Table 2.

Risk of bias assessment for all studies included in the review

Study	Risk of bias judgment	Comments
Gundermann et al. [28] Activation of mTORC1 signalling and protein synthesis in human muscle following blood flow restriction exercise is inhibited by rapamycin	Low (Cochrane RoB 2.0)	- Well-described method including time points of rapamycin administration vs. tracer material infusion, muscle biopsy and blood sampling times. Important given the time to reach peak concentration of rapamycin - No difference in measured parameters of phospho-mTOR, phospho-S6K1, phospho-Erk1/2, phospho-Mnk1 at baseline - Authors note limitations with providing a dosage that guarantees complete inhibition of mTORC1: dose used was much lower than that often used in animal (rodent) experiments
Dickinson et al. [29] Mammalian target of rapamycin complex 1 activation is required for the stimulation of human skeletal muscle protein synthesis by essential amino acids	Low (Cochrane RoB 2.0)	- No variation in administration of the essential amino acid solution between rapamycin and control groups - No difference in measured parameter of mixed muscle protein fractional synthesis rate at baseline (i.e., pre-essential amino acid solution administration)
Dickinson et al. [30] Rapamycin does not affect post-absorptive protein metabolism in human skeletal muscle	Low (Cochrane RoB 2.0)	- No difference in measured parameters of mixed muscle fractional synthesis rate and whole-body phenylalanine rate of appearance at baseline - Authors note limitations of analyzing basal protein metabolism later course, as study period only included the first 2 h post-rapamycin ingestion (i.e., close to the peak circulating level of rapamycin post-administration)
Drummond et al. [31] Rapamycin administration in humans blocks the contraction-induced increase in skeletal muscle protein synthesis	Low (Cochrane RoB 2.0)	- No difference in measured parameters of serum hormones (cortisol, insulin), amino acids (leucine, isoleucine, valine, phenylalanine), mixed muscle protein fractional synthesis rate at baseline - Authors note limitations with providing a dosage that guarantees complete inhibition of phosphorylation of mTOR, S6K1, rpS6 pre-exercise: dose used was much lower than that often used in animal experiments
Veasey-Rodrigues et al. [32] A pilot study of Temsirolimus and body composition	Serious (ROBINS-I)	- Participants were almost exclusively below ECOG 2 (n = 15/16), indicating relatively good baseline function - Participants had a wide age range (36–71) and various types of cancers. The cancer diagnoses distribution is documented and appears skewed toward cancers that occur exclusively in females (25% endometrial, 18% ovarian, 12% cervical) - Grouping patients based on toxicity status and allowing dose reductions and treatment interruptions introduce possible confounding. Notably, there does not appear to be a protocol to describe the exact dose reduction and the frequency of treatment interruptions
Gyawali et al. [33] Muscle wasting associated with the long-term use of mTOR inhibitors	Serious (ROBINS-I)	- Study includes primarily only one cancer (90% renal cell carcinoma) in the study population, making the study participants relatively comparable - Study eligibility did not adequately address many other confounders contributing to muscle wasting. Regarding dosage of intervention, dose-skipping was allowed due to side effects. Notably, the dose of everolimus and Temsirolimus may have differed among participants, but this information was not documented - Study did not document many other diseases or groups of diseases for exclusion that may influence muscle wasting
Campistol et al. [23] Bone metabolism in renal transplant patients treated with cyclosporine or sirolimus	High (Cochrane RoB 2.0)	- No variation in administration of glucocorticoids between sirolimus and control (cyclosporine) groups. This resulted in no difference in cumulative glucocorticoid dose between groups - Open-label design introduces bias - Dose of sirolimus was carefully titrated to achieve chosen steady-state concentration for each intervention period - High dropout rate ~ 28% (161 enrolled, but analysis only includes 115 patients) - Potential omission of data when discussing individual vs. pooled analyses of studies. The author states “the same pattern of differences between the groups was observed…(data not shown).”
Westenfeld et al. [24] Impact of sirolimus, tacrolimus and mycophenolate mofetil on osteoclastogenesis—implications for post-transplantation bone disease	Moderate (ROBINS-I)	- No differences in characteristics of kidney transplant patients at baseline - Intervention was commenced at different times between participants, resulting in varying treatment durations. An average/median time to treatment is not documented - Authors note that patients treated with sirolimus might be more likely to take a lower cumulative steroid dosage, i.e., the presence of another intervention (steroids) is different at baseline between sirolimus vs. control groups - Baseline confounding was also possible due to the large differences in the meantime from transplantation, i.e., different states of osteodystrophy depending on time from transplant and follow-up investigation and treatment since]
Sessa et al. [25] Immunosuppressive agents and bone disease in renal transplant patients with hypercalcemia	Moderate (ROBINS-I)	- Appropriate use of exclusion criteria to minimize confounders of co-morbid disease, particularly chronic inflammatory disorders, smoking, and excessive alcohol intake, all of which may contribute to bone disease - 5 groups were created, each with different immunosuppressive protocols. Notably, the mean age of each group is appreciably different. On age alone as a confounder, this would already make a comparison between groups complexes - Elaboration on whether certain medications were restricted was required to strengthen the study - It is unclear whether the start and follow up times of the intervention/study differed between participants/groups
Gnant et al. [26] Effect of Everolimus on Bone Marker Levels and Progressive Disease in Bone in BOLERO-2	High (Cochrane RoB 2.0)	- No major dropout rate in the full analysis set (ITT), but major dropout rates in the safety analysis set: 404/482 discontinued intervention in the everolimus group, 229/238 discontinued intervention in the placebo group. While the safety analysis set is not the basis for the conclusions in the study, the high dropout rates are noteworthy. The sirolimus group contains ~ 102% more participants for analysis vs. the placebo group - No differences in characteristics in bone metastases and bisphosphonate use at baseline - Possible detection bias where local investigators could order additional scans or surveys at their discretion - Sponsored by Novartis, a major pharmaceutical company
Hadji et al. [27] The impact of mammalian target of rapamycin inhibition on bone health in postmenopausal women with hormone receptor-positive advanced breast cancer receiving everolimus plus exemestane in the phase IIIb 4EVER trial	Moderate (ROBINS-I)	- No difference in measured parameters of CTX, osteocalcin, P1NP, PTH, 25-OH-vitamin D at baseline between groups - Study design allowed anti-resorptive therapy to be used among participants. Although well-addressed in the analysis, the participants received different treatment dosages, negatively impacting the analyses
Bryun et al. [34] Everolimus in patients with rheumatoid arthritis receiving concomitant methotrexate: a 3-month, double-blind, randomized, placebo-controlled, parallel-group, proof-of-concept study	High (Cochrane RoB 2.0)	- No difference in measured parameters of a tender or swollen joint count, pain, patient, physician global disease activity, HAQ physical function, ESR or CRP - Some concerns with subjectivity in measurement, e.g., visual pain scale - Unclear justification in the grouping of some results, i.e., those that dropped out early grouped with those that used steroids - The average cumulative dose of systemic steroids over the 12-week treatment period was not significantly different, but dose adjustments were also allowed at investigators' discretion, given specific blood results
Wen et al. [35] Low-dose sirolimus immunoregulation therapy in patients with active rheumatoid arthritis: a 24-week follow-up of the randomized, open-label, parallel-controlled trial	High (Cochrane RoB 2.0)	- Sirolimus group contains 110% more participants for analysis vs. the conventional group. Importantly, concurrent immunosuppressive agents appear similar in proportion between groups - Open-label design introduces bias - Participants were able to freely use other use prednisone and other immunosuppressive medications to meet the treat-to-target recommendations
Niu et al. [36] Sirolimus selectively increases circulating Treg cell numbers and restores the Th17/Treg balance in rheumatoid arthritis patients with low disease activity or in DAS28 remission who previously received conventional disease-modifying anti-rheumatic drugs	Serious (ROBINS-I)	- Appropriate use of a third group—the “healthy” group, which comprised participants with no known co-morbidities or immunosuppressive use. This appears to serve as an appropriate baseline for Treg cell numbers measured by flow cytometry - No difference in characteristics of rheumatoid arthritis patients at baseline. Importantly, concomitant immunosuppressive agents appear similar in proportion between rheumatoid arthritis patients treated with sirolimus vs. conventional treatment - Confounding was possible given how other immunosuppressive medications were allowed. In addition, the reviewing physician could freely change the dosages of these medications during the study