Clinical Lessons of MSC Therapy Over the Past 15 Years: A Systematic Review and Meta-Analysis

Background: Despite increasing clinical investigations in emphasizing the safety of MSC therapy in different populations with different diseases, recently, no article overall reviewed the side events in all populations. Aim: To evaluate the safety of MSC therapy in all populations receiving MSC therapy and explore the potential heterogeneities inuencing the clinical application of MSC. Methods: The databases of PubMed, Embase, Web of Science and Scopus were searched from onset until March 1 st , 2021. Results: All side events were displayed as odds ratio (OR) and 95% CI (condential intervals). Totally, 62 randomized clinical trials (RCTs) that enrolled 3546 participants diagnosed with various diseases (about 20 kinds of diseases) treated with intravenous or local implantation vs placebo, or no treatment were included. All studies were high quality, neither serious publication bias nor serious adverse events (such as death and infection) were discovered across included studies. The pooled analysis demonstrated that MSC administration was extremely associated with transient fever [OR, 3.65, 95% CI: 2.05 to 6.49, p<0.01], administration site conditions [OR, 1.98, 95% CI: 1.01 to 3.87, p=0.05], constipation [OR, 2.45, 95% CI: 1.01 to 5.97, p=0.05], fatigue [OR, 2.99, 95% CI: 1.06 to 8.44, p=0.04], and sleeplessness [OR, 5.90, 95% CI: 1.04 to 33.47, p=0.05]. Interestingly, MSC administration trended to lower rather than boost the incidence rate of arrhythmia [OR, 0.62, 95% CI: 0.36 to 1.07, p=0.09]. Conclusions: Conclusively, MSC administration was safe in different populations compared with the other placebo modalities. were puried from studies performed over the past 11 years. Only 2 studies were prospective and the rest were randomized controlled trials (RCTs) ranging from study phrase 1/2 to study phrase 3. Asia was ranked rst in the NO. of studies, and then North America and Europe. The MSC used in these studies were mainly isolated from bone marrow, adipose and umbilical cord. The injection dose ranged from 4×10 7 to 1.2 ×10 9 cells. The follow-up day was from 6 months to 2 years. and constipation were discovered. Many population characteristics, including age, analysis, cell type, disease, gender, location, study phrase, year, and administration method possibly impacted the occurrence of one side event. The safety of MSC administration should be under sustained observation despite the innovative therapy appears safe.


Introduction
Mesenchymal stromal cells (MSCs), a class of highly heterogeneous cells, that can be isolated from bone marrow, adipose, umbilical cord, and placenta, are primarily discovered in 1974 by Friedenstein [1] . Over these years, exogenous MSCs are amazingly found to have a therapeutic effect on many diseases (e.g., myocardial infarction, liver cirrhosis, limb ischemia and spinal cord injury) [2][3][4][5] .
Deferring from multipotent stem cells, the potency of MSCs is restricted but MSCs can be induced into osteoblasts, chondrocytes and adipocytes in vitro.
Universally, MSCs exert their favourable effects by immunomodulatory regulation and paracrine manners [6,7] . Clinically, MSCs have been applied in many refractory diseases, such as cerebral palsy [8] , spinal cord injury [9] and systemic lupus erythematosus [10] . However, MSCs easily ock together forming the core of the clot and leading to vascular disorders. Additional, MSCs are tumorigenic as a result of their reproductive capacity and potentially cause acute or chronic immunogenicity of the cells themselves as foreign matters [11][12][13] . A large number of studies, most of which are small samples, have been investigating the safety of MSCs transplantation, but no article overall reviews these studies to characterize the side events closely associated with MSCs administration over the past 9 years.
We performed this systematic review and meta-analysis to identify all treatment-related side events concerning MSCs administration and explore the safety of MSCs in clinical utilization.

Methods And Materials
Search results This meta-analysis was limited to published articles assessing the safety of MSC administration and was performed by searching PubMed, EMBASE, Web of ). The reference lists of the included articles were also browsed to identify potential studies. To perform a comprehensive search, we did not limit the "study type"; retrospective studies were excluded during the study selection process. The detailed database search strategy is provided in Additional le 1.

Article selection
Primarily, duplicates of all articles were excluded. Two participants initially screened all titles and abstracts to preclude articles that were irrelated to our research objectives. Then, we carefully read the full manuscripts and selected the eligible ones.

Eligibility criteria
The selection process strictly obeyed the PICOS (participants, interventions, comparison, outcome and study) principles and these principles were listed in Table 1.

Data extraction
Two skilled reviewers (YW and HX Y) independently extracted data from all articles according to pre-set criteria. We retrieved 12 characteristic entries from the original articles, including author, year, and study type, location, disease, and cell type, administration method, study phrase, and language, dose, follow-up day, and the NO. of the patient in each group. Con icts were resolved in consultation with a third referee.

Quality assessment
Risk of bias in individual study and across studies were performed by using the Cochrane Collaboration's tool for assessing the risk of bias.

Outcome de nition
Totally, we reported 17 side events appeared during MSC therapy, and of which 9 events (death, infection, diarrhea, central nervous system disorders, arrhythmia, urticaria/dermatitis, vascular disorders) were classi ed into major events, and 8 events (anemia, constipation, metabolism disorders, fatigue, nausea, seizure, sleepless, vomiting) were classi ed into minor events. One event would be considered as a major event if it was reported by more than 5 studies or life-threatening judged by our clinician; otherwise, it should be sorted into a minor event. Among these events, some events were not speci cally clinical symptoms but referred to a series of correlated symptoms, such as central nervous system disorders, vascular disorders, infection, arrhythmia, administration site conditions, metabolism and nutrition disorders. These side events were rede ned in Table 2. Other entries were retrieved from the original de nitions.

Statistical analysis
All data were synthesized by using R software version 4.0.3 (University of Auckland, New Zealand). All results presented in this article were presented as odds ratio (OR) with 95% CIs for outcomes. A random-effect model was used to analyze the data when heterogeneity was signi cant (p<0.05 or I 2 >50%); otherwise, a xed-effect model was used. Publication bias was tested by Egger's and Begg's tests were utilized to analyze the publication bias of the included articles with R software version 4.0.3 (meta package). Subgroup analysis was also conducted to seek potential heterogeneous factors.

Results
The items of this meta-analysis were reported according to the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) (Additional le 2).

Article selection process
Approximately, 2078 articles were identi ed after the initial search. 1898 irrelevant articles were eliminated through browsing titles and abstracts, and 118 articles were excluded due to unexpected outcomes and interventions. Finally, 62 clinical trials, including 2 trials from the reference list, were taken into the analysis even if the elimination of 2 systematic reviews ( Figure 1).

Baseline of included studies
The data were puri ed from studies performed over the past 11 years. Only 2 studies were prospective and the rest were randomized controlled trials (RCTs) ranging from study phrase 1/2 to study phrase 3. Asia was ranked rst in the NO. of studies, and then North America and Europe. The MSC used in these studies were mainly isolated from bone marrow, adipose and umbilical cord. The injection dose ranged from 4×10 7 to 1.2 ×10 9 cells. The follow-up day was from 6 months to 2 years.

Pooled analysis of all studies
Totally, 62 clinical trials, containing populations with different characteristics, were included into analysis ( Figure 2A). We discovered that MSC administration would not induce major side events, such as vascular disorders ( Figure 2B).

Subgroup analysis of all studies
Subsequently, we dissected potential factors, including administration (method), age, methodology (analysis) of the article, cell type, population (disease), gender proportion, location, study phrase and publication date (year), in uencing the major side events ( Figure 3A). We identi ed that the non-signi cance of death, infection and diarrhea, which were not treatment-related side events of MSC therapy, were not altered in the slightest by any of the analyzed factors. MSC therapy was demonstrated to reduce the incidence of arrhythmia in the population with the age <60 years (p<0.01), PP analysis (p=0.01) and beyond 5 years (p<0.01). Despite the non-signi cant central nervous system disorders (head and dizziness) proved by pooled analysis, AD-MSC (p<0.01), placenta MSC (p<0.01) and uc-MSC (p<0.01) were more easily to cause headache and dizziness. Meanwhile, a population with degenerative joint diseases (p<0.01) and digestive diseases (p<0.01) could have headache and dizziness symptoms while receiving MSC implantation. Urticaria signi cantly occurred when the data were analyzed by PP analysis exclusively (p<0.01). As for vascular disorders, Asian people more easily had vascular disorders (p<0.01) after MSC treatment. Administration site conditions preferred to occur in populations with the age < 60 years (p=0.02), heart related diseases (p=0.01), the male proportion >60% (p=0.08), in study phrase 1/2 (p=0.01), and within 5 years (p=0.05). Even though transient fever was conspicuously associated with MSC treatment, populations with the age > 60 years (p=0.86), the male proportion < 60% (p=0.7), receiving local implantation (p=0.76), in North America (p=0.82) and study phrase 1 (p=0.15), had a lower risk of transient fever over the period MSC of therapy.
In terms of the minor side events, only ve side events, including anemia, constipation, metabolism and nutrition disorders and nausea, were analyzed ( Figure   3B). Similarly, the interactions between the 9 predicted factors and seldomly reported side events were dissected. Contrary to pooled analysis, neither constipation nor fatigue was a signi cant side event in these subgroup analyses. Similar to pooled analysis, both metabolism and nutrition disorders and nausea were non-impacted by these factors and were non-signi cant side events. Interestingly, we found that populations with the age < 60 years trended to have transient anemia (p=0.07) post-MSC treatment.

Subgroup analysis of high-quality studies
We examined whether the potential factors signi cantly in uenced the terminal outcomes (7 major side events) reported by high-quality studies ( Figure 5). MSC administration would not directly lead to death, death, central nervous disorders (headache and dizziness), or vascular disorders. Populations with the age <60 years (p<0.01), receiving BMSC injection (p=0.04), in study phrase 3 (p=0.04), and beyond 5 years (p<0.01) seemed to have a lower incidence of arrhythmia and bene t from MSC administration. When it came to transient fever, MSC administration would not trigger fever in populations with the age > 60 years (p=0.86), the male proportion <60% (p=0.70), from Europe (p=0.82), in study phrase 2 (p=0.15), beyond 5 years (p=0.11), and receiving local implantation (p=0.76).

Sensitivity analysis
Leave-one meta-analysis was performed for administration site conditions, arrhythmia, death, dermatitis, diarrhea, transient fever, infection, central nervous system disorders, and vascular disorders, and fatigue, metabolism and nutrition disorders, anemia, constipation, and nausea from all studies (Additional le 3-16), and for administration site conditions, arrhythmia, death, transient fever, infection, central nervous system disorders, and vascular disorders from highquality studies (Additional le 17-23 ).

Publication bias and article quality
We assessed the article quality by using the Cochrane Collaboration's tool for assessing the risk of bias ( Figure 6). We concluded that most studies' design was suitable and high quality. Only 14 studies were considered as low quality because they had more than two entries marked as high risk and less than four entries evaluated as low risk. There was performance bias, selection bias, detection bias, and attrition bias potentially lowering the integral quality of included studies. Furtherly, we tested the publication bias for administration site conditions, arrhythmia, death, dermatitis, diarrhea, transient fever, infection, central nervous system disorders, and vascular disorders (Additional le 24-32) from all studies. Publication bias for administration site conditions, arrythmia, death, fever, infection, central nervous system disorders, and vascular disorders (Additional le 33-39) from high-quality studies were conducted as well.

Summary of evidence
The association between the side events and MSC administration is rst reported by M. Lalu [14] and the association between MSC administration and the infusional toxicity, organ system complications, infection, death was not explored due to limited clinical researches. However, aside from these side events above, which were analyzed in this systematic review, more side events are described in recent trials with the expansion of population. In addition to transient fever, which is the most frequently reported by researchers, other side events such as constipation, fatigue, administration site conditions and sleeplessness can be induced by MSC administration as well. As for arrhythmia, MSC seems to bene t patients with cardiac diseases.
We were unable to discover the conspicuous association between MSC administration and the rest side events (vascular disorders, urticaria/dermatitis, dizziness/headache, diarrhea, infection, death, anemia, metabolism and nutrition disorders, nausea, seizure, and vomiting). Neither, there was direct proof suggesting that the MSC administration is tumorigenic. Up to date, the malignance of MSC was merely reported by Ning [15] despite the potential of tumorigenesis of MSC.
After the elimination of the low-quality studies, eight side events were actually analyzed, including metabolism and nutrition disorders, infection, fever, and death, arrhythmia, dizziness/headache, vascular disorders, and administrations. Among these side events, transient fever was exclusively associated with MSC administration. Arrhythmia and administration site conditions trended to be signi cant after MSC administration. Other side events had no relevance to MSC administration.
Furtherly, we analyzed each side events in various sub-populations to dissect by which side event was determined. We discovered that age, gender proportion, location, and year, analysis, disease, study phrase, cell type, and administration method were the main factors impacting the nal side event. Take the de nite side event fever as an instance, the aged were not susceptible to MSC administration and this may be because of blunt reactions of the organism to acute in ammation triggered by MSC [16] . The female more easily suffers from transient fever and the estrogen level is under serious doubt [17] . The population in North America less undergo transient fever compared to other regional populations and this may suggest the racial discrepancy of MSC administration.

Strengths and weakness
This meta-analysis removes studies of low-grade evidence (retrospective study, single-arm study, and case) and included 62 prospective studies. All results suggest the strong association between MSC administration and transient fever, and administration site conditions. Moreover, more side events that were not reported before (e.g., anemia constipation and vomiting) are gradually being discovered [18][19][20] . Theoretically, the side events of MSC administration should be under stringent surveillance in case of occurrence of other side events that were not reported before along with the expansion of clinical trials. We also notice that the longest follow-up is 5 years, which may be a shorter time considering the fact that we are using cell products. We should be cautious that longer term events in the farther future possibly impend.
Our research has limitations. First, we synthesized the data across heterogeneous disease states. Despite subgroup analysis of disease, it was di cult to distinguish whether one side event was speci cally disease-related owing to the limited number of studies. Second, some studies presented their data in the form of abstract prior to formal publication which may impose an unknown effect on the interpretation of the outcomes. These data are di cult for us to obtain because many ongoing trials are in the middle stage and the performers would not like to release these data. Third, serval side events are merely comprehensive conceptions rather than speci c clinic symptoms and we contend that it is important to record these obscure descriptions (e.g., metabolism and nutrition disorders and gastrointestinal dysfunction). Fourth, we are not informed of whether cell dose is closely associated with these side events as a result of the lack of dose-dependent trials. If possible, a Bayesian network meta-analysis should be conducted to explore the puzzlement. Finally, tumorigenesis, which theoretically exists in MSC therapy, is rarely reported by researchers. And this interesting point should raise our attention.

Conclusions
We summarized all side events potentially related to the application of MSC and no serious safety signals other than transient fever, administration site conditions, sleeplessness, and constipation were discovered. Many population characteristics, including age, analysis, cell type, disease, gender, location, study phrase, year, and administration method possibly impacted the occurrence of one side event. The safety of MSC administration should be under sustained observation despite the innovative therapy appears safe. Ethics approval and consent to participate Not applicable.

Consent for publication
The publication of this manuscript is approved by all authors.

Competing interests
No con icts of interests are declared.

Principle
Inclusion criteria Exclusion criteria Population Any populations including the healthy people and the diseased people NA Intervention Using MSC as treatment, regardless of the administration methods (e.g. local implannation and injection) and sources of MSC (e.g. from the adipose, bomarrow marrow and gum ) Using NSC, ESC, olfactory neuron, schwann cell, h-IPS and stem cell from body uids (e.g. saliva, urine, serum and tears) etc but MSC as interventions

Comparison
Placeebo treatment, non-treatment or basic treatment both utilized in the control and the intervention groups Merely using traditional treatment (surgery and drug) in the control group but not in the intervention group Outcome 1) Any side events associated with MSC treatment; 2) one side event reported by more than one study; 3) regardless of the e cacy of MSC therapy for any diseases No side events reported Study 1) RCT; 2) prospective controlled study 1) Case report (series); 2) single arm study; 3) retrospective controlled study; 4) cross controlled study; 5) study protocol ESC, embryonic stem cell; MSC, mesenchymal stromal cells ; NSC, neuronal stem cell; h-IPS, human induced pluripotent stem cell; NA, not available.   Bar plot for events of all articles. This gure depicted the signi cance of major events (A) and minor events (B) of all included articles. The OR value of each pooled event is presented as mean and 95% con dential intervals. The signi cance of each event is marked by different colors. The more the color approaches the bottom of the p value bar, the occurrence of the event is signi cant. Scarcely reported event (reported by single one article) was not collected and considered as minor event. Bar plot for events of high-quality articles. This gure depicted the signi cance of major events (A) and minor events (B) of high-quality articles. The OR value of each pooled event is presented as mean and 95% con dential intervals. The signi cance of each event is marked by different colors. The more the color approaches the bottom of the p value bar, the occurrence of the event is signi cant. Scarcely reported event (reported by single one article) was not collected and considered as minor events.