Although cytotoxic chemotherapy has still been the mainstay for cancer treatment, advances in the knowledge of tumor biology and the molecular pathways involved in cancer cell proliferation have ushered the age of molecularly targeted agents for cancer treatment 
. In contrast with traditional cytotoxic agents, these agents offer the promise of improved efficacy and a more favorable toxicity prolife. However, unique common side effect profile of these agents including hypertension, rashes, and metabolic abnormalities has also been reported in clinical trials 
. The incidence and management algorithms for those common side effects have been well defined in previous researches, but there is much more challenging to appreciate the uncommon, yet serious, toxicities associated with these drugs.
The meta-analysis is a powerful statistical tool to estimate the incidence and risk of those uncommon serious drug-related toxicities and this approach has been utilized to demonstrate an increased risk in treatment related mortality with bevacizumab and VEGFR-TKIs in previous researches 
. To the best of our knowledge, this is the first meta-analysis to investigate the incidence and risk of FAE associated with the mTOR inhibitors everolimus and temsirolimus. Our meta-analysis included 3322 patients from 12 trials demonstrates the overall incidence rate of FAEs is 1.8% (95%CI: 1.3–2.5%), and there is a significant three-times increased risk of death with these agents. However, a non-significantly increased risk of mTOR inhibitor associated FAEs is observed in sub-group analysis according to the mTOR inhibitors, tumor types and controlled therapy, for which we suggest several possible explanations: the small number of events recorded; under-reporting of rare (<5%) adverse events; the fact that clinical trials are usually not designed specifically to address toxic events; and the small number of randomized controlled trials included.
As mTOR inhibitors find more clinical applications and are used to treat a more heterogeneous patient population than those found in clinical trials, efforts are still needed to limit the risk of FAEs. Patients receiving mTOR inhibitors should be carefully monitored for the evidence of infection, especially patients with underlying known chronic lung disease or risk factors of infection. What’s more, as the use of mTOR inhibitors could cause non-infectious pneumonitis, which is characterized by non-infectious, non-malignant, and non-specific inflammatory infiltrates 
. Therefore, high-resolution computed tomography scans might be performed for patients present with cough and/or dyspnoea and/or hypoxemia, and/or fever when receiving mTOR inhibitors 
. In addition, previous researches have demonstrated that pneumovax is effective in preventing both influenza (in 70–80% of people) and pneumococcal infection (in 60–70% of people) 
, thus it might be a potential effective therapy for preventing mTOR inhibitors related pneumovax in cancer patients. However, until now, there is no specifically designed study to investigate the role of pneumovax for these patients, and studies focus on this issue is still needed.
Besides antitumor properties, mTOR inhibitors, especially sirolimus (rapamycin), have been widely used as an immunosuppressant in solid organ transplantation to prevent immune-mediated graft rejection 
. Interesting, sirolimus-associated pneumonitis has also been observed in renal and heart transplant recipients 
, and two deaths in patients who received sirolimus after heart transplants have been reported 
. However, the overall incidence of treatment mortality associated mTOR inhibitors is very low, and the use of sirolimus in transplant recipients is safe and tolerable 
This meta-analysis has some limitations. First, determining whether FAEs are attributable to mTOR inhibitors is particularly difficult in our study. Despite recommendations in the CTCAE version three (and beyond), the attribution of fatal events to particular toxicities was lacking in the majority of studies. Some studies did not clearly differentiate disease-related from non-disease-related fatal events. The lack of consistent reporting likely, in part, reflects the real-world difficulties of assigning causality to patient deaths, when the precise cause of death is unknown, or the cause of death may be easily associated with either the disease under study or the treatment being explored (e.g., thromboembolic events). However, in the current analysis, identical rules were utilized for abstracting events on both the mTOR inhibitors and control arms (treatment emergent fatal adverse events that were not specifically attributed to disease progression) which should have impacted over- or under-reporting of events on the mTOR inhibitors and control arms equally. Second, the ability of this study to detect variants in the FAE rate on the basis of specific drug or malignancy was limited because of low statistical power. Given that the conserved mechanism of action and known toxicities among the two study drugs are similar, it is unlikely that significant differences in FAEs would arise between them if more studies were available for analysis. As more high-quality studies of mTOR inhibitors in different malignancies and clinical settings become available, further analyses could be preformed to confirm the trends observed here. Third, the process by which investigators attribute FAE causality is a variable practice since FAEs were not the primary end point of any of the included studies. In addition, a continuity correction of 0.5 subjects with an event is used, which may have slightly overestimated the actual event rate of individual trials. Fourthly, although FAEs are prospectively collected for each individual study, this analysis is retrospective, and there are potentially important differences among the studies, including differing tumor types, dosage and administration schedule of mTOR inhibitors, periods of study conduct and study investigators. All of these would increase the clinical heterogeneity among included trials, which also made the interpretation of a meta-analysis more problematic. Additionally, our study includes a mixed population of patients treated mTOR inhibitors-based combination therapy or mTOR inhibitors alone, and patients received placebo or non-placebo therapy were also included in our study. Therefore, the treatment design is not the same in all arms, and it might be another source of heterogeneity. Finally, it is not an individual patient data analysis, and meta-analyses based on published data tend to overestimate treatment effects compared with individual patient data analyses. In addition, it precludes a more comprehensive analysis such as adjusting for baseline factors and other differences that existed between the trials from which the data were pooled.
In summary, our study demonstrates that the use of mTOR inhibitors seems to increase the risk of FAEs in patients with advanced solid tumors, but one should be cautious when interpreting these results due to the limitations of our study. Additionally, as this class of drugs gains greater clinical use, clinicians should be aware of the risks of FAEs with the administration of mTOR inhibitors in solid cancer, and closely monitoring is recommended during the therapy.