Stanfield et al. 2026 — RAPA-EX-01 trial
The first randomized double-blind placebo-controlled trial designed to test whether once-weekly sirolimus combined with home-based exercise enhances functional gains in older adults — the “rapamycin cycling hypothesis” derived from preclinical work suggesting alternating mTORC1 inhibition (rest days) and activation (training days) could amplify training-induced adaptation while preserving autophagy. The trial was null on the primary endpoint and the sensitivity analyses favored placebo, with statistically significant attenuation of chair-stand gains in the per-protocol cohort. Adverse-event burden was higher in the sirolimus arm, including one possibly drug-related serious infection (pneumonia) after a single dose.
Together with the PEARL trial (null on visceral adiposity, 2025), RAPA-EX-01 marks the second major aging-rejuvenation RCT of rapamycin to fail or trend negative on its prespecified primary endpoint.
Design
- Registration: ACTRN12624000790549 (ANZCTR)
- Population: sedentary adults aged 65–85; mean age 72.2 yr; 47.5% female
- Allocation: 1:1 randomization, double-blind
- n=40 (20 sirolimus, 20 placebo)
- Intervention: sirolimus 6 mg orally once weekly for 13 weeks (vs. matched placebo)
- Co-intervention: standardized home-based exercise programme — chair-stands (resistance) + exercycle (endurance), 3×/week, both arms
- Primary outcome: change in 30-second chair-stand repetitions at week 13 (intention-to-treat; ANCOVA adjusted for baseline performance, age stratum, and sex)
- Sensitivity analyses (prespecified): complete-case (CC) and per-protocol (PP, defined as ≥75% dose adherence + exercise adherence)
- Secondary outcomes: grip strength, 6-minute walk distance (6MWD), SF-36 physical and mental component scores, C-reactive protein (CRP), four epigenetic age estimators
Primary outcome — 30-second chair-stand
Both arms improved chair-stand performance over 13 weeks (placebo: 14.30 ± 3.42 → 20.89 ± 5.92 reps; sirolimus: 13.75 ± 2.12 → 17.56 ± 4.18 reps).
Adjusted mean difference (sirolimus − placebo), ITT: −2.13 repetitions (95% CI −4.61 to 0.34; p=0.089; Cohen’s d = −0.53). Direction favors placebo; not statistically significant at the 0.05 threshold.
| Analysis | n (sirolimus / placebo) | Adjusted Δ | 95% CI | p | Cohen’s d |
|---|---|---|---|---|---|
| ITT | 20 / 20 | −2.13 reps | −4.61 to 0.34 | 0.089 | −0.53 |
| Complete-case | 16 / 19 | −2.46 reps | −4.87 to −0.06 | 0.045 | −0.64 |
| Per-protocol | 15 / 16 | −3.44 reps | −5.86 to −0.99 | 0.007 | −0.90 |
The CC and PP sensitivity analyses both reached statistical significance favoring placebo. The PP effect size (Cohen’s d = −0.90) is large.
Discontinuation: five participants in the sirolimus arm and three in the placebo arm discontinued early; their available data were retained in the ITT analysis.
Secondary outcomes
| Outcome | Adjusted Δ (sirolimus − placebo) | 95% CI | p | Direction |
|---|---|---|---|---|
| Grip strength (kg) | −1.19 | −3.52 to 1.18 | 0.344 | favors placebo, ns |
| 6MWD (m) | −4.87 | −28.97 to 19.71 | 0.706 | favors placebo, ns |
| SF-36 Physical Component | −2.76 | −8.81 to 3.32 | 0.376 | favors placebo, ns |
| SF-36 Mental Component | −1.22 | −4.16 to 1.91 | 0.455 | favors placebo, ns |
| C-reactive protein (mg/L) | +4.26 | −0.04 to 8.68 | 0.152 | sirolimus arm higher, ns |
CRP caveat: the +4.26 mg/L difference was driven by two outliers in the sirolimus arm with marked elevations (17 and 50 mg/L) at week 13. Excluding those two participants reduced the between-arm difference to <1 mg/L. The CRP signal should not be over-interpreted; immune-modulation interpretation requires biopsy-level mechanism work the trial did not perform.
Epigenetic age clocks
Four estimators were assessed; all showed mixed, non-significant differences between arms.
| Clock | Adjusted Δ (sirolimus − placebo) | 95% CI | p | Cohen’s d |
|---|---|---|---|---|
| PCGrimAge (years) | −2.28 | −4.93 to 0.44 | 0.098 | −0.112 |
| SystemsAge (years) | +1.04 | −0.69 to 2.55 | 0.24 | +0.137 |
| OMICmAge (years) | +0.28 | −0.74 to 1.40 | 0.592 | +0.002 |
| DunedinPACE | +0.0035 | −0.048 to 0.056 | 0.905 | +0.04 |
PCGrimAge (a principal-component refinement of GrimAge) showed the only suggestive trend (p=0.098, favoring sirolimus arm), but this contradicts the functional-outcome direction and the other three clocks. The authors characterize the panel as showing “mixed, non-significant trends.”
This is informative for the broader biological-age-measurement discussion: epigenetic clocks did not align with the functional outcomes that the trial was designed to detect, illustrating the open question of whether clocks measure something interventionally meaningful in short-duration trials.
Adverse events
- ≥1 AE: 17/20 (85%) in each arm
- Total AEs: 99 (sirolimus) vs. 63 (placebo); incidence rate ratio 1.57 (95% CI 0.86–2.87; p=0.14)
- Mean per-participant: 4.95 (sirolimus) vs. 3.15 (placebo)
- Most common AEs (Grade 1, mild): headache, fatigue, upper respiratory tract symptoms
- Possibly/probably drug-related: 35% (sirolimus) vs. 15% (placebo)
- SAEs: 1 in sirolimus arm (see below), 1 in placebo arm (not detailed in extraction)
Serious adverse event — pneumonia (Participant MDMR-00015)
- Onset: 2 October 2024
- Presentation: community-acquired pneumonia with nasal congestion and severe constipation
- Treatment: overnight hospitalization; IV antibiotics + steroids; constipation managed conservatively
- Outcome: symptoms resolved; participant withdrew from trial
- Exposure: had received only a single dose of sirolimus before the event
- Causality: “Given sirolimus’s immunosuppressive properties, a causal contribution to this serious AE cannot be excluded.”
This is mechanistically notable. Single-dose exposure to weekly sirolimus 6 mg is short — but a 62-h half-life means biologically active drug is still present at the time of infection presentation. The event is consistent with the documented immune-suppression risk profile that motivated rapamycin’s transplant indication.
Cycling hypothesis — what the trial tested and why it didn’t work
The “cycling hypothesis” proposed by preclinical models (Kaeberlein lab and others) holds that alternating mTORC1 activation (training days) and inhibition (rest days) could preserve or amplify training-induced muscle adaptations while still permitting autophagy-mediated cellular rejuvenation on rest days. Once-weekly 6 mg dosing was the schedule chosen to test this.
The result is the opposite — sirolimus attenuated training gains. The authors propose a pharmacokinetic explanation:
“The drug’s terminal half-life of approximately 62 h implies that biologically active concentrations persisted well into the subsequent training week.”
In other words, “once-weekly” dosing at 6 mg does not produce the on-off pulse the cycling hypothesis requires — drug persists across the training window, suppressing mTORC1 during exercise itself, blunting the training stimulus. This is a useful null test: the cycling hypothesis is not falsified in principle, but 6 mg/week is too high a dose to produce the cycling pharmacology. Lower doses or less-frequent dosing might still produce an on-off pulse compatible with the hypothesis.
Limitations (authors’ own)
- Exercise modality. Home-based body-weight resistance (chair-stands) has a lower ceiling for maximal strength development than gym-based heavy resistance training; the strength-gain magnitude available for either arm to demonstrate was constrained.
- Trial duration. 13 weeks is too short to assess longer-term effects of intermittent rapamycin combined with exercise, particularly at lower doses or less-frequent schedules.
- No mechanistic measurements. Muscle biopsies were not performed; circulating drug levels were not monitored. The “blunting” attribution to persistent mTORC1 inhibition is inferential, not directly demonstrated.
- Activity tracking. No accelerometry; daily diaries used to track non-exercise physical activity (NEPA). Authors report no systematic between-arm difference in extra-curricular activity, but this is self-report.
Implications for the rapamycin aging-rejuvenation thesis
- This is the second consecutive major aging-focused rapamycin RCT to report null/negative on its primary endpoint (after PEARL 2025 on visceral adiposity).
- Both trials used compounded or commercial sirolimus at 5–10 mg/week — the n=1 biohacker dose range that has been the de facto aging-community standard.
- Neither result falsifies the broader rapamycin-extends-mammalian-lifespan thesis (mouse ITP data is robust at 14–42 ppm in chow continuous; see Harrison 2009 and Miller 2014).
- They DO meaningfully constrain the dose × schedule × age-at-initiation × outcome space in which weekly sirolimus produces a clinically detectable aging benefit in older adults — at least for visceral fat mass (PEARL) and exercise-trained functional capacity (RAPA-EX-01).
- Authors’ own bottom line: “regular physical activity remains the unequivocal first-line strategy for preserving and improving function in older adults.”
Cross-references
- rapamycin — compound page; this is the third human RCT cited there (after Mannick 2014 and PEARL 2025)
- mtor — mechanism target
- mtor-inhibitors — class page
- hyperfunction-theory — rapamycin is the canonical hyperfunction-theory test molecule
- pearl-2025-rapamycin-healthspan — preceding null aging-RCT
- mannick-2014-everolimus-immune-aging — earlier positive immune-aging trial (everolimus, not sirolimus)
- grimage-2019 — PCGrimAge is a refinement of GrimAge
- dunedinpace-2022 — DunedinPACE used as secondary outcome