivating this functional follow-up |
| santo-andre-2023-n3pufa-muscle-meta-analysis | Later meta-analysis contradicting this trial’s muscle-mass finding |
Limitations and gaps
gap/contradictory-evidence — This trial’s positive thigh-muscle-volume and 1-RM strength findings are not supported by subsequent pooled meta-analytic evidence (Santo André 2023: SMD 0.07 for muscle mass, CI crossing zero; Cornish 2022: SMD 0.09 for lean mass, CI crossing zero). Both bodies of evidence are documented here rather than one being discarded.
gap/needs-replication — No independent single trial of comparable design has replicated the specific muscle-volume and 1-RM effect sizes reported here; the completer sample (n=44) is modest, and the muscle-volume and 1-RM confidence-interval lower bounds sit close to zero, adding uncertainty on top of the ~27% completer attrition.
gap/long-term-unknown — 6-month duration only; no data on whether effects (if real) persist, grow, or regress with longer supplementation, or on hard functional/fall/mortality endpoints.
Footnotes
Footnotes
-
doi:10.1016/j.advnut.2022.11.005 · PMID 36811583 · Santo André HC et al. · meta-analysis · omega-3 supplementation and muscle mass/function/strength in older adults · muscle mass SMD 0.07 (95% CI −0.02, 0.17); strength SMD 0.12 (95% CI 0.006, 0.24) ↩ ↩2
-
doi:10.3390/nu14112221 · PMID 35684018 · Cornish SM et al. · systematic review/meta-analysis · omega-3 supplementation and lean tissue mass in older adults · lean mass SMD 0.09 (95% CI −0.10, 0.28) ↩ ↩2
-
smith-2015-omega3-muscle-mass-function-older-adults (this page) · doi:10.3945/ajcn.114.105833 · PMID 25994567 · PMC4480667 · Smith GI et al. · rct · Am J Clin Nutr 2015;102(1):115–122 · n=60 randomized (44 completed) · healthy adults aged 60–85 · 3.36 g/day EPA+DHA vs. corn oil control · 6 months · confirmed against PMC open-access full text ↩ ↩2 ↩3 ↩4 ↩5 ↩6