Exercise and Weekly Sirolimus (Rapamycin) Use in Older Adults: A Randomized, Double-Blind, Placebo-Controlled Trial

We conducted a single‐centre, randomised, double‐blind, placebo‐controlled, parallel‐group exploratory trial at Aotearoa Clinical Trials Trust (Auckland, New Zealand). The protocol was approved by the New Zealand Northern B Health and Disability Ethics Committee (2024 FULL 20084), registered with ANZCTR (ACTRN12624000790549) and published in Trials [33]. All participants provided written informed consent. The trial timeline and schedule of assessments are detailed in Table 1.

We recruited community‐dwelling adults aged 65–85 years. To target a population likely to benefit from exercise initiation, inclusion required a sedentary lifestyle, defined as performing moderate intensity exercise for less than 15 min, three times per week. Exclusion criteria included uncontrolled diabetes (HbA1c ≥ 60 mmol/mol), significant renal (eGFR < 30 mL/min) or hepatic impairment, chronic corticosteroid use or contraindications to sirolimus. Participants taking rate‐limiting cardiac medications (e.g., beta‐blockers) were not excluded provided the condition was stable.

Participants were randomly assigned (1:1) to sirolimus (rapamycin) or placebo using a computer‐generated sequence stratified by age (65–74 and 75–85 years). The allocation sequence was generated by an independent statistician and implemented via a secure REDCap module; allocation remained concealed from participants, care providers and outcome assessors throughout the trial.

Participants self‐administered three #000 capsules containing either 6 mg sirolimus (rapamycin; 3 × 2 mg Rapamune tablets) or matched microcrystalline cellulose placebo once weekly for 13 weeks. To minimise interference with exercise‐induced anabolic signalling, dosing was prescribed for ‘Day 6’ of each training week, approximately 24 h after the final exercise session.

Both groups performed an identical home‐based exercise programme three times per week (Days 1, 3, and 5) as outlined in Table 2. The regimen consisted of two components. First, a resistance component (30‐s chair stand) where participants performed maximum repetitions of standing from a seated position in 30 s. Although external weight was not added, progressive overload was achieved via ‘density training’, where participants were instructed to increase the number of repetitions performed within the fixed time interval as their fitness improved. Second, an endurance component using a provided magnetic resistance exercycle. The protocol followed a fixed 13‐week progression (Table 3), starting at 10 min (Level 1 resistance) and titrating up to 25 min (Level 5 resistance) by Week 9. Intensity was prescribed via mechanical targets (Resistance Level 1–5 and RPM 70–80) rather than heart‐rate zones. This ensured that participants on beta‐blockers or with chronotropic incompetence received a consistent physiological work stimulus. Where participants were unable to complete the exercycle training programme due to difficulty, the programme could be adjusted whereby the resistance setting was lowered, followed by a reduction in speed. If a participant still could not complete the training programme, they aimed to ride for as long as they were able before moving to the cooldown phase. Participants recorded exercise activities in their Participant Diary, and specifically for the exercycle sessions, the participants included the length of training, the RPM and the resistance setting.

The primary outcome was the change in lower‐body functional performance measured by the 30CST from baseline to Week 13. Secondary functional outcomes included the change in 6‐min walk distance (6MWD) assessed on a 30‐m indoor course and dominant‐hand grip strength measured as the maximum of three trials using a hydraulic dynamometer. Health‐related quality of life was assessed using the SF‐36 Physical and Mental Component Summary scores. Exploratory mechanistic endpoints included CRP, which serves as a marker of inflammation, and epigenetic age acceleration derived from genome‐wide DNA methylation profiling (TruDiagnostic TruAge). Safety was assessed via the incidence of adverse events (AEs; CTCAE v5.0) and changes in safety laboratory parameters.

The sample size of 40 participants (20 per arm) was determined to provide reasonable precision for estimating effect sizes and variance components in an exploratory setting [34, 35]. This sample size yields approximately 80% power at a two‐sided alpha of 0.05 to detect a large effect size (Cohen's d ≈0.80–0.90) on the primary outcome using ANCOVA adjusted for baseline values. Additionally, this sample size enables detection of AE rate ratios between approximately 2.4 and 3.7, providing reasonable sensitivity to clinically important safety signals. The target of 40 participants also represented the maximum feasible enrolment given single‐site operational capacity and funding constraints. No interim analyses or stopping rules were planned.

All analyses were performed in R Version 4.4.0 and the R script is available as a supplementary file. The primary efficacy analysis was performed on an intention‐to‐treat (ITT) basis including all randomised participants. The Week 13 chair‐stand count was compared between arms using a linear regression model adjusted for sex, age group (< 75 vs. ≥ 75 years) and baseline values of chair‐stand, hand‐grip strength and 6MWD. Because heteroskedasticity was anticipated, standard errors and confidence intervals were estimated using a Rademacher wild bootstrap with 999 replicates. Missing outcome data were handled via multiple imputation using predictive mean matching (15 imputed datasets).

Secondary outcomes (grip strength, 6‐min walk, SF‐36 scores, CRP and epigenetic clocks) were analysed using the same baseline‐adjusted model and wild‐bootstrap approach. Prespecified sensitivity analyses included a complete‐case analysis (participants with paired baseline and Week 13 data) and a per‐protocol analysis (participants completing ≥ 75% of prescribed doses and exercise sessions). AEs were analysed using negative binomial regression to compare event counts between groups. As this was an exploratory trial, no adjustments were made for multiplicity.

Between 29 July 2024 and 30 September 2024, 45 individuals were assessed for eligibility; 40 participants were randomised (20 assigned to sirolimus, 20 to placebo). Baseline demographic and clinical characteristics were well balanced between groups (Table 4). Detailed baseline vital signs, laboratory parameters and individual SF‐36 domain scores are provided as Table S1. Follow‐up was completed on 14 January 2025. Five participants in the sirolimus (rapamycin) arm and three in the placebo arm discontinued the intervention early, but their available data were included in the ITT analysis (Figure 1).

Adherence to the protocol was high. Median medication adherence was 90.9% (IQR 80.8–100.0) in the placebo arm and 87.3% (IQR 76.6–100.0) in the sirolimus (rapamycin) arm. Exercise adherence was similarly robust, with mean session completion rates of 77.3% (SD 34.7%) and 72.6% (SD 38.1%), respectively (mean difference 4.7 percentage points; 95% CI −18.6 to 28.0; p = 0.69).

To address potential confounding by exercise intensity tolerance, we analysed daily session logs. Deviations from the prescribed workload (e.g., reducing resistance below the weekly target or stopping a session early due to fatigue) were rare, occurring in less than 1% of total scheduled sessions across both groups. Furthermore, the use of rate‐limiting cardiac medications (beta‐blockers) was balanced between arms (n = 2 in sirolimus; n = 2 in placebo), suggesting that chronotropic incompetence did not differentially affect the ability to achieve the prescribed mechanical workload targets.

Both groups improved their lower‐body functional performance over 13 weeks (Figure 2A). In the primary ITT analysis (n = 40), the baseline‐adjusted mean difference in 30‐s chair‐stand repetitions at Week 13 (sirolimus minus placebo) was −2.13 repetitions (95% CI −4.61 to 0.34; p = 0.089). This represented a small‐to‐medium negative effect size (Cohen's d = −0.53).

Prespecified sensitivity analyses, which excluded non‐adherent participants or those with missing data, indicated a significant attenuation of benefit in the treatment arm. The complete‐case analysis (16 sirolimus, 19 placebo) yielded a mean difference of −2.46 repetitions (95% CI −4.87 to −0.06; p = 0.045; Cohen's d = −0.64). The per‐protocol analysis (15 sirolimus, 16 placebo) showed a large negative effect, with a difference of −3.44 repetitions (95% CI −5.86 to −0.99; p = 0.007; Cohen's d = −0.90).

Secondary functional and patient‐reported outcomes followed a similar pattern, with point estimates favouring placebo, though differences did not reach statistical significance.

The adjusted mean difference for the 6MWD was −4.87 m (95% CI −28.97 to 19.71; p = 0.706; Cohen's d = −0.26). Grip strength showed a difference of −1.19 kg (95% CI −3.52 to 1.18; p = 0.344; Cohen's d = −0.40).

Self‐reported health‐related quality of life (SF‐36) showed negligible between‐group differences. The adjusted mean difference for the Physical Component Summary was −2.76 points (95% CI −8.81 to 3.32; p = 0.376), and for the Mental Component Summary was −1.22 points (95% CI −4.16 to 1.91; p = 0.455).

Exploratory analysis of CRP showed a mean difference of +4.26 mg/L (95% CI −0.04 to 8.68; p = 0.152) in the sirolimus arm. However, this was driven by two outliers in the treatment group with marked elevations (17 and 50 mg/L) at Week 13; excluding these participants reduced the difference to < 1 mg/L. Epigenetic age measures showed mixed, non‐significant trends (Table 5).

Overall, across all secondary outcomes the confidence intervals spanned zero and the direction of effect usually favoured placebo, supporting the null hypothesis that once‐weekly sirolimus (rapamycin) does not improve and may modestly attenuate functional or quality‐of‐life gains from exercise over 13 weeks. Detailed descriptive statistics and adjusted estimates for all primary and secondary outcomes are presented in Table 5.

Safety was assessed in all 40 randomised participants. Seventeen participants (85%) in each arm reported at least one AE. However, the total burden of events was numerically higher in the sirolimus arm (99 total events vs. 63 in placebo; incidence rate ratio 1.57; 95% CI 0.86–2.87; p = 0.14). The majority of events were mild (CTCAE Grade 1), with headache, fatigue and upper respiratory tract symptoms being the most common (Table 6). Importantly, events adjudicated as possibly or probably related to the study drug were more frequent in the sirolimus arm (35% vs. 15%).

One serious AE occurred in a participant in the sirolimus (rapamycin) arm (ID MDMR‐00015) who developed community‐acquired pneumonia with nasal congestion and severe constipation on 2 October 2024. The participant was hospitalised overnight and treated with intravenous antibiotics and steroids; constipation was managed conservatively. Symptoms resolved, but the participant withdrew from the trial. No other participants discontinued because of AEs. Of note, this participant had only received a single dose of sirolimus (rapamycin); however, given sirolimus (rapamycin)'s immunosuppressive properties, a causal contribution to this serious AE cannot be excluded.

Analysis of safety laboratory parameters revealed several statistically significant but clinically modest shifts in the sirolimus arm (Table 6). Mean corpuscular volume was lower (adjusted difference −2.90 fL; p < 0.001), whereas platelet count (+17.6 × 10⁹/L; p = 0.025), alkaline phosphatase (+5.56 U/L; p = 0.012), LDL cholesterol (+0.32 mmol/L; p = 0.036) and HbA1c (+1.74 mmol/mol; p = 0.030) were slightly elevated compared with placebo. Bilirubin and albumin trended lower, whereas insulin‐like growth factor 1 (IGF‐1) trended higher, though these did not reach statistical significance. As noted in the secondary outcomes, CRP levels were driven by outliers and showed no consistent pattern.

This randomised, double‐blind, exploratory trial evaluated whether once‐weekly sirolimus (rapamycin) 6 mg, timed to avoid peak training windows, could enhance functional adaptations to a 13‐week resistance and endurance programme in sedentary older adults.

Contrary to the ‘cycling hypothesis’ derived from preclinical models, we found no evidence of benefit. Instead, sensitivity analyses (complete‐case and per‐protocol) revealed a statistically significant blunting of functional gains in the sirolimus arm. Specifically, participants taking the drug performed fewer chair‐stand repetitions and showed non‐significant trends toward reduced walking distance and grip strength compared with placebo. Safety monitoring identified a higher burden of minor AEs and one serious infection (pneumonia) in the treatment arm.

This observation is consistent with classic rodent overload studies in which rapamycin largely prevented compensatory hypertrophy and with acute human data showing that a single large dose abolishes the post‐exercise rise in muscle protein synthesis [4, 5, 7]. This parallels findings from trials of metformin (an indirect mTORC1 inhibitor), where concurrent administration attenuated improvements in muscle mass and strength in older adults [36, 37].

Notably, very recent preclinical data in mice showed no blunting of exercise adaptations with rapamycin; differences in species, sex, dose or timing relative timing likely explain the apparent discrepancy [8].

Although our trial is the first to combine once weekly sirolimus (rapamycin) with exercise in older adults, its findings differ from those of the PEARL study, which administered rapamycin weekly without an exercise programme and reported modest improvements in female lean body mass and pain [28]. Differences in trial duration (48 weeks vs. 13 weeks), endpoints (body composition and symptoms versus functional performance) and the absence of an anabolic stimulus in PEARL likely account for these divergent results. Together, the evidence suggests sirolimus (rapamycin)'s effects are context dependent: possibly beneficial in sedentary settings over longer durations, but counterproductive when overlapping with short term exercise training.

The failure to achieve a ‘best of both worlds’ effect (the concurrent realisation of sirolimus‐induced geroprotection and exercise‐induced muscle hypertrophy) likely stems from a pharmacokinetic mismatch. The ‘cycling hypothesis’ relies on a clear separation between the catabolic/autophagic phase (drug effect) and the anabolic phase (recovery from exercise). Although we dosed sirolimus 24 h after the final weekly exercise session, the drug's terminal half‐life of approximately 62 h implies that biologically active concentrations persisted well into the subsequent training week [38]. Consequently, mTORC1 (the master regulator of translation initiation) likely remained partially inhibited during the critical post‐exercise windows of the following sessions, thereby dampening the hypertrophic response.

Although weekly sirolimus is often discussed as a potential longevity therapeutic, our data warrant caution regarding its short‐term safety profile in active older adults. The treatment arm experienced a 57% higher incidence rate of AEs, driven by minor infections and constitutional symptoms. The occurrence of community‐acquired pneumonia requiring hospitalisation in one participant, though singular, aligns with the drug's known immunosuppressive mechanism.

Strengths of this trial include the double‐blind, placebo‐controlled design, high adherence to the home‐based protocol, and the use of validated functional endpoints. We also addressed potential confounders such as beta‐blocker use and exercise intensity tolerance, finding no evidence that these factors biased the results.

A key limitation was the home‐based nature of the exercise intervention. Unlike gym‐based training with external weights, our chair‐stand protocol relied on body weight. Although we employed ‘density training’ (increasing repetition volume within a fixed time) to ensure progressive overload, this approach may have a lower ceiling for maximal strength development than heavy resistance training. Furthermore, the trial was limited to 13 weeks, so the longer‐term effects of combining sirolimus (rapamycin) with exercise, particularly with lower doses or less frequent administration, remain unknown. Finally, we did not perform muscle biopsies or pharmacokinetic monitoring, so our mechanistic attribution of the ‘blunting’ effect to persistent mTORC1 inhibition remains inferential.

Finally, although we did not use accelerometry to quantify total non‐exercise physical activity (NEPA), participants recorded additional voluntary physical activity (e.g., walking, gardening) in daily diaries; review of these records suggests no systematic difference in extra‐curricular activity between groups.

Future research should explore regimens that lengthen the time between doses (e.g., every 3–6 weeks). Given sirolimus's ~62‐h terminal half‐life, this interval would allow for near‐complete drug clearance, ensuring mTORC1 signalling can recover to support exercise adaptations. Furthermore, a longer interdose interval may mitigate the risk of continuous immunosuppression and AEs, such as the serious infection observed in this trial, by allowing cyclical recovery of immune function. Testing such schedules would require at least 12 months of follow‐up. Sirolimus (rapamycin)'s putative benefits on epigenetic aging, chronic inflammation and body composition likely accrue slowly; short trials risk missing meaningful divergences. Moreover, although our initial data suggests that sirolimus (rapamycin) 6 mg once weekly may blunt early training gains, it may plausibly flatten the longer‐term slope of functional decline. Observing function over a full year would reveal whether longer interdose intervals lead to net preservation of strength and endurance or merely shift the timing of adaptations.