Long-Term Effects of Autologous Bone Marrow Stem Cell Treatment in Acute Myocardial Infarction: Factors That May Influence Outcomes

Abstract

Aims

To investigate whether there are important sources of heterogeneity between the findings of different clinical trials which administer autologous stem cell treatment for acute myocardial infarction (AMI) and to evaluate what factors may influence the long-term effects of this treatment.

Methods and Results

MEDLINE (1950-January 2011), EMBASE (1974-January 2011), CENTRAL (The Cochrane Library 2011, Issue 1), CINAHL (1982-January 2011), and ongoing trials registers were searched for randomised trials of bone marrow stem cells as treatment for AMI. Hand-searching was used to screen recent, relevant conference proceedings (2005–2010/11). Meta-analyses were conducted using random-effects models and heterogeneity between subgroups was assessed using chi-squared tests. Planned analyses included length of follow-up, timing of cell infusion and dose, patient selection, small trial size effect, methodological quality, loss of follow-up and date of publication. Thirty-three trials with a total of 1,765 participants were included. There was no evidence of bias due to publication or time-lag, methodological quality of included studies, participant drop-out, duration of follow-up or date of the first disclosure of results. However, in long-term follow-ups the treatment seemed more effective when administered at doses greater than 108 cells and to patients with more severe heart dysfunction.

Conclusions

Evaluation of heterogeneity between trials has not identified significant sources of bias in this study. However, clinical differences between trials are likely to exist which should be considered when undertaking future trials.

About the Authors:

David M. Clifford

Current address: University of Melbourne, Melbourne, Australia

Affiliations Stem Cell Research Laboratory, NHS-Blood and Transplant, John Radcliffe Hospital, Oxford, United Kingdom, Nuffield Department of Clinical Laboratory Sciences, University of Oxford, Oxford, United Kingdom

Sheila A. Fisher

Affiliation: Systematic Review Initiative, Clinical Research Group, NHSBT-Oxford, John Radcliffe Hospital, Oxford, United Kingdom

Susan J. Brunskill

Affiliation: Systematic Review Initiative, Clinical Research Group, NHSBT-Oxford, John Radcliffe Hospital, Oxford, United Kingdom

Carolyn Doree

Affiliation: Systematic Review Initiative, Clinical Research Group, NHSBT-Oxford, John Radcliffe Hospital, Oxford, United Kingdom

Anthony Mathur

Affiliations Department of Clinical Pharmacology, William Harvey Research Institute, London, United Kingdom, Barts and the London NIHR Biomedical Research Unit, London, United Kingdom

Mike J. Clarke

Affiliation: All-Ireland Hub For Trials Methodology Research, Queen’s University, Belfast, United Kingdom

Suzanne M. Watt

Affiliations Stem Cell Research Laboratory, NHS-Blood and Transplant, John Radcliffe Hospital, Oxford, United Kingdom, Nuffield Department of Clinical Laboratory Sciences, University of Oxford, Oxford, United Kingdom

Enca Martin-Rendon

* E-mail: [email protected]

Affiliations Stem Cell Research Laboratory, NHS-Blood and Transplant, John Radcliffe Hospital, Oxford, United Kingdom, Nuffield Department of Clinical Laboratory Sciences, University of Oxford, Oxford, United Kingdom

Introduction

Although advanced therapies have improved short-term survival following acute myocardial infarction (AMI), the incidence of heart failure is steadily increasing worldwide [1]. Current treatments do not address the substantial loss of tissue through injury nor cell death incurred during AMI [2]. In the last decade, autologous bone marrow stem cell (BMSC) treatment has aimed to complement thrombolytic therapies and primary angioplasty in the treatment of AMI (for review see [3]). The hypothesis has been that BMSC would improve heart function delaying the progression of the disease. There is now a substantial body of evidence from randomised trials to assess the effects of this treatment, and a recent update of a Cochrane review by several of the authors of this paper has systematically reviewed this evidence [4].

The first clinical trials were designed to test the safety and feasibility of this new treatment, but were not necessarily powered to assess its efficacy and long-term effects on survival free of major associated cardiac events [5], [6], [7], [8], [9]. To date the treatment appears safe and associated with low mortality and morbidity rates (for review see [3], [10]). However, there is controversial evidence that a beneficial effect on global heart function is significant and persist long-term (for review see [5], [11] and references therein). Clinical evidence from randomised trials of intracoronary infusion of BMSC post-AMI have been evaluated previously in several meta-analyses [10], [12], [13]. The major limitations in the field that may contribute to these conflicting results among trials include the small trial sizes and differences in patient selection, participants lost to follow-up, cell isolation protocols, cell dose/type, timing of cell infusion, route of delivery and methodologies used to measure surrogates; as well as variation in data acquisition and data analysis protocols. In addition, new interventions generally raise concerns that early optimism is fuelled by extreme results in early disclosure just to be contradicted by later results [14], [15], [16]. The rationale of the underlying Cochrane systematic review was to evaluate the efficacy of any dose of autologous BMSC administered to patients with a diagnosis of AMI following revascularisation [4]. In our previous study, sub-group analyses were planned to assess the effect of using different methods (e.g. magnetic resonance imaging (MRI), echocardiography, left ventricular angiography, single positron emission computed tomography (SPECT) or radionucleide ventriculography (RNV)) to measure heart function [4]. The aim of the present study was to conduct further risk of bias and sub-group analyses to explore whether the overall estimate of treatment effect size is a reliable guide to its effect and therefore to address some of the limitations in the field. Here, pre-planned analyses included (i) small trial size effect, (ii) trial quality and participants lost to follow-up, (iii) length of follow-up in the trial design, (iv) date of publication bias, (v) timing of cell infusion, (vi) cell dose/type, (vii) route of delivery and (viii) differences in patient selection.

Methods

Eligibility

Inclusion criteria: (i) randomized trials, (ii) participants with a clinical diagnosis of AMI, (iii) within a month of receiving re-vascularisation by percutaneous coronary intervention (PCI) or thrombolytic therapy or both, (iv) any dose of autologous BMSC, (v) any route of administration, (vi) in the comparator arm participants did not receive BMSC and (vii) any co-interventions provided they were equally applied to each trial arm.

Search Strategy

The search strategy is detailed elsewhere [4]. Briefly, databases were searched through to January 31st 2011 for randomised trials in which BMSC were administered as treatment for AMI, including MEDLINE (1950–2011), EMBASE (1974–2011), CENTRAL (The Cochrane Library 2011, Issue 1), CINAHL (1982–2011), PubMed, Lilacs, and the Transfusion Evidence Library. Ongoing trial registers (ClinicalTrials.gov, the ISRCTN Register and the WHO International Clinical Trials Platform Registry) were also searched. Searches were combined with adaptations of the Cochrane highly sensitive RCT search filter in MEDLINE, EMBASE and CINAHL [17]. No restrictions by language, year of publication or publication status were imposed. Proceedings from the American Heart Association (2005–2010) and International Society of Stem Cell Research (2005–2011) conferences and the reference lists of identified studies and relevant review articles were hand searched for additional studies.

Data Extraction

For each eligible trial, the study and patient population characteristics, the nature of the intervention and comparator, and the outcomes assessed were extracted. The quality of the studies was assessed on the bases of generation of random sequence, concealment of treatment allocation, blinding of outcome assessment and adequacy of follow-up [18]. Eligibility screening, data extraction and assessment of methodological quality were undertaken independently by a total of three reviewers, such that at least two reviewers looked at each potentially eligible trial. Where a trial had used several methods for outcome assessment (e.g. echocardiography, MRI, SPECT, RNV or left ventricular angiography), MRI data were preferentially included in our analysis.

Statistical Analysis

Outcome data were analysed quantitatively using RevMan 5 and presented as relative risk (RR) for dichotomous outcomes or weighted mean difference (WMD) for continuous outcomes with 95% confidence interval (CI), two-sided significance tests are reported. Meta-analyses were undertaken using random effects models, due to the high degree of heterogeneity present in these studies [4], [10]. Statistical heterogeneity was examined using the I2 statistic [19] and the chi-squared test.

Sensitivity Analysis

Bias related to study size (such as publication bias [20]) was assessed by Funnel plots with Egger’s test used to assess asymmetry. Sensitivity analyses were undertaken for all relevant included data to assess the influence of (i) the methodological quality of the trials, (ii) the length of follow-up, (iii) participant drop-out and (iv) publication date. These analyses were specified before they were conducted. In the first instance, trials where the generation of random sequence was rated adequate (marked YES) were analysed separately from those where the generation of random sequence was rated as unclear (marked UNCLEAR) or inadequate (marked NO) as a possible explanation for observed statistical heterogeneity. To assess the effect of length of follow-up on results, trials with short term follow-up periods that have been followed up long term were analyzed separately from those with no long-term follow-up. The influence of participants drop-out was determined by analyzing separately trials with less than 20% drop-out, randomised trials with 20 to 50% drop-out and randomised trials with greater than 50% drop-out for major outcomes measured as dichotomous data (e.g. mortality, reinfarction and target vessel revascularisation). Finally, trials were sub-grouped on the basis of their by start date, end date, publication of the main/full article and disclosure or publication of the first results on the primary outcome (LVEF), to assess the possibility of a relationship between publication date and effect size. Other potential reasons for observed heterogeneity were explored via sub-group analysis, with particular emphasis placed on clinical, treatment and outcome measurement differences among the included studies. Sub-groups were stratified by the timing of BMSC transplantation from onset of AMI, dose of BMSC administered, route of administration and baseline LVEF. Differences in effect size between subgroups were assessed using chi-squared tests for heterogeneity between sub-groups as implemented in RevMan 5.

Results

Description of the Included Studies and Summary of Previous Findings

The search strategy followed has been described in detail elsewhere [4]. A total of 2,169 citations were identified in the initial search of which thirty-three were primary references to eligible studies (Figure S1). The characteristics of all included studies are detailed in Tables S1 and S5. The thirty-three included randomised trials represent thirty-nine treatment comparisons where BMSC was compared with control in 1,765 patients. Treatment comparisons were defined following the criteria previously described (for review see also [4]). Clinical outcomes and efficacy of BMSC treatment following AMI are fully described in detail in our previous study [4] and are summarised here and in supplementary material for clarity. BMSC administration within a month of AMI has no significant effect on mortality, morbidity, or adverse events. Cumulative figures are presented as ≤61 months follow-up in supplementary data (Table S2 and reference to studies in Table S5). The statistical heterogeneity in this case was negligible (I2 = 0–11%). However, a statistical power calculation showed that a study with over 5,000 participants in each randomised group would be required in order to achieve 80% power to detect any significant difference in mortality rates given the relatively low incidence of death within the follow-up period for these trials (2.8% in BMSC and 3.6% in Control groups, respectively). BMSC treatment significantly improved left ventricular volumes and ejection fraction in short- and long-term follow-up periods. However, a considerable degree of statistical heterogeneity (I2>75%) was observed in both infarct size and LVEF comparisons. A summary is presented in supplementary data (Table S3 and reference to studies Table S5). As global LVEF and infarct size have been used as surrogates in many of the included trials, this study was design to explore the observed heterogeneity using these two outcome measures.

Exploring Heterogeneity: Risk of Bias

In order to explore the heterogeneity observed above and to dispel concerns raised with novel interventions, risks of bias were assessed according to the criteria described in the Methods.

1. Risk of publication bias and small study effect. The possibility of publication bias and study size was assessed by Funnel plot and the Egger’s test (Figure 1A&B). The analysis showed no significant small study effects (p = 0.726) and, therefore, little evidence of publication bias.

2. Methodological quality of included studies. The quality assessment of the included studies is summarized in supplementary data (Table S4). Sensitivity analysis to estimate the effect of randomization on LVEF was not required as all included studies followed an adequate method of sequence generation during randomization (Table S4) and this could be assessed because of the high quality of reporting of these relatively recent trials.

3. Loss of follow-up bias on mortality and morbidity. In 37 trials, 80% or more (ranging from 80%–100%) of randomized participants were analyzed by their randomized treatment group. One trial did not report loss of follow-up [21]. In the remaining trial, only 63.64% of randomized participants were included in the analysis [22]. Sensitivity analyses excluding this trial from the meta-analysis [22] had a negligible effect on the effects on mortality during short term (RR 0.80, 95% CI 0.40 to 1.61, p = 0.52 compared to RR 0.75, 95% CI 0.39 to 1.46, p = 0.40) and long term (RR 0.63, 95% CI 0.20 to 2.00, p = 0.43 compared to RR 0.59, 95% CI 0.22 to 1.56, p = 0.29) follow-up. Similar results were observed for incidence of re-infarction, restenosis, hospital readmission and target vessel revascularization, suggesting a negligible risk of bias. Here the original analyses with all included studies are presented (Table S2).

4. Length of follow-up bias on LVEF. For this purpose, the 36 trials that reported short-term LVEF data were divided into two groups: 22 trials were followed-up only for <12 months (Figure 2A) whilst the remaining 14 trials were followed up for 12–61 months (Figure 2B). BMSC treatment effect on LVEF in trials followed up for <12 months (WMD 3.56%, 95% CI 1.74 to 5.37, p = 0.0001- Figure 2A) was not statistically different from the treatment effect in trials followed-up for 12–61 months (WMD 2.71%, 95% CI 1.35 to 4.06, p<0.0001- Figure 2B). These results suggest that those trials with long-term follow-up are representative of all included studies (Table S1).

5. Disclosure or publication date bias on LVEF. The influence of study start date or end date and main publication date on the primary outcome (LVEF) was estimated by sorting the included studies according to those dates. Interestingly, no significant effect was observed in any of these comparisons, indicating a negligible risk of bias on treatment effect. However, when trials were grouped by the year when the first results were disclosed, the studies that disclosed their results first (in 2004) [6], [23] showed an average greater effect on LVEF in favour of the treatment than the studies that were designed or reported later (Figure 3A). If we exclude these early trials from the meta-analysis, the overall estimate for the effect on LVEF was reduced (WMD 2.80%, 95% CI 1.83 to 3.77, p<0.001- Figure 3B) compared to the pool of all included studies (WMD 3.26%, 95% CI 2.12 to 4.40, p<0.001- Figure 3A). However, the difference was not substantial and the inclusion of trials with the early, most promising results showed low risk of bias.

[Figure omitted. See PDF.]

Figure 1. Assessment of risk of bias due to publication and study size on LVEF.

(A) Funnel plot and (B) Egger’s test. No significant risk of publication bias or small study effects was observed.

https://doi.org/10.1371/journal.pone.0037373.g001

[Figure omitted. See PDF.]

Figure 2. Forest plot of Weighted Mean Difference [WMD, with 95% CI (confidence interval)] in left ventricular ejection fraction (LVEF) in short-term follow-up.

(A) Twenty-two randomised trials reporting only short-term follow-up and (B) the remaining 14 trials that reported long-term outcome data as well as short-term data. BMSC treatment significantly improved LVEF in trials with short-term follow-up (3.56%, 95% CI 1.74 to 5.37, p<0.0001) as well as in trials with short- and long-term follow-up (2.71%, 95% CI 1.35 to 4.06, p<0.0001).

https://doi.org/10.1371/journal.pone.0037373.g002

[Figure omitted. See PDF.]

Figure 3. Forest plot of Weighted Mean Difference [WMD, with 95% CI (confidence interval)] in left ventricular ejection fraction (LVEF) in short-term follow-up sub-grouped by the year of the first results’ disclosure.

(A) Including early studies reporting data in 2004 (3.26%, 95% CI 2.12 to 4.40, p<0.00001).and (B) excluding studies with early reporting in 2004 (2.80%, 95% CI 1.83 to 3.77, p<0.00001). BMSC treatment significantly improved LVEF in both meta-analyses. WMD had overlapping CI and were not significantly different.

https://doi.org/10.1371/journal.pone.0037373.g003

Exploring Heterogeneity: Sub-group Analysis

Planned sub-group analyses were carried out to assess the impact of (i) the timing of the BMSC transplantation following AMI, (ii) the dose of BMSC administered, (iii) the route of administration and (vi) the baseline LVEF on infarct size and LVEF for long term follow-up (Table 1). Timing of administration of BMSC infusion was sub-grouped into ≤7days and >7days, to reflect the median delay to BMSC infusion from AMI across included studies. Our previous work suggested that doses of BMSC >108 would be required to observe a significant change in LVEF in the treated arm compared with the control arm [10]. Therefore, for the purpose of this study, trials were divided into two groups according to dose: ≤108 BMSC and >108 BMSC. All included trials where the route of administration is detailed in their methods administered BMSC via the infarct related coronary artery. Only one trial compared venus and arterial delivery of BMSC [24]. Therefore, the pre-planned analysis subgrouping the trials by route of delivery was deemed not appropriate in the present study. Finally, further analyses were carried out dividing the trials in two groups to reflect the median value of baseline LVEF in the included trials: ≤40% and >40% baseline LVEF.

[Figure omitted. See PDF.]

Table 1. Long term follow up (≥12–61months) for infarct size and LVEF: sub-group analysis.

https://doi.org/10.1371/journal.pone.0037373.t001

1. Timing of BMSC infusion. Table 1 shows statistically significant changes in both infarct size (WMD = −5.2%, p = 0.007) and LVEF (WMD = 4.8%, p = 0.0003), in favour of BMSC, when the treatment was administered within 7 days post-AMI. At present, there are no data available to assess the long term effect on BMSC on infarct size for treatment administered after 7 days, however a significant difference in LVEF in favour of BMSC was maintained when the treatment was administered later than 7 days (WMD = 5.9%, p = 0.01). No significant differences between subgroups were observed for long-term follow-up LVEF (p = 0.68).

2. BMSC dose. BMSC treatment had a significant effect on infarct size (WMD = −4.3%, p = 0.005) and LVEF (WMD = 4.7%, p = 0.0001) after long-term follow-up when doses >108 BMSC were administered. In contrast, the infarct size and LVEF showed no significant improvement when doses ≤108 BMSC were administered.

3. Baseline LVEF. Furthermore, the long-term reduction in infarct size in favour of BMSC treatment was statistically significant when the treatment was administered to participants with baseline LVEF ≤40% (WMD = −5.1%, p = 0.006) whereas no significant effect was observed in participants with baseline LVEF >40% (WMD = −1.4%, p = 0.23). The difference in effect sizes these between subgroups was marginally significant (p = 0.08). The effect of BMSC on LVEF was also greater and more statistically significant when participants had LVEF ≤40% at baseline (WMD = 5.6%, p<0.0001) compared with participants with baseline LVEF >40% (WMD = 2.4%, p = 0.006), the difference in effect size between these subgroups was clearly significant for this outcome (p = 0.04).

Taken together, these data indicate that the timing of BMSC transplantation following AMI, the dose of BMSC administered and the baseline LVEF are factors that may contribute to the clinical heterogeneity observed among the studies included in this meta-analysis.

Discussion

The present study was designed: (i) to assess potential risks of bias and diversity amongst different randomised trials to address major limitations in the field and (ii) to evaluate what factors may influence the long-term effect of BMSC treatment. This meta-analysis confirms the findings of our previous study [4] that BMSC treatment moderately improves heart function and has as yet not been associated with any significant safety concerns, but does not decrease mortality or morbidity significantly in long-term follow-up (with the caveat that there have been no studies designed to address mortality). Our power calculation estimates that a trial with several thousand participants would be needed to detect significant differences in mortality between treatment and control groups. Recently, the first trial to address mortality associated with BMSC treatment in patients who suffered AMI has been designed and funded by the European Union (EU FP7– BAMI). The BAMI trial will be administering unfractionated bone marrow mononuclear cells to patients who have suffered from AMI, similarly to the majority of trials included in the present study. The results of this trial should allow us to directly answer the question of whether BMSC therapy following AMI can change the prognosis of this disease.

As the results of the present study are not always consistent with those of large randomised trials [7], [8], [25], [26], [27], we have conducted extensive sensitivity analyses to evaluate potential factors that may account for the discrepancies observed. Major limitations in the field have been enumerated earlier in the Introduction. Importantly, the underlying Cochrane systematic review for this study is unique and superior to others previously reported in a number of ways and has allowed us to address some of those issues. Firstly, it is based on a comprehensive search strategy and a protocol approved by the Cochrane Collaboration prior to starting the search. Although the majority of randomised trials were identified through searching the main databases (MEDLINE, EMBASE, CENTRAL and CINAHL), hand-searching identified several crucial references used to assess risk of bias that would have otherwise been missed [7], [21], [26], [28], [29], [30], [31]. Secondly, the robustness of the original systematic review has been tested extensively within this study through a comprehensive sensitivity analysis and risk of bias assessment to explore the presence and impact of heterogeneity. Finally, the large number of trials included here allows particularly powerful assessments of heterogeneity through planned sub-group analysis.

It has been suggested that measurement of surrogates such as LVEF by different methodology may be a risk of bias because of the known limitations of some methods [32]. Analysis of the included studies sub-grouped by the methodology used to measure surrogate outcomes has been addressed elsewhere [4]. It has also been suggested that stopping or reporting early results in randomised trial may affect the perception that the public has of the treatment efficacy in novel interventions [33]. As early results are more limited than the results of a final analysis, concerns have been raised as treatment effects seen early may either not be real or may be overly optimistic. In this study, we have had a unique opportunity to assess whether factors such as short versus long follow-up or early reporting of results may contribute to bias amongst the pooled results of all included trials. The risk of bias due to length of follow-up for LVEF was negligible indicating that those trials with long term follow-up data are representative of all included studies. This conclusion may have implications interpreting early results from trials with long-term follow-up or trials with short term follow-up only. In addition, low risk of bias was observed when studies were grouped by study start date, study end date or publication of main reference to the studies. There were no significant differences of treatment effect on LVEF when two early studies [5], [6] where either included or excluded from the meta-analysis and negligible risk of bias due to quality of included studies or loss of participant’s follow-up. This may also imply that trial design and methodology has not changed drastically amongst trials in the last decade for significant differences to be observed. Although no significant risk of bias was observed in the present study, one cannot exclude the possibility that discrepancies between the different studies may be explained by variability in factors such as cell isolation, data acquisition or data analysis protocols amongst others. Hence the importance of agreeing to standardised protocols in the future. Part of the remit of the BAMI trial (EU-BAMI mentioned above) is to consider the methodologies used to date and produce a standardised technique for bone marrow processing and delivery.

Here, a parallel significant improvement on LVEF and reduction of infarct size was observed. Although caution is advisable in interpreting results from surrogate outcomes, the moderate improvement in LVEF over short- (3.26%) and long-term (3.91%) follow-up is similar to that obtained in previous trials where AMI patients were treated with a combination of thrombolytic therapy and PCI [34], [35]. In the CADILLAC trial, improvement of LVEF correlated with better long-term survival rate [35], [36]. The results of the CADILLAC trial were also consistent with those of the Netherland’s trial, where thrombolytic therapy was administered to patients suffering from AMI [37]. Consequently, the moderate but significant improvement in LVEF in favour of BMSC treatment reported in the present study could be clinically very relevant providing that limitations such as study size could be overcome. Improvement in long-term survival has been suggested by the results of two recent randomised trials [28], [38]. When sub-group analyses were conducted, greater effects on infarct size were observed when BMSC were administered earlier (≤7days). Effects on both infarct size and LVEF were greater when BMSC were administered at doses >108 and to patients with larger infarcts or lower baseline LVEF. This is in agreement with previous published results [10], [39]. Administering cells earlier may reduce infarct size and reduce damage during ventricular remodelling thus preventing or delaying the onset of heart failure. The requirement for a larger dose of BMSC to reduce infarct size and improve LVEF can be explained by the low rates of cell retention in the heart after BMSC infusion [29], [40]. This supports the idea that the treatment may have a paracrine effect [41]. A number of randomised trials are currently addressing the effects of timing of stem cell transplantation [42], [43], cell dose [44], [45], [46], cell delivery [24] and cell composition [27] on global left ventricular function. Although global LVEF has been a primary surrogate measured in the majority of included trials, the results presented here should be carefully considered, as there is still no evidence of clinical efficacy.

In summary, we have addressed some of the limitations present in the field here and elsewhere [4]. However, other limitations such as small study sizes, patient-related factors and variability in protocols still remain. This study shows that risk of bias due to publication, quality of the studies, loss of follow-up, duration of follow-up and date of disclosure of early results is minor among randomised trials that administer BMSC as treatment for AMI. BMSC treatment significantly reduces infarct size and improves LVEF long-term. Factors such as timing of BMSC transplantation, cell dose and baseline LVEF could affect the successful outcome of this treatment. An attempt has now been made with the design of the BAMI trial to standardise the techniques of BMNC isolation and delivery to man and to measure clinically significant end-points such as mortality.

Supporting Information

[Figure omitted. See PDF.]

Figure S1.

PRISMA diagram.

https://doi.org/10.1371/journal.pone.0037373.s001

(DOC)

Table S1.

Characteristics of the included studies.

https://doi.org/10.1371/journal.pone.0037373.s002

(DOC)

Table S2.

Relative risk of dichotomous clinical outcomes.

https://doi.org/10.1371/journal.pone.0037373.s003

(DOC)

Table S3.

Weighted mean differences of continuous outcomes measured.

https://doi.org/10.1371/journal.pone.0037373.s004

(DOC)

Table S4.

Quality assessment of included studies.

https://doi.org/10.1371/journal.pone.0037373.s005

(DOC)

Table S5.

Reference to studies.

https://doi.org/10.1371/journal.pone.0037373.s006

(DOC)

Acknowledgments

We thank Drs. Fengjuan Lu and Huajun Zhang for the translation of papers from Mandarin to English, Sarina Agkatsev for the translation from German to English, Dr. Sally Hopewell for advice with the publication bias tests.

Author Contributions

Conceived and designed the experiments: SJB AM MC EMR. Performed the experiments: DMC SAF CD SJB EMR. Analyzed the data: DMC SAF EMR. Contributed reagents/materials/analysis tools: SMW. Wrote the paper: DMC EMR. Approved final version of the manuscript: DMC SAF CD SJB AM MC SMW EMR.

Citation: Clifford DM, Fisher SA, Brunskill SJ, Doree C, Mathur A, Clarke MJ, et al. (2012) Long-Term Effects of Autologous Bone Marrow Stem Cell Treatment in Acute Myocardial Infarction: Factors That May Influence Outcomes. PLoS ONE7(5): e37373. https://doi.org/10.1371/journal.pone.0037373

References

1. Velagaleti RS, Pencina MJ, Murabito JM, Wang TJ, Parikh NI, et al. (2008) Long-term trends in the incidence of heart failure after myocardial infarction. Circulation 118: 2057–2062.RS VelagaletiMJ PencinaJM MurabitoTJ WangNI Parikh2008Long-term trends in the incidence of heart failure after myocardial infarction.Circulation11820572062

2. McMurray JJ, Pfeffer MA (2005) Heart failure. Lancet 365: 1877–1889.JJ McMurrayMA Pfeffer2005Heart failure.Lancet36518771889

3. Wollert KC, Drexler H (2010) Cell therapy for the treatment of coronary heart disease: a critical appraisal. Nat Rev Cardiol 7: 204–215.KC WollertH. Drexler2010Cell therapy for the treatment of coronary heart disease: a critical appraisal.Nat Rev Cardiol7204215

4. Clifford DM, Fisher SA, Brunskill SJ, Doree C, Mathur A, et al. (2012) Stem cell treatment for acute myocardial infarction. Cochrane Database Syst Rev 2: CD006536.DM CliffordSA FisherSJ BrunskillC. DoreeA. Mathur2012Stem cell treatment for acute myocardial infarction.Cochrane Database Syst Rev2CD006536

5. Wollert KC, Meyer GP, Lotz J, Ringes-Lichtenberg S, Lippolt P, et al. (2004) Intracoronary autologous bone-marrow cell transfer after myocardial infarction: the BOOST randomised controlled clinical trial. Lancet 364: 141–148.KC WollertGP MeyerJ. LotzS. Ringes-LichtenbergP. Lippolt2004Intracoronary autologous bone-marrow cell transfer after myocardial infarction: the BOOST randomised controlled clinical trial.Lancet364141148

6. Chen SL, Fang WW, Ye F, Liu YH, Qian J, et al. (2004) Effect on left ventricular function of intracoronary transplantation of autologous bone marrow mesenchymal stem cell in patients with acute myocardial infarction. Am J Cardiol 94: 92–95.SL ChenWW FangF. YeYH LiuJ. Qian2004Effect on left ventricular function of intracoronary transplantation of autologous bone marrow mesenchymal stem cell in patients with acute myocardial infarction.Am J Cardiol949295

7. Janssens S, Dubois C, Bogaert J, Theunissen K, Deroose C, et al. (2006) Autologous bone marrow-derived stem-cell transfer in patients with ST-segment elevation myocardial infarction: double-blind, randomised controlled trial. Lancet 367: 113–121.S. JanssensC. DuboisJ. BogaertK. TheunissenC. Deroose2006Autologous bone marrow-derived stem-cell transfer in patients with ST-segment elevation myocardial infarction: double-blind, randomised controlled trial.Lancet367113121

8. Lunde K, Solheim S, Aakhus S, Arnesen H, Abdelnoor M, et al. (2006) Intracoronary injection of mononuclear bone marrow cells in acute myocardial infarction. N Engl J Med 355: 1199–1209.K. LundeS. SolheimS. AakhusH. ArnesenM. Abdelnoor2006Intracoronary injection of mononuclear bone marrow cells in acute myocardial infarction.N Engl J Med35511991209

9. Schachinger V, Erbs S, Elsasser A, Haberbosch W, Hambrecht R, et al. (2006) Intracoronary bone marrow-derived progenitor cells in acute myocardial infarction. N Engl J Med 355: 1210–1221.V. SchachingerS. ErbsA. ElsasserW. HaberboschR. Hambrecht2006Intracoronary bone marrow-derived progenitor cells in acute myocardial infarction.N Engl J Med35512101221

10. Martin-Rendon E, Brunskill SJ, Hyde CJ, Stanworth SJ, Mathur A, et al. (2008) Autologous bone marrow stem cells to treat acute myocardial infarction: a systematic review. Eur Heart J 29: 1807–1818.E. Martin-RendonSJ BrunskillCJ HydeSJ StanworthA. Mathur2008Autologous bone marrow stem cells to treat acute myocardial infarction: a systematic review.Eur Heart J2918071818

11. Janssens SP (2011) Cardiac bone marrow cell therapy: the proof of the pudding remains in the eating. Eur Heart J 32: 1697–1700.SP Janssens2011Cardiac bone marrow cell therapy: the proof of the pudding remains in the eating.Eur Heart J3216971700

12. Abdel-Latif A, Bolli R, Tleyjeh IM, Montori VM, Perin EC, et al. (2007) Adult bone marrow-derived cells for cardiac repair: a systematic review and meta-analysis. Arch Intern Med 167: 989–997.A. Abdel-LatifR. BolliIM TleyjehVM MontoriEC Perin2007Adult bone marrow-derived cells for cardiac repair: a systematic review and meta-analysis.Arch Intern Med167989997

13. Lipinski MJ, Biondi-Zoccai GG, Abbate A, Khianey R, Sheiban I, et al. (2007) Impact of intracoronary cell therapy on left ventricular function in the setting of acute myocardial infarction: a collaborative systematic review and meta-analysis of controlled clinical trials. J Am Coll Cardiol 50: 1761–1767.MJ LipinskiGG Biondi-ZoccaiA. AbbateR. KhianeyI. Sheiban2007Impact of intracoronary cell therapy on left ventricular function in the setting of acute myocardial infarction: a collaborative systematic review and meta-analysis of controlled clinical trials.J Am Coll Cardiol5017611767

14. Ioannidis JP, Trikalinos TA (2005) Early extreme contradictory estimates may appear in published research: the Proteus phenomenon in molecular genetics research and randomized trials. J Clin Epidemiol 58: 543–549.JP IoannidisTA Trikalinos2005Early extreme contradictory estimates may appear in published research: the Proteus phenomenon in molecular genetics research and randomized trials.J Clin Epidemiol58543549

15. Hopewell S, Clarke M, Stewart L, Tierney J (2007) Time to publication for results of clinical trials. Cochrane Database Syst Rev. MR000011 p.S. HopewellM. ClarkeL. StewartJ. Tierney2007Time to publication for results of clinical trials.Cochrane Database Syst RevMR000011

16. Trikalinos TA, Churchill R, Ferri M, Leucht S, Tuunainen A, et al. (2004) Effect sizes in cumulative meta-analyses of mental health randomized trials evolved over time. J Clin Epidemiol 57: 1124–1130.TA TrikalinosR. ChurchillM. FerriS. LeuchtA. Tuunainen2004Effect sizes in cumulative meta-analyses of mental health randomized trials evolved over time.J Clin Epidemiol5711241130

17. Robinson KA, Dickersin K (2002) Development of a highly sensitive search strategy for the retrieval of reports of controlled trials using PubMed. Int J Epidemiol 31: 150–153.KA RobinsonK. Dickersin2002Development of a highly sensitive search strategy for the retrieval of reports of controlled trials using PubMed.Int J Epidemiol31150153

18. Higgins J, Green S (2008) Cochrane Handbook for Systematic Reviews of Interventions 5.0.0; Higgins J, Green S, editors. Chichester, UK: John Wiley & Sons Ltd.. J. HigginsS. Green2008Cochrane Handbook for Systematic Reviews of Interventions 5.0.0; Higgins J, Green S, editors.Chichester, UKJohn Wiley & Sons Ltd.

19. Higgins JP, Thompson SG (2002) Quantifying heterogeneity in a meta-analysis. Stat Med 21: 1539–1558.JP HigginsSG Thompson2002Quantifying heterogeneity in a meta-analysis.Stat Med2115391558

20. Hopewell S, Loudon K, Clarke MJ, Oxman AD, Dickersin K (2009) Publication bias in clinical trials due to statistical significance or direction of trial results. Cochrane Database Syst Rev. MR000006 p.S. HopewellK. LoudonMJ ClarkeAD OxmanK. Dickersin2009Publication bias in clinical trials due to statistical significance or direction of trial results.Cochrane Database Syst RevMR000006

21. Fernandez-Pereira C, Vigo C, Bataglia S, De la Hoz R, Vetulli H, et al. (2006) Autologous bone marrow stem cell transplant after myocardial infarction. In-hospital and long-term follow-up results of the randomised argentina trial (STAR AMI). Eur Heart J August. 279 p.C. Fernandez-PereiraC. VigoS. BatagliaR. De la HozH. Vetulli2006Autologous bone marrow stem cell transplant after myocardial infarction.In-hospital and long-term follow-up results of the randomised argentina trial (STAR AMI). Eur Heart J August279

22. Zhukova NS, Staroverov , II , Stukalova OV, Samoleinko LE, Romanov Iu A, et al. (2009) An experience of the use of stem cells in the treatment of patients with myocardial infarction and low ejection fraction. Kardiologiia 49: 19–24.NS ZhukovaStaroverovIIOV StukalovaLE SamoleinkoA. Romanov Iu2009An experience of the use of stem cells in the treatment of patients with myocardial infarction and low ejection fraction.Kardiologiia491924

23. Meyer GP, Wollert KC, Lotz J, Steffens J, Lippolt P, et al. (2006) Intracoronary bone marrow cell transfer after myocardial infarction: eighteen months’ follow-up data from the randomized, controlled BOOST (BOne marrOw transfer to enhance ST-elevation infarct regeneration) trial. Circulation 113: 1287–1294.GP MeyerKC WollertJ. LotzJ. SteffensP. Lippolt2006Intracoronary bone marrow cell transfer after myocardial infarction: eighteen months’ follow-up data from the randomized, controlled BOOST (BOne marrOw transfer to enhance ST-elevation infarct regeneration) trial.Circulation11312871294

24. Nogueira FB, Silva SA, Haddad AF, Peixoto CM, Carvalho RM, et al. (2009) Systolic function of patients with myocardial infarction undergoing autologous bone marrow transplantation. Arq Bras Cardiol 93: 374–379, 367–372: FB NogueiraSA SilvaAF HaddadCM PeixotoRM Carvalho2009Systolic function of patients with myocardial infarction undergoing autologous bone marrow transplantation.Arq Bras Cardiol 93: 374–379,367–372

25. Hirsch A, Nijveldt R, van der Vleuten PA, Tijssen JG, van der Giessen WJ, et al. (2011) Intracoronary infusion of mononuclear cells from bone marrow or peripheral blood compared with standard therapy in patients after acute myocardial infarction treated by primary percutaneous coronary intervention: results of the randomized controlled HEBE trial. Eur Heart J 32: 1736–1747.A. HirschR. NijveldtPA van der VleutenJG TijssenWJ van der Giessen2011Intracoronary infusion of mononuclear cells from bone marrow or peripheral blood compared with standard therapy in patients after acute myocardial infarction treated by primary percutaneous coronary intervention: results of the randomized controlled HEBE trial.Eur Heart J3217361747

26. Roncalli J, Mouquet F, Piot C, Trochu JN, Le Corvoisier P, et al. (2011) Intracoronary autologous mononucleated bone marrow cell infusion for acute myocardial infarction: results of the randomized multicenter BONAMI trial. Eur Heart J 32: 1748–1757.J. RoncalliF. MouquetC. PiotJN TrochuP. Le Corvoisier2011Intracoronary autologous mononucleated bone marrow cell infusion for acute myocardial infarction: results of the randomized multicenter BONAMI trial.Eur Heart J3217481757

27. Tendera M, Wojakowski W, Ruzyllo W, Chojnowska L, Kepka C, et al. (2009) Intracoronary infusion of bone marrow-derived selected CD34+CXCR4+ cells and non-selected mononuclear cells in patients with acute STEMI and reduced left ventricular ejection fraction: results of randomized, multicentre Myocardial Regeneration by Intracoronary Infusion of Selected Population of Stem Cells in Acute Myocardial Infarction (REGENT) Trial. Eur Heart J 30: 1313–1321.M. TenderaW. WojakowskiW. RuzylloL. ChojnowskaC. Kepka2009Intracoronary infusion of bone marrow-derived selected CD34+CXCR4+ cells and non-selected mononuclear cells in patients with acute STEMI and reduced left ventricular ejection fraction: results of randomized, multicentre Myocardial Regeneration by Intracoronary Infusion of Selected Population of Stem Cells in Acute Myocardial Infarction (REGENT) Trial.Eur Heart J3013131321

28. Grajek S, Popiel M, Gil L, Breborowicz P, Lesiak M, et al. (2010) Influence of bone marrow stem cells on left ventricle perfusion and ejection fraction in patients with acute myocardial infarction of anterior wall: randomized clinical trial: Impact of bone marrow stem cell intracoronary infusion on improvement of microcirculation. Eur Heart J 31: 691–702.S. GrajekM. PopielL. GilP. BreborowiczM. Lesiak2010Influence of bone marrow stem cells on left ventricle perfusion and ejection fraction in patients with acute myocardial infarction of anterior wall: randomized clinical trial: Impact of bone marrow stem cell intracoronary infusion on improvement of microcirculation.Eur Heart J31691702

29. Karpov RS, Popov SV, Markov VA, Suslova TE, Ryabov VV, et al. (2005) Autologous mononuclear bone marrow cells during reparative regeneratrion after acute myocardial infarction. Bull Exp Biol Med 140: 640–643.RS KarpovSV PopovVA MarkovTE SuslovaVV Ryabov2005Autologous mononuclear bone marrow cells during reparative regeneratrion after acute myocardial infarction.Bull Exp Biol Med140640643

30. Piepoli MF, Vallisa D, Arbasi M, Cavanna L, Cerri L, et al. (2010) Bone marrow cell transplantation improves cardiac, autonomic, and functional indexes in acute anterior myocardial infarction patients (Cardiac Study). Eur J Heart Fail 12: 172–180.MF PiepoliD. VallisaM. ArbasiL. CavannaL. Cerri2010Bone marrow cell transplantation improves cardiac, autonomic, and functional indexes in acute anterior myocardial infarction patients (Cardiac Study).Eur J Heart Fail12172180

31. Li ZQ, Zhang M, Jing YZ, Zhang WW, Liu Y, et al. (2007) The clinical study of autologous peripheral blood stem cell transplantation by intracoronary infusion in patients with acute myocardial infarction (AMI). Int J Cardiol 115: 52–56.ZQ LiM. ZhangYZ JingWW ZhangY. Liu2007The clinical study of autologous peripheral blood stem cell transplantation by intracoronary infusion in patients with acute myocardial infarction (AMI).Int J Cardiol1155256

32. Arnesen H, Lunde K, Aakhus S, Forfang K (2007) Cell therapy in myocardial infarction. Lancet 369: 2142–2143.H. ArnesenK. LundeS. AakhusK. Forfang2007Cell therapy in myocardial infarction.Lancet36921422143

33. Kumar A, Galeb S, Djulbegovic B Treatment of patients with multiple myeloma: an overview of systematic reviews. Acta Haematol 125: 8–22.Kumar A, Galeb S, Djulbegovic B Treatment of patients with multiple myeloma: an overview of systematic reviews.Acta Haematol125822

34. Montalescot G, Barragan P, Wittenberg O, Ecollan P, Elhadad S, et al. (2001) Platelet glycoprotein IIb/IIIa inhibition with coronary stenting for acute myocardial infarction. N Engl J Med 344: 1895–1903.G. MontalescotP. BarraganO. WittenbergP. EcollanS. Elhadad2001Platelet glycoprotein IIb/IIIa inhibition with coronary stenting for acute myocardial infarction.N Engl J Med34418951903

35. Stone GW, Grines CL, Cox DA, Garcia E, Tcheng JE, et al. (2002) Comparison of angioplasty with stenting, with or without abciximab, in acute myocardial infarction. N Engl J Med 346: 957–966.GW StoneCL GrinesDA CoxE. GarciaJE Tcheng2002Comparison of angioplasty with stenting, with or without abciximab, in acute myocardial infarction.N Engl J Med346957966

36. Halkin A, Singh M, Nikolsky E, Grines CL, Tcheng JE, et al. (2005) Prediction of mortality after primary percutaneous coronary intervention for acute myocardial infarction: the CADILLAC risk score. J Am Coll Cardiol 45: 1397–1405.A. HalkinM. SinghE. NikolskyCL GrinesJE Tcheng2005Prediction of mortality after primary percutaneous coronary intervention for acute myocardial infarction: the CADILLAC risk score.J Am Coll Cardiol4513971405

37. Simoons ML, Vos J, Tijssen JG, Vermeer F, Verheugt FW, et al. (1989) Long-term benefit of early thrombolytic therapy in patients with acute myocardial infarction: 5 year follow-up of a trial conducted by the Interuniversity Cardiology Institute of The Netherlands. J Am Coll Cardiol 14: 1609–1615.ML SimoonsJ. VosJG TijssenF. VermeerFW Verheugt1989Long-term benefit of early thrombolytic therapy in patients with acute myocardial infarction: 5 year follow-up of a trial conducted by the Interuniversity Cardiology Institute of The Netherlands.J Am Coll Cardiol1416091615

38. Assmus B, Rolf A, Erbs S, Elsasser A, Haberbosch W, et al. (2010) Clinical outcome 2 years after intracoronary administration of bone marrow-derived progenitor cells in acute myocardial infarction. Circ Heart Fail 3: 89–96.B. AssmusA. RolfS. ErbsA. ElsasserW. Haberbosch2010Clinical outcome 2 years after intracoronary administration of bone marrow-derived progenitor cells in acute myocardial infarction.Circ Heart Fail38996

39. Brunskill SJ, Hyde CJ, Doree CJ, Watt SM, Martin-Rendon E (2009) Route of delivery and baseline left ventricular ejection fraction, key factors of bone-marrow-derived cell therapy for ischaemic heart disease. Eur J Heart Fail 11: 887–896.SJ BrunskillCJ HydeCJ DoreeSM WattE. Martin-Rendon2009Route of delivery and baseline left ventricular ejection fraction, key factors of bone-marrow-derived cell therapy for ischaemic heart disease.Eur J Heart Fail11887896

40. Schachinger V, Aicher A, Dobert N, Rover R, Diener J, et al. (2008) Pilot trial on determinants of progenitor cell recruitment to the infarcted human myocardium. Circulation 118: 1425–1432.V. SchachingerA. AicherN. DobertR. RoverJ. Diener2008Pilot trial on determinants of progenitor cell recruitment to the infarcted human myocardium.Circulation11814251432

41. Gnecchi M, Zhang Z, Ni A, Dzau VJ (2008) Paracrine mechanisms in adult stem cell signaling and therapy. Circ Res 103: 1204–1219.M. GnecchiZ. ZhangA. NiVJ Dzau2008Paracrine mechanisms in adult stem cell signaling and therapy.Circ Res10312041219

42. Traverse JH, Henry TD, Vaughan DE, Ellis SG, Pepine CJ, et al. (2009) Rationale and design for TIME: A phase II, randomized, double-blind, placebo-controlled pilot trial evaluating the safety and effect of timing of administration of bone marrow mononuclear cells after acute myocardial infarction. Am Heart J 158: 356–363.JH TraverseTD HenryDE VaughanSG EllisCJ Pepine2009Rationale and design for TIME: A phase II, randomized, double-blind, placebo-controlled pilot trial evaluating the safety and effect of timing of administration of bone marrow mononuclear cells after acute myocardial infarction.Am Heart J158356363

43. Traverse JH, Henry TD, Vaughan DE, Ellis SG, Pepine CJ, et al. (2010) LateTIME: a phase-II, randomized, double-blinded, placebo-controlled, pilot trial evaluating the safety and effect of administration of bone marrow mononuclear cells 2 to 3 weeks after acute myocardial infarction. Tex Heart Inst J 37: 412–420.JH TraverseTD HenryDE VaughanSG EllisCJ Pepine2010LateTIME: a phase-II, randomized, double-blinded, placebo-controlled, pilot trial evaluating the safety and effect of administration of bone marrow mononuclear cells 2 to 3 weeks after acute myocardial infarction.Tex Heart Inst J37412420

44. Meluzin J, Janousek S, Mayer J, Groch L, Hornacek I, et al. (2008) Three-, 6-, and 12-month results of autologous transplantation of mononuclear bone marrow cells in patients with acute myocardial infarction. Int J Cardiol 128: 185–192.J. MeluzinS. JanousekJ. MayerL. GrochI. Hornacek2008Three-, 6-, and 12-month results of autologous transplantation of mononuclear bone marrow cells in patients with acute myocardial infarction.Int J Cardiol128185192

45. Yao K, Huang R, Sun A, Qian J, Liu X, et al. (2009) Repeated autologous bone marrow mononuclear cell therapy in patients with large myocardial infarction. Eur J Heart Fail 11: 691–698.K. YaoR. HuangA. SunJ. QianX. Liu2009Repeated autologous bone marrow mononuclear cell therapy in patients with large myocardial infarction.Eur J Heart Fail11691698

46. Quyyumi AA, Waller EK, Murrow J, Esteves F, Galt J, et al. (2011) CD34(+) cell infusion after ST elevation myocardial infarction is associated with improved perfusion and is dose dependent. Am Heart J 161: 98–105.AA QuyyumiEK WallerJ. MurrowF. EstevesJ. Galt2011CD34(+) cell infusion after ST elevation myocardial infarction is associated with improved perfusion and is dose dependent.Am Heart J16198105

47. Cao F, Sun D, Li C, Narsinh K, Zhao L, et al. (2009) Long-term myocardial functional improvement after autologous bone marrow mononuclear cells transplantation in patients with ST-segment elevation myocardial infarction: 4 years follow-up. Eur Heart J 30: 1986–1994.F. CaoD. SunC. LiK. NarsinhL. Zhao2009Long-term myocardial functional improvement after autologous bone marrow mononuclear cells transplantation in patients with ST-segment elevation myocardial infarction: 4 years follow-up.Eur Heart J3019861994

48. Jin B, Yang Y, Shi H, Luo X, Li Y, et al. (2008) Autologous intracoronary mononuclear bone marrow cell transplantation for acute anterior myocardial infarction: Outcomes after 12-month follow-up. Journal of Clinical Rehabilitative Tissue Engineering Research 12: 2267–2271.B. JinY. YangH. ShiX. LuoY. Li2008Autologous intracoronary mononuclear bone marrow cell transplantation for acute anterior myocardial infarction: Outcomes after 12-month follow-up.Journal of Clinical Rehabilitative Tissue Engineering Research1222672271

49. Suarez de Lezo J, Herrera CP, M, Romero M, Pavlovic D, Segura J, et al. (2007) Tratamiento regenerativo en pacientes con infarto agudo anterior revascularizado y funcion ventricular deprimida. Rev Esp Cardiol 60: 357–365.J. Suarez de LezoCP, M. HerreraM. RomeroD. PavlovicJ. Segura2007Tratamiento regenerativo en pacientes con infarto agudo anterior revascularizado y funcion ventricular deprimida.Rev Esp Cardiol60357365

50. Penicka M, Horak J, Kobylka P, Pytlik R, Kozak T, et al. (2007) Intracoronary injection of autologous bone marrow-derived mononuclear cells in patients with large anterior acute myocardial infarction: a prematurely terminated randomized study. J Am Coll Cardiol 49: 2373–2374.M. PenickaJ. HorakP. KobylkaR. PytlikT. Kozak2007Intracoronary injection of autologous bone marrow-derived mononuclear cells in patients with large anterior acute myocardial infarction: a prematurely terminated randomized study.J Am Coll Cardiol4923732374