OUP user menu

Intracoronary bone marrow cell transfer after myocardial infarction: 5-year follow-up from the randomized-controlled BOOST trial

Gerd P. Meyer, Kai C. Wollert, Joachim Lotz, Jens Pirr, Ulrike Rager, Peter Lippolt, Andreas Hahn, Stephanie Fichtner, Arnd Schaefer, Lubomir Arseniev, Arnold Ganser, Helmut Drexler
DOI: http://dx.doi.org/10.1093/eurheartj/ehp374 2978-2984 First published online: 22 September 2009


Aims We assessed whether a single intracoronary infusion of autologous bone marrow cells (BMCs) can have a sustained impact on left ventricular ejection fraction (LVEF) in patients after ST-elevation myocardial infarction (STEMI). In the BOne marrOw transfer to enhance ST-elevation infarct regeneration (BOOST) trial, 60 patients with STEMI and successful percutaneous coronary intervention were randomized to a control and a cell therapy group. As previously reported, BMC transfer led to an improvement of LVEF by 6.0% at 6 months (P = 0.003) and 2.8% at 18 months (P = 0.27).

Methods and results Left ventricular ejection fraction and clinical status were re-assessed in all surviving patients after 61 ± 11 months. Major adverse cardiac events occurred with similar frequency in both groups. When compared with baseline, LVEF assessed by magnetic resonance imaging at 61 months decreased by 3.3 ± 9.5% in the control group and by 2.5 ± 11.9% in the BMC group (P = 0.30). Patients with an infarct transmurality > median appeared to benefit from BMC transfer throughout the 61-month study period (P = 0.040).

Conclusion A single intracoronary application of BMCs does not promote a sustained improvement of LVEF in STEMI patients with relatively preserved systolic function. It is conceivable that a subgroup of patients with more transmural infarcts may derive a sustained benefit from BMC therapy. However, this needs to be tested prospectively in a randomized trial.

  • Acute myocardial infarction
  • Cell therapy
  • Magnetic resonance imaging


Based on experimental studies suggesting that cardiac transfer of unfractionated bone marrow cells (BMCs) or bone marrow-derived stem and progenitor cells can promote functional improvements after acute myocardial infarction (AMI),13 several randomized clinical trials have explored the hypothesis that an intracoronary infusion of autologous BMCs may enhance the recovery of left ventricular (LV) systolic function in patients after AMI.47 The combined experience from these studies suggests that intracoronary delivery of unselected BMCs a few days after myocardial reperfusion is feasible and, probably, safe. The outcome of these trials has been mixed, however, with some studies, including the initial BOne marrOw transfer to enhance ST-elevation infarct regeneration (BOOST) trial, showing significant improvements in global and regional LV systolic function,4,5 one trial showing improvements in regional function only,6 and one trial reporting no significant improvements at all.7 Although the reasons for these heterogeneous results are difficult to resolve, it has been hypothesized that differences between these studies in cell preparation methods and the timing of cell transfer may have had an influence on the findings.58 Recent meta-analyses of published randomized and non-randomized studies, involving a total of ∼1000 patients, support the notion that BMC transfer may contribute to modest improvements in cardiac function after AMI.911

All randomized trials have covered only relatively short follow-up times, ranging from 4 to 18 months.47,1214 The critical question whether a single intracoronary infusion of BMCs can have a sustained impact on LV function after AMI has never been addressed. Moreover, patient subgroups that may benefit most from BMC transfer have not been defined. In that regard, the 18-month follow-up data from BOOST trial have suggested that patients with greater infarct transmurality may have the greatest potential for functional improvement.12

In the present investigation, we have re-assessed the clinical status and LV systolic function 61 months after intracoronary BMC transfer in the BOOST study population. Our data provide the longest available follow-up on the safety and efficacy of intracoronary BMC transfer after AMI.


Study design

The study protocol and the 6- and 18-month follow-up results of the BOOST trial have been published.4,12 In brief, patients with a first ST-elevation myocardial infarction (STEMI), who had undergone percutaneous coronary intervention (PCI) with stent implantation of the infarct-related artery and who demonstrated hypokinesia or akinesia involving more than two-thirds of the LV anteroseptal, lateral, and/or inferior wall, as shown by LV angiography immediately after PCI, were eligible for the trial. Patients were randomized 1:1 to the control (n = 30) and BMC transfer groups (n = 30). Baseline magnetic resonance imaging (MRI) was performed 3.5 ± 1.5 days after PCI. In the BMC transfer group, BMCs were harvested, processed by gelatine-polysuccinate density gradient sedimentation, and infused into the infarct-related artery 4.8 ± 1.3 days after PCI. Cardiac MRI was repeated, and interim hospitalizations, medical procedures, and serious adverse events were documented in all surviving patients 6 ± 1, 18 ± 6, and 61 ± 11 months after PCI. No patient was lost to follow-up. All patients were treated according current practice guidelines. The trial was approved by the Ethics Committee at Hannover Medical School. The study complies with the Declaration of Helsinki, written informed consent was obtained from the patients, and the study has been registered at www.clinicaltrials.gov (registration number, NCT00224536).

Cardiac magnetic resonance imaging

The MRI technique was identical throughout the entire study period, so was the software package for the MRI analyses, and the responsible MRI lead investigator (J.L.). Magnetic resonance imaging was performed with a 1.5 Tesla scanner using ECG gating and a four-element phased array surface coil during repeated breath-holds. Sequence parameters, image planning, and methods for evaluation of LV volumes, and global and regional LV function have been reported.4,12 For infarct size determination, LV myocardial tissue showing late contrast enhancement (LE) was quantified. Late contrast enhancement volumes were quantified by manual contour tracing at baseline and at all follow-up time points, as described.4 Within the LV myocardial sector displaying LE, infarct transmurality was defined as the ratio of the hyperenhanced (mostly subendocardial) region to the hyperenhanced plus non-hyperenhanced (mostly subepicardial) regions. Contrast media for LE assessment could not be applied in four patients at the 61-month follow-up, due to contraindication (n = 2) or patients' preference (n = 2). Exploratory subgroup analyses at the 6- and 18-month follow-up suggested that patients with more transmural infarcts had the highest potential of functional improvement. We therefore repeated this subgroup analysis after 5 years. It should be pointed out that this subgroup analysis was not pre-specified in the original BOOST study protocol.

Study endpoints and statistical analysis

Data are presented as numbers (percentage) or mean ± SD. Homogeneity of treatment groups at baseline was assessed using Student's unpaired t-test for continuous variables showing no marked deviations from the normal distribution. For other continuous variables or ordinal baseline data, the Wilcoxon rank-sum test was used. Categorical baseline data were investigated using χ2 tests. Global left ventricular ejection fraction (LVEF) change from baseline to follow-up was the pre-specified primary endpoint of the BOOST trial. Changes in LV end-diastolic volume (LVEDV) index, LV end-systolic volume (LVESV) index, LV mass index, regional wall motion and wall thickening, and volume of LE represented secondary endpoints. Global LVEF changes in the two study groups were compared by using a mixed effect model (SAS 9.1 PROC MIXED), which included time, group (BMC treatment), time by group interaction, and LVEF at baseline as fixed effects and a random intercept. Secondary endpoints were analysed using the same methods. Two patients in each group died before the final follow-up examination; the last value carried-forward method was used in these patients. Tests were performed two-sided at a significance level of 0.05.


Patient population and safety issues

We randomized 60 patients, 56 ± 14 years of age, with a first STEMI to either a control or a BMC transfer group. A detailed characterization of the study population has been published.4 Four patients died during follow-up. The last available MRI data of these patients were carried forward to the 5-year follow-up time point. One patient with a new contraindication against MRI underwent contrast echocardiography to assess LV volumes and mass. For this patient, the last available late enhancement volume was carried forward. In four additional patients, MRI contrast application was not possible because of patient preference (n = 2) or new contraindications (n = 2) at the 5-year follow-up. These patients underwent a complete MRI study to assess LV volumes and mass. Late enhancement volumes were carried forward in these patients from the 18-month time point. At 61 months, 26 control and 27 BMC transfer patients were treated with aspirin and/or clopidogrel, 24 and 25 patients with an ACE-inhibitor or AT1-receptor antagonist, 25 and 25 with a beta-blocker, 24 and 24 with a statin, and 2 and 1 with an aldosterone antagonist. No significant differences in adverse events or New York Heart Association class were noted after 61 months of follow-up (Table 1). One case of cancer was diagnosed during follow-up in the control group and none in the BMC transfer group. Magnetic resonance imaging and echocardiography did not reveal any evidence of intramyocardial tumour formation or calcifications.

View this table:
Table 1

Safety endpoints at 61-month follow-up

Control group (n = 30)BMC group (n = 30)
All-cause mortality2 (7%)2 (7%)
Cardiovascular mortality1 (3%)1 (3%)
Recurrent myocardial infarction1 (3%)1 (3%)
Heart failure hospitalization3 (10%)2 (7%)
Composite of death, myocardial infarction, and heart failure hospitalization6 (20%)5 (17%)
Target vessel revascularization4 (13%)6 (20%)
Documented ventricular tachycardia or syncope1 (3%)1 (3%)
NYHA class1.6 ± 0.81.7 ± 0.8
  • Data are shown as numbers (percentages). BMC, bone marrow cell; NYHA, New York Heart Association; P = non-significant for all inter-group comparisons.

Functional outcome

Changes in LVEF from baseline to 61-month follow-up are presented in Figure 1 and in Table 2. Six months after cell transfer, the BMC transfer group showed a significantly better recovery of LVEF when compared with the control group (6.7 vs. 0.7 percentage-points, respectively; P = 0.003); this difference was attenuated at 18 months (5.9 vs. 3.1 percentage-points, respectively; P = 0.27), and virtually absent at 61 months (−2.5 vs. −3.3 percentage-points, respectively; P = 0.30). Reflecting this early and transient improvement in LVEF in the BMC transfer group, there was a significant deviation from a parallel time course between both groups (P = 0.016; Table 2).

Figure 1

Global left ventricular ejection fraction at baseline, 6-, 18-, and 61-month follow-up. The last value carried-forward method was applied in two deceased patients from each group (dashed lines).

View this table:
Table 2

Left ventricular volume and mass indices, global left ventricular ejection fraction, late contrast enhancement, and regional wall motion/thickening

Control group (n = 30)BMC group (n = 30)
Δ vs. baselineΔ vs. baseline
LVEDV index (mL/m2 BSA)
 Baseline81.4 ± 16.984.2 ± 17.2
 6 months84.9 ± 21.93.4 ± 11.191.7 ± 26.07.6 ± 20.0
 18 months85.0 ± 24.23.6 ± 15.190.3 ± 26.56.1 ± 20.3
 61 months89.0 ± 24.77.5 ± 17.398.0 ± 30.713.9 ± 25.4
 Change from baseline (P-values)
  Overall time effect at 61 months0.012
  Group difference at 61 months0.29
  Deviation from parallel time course0.57
LVESV index (mL/m2 BSA)
 Baseline40.6 ± 16.943.0 ± 14.7
 6 months42.6 ± 23.52.0 ± 11.142.4 ± 23.9−0.6 ± 14.9
 18 months41.0 ± 24.70.4 ± 12.542.5 ± 25.0−0.5 ± 16.5
 61 months48.7 ± 27.28.1 ± 16.455.1 ± 30.312.1 ± 23.1
 Change from baseline (P-values)
  Overall time effect at 61 months0.001
  Group difference at 61 months0.42
  Deviation from parallel time course0.94
LVEF (%)
 Baseline51.3 ± 9.350.0 ± 10.0
 6 months52.0 ± 12.40.7 ± 8.156.7 ± 12.56.7 ± 6.5
 18 months54.4 ± 13.03.1 ± 9.655.9 ± 14.75.9 ± 8.9
 61 months48.1 ± 12.9−3.3 ± 9.547.5 ± 16.7−2.5 ± 11.9
 Change from baseline (P-values)
  Overall time effect at 61 months<0.001
  Group difference at 61 months0.30
  Deviation from parallel time course0.016
LV mass index (g/m2 BSA)
 Baseline78.2 ± 18.382.7 ± 18.7
 6 months71.7 ± 14.2−6.5 ± 12.871.9 ± 14.6−10.8 ± 10.6
 18 months71.2 ± 11.8−7.0 ± 14.873.7 ± 17.7−9.0 ± 14.5
 61 months72.2 ± 15.3−6.0 ± 16.976.9 ± 17.7−5.8 ± 14.6
 Change from baseline (P-values)
  Overall time effect at 61 months<0.001
  Group difference at 61 months0.44
  Deviation from parallel time course0.43
LE volume (mL)
 Baseline30.3 ± 17.433.0 ± 21.1
 6 months19.8 ± 9.8−10.5 ± 10.618.9 ± 12.2−14.1 ± 13.0
 18 months20.2 ± 10.3−10.1 ± 13.120.2 ± 14.4−12.8 ± 11.8
 61 months15.7 ± 11.0−14.6 ± 12.118.0 ± 12.6−15.0 ± 12.2
 Change from baseline (P-values)
  Overall time effect at 61 months<0.001
  Group difference at 61 months0.34
  Deviation from parallel time course0.41
Wall thickening, infarct region (%)
 Baseline21.8 ± 11.829.0 ± 18.4
 6 months38.0 ± 32.116.2 ± 30.347.2 ± 26.118.2 ± 22.9
 18 months36.5 ± 25.114.8 ± 22.144.5 ± 24.815.6 ± 24.3
 61 months35.4 ± 27.213.7 ± 24.533.4 ± 26.54.4 ± 29.2
 Change from baseline (P-values)
  Overall time effect at 61 months<0.001
  Group difference at 61 months0.084
  Deviation from parallel time course0.95
Wall thickening, border zone (%)
 Baseline51.0 ± 15.054.0 ± 24.1
 6 months51.3 ± 22.00.3 ± 19.864.3 ± 28.810.3 ± 22.6
 18 months58.0 ± 25.57.0 ± 24.764.1 ± 30.210.1 ± 24.5
 61 months41.2 ± 21.4−9.8 ± 24.442.3 ± 23.0−11.6 ± 27.9
 Change from baseline (P-values)
  Overall time effect at 61 months0.025
  Group difference at 61 months0.25
  Deviation from parallel time course0.27
Wall motion, infarct region (%)
 Baseline3.9 ± 1.84.4 ± 1.9
 6 months4.9 ± 2.91.0 ± 2.55.9 ± 2.51.5 ± 2.1
 18 months4.5 ± 2.60.6 ± 2.75.2 ± 2.20.8 ± 2.1
 61 months4.4 ± 2.50.4 ± 2.44.8 ± 2.40.3 ± 2.4
 Change from baseline (P-values)
  Overall time effect at 61 months<0.001
  Group difference at 61 months0.37
  Deviation from parallel time course0.67
Wall motion, border zone (%)
 Baseline6.8 ± 1.67.0 ± 1.7
 6 months6.8 ± 2.1−0.1 ± 2.28.0 ± 2.11.0 ± 1.9
 18 months6.6 ± 2.0−0.2 ± 2.67.1 ± 1.80.1 ± 1.7
 61 months5.6 ± 1.9−1.2 ± 2.26.2 ± 2.2−0.8 ± 2.4
 Change from baseline (P-values)
  Overall time effect at 61 months0.009
  Group difference at 61 months0.99
  Deviation from parallel time course0.17
  • BMC, bone marrow cell; BSA, body surface area; LV, left ventricular; LVEDV, LV end-diastolic volume; LVESV, LV end-systolic volume; LVEF, LV ejection fraction; LE, late contrast enhancement. At baseline, there were no significant inter-group differences in LVEDV index (P = 0.54), LVESV index (P = 0.57), LVEF (P = 0.59), LV mass index (P = 0.35), and LE volume (P = 0.59). The P-values refer (i) to the overall time effect that addresses the temporal evolution from baseline to 61-month follow-up in the entire study cohort (control and BMC transfer patients combined), (ii) to differences between both groups at 61 months, and (iii) to any deviations from a parallel time course between the control and the BMC transfer groups.

In the entire study population, LVEDV and LVESV indices increased, whereas LV mass index and LE volumes decreased during 61 months of follow-up (Table 2). For these endpoints, no significant differences in the time course or in the absolute difference at 61 months were observed between the control and BMC transfer groups (Table 2). For the control and BMC groups, there was a decrease in regional wall motion and wall thickening in the infarct border zone and the infarct region from the 18-month to the 61-month follow-up (Table 2). The improvement in regional contractility in the infarct border zone that was observed in the BMC group in comparison to the control group at 6 months4 was no longer present at 61 months.

Left ventricular ejection fraction change in relation to infarct transmurality at baseline

Exploratory subgroup analyses at the 18-month follow-up have suggested that patients with greater infarct transmurality may have the highest potential for functional improvement.12 This hypothesis was further explored at 61 months. Transmural extent of the infarct at baseline was 61.0 ± 16.5 and 61.4 ± 18.5% in the control and BMC transfer groups, respectively (P = 0.93). As shown in Table 3, patients with an infarct transmurality at baseline greater than the median appeared to benefit from BMC transfer throughout the 61-month study period. Conversely, no such improvement was observed in patients with an infarct transmurality that was smaller than the median. The patient subgroup with an infarct transmurality greater than the median was characterized by greater infarct sizes at baseline (41.2 ± 19.8 vs. 22.1 ± 13.0 mL in the subgroup with an infarct transmurality smaller than the median; P < 0.001), and a more severely depressed baseline LVEF (47.8 ± 9.4 vs. 53.5 ± 9.1%; P = 0.019).

View this table:
Table 3

Left ventricular ejection fraction change in relation to infarct transmurality at baseline

LVEF change from baseline (%)
Infarct transmurality ≤ median, n = 30Infarct transmurality > median, n = 30
6 months18 months61 months6 months18 months61 months
Control group5.6 ± 8.08.0 ± 10.12.6 ± 7.7−3.6 ± 5.5−1.2 ± 7.0−8.4 ± 7.9
BMC group7.0 ± 5.95.1 ± 7.7−4.0 ± 12.66.3 ± 7.36.8 ± 9.70.7 ± 11.2
BMC treatment effect1.4−2.9−
  • BMC, bone marrow cell; LVEF, LV ejection fraction. Patients were dichotomized according to their transmural infarct extension at baseline. P-values were calculated using unpaired Student's t-test. Infarct transmurality ≤ median: control group, n = 14; BMC group, n = 16; infarct transmurality > median: control group, n = 16; BMC group, n = 14.


Randomized clinical trials addressing the effects of BMC transfer after AMI have covered only limited time frames ranging from 4 to 18 months.47,1214 The present 5-year data from the BOOST trial represent the longest available clinical and MRI follow-up from any randomized trial in the field. Although we have observed a substantial improvement in LVEF at 6 months,4 a subsequent analysis at the 18-month time point12 and the present analysis show that a single dose of BMCs was not able to promote a sustained improvement of LV systolic function. We observed no significant differences in mortality and other clinical endpoints between the control and BMC transfer groups at 5 years. The BOOST trial, however, was not powered to assess clinical outcome; so these data provide reassurance concerning the safety of the procedure rather than insight into possible effects on clinical events. Large randomized trials would be required to explore the impact of BMC transfer on clinical outcome.15

The mean baseline LVEF in the BOOST study population was 51 ± 10%. Considering that LVEF values of ∼67 ± 5% have been obtained by MRI in healthy volunteers,16 the BOOST population represents a patient cohort with only a moderate impairment of LV systolic function. Still, the myocardial damage accrued during the infarct was sufficient to initiate a remodelling process with significant increases in LVEDVs and decreases in LVEF at 5 years. Although our data, which included repetitive MRI assessments of LV volumes and function at 6, 18, and 61 months, clearly show that BMC transfer does not promote a sustained improvement of LVEF in the overall study population, there may be a signal that patients with more severe infarct damage may derive a long-term benefit from BMC therapy. Subgroup analyses at the 6- and 18-month follow-up have indicated that patients with a greater infarct transmurality may respond to BMC therapy with a prolonged treatment effect concerning LVEF.12 This was now confirmed after 5 years. Notably, patients with greater infarct transmurality had larger infarct sizes and more severely depressed LVEF at baseline, indicating that these patients suffered greater infarct damage. Considering the size of the BOOST trial and the fact that this analysis was not pre-specified, our results from this subgroup analysis should be viewed as hypothesis-generating only. It is interesting to note, however, that subgroup analyses from the Reinfusion of Enriched Progenitor Cells and Infarct Remodeling in Acute Myocardial Infarction (REPAIR-AMI) trial, which used LV angiography to assess LVEF before and 4 months after BMC delivery, have indicated that patients with more severely depressed LVEF at baseline may derive the greatest benefit from BMC therapy.5 Furthermore, patients with more severely depressed baseline LVEF benefited more from BMC transfer in terms of LVEF improvement also in the Myocardial Regeneration by Intracoronary Infusion of Selected Population of Stem Cells in Acute Myocardial Infarction (REGENT) trial.21 Accordingly, future studies in this area may want to focus on patients with more substantial infarct damage. Moreover, pharmacological and/or genetic strategies should be considered to enhance cell retention after intracoronary delivery17 and/or to improve the functionality of the cells.1820


Institutional funding of the Department of Cardiology/Angiology, Hannover Medical School.

Conflict of interest: L.A. is head of production, Cytonet-group, the company that performed bone marrow cell preparations during the trial. L.A. has not been involved in MRI data collection or data analysis.


  • These authors contributed equally to this work.


View Abstract