OUP user menu

Effects of intracoronary injection of mononuclear bone marrow cells on left ventricular function, arrhythmia risk profile, and restenosis after thrombolytic therapy of acute myocardial infarction

Heikki V. Huikuri, Kari Kervinen, Matti Niemelä, Kari Ylitalo, Marjaana Säily, Pirjo Koistinen, Eeva-Riitta Savolainen, Heikki Ukkonen, Mikko Pietilä, Juhani K.E. Airaksinen, Juhani Knuuti, Timo H. Mäkikallio
DOI: http://dx.doi.org/10.1093/eurheartj/ehn436 2723-2732 First published online: 9 October 2008


Aims To assess the efficacy and safety of bone marrow cell (BMC) therapy after thrombolytic therapy of an acute ST-elevation myocardial infarction (STEMI).

Methods and results Patients with STEMI treated with thrombolysis followed by percutaneous coronary intervention (PCI) 2–6 days after STEMI were randomly assigned to receive intracoronary BMCs (n = 40) or placebo medium (n = 40), collected and prepared 3–6 h prior PCI and injected into the infarct artery immediately after stenting. Efficacy was assessed by the measurement of global left ventricular ejection fraction (LVEF) by left ventricular angiography and 2-D echocardiography, and safety by measuring arrhythmia risk variables and restenosis of the stented vessel by intravascular ultrasound. At 6 months, BMC group had a greater absolute increase of global LVEF than placebo group, measured either by angiography (mean ± SD increase 7.1 ± 12.3 vs. 1.2 ± 11.5%, P = 0.05) or by 2-D echocardiography (mean ± SD increase 4.0 ± 11.2 vs. −1.4 ± 10.2%, P = 0.03). No differences were observed between the groups in the adverse clinical events, arrhythmia risk variables, or the minimal lumen diameter of the stented coronary lesion.

Conclusion Intracoronary BMC therapy is associated with an improvement of global LVEF and neutral effects on arrhythmia risk profile and restenosis of the stented coronary lesions in patients after thrombolytic therapy of STEMI.

  • Stem cells
  • Acute coronary syndrome
  • Arrhythmias
  • Restenosis


Recent studies have suggested that autologous bone marrow cells (BMCs) may improve global left ventricular ejection fraction (LVEF) when administered after primary coronary intervention (PCI) for patients with an acute ST-elevation myocardial infarction (STEMI).17 BMC therapy has also been suggested to be safe in STEMI patients,17 although some experimental and clinical studies have suggested that progenitor cells may increase electrical instability and increase restenosis of the stented coronary vessel.811

Although thrombolysis is still an established treatment of acute STEMI worldwide, no previous studies have assessed the effects of BMC therapy in thrombolysis-treated STEMI patients. Therefore, we designed a randomized, placebo-controlled, double blind, two-centre FINnish stem CELL study (FINCELL) aimed at assessing the efficacy and safety of BMC therapy in patients with acute STEMI treated with thrombolytic therapy followed by PCI 2–6 days later. The efficacy was assessed by measuring the global LVEF by angiography and 2-D echocardiography, and safety was assessed by measuring various arrhythmia risk variables and by quantifying the degree of restenosis of the stented coronary lesions and the peristent coronary segments by intravascular ultrasound (IVUS).


Patients and procedures

A total of 522 consecutive patients with STEMI treated with intravenous thrombolytic therapy and admitted to the University Hospital of Oulu and the University Hospital of Turku, Finland, between October 2004 and February 2007 were screened for eligibility (Figure 1). The patients were considered eligible for the trial, if the age was <75 years, there was an electrocardiogram (ECG) evidence of STEMI, elevated troponin levels, and thrombolytic therapy given within 12 h after the onset of symptoms, no need for urgent PCI immediately after thrombolysis, no cardiogenic shock, no rescue PCI due to chest pain, haemodynamic instability, or lack of resolution of ST-segment elevations after thrombolysis, no need for coronary artery bypass graft surgery, no refusal of the patient to participate, and no severe coexisting condition that interfered with the ability of the patient to comply with the protocol. Written informed consent was obtained from the patients within 2 days after thrombolytic therapy. After exclusions, a total of 80 patients were included in the study (Figure 1). The study protocol conformed to the Declaration of Helsinki and was approved by the appropriate regional Ethic Committees.

Figure 1

Flow diagram of the trial. BMC, bone marrow cell.

Study design

The day of thrombolysis of acute STEMI was defined as day 0. Thrombolysis was given by using intravenous reteplase in the emergency room of both hospitals for patients presenting with ECG evidence of STEMI and with onset of symptoms <12 h at admission. After considered eligible for the study and after informed consent, the cardiologist either at Oulu or at Turku informed the laboratory nurse of the Clinical Research Laboratory of Oulu University, who drew the randomization code for each patient using a computer-generated random-permuted block design with variable block sizes and selected on the basis of whether suspension containing BMCs or placebo medium was given to each patient. The laboratory nurse in Turku was informed by a telephone call from Oulu about the randomization and type of treatment. The laboratory nurses at both sites prepared the suspension containing either BMCs or placebo on the basis of randomization, but did not participate in any other parts of the research protocol. The patients and investigators performing the PCI procedures and analysis of data were unaware of the randomization throughout the study. Consecutively numbered, sealed envelopes were provided and stored in the Clinical Research Laboratory of the University of Oulu and were opened after all baseline and 6 month data were analysed from all patients.

The efficacy was assessed by measuring the global LVEF by angiography and 2-D echocardiography, and safety was assessed by quantifying the degree of restenosis of the stented coronary lesions and the peristent coronary segments by IVUS and by measuring various arrhythmia risk variables.

Cell preparation, administration, and measurement of cell function

A total of 80 mL of bone marrow was aspirated into heparin-treated syringes from the posterior iliac crest under local anaesthesia in the morning of the PCI day. Mononuclear cells were immediately isolated from the aspirate using density gradient centrifugation on Ficoll-Hypaque. After being washed twice with heparinized physiological saline, the mononuclear cells were suspended in 10 mL of medium containing 5 mL of patient’s own serum and heparinized physiological saline. Then the BMC suspension was filtered through 100 µm nylon mesh (BD Falcon™ Cell Strainer, BD Biosciences, Erembodegem, Belgium) and subjected to quality-control procedures, i.e. microbial culture for sterility and flow cytometer analysis for CD34+ cell counting12 in the accredited laboratory of the Oulu University Hospital including both outside and inside quality control. The BMC separation procedure took about 3 h and intracoronary injection of the cells was performed within 3 h after the procedure. The placebo medium contained the medium without the cells.

To assess the validity of our cell preparation system, we assessed the BMC function and viability using a colony-forming unit assay as follows: The BM-MNCs (including 2 × 103 CD34+ cells per plate) were seeded in methylcellulose plates in triplicate (Methocult GF H84434, Stem Cell Technologies, Vancouver, Canada), with PHA-stimulated human leucocyte-conditioned medium and foetal bovine serum as growth factors. Culture plates were incubated at 37°C in a humidified atmosphere with 5% CO2 for 14 days. Thereafter, the granulocyte-macrophage colony-forming units (CFU-GM, colonies >50 cells) were classified and counted under phase-contrast microscopy and the mean colony number calculated. In these analyses, a mean of 450 CFUs were formed per plate, indicating a high functional activity of the isolated BMCs.

PCI of the culprit coronary lesion supplying the infarct area was performed by standard techniques with the implantation of paclitaxel drug-eluting stents for all patients. After stenting, the medium containing the BMCs or placebo medium was injected intracoronarily through over the wire balloon by using intermittent balloon inflation in the stent at the time of injection.

Measurement of left ventricular ejection fraction

Left ventricular angiograms were performed at the time of baseline cardiac catheterization and PCI and repeated in identical standard projections at 6 months after STEMI. An experienced investigator in a central core laboratory quantitatively analysed the left ventricular angiograms with Philips Integris BH5000 system (Philips Medical System Netherland B.V., The Netherlands). Left ventricular volumes and LVEF were calculated with the use of biplane area-length method including the LV outflow tract in the measurements.13

A 2-D echocardiogram was performed within 2 days after the PCI and at 6 months after STEMI. The LVEF was measured by an experienced investigator in the core laboratory, using the technique described previously.14 Transthoracic echocardiography was performed in the left lateral position using a Philips Sonos 7500 echo system with a 1.8–4.4 MHz transducer (Philips, Amsterdam, The Netherlands). LVEF was calculated from the end-diastolic and end-systolic volumes (modified Simpson’s rule) obtained from the apical four-chamber and two-chamber views according to the recommendations of European Association of Echocardiography.15

Intravascular ultrasound

IVUS was performed immediately after the baseline PCI and at 6 months after STEMI. Intracoronary nitroglycerin 0.1 mg was given before IVUS imaging. Automatic pullback images using 0.5 mm/s speed were obtained with a 40 MHz intracoronary transducer (Atlantis™ SR Pro Coronary Imaging Catheter, Boston Scientific Corp., MN, USA). The imaging included the stented segment and the 5 mm segments proximal and distal to the stented segment. The images were analysed by an experienced observer with Galaxy™ Intravascular Ultrasound System (Boston Scientific Corp.) in the core laboratory. The measurements were done according to the ACC clinical expert consensus document.16 The following measurements were performed: minimum lumen diameter (MLD) and the minimum lumen area (MLA) of the stented segment, and the proximal and distal 5 mm persistent regions.

Measurement of arrhythmia risk variables

Holter monitoring

The patients underwent a 24 h ambulatory three-channel ECG recording (Medilog AR12, Oxford Instruments Medical Ltd, Surrey, UK) within 2 days after PCI and at 6 months after STEMI. The standard deviation of all N–N intervals was used as a measure of heart rate variability.14 Total number and frequency of premature ventricular beats and episodes of non-sustained ventricular tachycardia (three or more consecutive beats) were also analysed. Holter recordings were analysed by an experienced technician in a central core laboratory.

Microvolt T-wave alternans

Symptom-limited maximal exercise tests and microvolt T-wave alternans analysis were performed within 2 days after PCI and at 6 months after STEMI. Maximal work load and heart rate were analysed from the exercise ECG. The presence of microvolt T-wave alternans was assessed using CH2000 system (Cambridge Heart, Inc., Bedford, MA, USA) and high-resolution electrodes with standard 12-lead ECG and three orthogonal (X, Y, Z) ECG lead positions and interpreted in a core laboratory as described earlier.17

Signal-averaged electrocardiogram

Signal-averaged ECGs (SAECGs) were performed within 2 days after PCI and at 6 months after STEMI using the LP Plus system (Fidelity Medical Ltd, Haifa, Israel) capable of measuring averaged high-frequency QRS complexes. Standard criteria of the filtered high-frequency QRS complex duration were used to define the presence of late potentials.14,17 All SAECGs were analysed in a core laboratory by an experienced investigator.


There were three primary endpoints: (i) absolute change in global LVEF from baseline to 6 months assessed either by 2-D echocardiography or left ventricular angiography, (ii) absolute changes in the measures obtained by IVUS, and (iii) changes in arrhythmia risk variables from baseline to 6 months.

Statistical analysis

For sample size calculation, the study was designed to have 80% power to detect a significant difference in the change of LVEF between baseline and 6 months at an alpha-error level of 5%. A significant difference in the change of LVEF between the groups was defined to be 5% units between the groups. With an estimated standard deviation of 8, we calculated that 41 patients will be needed to be enrolled in each group. The analyses were performed according to modified intention-to-treat principle, i.e. all randomized patients were included in the analyses except for those with missing data.

Analysis of covariance was used to compare the differences in the primary and secondary endpoints between the groups including each baseline variable as a covariate. Values for continuous variables that approximated normal distribution are presented as means ± SD, and values for variables that were not normally distributed are also presented as medians. Categorical variables were analysed with χ2 test or Fisher’s exact test, as appropriate. All tests were two-sided, and P-values <0.05 were considered significant. Analyses were performed with SPSS software, version 15.0.1.


Patient characteristics and procedural results

A total of 80 patients, 40 in each group, were randomized into the study. Table 1 lists the characteristics of the 40 patients in both groups. The two groups were well matched regarding all baseline characteristics, including the pharmacological therapy at the time of discharge from hospital and at 6 months after STEMI.

View this table:
Table 1

Characteristics of the patients and major complications

BMC (n = 40)Placebo (n = 40)
Age (years)60 ± 1059 ± 10
Male sex (%)9085
Hypertension [n (%)]16 (40)13 (33)
Diabetes mellitus [n (%)]5 (13)3 (8)
Previous angina [n (%)]8 (20)7 (18)
Current smoking [n (%)]8 (20)7 (18)
Time delay to thrombolysis (h)
 Mean2.8 ± 2.33.1 ± 3.9
Time delay from thrombolysis to PCI (h)
 Mean48 ± 1244 ± 13
Troponin I 2 days after AMI (µg/L)
 Mean5.4 ± 7.24.2 ± 5.9
Infarct-related vessel [n (%)]
 Left anterior descending coronary artery20 (50)20 (50)
 Circumflex coronary artery6 (15)8 (20)
 Right coronary artery14 (35)12 (30)
TIMI flow before PCI [n (%)]
 05 (13)1 (3)
 12 (5)4 (10)
 22 (5)8 (20)
 331 (77)27 (67)
TIMI flow after PCI [n (%)]
 01 (2)1 (2)
 339 (98)39 (98)
%Stenosis of the infarct-related artery before PCI
 Mean85 ± 1879 ± 19
Severity of CAD [n (%)]
 One-vessel19 (48)25 (62)
 Two-vessel15 (37)13 (33)
 Three-vessel6 (15)2 (5)
Number of injected BMCs
 Number of mononuclear cells (×106)
  Mean402 ± 196
 Number of CD34+ cells (×106)
  Mean2.6 ± 1.6
Time to cell/placebo injection from thrombolysis (h)
 Mean70 ± 3680 ± 36
Medication at discharge [n (%)]
 Aspirin40 (100)40 (100)
 Clopidogrel40 (100)40 (100)
 Warfarin4 (10)5 (13)
 Beta-blocker38 (95)39 (97)
 ACE-inhibitor/AT II receptor blocker34 (85)33 (83)
 Statin40 (100)40 (100)
 Diuretic6 (15)10 (25)
n = 39n = 38
Medication at 6 month follow-up [n (%)]
 Aspirin39 (100)38 (100)
 Clopidogrel38 (97)38 (100)
 Warfarin2 (5)1 (3)
 Beta-blocker36 (92)38 (100)
 ACE-inhibitor/AT II receptor blocker31 (79)30 (79)
 Statin39 (100)38 (100)
 Diuretic4 (10)5 (13)
Major complications [n (%)]
 Before discharge from hospital
  Major bleeding1 (3)3 (8)
  No reflow of the infarct  artery after PCI1 (3)0
  Stent thrombosis1 (3)3 (8)
 During the 6 month follow-up
  CHF needing hospitalization01 (3)
  Recurrent AMI02 (5)
  Death01 (3)
  PCI/CABG during follow-up3 (8)3 (8)
  • Values are means ± standard deviation and/or medians.

  • ACE, angiotensin-converting enzyme; AMI, acute myocardial infarction; AT, angiotensin receptor; CABG, coronary artery bypass grafting; CAD, coronary artery disease.

There were three mild self-terminating vasovagal reactions during the bone marrow aspiration. No other procedural complications occurred related to the aspiration. The median number of mononuclear cells injected was 360 × 106, and the median number of CD34+ cells was 2.6 × 106. One patient in the BMC group had ‘no reflow’ phenomenon after the stenting of the infarct artery. Subacute stent thrombosis needing a repeat PCI occurred in three patients in the placebo group and in one in the BMC group (Table 1).

Left ventricular function

Adequate contrast opacification of left ventricular angiograms both at baseline and at 6 months were available for 36 patients in each group. The absolute change of global LVEF was greater in the BMC than in the placebo group (7.1 ± 12.3 vs. 1.2 ± 11.5%, P = 0.05) (Table 2 and Figure 2).

Figure 2

Left ventricular ejection fraction (LVEF) at baseline and 6 months after myocardial infarction measured by left ventricular angiography. The lines represent the change in left ventricular ejection fraction for individual patients, and the bars the means and standard deviations.

View this table:
Table 2

Quantitative measures of left ventricular angiography, 2-D echocardiography, and intravascular ultrasound

BMC groupPlacebo groupTreatment effect (95% CI)aP-value
LV angiographyn = 36n = 36
  LVEDV (mL)148 ± 51146 ± 47
  LVESV (mL)62 ± 3157 ± 28
  LVEF (%)59 ± 1162 ± 12
 6 months
  LVEDV (mL)152 ± 53154 ± 38
  LVESV (mL)52 ± 3256 ± 21
  LVEF (%)66 ± 1063 ± 14
 Absolute difference
  LVEDV (mL)5.4 ± 37.18.2 ± 34.3−1.3 (−16.8 to −14.3)0.63
  LVESV (mL)−10.0 ± 30.3−1.2 ± 11.5−5.7 (−17.0 to −5.7)0.26
  LVEF (%points)7.1 ± 12.31.2 ± 11.55.0 (0.02–9.8)0.05
2-D echocardiographyn = 39n = 38
  LVEF (%)56 ± 1057 ± 10
 6 months
  LVEF (%)60 ± 856 ± 10
 Absolute difference
  LVEF (%points)4.0 ± 11.3−1.4 ± 10.14.4 (0.5–8.3)0.03
IVUSn = 28n = 30
  MLD of the stented segment (mm)2.7 ± 0.52.7 ± 0.5
  MLA of the stented segment (mm2)6.8 ± 2.47.1 ± 2.6
  MLD of the proximal persistent region (mm)3.3 ± 0.63.0 ± 0.7
  MLD of the distal persistent region (mm)2.6 ± 0.42.6 ± 0.6
 6 months
  MLD of the stented segment (mm)2.4 ± 0.82.5 ± 0.6
  MLA of the stented segment (mm2)5.9 ± 2.76.2 ± 2.3
  MLD of the proximal persistent region (mm)2.9 ± 1.02.8 ± 0.9
  MLD of the distal persistent region (mm)2.4 ± 0.92.6 ± 0.8
 Absolute difference
  MLD of the stented segment (mm)−0.23 ± 0.51−0.31 ± 0.60.06 (–0.23 to –0.36)0.66
  MLA of the stented segment (mm2)−0.68 ± 1.51−1.20 ± 2.060.30 (−0.63 to –1.24)0.52
  MLD of the proximal persistent region (mm)−0.19 ± 0.70−0.19 ± 0.610.01 (−0.35 to –0.38)0.94
  MLD of the distal persistent region (mm)−0.12 ± 0.60−0.0 ± 0.67−0.11 (−0.47 to −0.24)0.52
  • Values are means ± standard deviation and/or medians.

  • CI, confidence interval; LVEDV, left ventricular end-diastolic volume; LVESV, left ventricular end-systolic volume.

  • aTreatment effect is estimated according to the analysis of covariance after adjustment for the baseline value of each primary and secondary endpoints.

Global LVEF could be reliably analysed by 2-D echocardiography in all 39 patients randomized to the BMC group at baseline and at 6 months and in all except one patient who had died in the placebo group (n = 38). The absolute change of LVEF measured by echocardiography was significantly greater in the BMC group than in the placebo group (P = 0.03) (Table 2).

Intravascular ultrasound

The MLA of the stented vessel decreased significantly in both groups from baseline to 6 months (Table 2). However, no differences were observed in the change of the lumen area between the groups. The other measures of IVUS did not change significantly in either group, and no significant differences, or even trends in the differences, were observed between the groups.

Arrhythmia risk variables

After exclusions of recordings with noise, frequent ectopic beats, and atrial fibrillation, heart rate variability could be reliably analysed in 33 patients in the BMC group and in 30 in the placebo group. Standard deviation of N–N intervals increased in both groups from baseline to 6 months (Table 2). The magnitude of increase of the heart rate variability did not differ between the groups (Table 3). Similarly, no differences were observed between the study groups in the number of ventricular premature beats or episodes of non-sustained ventricular tachycardia either at baseline or at 6 months.

View this table:
Table 3

Quantitative measures of Holter recordings, signal-averaged electrocardiogram, and exercise sress test including T-wave alternans

BMC groupPlacebo groupTreatment effect (95% CI)aP-value
Holter recordingsn = 39n = 38
  SDNN (ms)
   Mean104 ± 32100 ± 37
  VPDs (per h)
   Median0.7 (0.2–3.9)0.6 (0.1–5.2)
  nsVT [n (%)]4 (10)2 (5)
 6 months
  SDNN (ms)
   Mean145 ± 47151 ± 42
  VPDs (per h)
   Median0.9 (0.3–6.7)0.6 (0.1–10.1)
  nsVT [n (%)]2 (5)3 (8)
 Absolute difference
  SDNN (ms)43 ± 4653 ± 38−8.4 (−29.1 to −12.4)0.42
   VPDs (per h) (median)0.4 (−0.6 to −0.4)0 (−0.9 to –1.4)−7.8 (−32.7 to −17.0)0.53
  nsVT (n)−210.49
SAECGn = 29n = 29
  QRS duration (ms)100 ± 18101 ± 21
  RMS last 40 ms (µV)48 ± 3241 ± 29
  Duration <40 µV (ms)34 ± 1537 ± 23
 6 months
  QRS duration (ms)96 ± 1396 ± 10
  RMS last 40 ms (µV)44 ± 2645 ± 29
  Duration <40 µV (ms)29 ± 828 ± 9
 Absolute difference
  QRS duration (ms)−3.7 ± 17.1−4.4 ± 20.60.37 (−5.4 to –6.2)0.89
  RMS last 40 ms (µV)−4.7 ± 26.8−2.4 ± 26.2−3.7 (−15.4 to −8.0)0.53
  Duration <40 µV (ms)−4.8 ± 14.8−6.9 ± 24.40.49 (−4.1 to –5.1)0.83
Exercise stress testn = 27n = 27
  METs6.1 ± 1.46.0 ± 1.6
  Maximum heart rate (b.p.m.)126 ± 15126 ± 200.41 (−1.11 to −0.32)0.26
  T-wave alternans
   Positive (n)43−6.0 (−13.8 to −1.90)0.13
   Negative (n)16170.75
   Indeterminate/incomplete (n)77
 6 months
  METs6.9 ± 1.56.9 ± 1.7
  Maximum heart rate (b.p.m.)125 ± 13131 ± 17
  T-wave alternans
   Positive (n)45
   Negative (n)1715
   Indeterminate/incomplete (n)67
 Absolute difference
  METs0.5 ± 1.21.0 ± 1.40.41 (−1.11–0.12)0.26
  Max heart rate (b.p.m.)−1.0 ± 13.94.6 ± 20.8−6.0 (−13.8–1.90)0.13
  T-wave alternans
   Positive (n)020.75
   Negative (n)1−2
   Indeterminate/incomplete (n)−10
  • Values are means ± standard deviation and/or medians. Inter-quartile range in parentheses for medians of VPS (per h).

  • METs, maximum exercise tolerance = (12 × load + 3.5 × weight) / (3.5 × weight); nsVT, non-sustained ventricular tachycardia; RMS, root mean square voltage; SDNN, standard deviation of N–N intervals; VPDs, ventricular premature contractions.

  • aTreatment effect is estimated according to the analysis of covariance after adjustment for the baseline value of each primary and secondary endpoints.

After exclusions of ECGs with bundle branch block, noise, frequent ectopy, and atrial fibrillation, SAECGs could be reliably analysed in 29 patients in both groups at baseline and at 6 months. No differences, or even trends, were observed in the changes of the SAECG measures between the groups (Table 3).

The changes in exercise test variables from baseline to 6 months did not differ between the groups (Table 3). Similarly, the prevalence of patients with positive microvolt T-wave alternans test did not differ between the groups (Table 3).

Clinical outcomes

There were no significant differences in the major clinical endpoints between the groups during the 6 month follow-up (Table 1). One patient in the placebo group died owing to progressive heart failure within few weeks after STEMI, and two patients experienced recurrent myocardial infarction in the placebo group. Both recurrent infarctions were caused by subacute stent thrombosis. No serious clinical arrhythmias occurred in either group, but one patient received a prophylactic implantable cardioverter–defibrillator in the BMC group. This patient had had a ‘no-reflow’ of the infarct artery, observed as no flow of contrast medium into the distal artery in coronary angiography performed after PCI, and subsequently had a total occlusion of the infarct artery confirmed by a repeated coronary angiogram 4 days after PCI, and a significant deterioration of LVEF.


This is the first study showing that intracoronary administration of BMC is safe after thrombolytic therapy followed by PCI of patients with STEMI and is associated with an improvement of global LVEF. Meticulous analysis of arrhythmia risk factors by several non-invasive risk stratification tests and coronary arteries with the use of IVUS showed that the arrhythmia risk profile and restenosis of the stented coronary arteries remained similar in the BMC-treated and placebo groups during a 6 month follow-up.

The present results are in line with a previous double-blind, placebo-controlled randomized trial7 and with some smaller non-blinded trials,16 which have shown an improvement in the global LVEF among patients treated with intracoronary BMC therapy. A distinct difference between the present and previous studies is that we included patients treated with thrombolysis followed by later PCI, whereas all prior trials have assessed the efficacy of BMC therapy in patients treated with primary PCI. In two previously reported studies, the intracoronary administration of BMC did not improve the global LVEF.18,19 In a negative study by Janssens et al.,18 the BMCs were administered within 24 h after primary PCI compared with 2–6 days’ time delay in our study. A larger randomized study suggested that the BMC should ideally be administrated more than 4 days after STEMI to gain the best benefit from this therapy.7 In another negative study by Lunde et al.,19 the numbers of injected CD34+ cells and total amount of mononuclear cells were significantly smaller compared with the present study. One previous study and a meta-analysis have shown an overall trend towards a significant association between the injected cell volume and the improvement of global LVEF.6,20 Furthermore, the method of cell preparation and storage may also partly influence the discrepant results.21 In the present study, the BMCs were injected immediately after the bone marrow aspiration without storage and the colony-forming function of the cells was confirmed.

Unlike in few previous studies,7,18,19 we could not demonstrate an improvement of global LVEF in the placebo group, probably due to a higher baseline LVEF in our study cohort than in previous trials. Improvement of LVEF has been shown to occur predominantly among the patients with markedly deteriorated ventricular function early after STEMI.22 Furthermore, we measured the baseline LVEF relatively late after the infarction, at the time period when the stunning of the infarcted myocardium causing impairment of LVEF is not as prominent as in the early phase of STEMI. The incidence of stent thrombosis was also relatively high, but this was not unexpected, given the thrombotic environment at the time of drug-eluting stent deployment and inclusion of patients with thrombolysis in the present trial. It has been shown that thrombolytic therapy preceding PCI predisposes to ischaemic cardiac complications, such as early re-infarction and urgent target vessel revascularization.23

A previous in vitro study has suggested that bone marrow-derived cells may cause pro-arrhythmic effects by increasing electrical heterogeneity.8 Similarly, autologous transplantation of skeletal myoblasts has been shown to increase the vulnerability to life-threatening arrhythmias.24 However, no differences were observed here between the BMC and placebo groups in heart rate variability, ambient arrhythmias, measures of SAECG, or microvolt T-wave alternans. All these risk variables have been shown to provide information on the risk of future clinical arrhythmic events after myocardial infarction.25 Notably, there was no evidence of delayed conduction of the infarcted myocardium in the BMC-treated group, analysed by the SAECGs, which would suggest to the creation of an arrhythmic substrate by the BMCs.

Experimental studies have also suggested that bone marrow-derived progenitor cells might contribute to neointima formation in transplant atherosclerosis, and one clinical study found an increased restenosis rate after intracoronary infusion of mobilized progenitor cells.911 By using the IVUS technique, we could not demonstrate increased restenosis of the stented segments among the patients treated with BMC. It should be noted that delayed restenosis has been observed using drug-eluted stents. Therefore, 6 month follow-up might have been too short to demonstrate such differences between BMC and placebo groups.

A potential limitation of the study is that the global LV function was quite preserved in our patient population. We used left ventricular angiography and 2-D echocardiography instead of magnetic resonance imaging in the assessment of the effects of BMC therapy on cardiac function. However, although regional wall motion analysis is likely more sensitive to detect subtle changes, global LV function is clinically more relevant providing also prognostic information. A relatively small sample size may also limit the generalization of the results; for example, IVUS and some arrhythmia risk variables could not be performed or analysed in all patients. However, it is unlikely that a larger patient sample would have changed the results, because we did not observe even non-significant trends in these variables between the groups.

Another potential confounding factor was the usage of paclitaxel stents in all patients, which may influence the rate of restenosis caused by BMC therapy as well as the function of the injected cells itself. The latter factor is unlikely, however, as the cells were injected distally from the stents using balloon inflation of the coronary artery at the time of injection.

In summary, the intracoronary administration of BMC during the PCI performed 2–6 days after STEMI in thrombolysis-treated patients is safe as regards adverse clinical events, arrhythmia risk profile, and restenosis, and it improves global LVEF measured at 6 months after infarction. In accordance with a previous randomized trial, the patients with a more marked deterioration of the baseline LVEF had the largest benefit from the BMC therapy. Future studies should perhaps address the efficacy and safety of this therapy among patients with larger infarcts, especially in those who cannot be treated by either early PCI or thrombolysis owing to a long time delay from the onset of symptoms to first medical contact.


The study was supported by grants from the Medical Council of the Academy of Finland, Helsinki, Finland, from the Finnish Foundation for Cardiovascular Research, Helsinki, Finland, and the Foundation for the Northern Health Support, Boston Scientific Sverige AB, Stockholm, Sweden.

Conflict of interest: none declared.


The following other investigators, research nurses, and committee members who are not mentioned in the author list participated in the FINCELL study. Oulu: Miia Hyytinen-Oinas, Jarmo Lumme, Kirsi Majamaa-Voltti, Pekka Raaatikainen, Paavo Uusimaa, Olli-Pekka Piira, Timo Siitonen, K. Kirsi Kvist-Mäkelä, Pirkko Huikuri, Päivi Karjalainen, P. Päivi Kastell, Anne Lehtinen; Turku: T. Vasankari, R. Lautamäki, H. Tuunanen, T. Pelliniemi .

ECHO Committee: Miia Hyytinen-Oinas, Kari Ylitalo.

Arrhythmia Risk Factor Committee: Timo Mäkikallio, Heikki Huikuri.

IVUS Committee: Matti Niemelä, Kari Kervinen.

Cell Preparation Committee: Kirsi Kvist-Mäkelä, Marjaana Säily, Eeva-Riitta Savolainen, H. Tuunanen, T. Pelliniemi.


  • Members of the FINnish study of autologous bone marrow-derived stem CELLs in acute myocardial infarction (FINCELL) group are listed in the Appendix.

  • ClinicalTrials.gov number, NCT00363324.


View Abstract