OUP user menu

Impact of intracoronary bone marrow cell transfer on diastolic function in patients after acute myocardial infarction: results from the BOOST trial

Arnd Schaefer, Gerd P. Meyer, Martin Fuchs, Gunnar Klein, Marina Kaplan, Kai C. Wollert, Helmut Drexler
DOI: http://dx.doi.org/10.1093/eurheartj/ehi817 929-935 First published online: 1 March 2006


Aims We have recently shown in the randomized-controlled BOne marrOw transfer to enhance ST-elevation infarct regeneration (BOOST) trial that intracoronary autologous bone marrow cell (BMC) transfer improves left ventricular (LV) ejection fraction recovery in patients after acute myocardial infarction (AMI). However, the impact of BMC therapy on LV diastolic function in patients after AMI has remained uncertain.

Methods and results Using (tissue) Doppler echocardiography, we evaluated the effects of BMC transfer on LV diastolic function in patients enrolled in the BOOST trial. After successful primary percutaneous coronary intervention (PCI) for acute ST-elevation myocardial infarction (MI), patients were randomized to a control (n=29) or BMC transfer group (n=30). Diastolic function was determined 4.5±1.5 days after PCI, at 6 months, and at 18 months by measuring transmitral flow velocities (E/A ratio), diastolic myocardial velocities (Ea/Aa ratio), isovolumic relaxation time (IVRT), and deceleration time (DT). All analyses were performed in a blinded fashion. There was an overall effect of BMC transfer on E/A [0.33±0.12; 95% confidence interval (CI): 0.09–0.57; P=0.008] and Ea/Aa ratios (0.29±0.14; 95% CI: 0.01–0.57; P=0.04). In contrast, we found no effect of BMC transfer on DT (−5±14 ms; 95% CI: −33 to 22; P=0.70), IVRT (−7±7 ms; 95% CI: −20 to 6; P=0.29), and E/Ea ratio (0.35±0.14; 95% CI: −0.92 to 1.62; P=0.57).

Conclusion Intracoronary autologous BMC transfer improves echocardiographic parameters of diastolic function in patients after AMI.

  • Bone marrow cell therapy
  • Diastolic function
  • Acute myocardial infarction


Development of left ventricular (LV) diastolic dysfunction is a frequent complication after acute myocardial infarction (AMI) and is associated with an increased risk for heart failure,1,2 even if LV systolic function is well preserved.3,4 Currently, treatment strategies for patients with diastolic dysfunction or diastolic heart failure remain poorly defined.5

In this context, experimental studies suggesting that myocardial transfer of specific bone marrow-derived stem and progenitor cell populations may enhance recovery of systolic and diastolic function after AMI have created a lot of excitement.611 Building on these experimental findings and pioneering clinical safety and feasibility trials,12,13 we have recently conducted the randomized BOOST trial assessing the impact of intracoronary autologous bone marrow cell (BMC) transfer on LV systolic function in patients recovering from an acute ST-elevation myocardial infarction (MI).14 In the BOOST trial, BMC transfer significantly enhanced regional systolic wall motion and global LV ejection fraction (LVEF) as shown by magnetic resonance imaging (MRI).14 The impact of BMC transfer on LV diastolic function in patients after AMI has not been addressed in clinical trials so far. Using (tissue) Doppler echocardiography, we evaluated the long-term effects of BMC transfer on diastolic function in patients enrolled in the BOOST trial.


Study protocol

The study protocol of the BOOST trial has been approved by the institutional review board at Hannover Medical School. The study protocol has been described elsewhere in detail.14 Briefly, patients were eligible for inclusion in the trial, if they were admitted within 5 days after symptom onset of a first ST-elevation MI, had undergone successful percutaneous coronary intervention (PCI) with stent implantation of the infarct-related artery, and demonstrated hypokinesia or akinesia involving more than two-thirds of the LV anteroseptal, lateral, and/or inferior wall, as revealed by angiography performed immediately after PCI. Patients were randomized in a 1:1 fashion to the control and BMC transfer groups. The baseline echocardiographic examination was performed 4.5±1.5 days after PCI. After echocardiography, bone marrow (128±6 mL) was harvested from patients randomized to the BMC transfer group, subjected to 4% gelatine–polysuccinate sedimentation, and infused into the infarct-related artery the same day (25±2×109 nucleated BMCs).14 Echocardiographic follow-up examinations were performed after 6 and 18 months.


Because of ethical considerations, a sham bone marrow aspiration and a sham left heart catheterization were not performed in patients randomized to the control group.14 Importantly, however, echocardiography and MRI analyses were performed by two investigators blinded for treatment assignments (A.S. and M.F.).

A Philips HDI 5000 CV ultrasound system with a 2–4 MHz transducer and second harmonic imaging was used. Echocardiographic examinations were performed according to the recommendations for the assessment of systolic and diastolic function/diameter and valvular heart disease issued by the American Society of Echocardiography (ASE).1517

LV diastolic function was assessed using transmitral inflow parameters [transmitral peak early (E) and peak late velocities (A), E-wave deceleration time (DT), and E/A ratio]. Isovolumic relaxation time (IVRT) was recorded from the apical four-chamber view by simultaneous recording of LV outflow tract and mitral flows. Doppler tissue imaging (DTI) recordings were obtained from the lateral mitral valve annulus. Early diastolic (Ea) and late diastolic (Aa) velocities were measured, and Ea/Aa and E/Ea ratios were calculated. All Doppler-derived parameters were measured and averaged during expiration from five consecutive beats. LVEF was calculated by the biplane disc summation method according to the modified Simpson's rule using the apical four- and two-chamber views.18,19

Aortic and mitral regurgitation were classified according to the ASE recommendations as being mild, moderate, or severe.15 The presence or absence of pericardial effusion was assessed from parasternal, apical, and subcostal views. The presence or absence of LV thrombi was determined from the apical view.

To determine inter-observer variability, echocardiographic recordings from 10 patients were assessed by two independent observers (A.S. and M.F.). Intra-observer variability was determined by one observer re-assessing the same echocardiographic recordings from 10 patients twice (4 weeks apart).

Cardiac MRI

Evaluation of LVEF by MRI has previously been described in detail.14 Briefly, MRI was done with the patient in supine position in a 1.5 T scanner (CV/i, General Electric, Munich, Germany) using electrocardiogram gating and a four-element-phased array receiver coil.

Statistical analyses

In BOOST, we calculated that we would need 30 patients in each group to achieve a power of at least 80% to detect a difference in global LVEF change of 5 percentage points between the two study groups, with a two-sided significance level of P<0·05 and a common standard deviation of 6.5 percentage points for the global LVEF change from baseline to 6 months' follow-up.14

To analyse the overall treatment effect of BMC transfer on diastolic function, repetitive measurements were compared using the general linear model repeated measures procedure [GLM, repeated measures analysis of variance (ANOVA)] (SPSS version 12.0). In this model, BMC treatment was used as fixed factor (between-subject factor group), time as within factor and echocardiographic measurements at 6 and 18 months as dependent variable. In the GLM model, no random effects were included (fixed model analysis). Estimated mean treatment effects with their corresponding confidence intervals (CIs) were determined by pairwise comparison of the mean values with Bonferroni correction for multiple comparisons within the repeated measures design. As treatment assignment was started after baseline echocardiography, these analyses were restricted to 6 and 18 months of data. One-way ANOVA was performed to evaluate differences among the mean values between both groups (independent variable) at baseline, 6 months, and 18 months (dependent variable) (Figure 1). SPSS was used to calculate post hoc power (‘observed power’) for all echocardiographic diastolic parameters.

Figure 1 (A) Time course of E/A ratio [ratio of transmitral peak early (E) and peak late velocities (A)]. *P=0.02 for between-group comparison at 6 months and **P=0.008 for between-group comparison at 18 months (one-way ANOVA). (B) Time course of Ea/Aa [ratio of early diastolic (Ea) and late diastolic (Aa) mitral annulus velocities]. ***P=0.02 for between-group comparison at 18 months.

Pearson's correlation was determined for comparison of LVEF (%) as determined by MRI and the echocardiographic data. The analysis of Bland and Altman was used to evaluate intra- and inter-observer variabilities. Categorical variables were compared using the χ2 test. All data are expressed as means±SEM. All tests were two-tailed and P-values less than 0.05 were considered to indicate statistical significance.


Study population

Patients’ baseline characteristics are summarized in Table 1; a detailed report of the study population has been provided previously.14 Thirty patients were randomized to each group. One patient from the control group died from progressive heart failure 9 months after randomization. Therefore, complete echocardiographic follow-up data were obtained in 30 patients in the BMC group and 29 patients in the control group. Only these patients were included in the study. There were no significant differences concerning time from symptom onset to primary PCI or infarct territory and infarct size. No patient presented with atrial fibrillation during echocardiography. All patients received optimal post-infarction medical therapy at baseline and throughout the 18-month follow-up period; statin use was somewhat lower in the BMC group at 18 months (Table 1). There were no significant differences in systolic or diastolic blood pressure and heart rate between the two groups (Table 2).

View this table:
Table 1

Patients' characteristics

Control group (n=29)BMC group (n=30)P-value
Age (years)59±353±30.15
Men21 (71%)20 (67%)0.54
Diabetes mellitus2 (7%)3 (10%)0.93
Hypertension12 (41%)9 (30%)0.14
Hyperlipidaemiaa6 (21%)9 (30%)0.39
Current cigarette use16 (55%)18 (60%)0.83
Medication at primary discharge
 Aspirin and clopidogrelb28 (97%)30 (100%)0.31
 ACE-inhibitors or ARB29 (100%)30 (100%)
 Beta-blockers29 (100%)29 (97%)0.31
 Statins29 (100%)30 (100%)
Medication at 18 months' follow-up
 Aspirin or clopidogrel29 (100%)30 (100%)
 ACE-inhibitors or ARB28 (97%)28 (93%)0.60
 Beta-blockers29 (100%)27 (90%)0.10
 Statins29 (100%)25 (83%)0.03

ARB, angiotensin receptor blockers.

aPatients not receiving aspirin or clopidogrel were treated with phenprocoumon.

bSerum cholesterol >5.2 mmol/L.

View this table:
Table 2

Diastolic function, blood pressure, LVEF and heart rate

Control group (n=29)BMC group (n=30)P-value between group
 6 months0.9±0.11.3±0.10.02
 18 months0.9±0.041.2±0.10.008
 6 months1.2±0.11.5±0.10.13
 18 months0.9±0.11.3±0.10.02
 6 months7±0.48±0.40.71
 18 months9±0.48±0.10.48
DT (ms)
 6 months205±12197±110.64
 18 months210±11204±100.45
IVRT (ms)
 6 months123±6105±70.18
 18 months127±3126±60.74
BP (mmHg)
 6 months121/76±2/1120/72±3/10.45/0.74
 18 months116/75±2/2119/74±2/10.51/0.61
LVEF (%)
 6 months50±249±20.51
 18 months52±352±20.44
Heart rate (b.p.m.)
 6 months67±268±20.51
 18 months68±269±20.62

E/A denotes ratio of transmitral peak early (E) and peak late velocities (A); Ea/Aa, ratio of early diastolic (Ea) and late diastolic (Aa) mitral annulus velocities; E/Ea, ratio of E-wave of mitral inflow to Ea mitral annular velocity. BP, arterial blood pressure (systolic and diastolic values are shown); P-value denotes the comparison of the mean values between control and BMC groups at baseline, 6 months, and 18 months (one-way ANOVA).

Parameters of diastolic function

There was an overall effect of BMC transfer on E/A (P=0.008) and Ea/Aa ratios (P=0.04) (Table 3). For E/A ratio, we found no effect of time (P=0.39) and no interaction between group and time (P=0.70). For Ea/Aa ratio, we found an effect of time (P=0.001), but no interaction between group and time (P=0.80).

View this table:
Table 3

Estimated effects of BMC on diastolic parameter

Mean differenceaP-value
E/A0.33±0.12 (0.09–0.57)0.008
Ea/Aa0.29±0.14 (0.01–0.57)0.04
E/Ea0.35±0.14 (−0.92 to 1.62)0.57
DT (ms)−5±14 (−33 to 22)0.70
IVRT (ms)−7±7 (−20 to 6)0.29

E/A, ratio of transmitral peak early (E) and peak late velocities (A); Ea/Aa, ratio of early diastolic (Ea) and late diastolic (Aa) mitral annulus velocities; 95% CIs are between parentheses.

aWith Bonferroni correction for multiple comparisons.

In the control group, E/A ratio was significantly lower at 6 and 18 months and Ea/Aa ratio was significantly reduced at 18 months when compared with the BMC group (Table 2).

There was no effect of BMC transfer on DT (P=0.70), IVRT (P=0.29), and E/Ea ratio (P=0.57). In addition, there was no interaction between time and group for DT (P=0.79), IVRT (P=0.30), and E/Ea ratio (P=0.47). IVRT and E/Ea were affected by time (P=0.03; P=0.001). Comparison between both groups at 6 and 18 months showed no differences for DT, IVRT, and E/Ea ratio. The estimated effects of BMC transfer on all diastolic parameters are summarized in Table 3.

Exclusion of patients (n=3 in both groups) with a restrictive filling pattern (E/A ratio>1.7) did not significantly alter the results (data not shown). When we restricted our analyses to patients with hypertension at onset, no significant effects of BMC therapy on E/A ratio change were detectable (0.15±0.16; 95% CI: −0.19 to 0.50; P=0.37).

In contrast, patients without hypertension at baseline displayed a persisting improvement of E/A ratio by BMC transfer (0.43±0.16; 95% CI: 0.08–0.77; P=0.01).

Retrospectively, the power of E/A, Ea/Aa, DT, and IVRT tests were 0.8, 0.5, 0.1, and 0.2, respectively.

Representative tissue Doppler recordings from both groups (at baseline and 18 months) are shown in Figure 2.

Figure 2 Representative tissue Doppler recordings of the lateral mitral annulus in a patient from the control (A and B) and a patient from the BMC group (C and D) at baseline (A and C) and 18 months (B and D). At baseline, both patients had an Ea/Aa ratio >1; at 18 months Ea/Aa ratio declined to 0.5 in the control patient and remained >1 in the BMC patient.

Cardiac dimensions, systolic LV function, LVEDP, and comparison with MRI

As shown in Tables 3 and 4, echocardiographic measures of LVEDV, LVESV, and LVEF were not different between both groups at baseline, 6 months, and 18 months. IVSD and PWD decreased in both groups. In the whole study population, there was a correlation of LVEF (%) as determined by MRI (previously published data)14 and by echocardiography (r=0.6, P<0.001). This correlation was also detectable in patients with large anterior MI and involvement of the apical segments (which are difficult to evaluate by echo) and in patients with inferior/lateral MI (r=0.6, P<0.001; r=0.5, P<0.001). Further analyses revealed a weak statistical correlation between the changes from baseline to 6 months’ follow-up of LVEF as determined by MRI and E/A ratio as determined by echocardiography (r=0.30, P=0.02). This was not the case for the changes from baseline to 6 months’ follow-up of MRI–LVEF and Ea/Aa ratio (r=0.09, P=0.48).

View this table:
Table 4

LV volumes, dimensions, and diastolic parameters

Control group (n=29)BMC group (n=30)P-value (between groups)
Baseline6 months18 monthsBaseline6 months18 monthsBaseline6 months18 months
LVEDV (mL)106±5109±5109±6114±8116±7116±70.410.510.50
LVESV (mL)55±354±553±559±659±658±70.530.530.52
IVSD (mm)12±0.312±0.311±0.412±0.311±0.310±
PWD (mm)11±0.310±0.39±0.311±0.310±0.39±
E (m/s)0.76±30.63±30.63±30.75±30.78±40.73±30.690.010.04
A (m/s)0.66±40.67±30.68±30.63±30.61±40.63±
Ea (cm/s)9.8±0.58.5±0.57.1±0.410.0±0.69.8±0.59.3±0.60.350.040.02
Aa (cm/s)6.9±0.47.1±0.47.5±0.56.8±0.56.6±0.87.3±0.40.710.080.32

LVEDV, left ventricular end-diastolic volume; LVESV, left ventricular end-systolic volume; IVSD, diastolic interventricular septal thickness; PWD, diastolic posterior wall thickness; E, transmitral peak early velocity; A, transmitral peak late velocity; Ea, early diastolic mitral annulus velocity; Aa, late diastolic mitral annulus velocity (one-way ANOVA).

Pericardial effusion, valvular disease, myocardial calcification, and LV thrombus

As shown in Table 5, there were no significant differences between the two groups with regard to the prevalence of mitral or aortic regurgitation or stenosis, LV thrombus, or pericardial effusion. At 18 months, no patient from the BMC (or control) group presented with echocardiographic signs of intramyocardial calcification or tumour formation.

View this table:
Table 5

Valvular disease, LV thrombus, and pericardial effusion

Control group (n=29)BMC group (n=30)P-value
Baseline18 monthsBaseline18 months
Mitral regurgitation
 Moderate or severe1 (3%)02 (7%)3 (10%)0.10
Aortic regurgitation
 Mild4 (14%)7 (24%)8 (27%)7 (23%)0.94
 Moderate or severe0001 (3%)0.31
Aortic/mitral stenosis0000
LV thrombus1 (3%)1 (3%)01 (3%)0.98
Pericardial effusion5 (17%)06 (20%)1 (3%)0.31

Intra-observer and inter-observer variabilities

Mean differences of inter-observer and intra-observer variabilities determined by the Bland–Altman plot are summarized in Table 6.

View this table:
Table 6

Intra-observer and inter-observer variabilities

Intra-observer variabilityInter-observer variability
E/A0.01 (−0.01 to 0.03)−0.01 (−0.1 to 0.07)
Ea/Aa0.01 (−0.01 to 0.03)−0.01 (−0.07 to 0.05)
DT (ms)−8 (−17.0 to 1.9)10 (1.5 – 19.0)
IVRT (ms)−0.9 (−4.4 to 2.6)3.3 (−3.3 to 9.9)
LVEF (%)−1.4 (−4.8 to 1.9)−1.1 (−5.6 to 3.3)

Mean difference between intra-observer and inter-observer variabilities according to the Bland and Altman plot. 95% CIs are between parentheses. E/A, ratio of transmitral peak early (E) and peak late velocities (A); Ea/Aa, ratio of early diastolic (Ea) and late diastolic (Aa) mitral annulus velocities.


This is the first randomized study to evaluate the effects of BMC transfer on diastolic function in patients recovering from AMI. The main finding from our study is that BMC therapy improves echocardiographic parameters of diastolic function after AMI.

Heart failure is the leading cause for hospitalization and the most costly cardiovascular disease among patients over 65 years in westernized countries.20,21 Epidemiological studies indicate that up to 40% of these cases are related to diastolic heart failure. These patients typically have a (near) normal LV systolic function and heart failure is thought to be caused by diastolic dysfunction.22,23 Diastolic function is determined by a process of active relaxation and by the passive elastic properties of the LV.24 Assessment of diastolic function by echocardiography is performed by measuring transmitral flow pattern (E/A), diastolic myocardial velocities (Ea/Aa; Tissue Doppler Imaging), IVRT, and DT. According to these parameters, diastolic dysfunction is classified into four stages. Mild diastolic dysfunction or impaired relaxation (stage I) is defined by an E/A ratio <1, a prolongation of DT >220 ms, an increase of IVRT >100 ms, and a normal atrial pressure with or without reduced LV compliance.17,2529 Our study shows that E/A and Ea/Aa ratios decreased significantly in the control group when compared with the BMC group. There was no significant effect of BMC therapy on DT and IVRT. However, IVRT was prolonged (>100 ms) and DT in the upper normal range in both groups. Prolongation of IVRT is observed at the earliest stage of diastolic dysfunction and is indicative of an impairment of the energy-dependent process of active LV relaxation.17,28,30 The E/Ea ratio, recently shown to be an indicator of elevated filling pressures and decreased survival after AMI,3133 was not elevated in both groups during follow-up. These data indicate that patients from the control group developed stage I diastolic dysfunction after AMI (decreased E/A ratio, prolongation of IVRT, and no change in E/Ea ratio), whereas patients from the BMC transfer group developed a very mild form of early diastolic dysfunction (prolongation of IVRT only).

Subgroup analyses of patients with hypertension suggested that these patients might benefit less from BMC therapy. However, given the size of our trial, such post hoc subgroup analyses have to be viewed with great caution.

Direct intramyocardial injection of unselected, nucleated BMCs 2 months after AMI has been shown to improve systolic and diastolic functional indices in rats. Notably, this improvement in diastolic function was associated with reduced LV collagen density.34 Similarly, application of bone marrow-derived angioblasts or mesenchymal stem cells has been shown to reduce LV collagen deposition in rats after AMI.35,36 Therefore, it is possible that the improvement of echocardiographic diastolic parameters that we observed after intracoronary BMC transfer in the BOOST trial may be related to a reduction in intramyocardial collagen deposition. In addition, considering the (weak) statistical correlation of systolic functional improvement (as determined by MRI) and the changes in E/A ratio, it is conceivable that improvement of diastolic parameters may in part be related to changes of systolic function.

We have recently shown by MRI that LVEF, the primary endpoint of the BOOST trial, increased by 6.7±6.5 percentage points after 6 months in the BMC transfer group vs. 0.7±8.1 percentage points in the control group (P=0.0026).14 Quite remarkably, in the current echocardiographic study, LVEF was not significantly different between the two groups. This may be related to the high inter- and intra-observer variabilities that37,38 have been observed when echocardiography was used to determine LVEF. In this context, it has been estimated that detection of a change in LVEF of 3 percentage points by echocardiography in patients with heart failure would require 115 patients when compared with 14 patients with MRI.37 The difference in LVEF outcome between our previous MRI and the present echocardiographic studies may also reflect intrinsic limitations of echocardiography to detect LVEF changes in patients with regional wall motion abnormalities, particularly apical dyskinesia.39,40 The correlation of LVEF as determined by MRI and by echocardiography was evident in patients with large anterior MIs and involvement of the apical segments and in patients with inferior/lateral MI. Thus, the limitation of echocardiography to reliably evaluate LVEF in patients with apical MIs does not appear to explain the inability of echocardiography to detect significant improvements of LVEF after 6 months.

One patient in the control group died at 9 months because of progressive heart failure. This patient showed E/A and Ea/Aa ratios <1 at baseline and at 6 months. However, we have repeated all our statistical analyses assuming that this patient's diastolic function remained stable between 6 and 18 months (last value carry forward method); this did not change the overall results and conclusions from our study.

Direct injection of filtered nucleated BMCs into the acutely infarcted myocardium in rats has been shown to induce intramyocardial calcifications.41,42 However, there was no evidence for intramyocardial calcifications in our patients receiving gelatine gradient-purified BMCs. In addition, we found no increases in the prevalence of pericardial effusions, valvular disease, or LV thrombi. No intramyocardial tumour formation was observed after 18 months.

There are a few potential limitations of this study that have to be addressed. First, in addition to the echocardiographic parameters of diastolic function that we studied, measurement of pulmonary venous (PV) flow has been recommended as an additional measure. However, as previously noted, it is not possible to obtain adequate PV flow measurements in as many as 40% of transthoracic echocardiographic studies.17 As we could register accurate PV Doppler signals in only about half of our patients during the initial echocardiographic study, we did not include this parameter in our analyses. Second, as diastolic function was not a pre-specified endpoint of the BOOST trial, this echocardiographic study represents an exploratory analysis of the data. Third, larger trials will have to assess whether improvements in systolic and diastolic function after BMC transfer translate into clinical benefit after AMI. Such trials are currently underway. Fourth, the underlying mechanisms of the effect of BMC therapy on echocardiographic parameters of diastolic function remain to be investigated. Fifth, no controlled invasive measurements of pressure–volume relation were performed in BOOST. Further studies are required to evaluate whether improvement of diastolic function is accompanied by a decrease of diastolic LV pressure after BMC therapy.

In conclusion, intracoronary autologous BMC transfer improves echocardiographic parameters of diastolic function in patients after AMI.

Conflict of interest: none declared.


  • Both authors contributed equally to this study.


View Abstract