RESULTSFourteen studies were identi

RESULTS
Fourteen studies were identified that compared intramyocardial
injection of BMSCs in addition to standard revascularization
with standard surgical revascularization
procedures alone in patients with chronic ischemic heart
disease (Figure 1). After more detailed evaluation, 8 studies
were excluded from the meta-analysis: Seven studies used catheter-based endocardial or intracoronary stem cell application
methods, and no CABG was performed in one study.
Because this restrictive selection process was aimed at
emerging as far as possible congruent, and consequently
comparable available studies, 6 studies (4 RCTs9,11,15,16
and 2 cohort studies8,10) with a total of 179 patients were
included. Four of the 6 studies had a Jadad score of 4, and
2 studies had a Jadad score of 3 (Table 1). The interreviewer
agreement on study eligibility was 100%.
Study Characteristics
The enrolled patients’ average age ranged from 48.6 to
64.9 years in the trials, and each trial enrolled mostly
men. Notably, the sample size in each study ranged from
20 to 40 patients (median, 32), and the follow-up duration
was 3 to 6 months (median, 6 months). The proportion of
followed-up patients was more than 98.5%. There was
considerable heterogeneity in the timing of CABG with
combined cell transplantation after myocardial infarction,
ranging from 7 to 84 weeks. In the selected studies,
LVEF and LVEDV were evaluated with echocardiography
8-11,16 or cardiac magnetic resonance imaging.15 Assessment
of baseline LVEF was performed preoperatively
after enrollment for the respective study. There was no significant
difference between the control and the treatment
groups regarding the number of bypasses. One study did
not report the number of grafts performed.8 All studies
targeted the infarct border zone for stem cell injection;
cells were also injected into the infarct area in only 1
study.10
Meta-Analysis and Efficacy
Homogeneity test revealed rejection of null hypothesis of
homogeneity for LVEFchange (Cochran’s c2 test, P

Hasil (Bahasa Indonesia) 1:

RESULTSFourteen studies were identified that compared intramyocardialinjection of BMSCs in addition to standard revascularizationwith standard surgical revascularizationprocedures alone in patients with chronic ischemic heartdisease (Figure 1). After more detailed evaluation, 8 studieswere excluded from the meta-analysis: Seven studies used catheter-based endocardial or intracoronary stem cell applicationmethods, and no CABG was performed in one study.Because this restrictive selection process was aimed atemerging as far as possible congruent, and consequentlycomparable available studies, 6 studies (4 RCTs9,11,15,16and 2 cohort studies8,10) with a total of 179 patients wereincluded. Four of the 6 studies had a Jadad score of 4, and2 studies had a Jadad score of 3 (Table 1). The interrevieweragreement on study eligibility was 100%.Study CharacteristicsThe enrolled patients’ average age ranged from 48.6 to64.9 years in the trials, and each trial enrolled mostlymen. Notably, the sample size in each study ranged from20 to 40 patients (median, 32), and the follow-up durationwas 3 to 6 months (median, 6 months). The proportion offollowed-up patients was more than 98.5%. There wasconsiderable heterogeneity in the timing of CABG withcombined cell transplantation after myocardial infarction,ranging from 7 to 84 weeks. In the selected studies,LVEF and LVEDV were evaluated with echocardiography8-11,16 or cardiac magnetic resonance imaging.15 Assessmentof baseline LVEF was performed preoperativelyafter enrollment for the respective study. There was no significantdifference between the control and the treatmentgroups regarding the number of bypasses. One study didnot report the number of grafts performed.8 All studiestargeted the infarct border zone for stem cell injection;cells were also injected into the infarct area in only 1study.10Meta-Analysis and EfficacyHomogeneity test revealed rejection of null hypothesis ofhomogeneity for LVEFchange (Cochran’s c2 test, P<.001,Figure 2) and a nonsignificant result for LVEDVchange(Cochran’s c2 test, P ¼ .485, Figure 3). Therefore,a random-effects model for LVEFchange and a fixed-effectsmodel for LVEDVchange were applied.Our meta-analysis resulted in a significant differenceregarding the overall change of the LVEF (LVEFchange)from baseline to follow-up between BMSC therapy andcontrol, favoring the BMSC therapy group (5.40%; 95%CI, 1.36–9.44; P ¼ .009; Figure 2).Because of evidence for heterogeneity, a meta-regressionanalysis was realized for LVEFchange. The regression coefficientsare the estimated change in the LVEFchange perunit change in the covariates. LVEFchange is estimated to increaseby 0.64 per year of age and to decrease by 0.50 perunit increase of baseline LVEF. Although the associationbetween LVEFchange and age is of strong statistical evidence(z-test for regression coefficient, P ¼ .021), the associationbetween LVEFchange and baseline LVEF could not be detectedsignificantly (P ¼ .248).Moreover,BMSC therapy showed a trend toward a reductionin LVEDV compared with the control group, with theoverall change of LVEDV (LVEDVchange) from baselineto follow-up favoring the BMSC therapy group (9.55 mL;95% CI,2.82 to 21.92; P ¼ .13; Figure 3).To exclude potential publication bias, Begg’s rank correlationtest and Egger’s weighted regression test for publicationbias were performed. No publication bias was evidentfor the 6 studies included in LVEF meta-analysis (Begg’srank correlation P ¼ .573 for mean difference LVEFchange;Egger’s weighted regression P ¼ .126 for mean differenceLVEFchange) or the 3 studies included in the LVEDVmeta-analysis (Begg’s rank correlation P ¼ .117 for meandifference LVEDVchange; Egger’s weighted regressionP ¼ .415 for mean difference LVEDVchange).Major Adverse Cardiovascular EventsMACE rates in all 6 studies are described in Table 4.Furthermore, we analyzed the risk ratios of MACE betweentheBMSC treatment and the control groups (Figures 4 and 5).We found that ventricular arrhythmias and the composite ofother cardiovascular events were not significantly differentbetween BMSC therapy and controls (RRVA: 0.951; 95%CI, 0.389–2.325; P ¼ .913; RRCE ¼ 1.134; 95% CI,0.28–4.6; P ¼ .86), supposing homogeneity (c2VA ¼ 1.122,df ¼ 5, P ¼ .952; c2CE ¼ 1.504, df ¼ 5, P ¼ .913).DISCUSSIONThis systematic review and meta-analysis, the first, to ourknowledge, summarizing the available evidence of intramyocardialBMSC transplantation during CABG in patientswith chronic ischemic heart disease, indicates that BMSCtransplantation in addition to CABG is safe and leads tobenefits compared with those achieved by CABG alone.Our results suggest a potential improvement of heart functionafter intramyocardial BMSC transplantation, indicated by improved LVEF. A trend toward reduced LVEDV in theBMSC transplantation group suggests a decrease of cardiacremodeling, however, without reaching statistical significance.In terms of the functional outcome after BMSCtransplantation in combination with CABG, our metaanalysisof available studies shows an improvement ofLVEF that is significantly higher in the stem cell-treatedgroups compared with groups treated with CABG only.Meta-regression analysis performed indicates that beneficialfunctional effects of stem cell therapy in addition toCABG surgery are positively correlated with the amountof preoperative LVEF depression. This underlines the factthat surgical stem cell therapy should primarily focus on patientswith indication for CABG surgery and reduced leftventricular function due to chronic ischemia. Notably, theage of patients was found to significantly correlate withfunctional effects of cell therapy. Because intrinsic myocardialregeneration takes places, but is reduced during a normallife span,17 it is conceivable that, in addition torevascularization procedures, elderly patients especiallymight profit from cell therapy aimed at stimulating regenerativeprocesses.The improvement of LVEF (mean difference LVEFchangeof 5.40%) (Figure 2) in our meta-analysis tends to be higherthan in a recently published meta-analysis of intracoronarybone marrow transplantation in the setting of acute myocardialinfarction.12 This difference in functional improvementmight be due to either the different time point of BMSCtransplantation (acute myocardial infarction vs chronic ischemicheart disease due to old infarction) or the differingroute of cell delivery (intracoronary vs intramyocardialcell application). Early injection after infarction could bebeneficial to prevent a large fibrotic scare. On the otherhand, because myocardial infarction leads to severe impairmentof heart function associated with rhythmic instabilityand poorer tolerance of additional treatment, it might bereasonable to wait for the acute phase to pass until theinfarction zone is consolidated. Furthermore, cell transplantationshould be more effective after the postischemicinflammatory reaction has subsided.6,18 Stem celltransplantation within the ‘‘hot’’ phase of post-infarction inflammationmight lead them to take part in the inflammation
cascade rather than in the formation of vessels.19 In regard
to the route of stem cell delivery, the amount of stem cell recruitment
after intravascular cell application is an obvious
problem, because preclinical reports describe the myocardial
persistence of mononuclear cells from bone marrow
after intravascular application as ‘‘poor.’’20 Local intramyocardial
surgical injection might overcome the problem. In
contrast with intracoronary application, stem cells can be
delivered directly into the target area of the myocardium
without depending on sufficient cell migration across the
endothelial barrier, most likely resulting in a decrease of immediate
washout and remote organ engraftment of injected
cells.21 However, except for one study,15 all included trials
lack placebo injections in the control groups and none of the
studies were performed in a double-blinded way. In addition
to these methodological advisements, there are further limitations
that have to be considered when interpreting the results
of our meta-analysis. Because we aimed for evaluating
specific surgical stem cell application in combination with
CABG, the number of included studies was rather small,
gaining an enhanced comparability of studies in return
and thus an increased power compared with individual studies.
Nevertheless, some different study aspects remained.
The methods for evaluating LVEF and LVEDV, the interval
between myocardial infarction and CABG, the baseline
LVEF, the amount, and the specific type of BMSC used varied
among the included studies (Table 1). All of these form
potential sources of heterogeneity. The differing baseline
LVEF (LVEFbaseline Table 3) from 29.2%  2.6%8 to
48.0%  8.0%10 especially limits the interpretation of
LVEF data. Future multicenter trials can overcome this
problem only by setting up strict inclusion criteria regarding
preoperative LVEF. Moreover, 3 of the included 6 studies
did not use a specified population of BMSCs but injected
a semi-enriched population of bone marrow mononuclear
cells. In addition to hematopoietic and mesenchymal stem
cells, this population contains leucocytes, making it difficult
to attribute functional effects to a certain cell type. Furthermore,
because half of the included articles lack
complete dimensional left ventricular data, LVEDV change was analyzed as qualified in 3 comparable studies.