Bone Marrow Cells and CABG for Myocardium Regeneration
Bone Marrow Cells and CABG for Myocardium Regeneration
Fifty-seven patients were male and three were female. The mean age was 62.7 ± 10.6 years in the placebo group and 61.9 ± 7.3 years in the CD133 group, and all other pre-operative characteristics were also similar between groups. The last documented myocardial infarction event was 2.3 months ago (median, range: 21 days to 28.5 years), and the average LVEF was 26.2 ± 5.6% (range: 15–35%) (Table 2). Nine patients in the CD133 group and nine in the placebo group had the last documented myocardial infarction <3 months before surgery. The decision for surgical re-vascularization was made in accordance with the guidelines on the management of stable angina pectoris of the European Society of Cardiology. Fifty-six patients had three-vessel disease and 4 had two-vessel disease. After surgery, re-vascularization was incomplete in five patients. In three of those, an occluded right coronary artery could not be grafted and, in two, the occluded LAD was not graftable. Two of the incompletely re-vascularized patients were in the CD133 group and three were in the placebo group. By gadolinium first-pass imaging, the proportion of LV myocardial segments with perfusion deficit at rest was similar between the groups (placebo, 23%; CD133, 23%).
Six-month MRI follow-up was performed in 48 patients, 22 patients in the placebo group and 26 in the CD133 group. Three patients refused the examination, two had died early post-operatively, one developed dyspnoea in the MRI scanner, and two patients had post-operative pacemaker implantation. At 6-month follow-up, ejection fraction at rest as determined by MRI was 31 ± 7% in the CD133 group and 33 ± 8% in the placebo group (P = 0.3), with an average inter-group difference of −2.1% (95% CI −6.3 to 2.1).
Mortality Thirty-day mortality was 3.3%, with two deaths in the placebo group and none in the CD133 group. In one patient with pre-operative LVEF of 20%, the LV failed despite intra-aortic balloon pump and catecholamine treatment, and implantation of a mechanical assist device was necessary. Later, the patient developed Candida sepsis with and died on POD 8. The other patient developed pneumonia, mediastinitis, and sepsis, and died on POD 22 although LVEF had increased from 25 to 35%. Beyond the follow-up period, two more patients died 31 and 34 months post-operatively, one in the CD133 and one in the placebo group. The causes of death were given as heart failure and/or sudden cardiac death but autopsies were not performed.
Adverse Events The similarity of AE and SAE frequency between the groups demonstrates the safety of the cell treatment Table 3. The overall high incidence of AE reflects the invasiveness of CABG surgery in patients with severely impaired heart function.
Magnetic Resonance Imaging The pre-operative LV function parameters were similar between the groups. Overall, global LV function improved after surgery (Figure 2). However, in line with the LVEF data (see above), there were no significant between-group differences in LVEDD or LVEDV measured at 6-month follow-up (placebo vs. CD133: 59 ± 7 vs. 61 ± 6 mm, P = 0.4; 218 ± 50 vs. 224 ± 57 mL, P = 0.7). Similarly, there was no significant difference in the change of LVEF (2.3% (−2.0 to 6.6), P = 0.3) (Figure 2D). Per qualitative assessment of regional contractile function in the placebo group, akinesia was documented in 97 LV regions pre-operatively and 72 regions post-operatively. It improved in 42 regions, worsened in 17. In the CD133 group, this pattern was similar (106 akinetic regions pre-operatively, 84 post-operative, improvement in 39 (P = 0.8), worsening in 17 (P = 0.9). The proportions of LV myocardial segments with a perfusion deficit that changed after surgery were different between groups. In placebo patients, perfusion deficit at rest was unchanged in 19% of all analysed segments, better in 2% and worse in 7%. In the CD133 group, the perfusion deficit was unchanged in 15% (P = 0.2 vs. placebo), improved in 9% (P = 0.001, OR 4.0, 95% CI 1.6–9.8), and worse in 1% of all analysed segments (P = 0.0007, OR 0.2, 95% CI 0.07–0.5). However, the perfusion improvement in the CD133 group did not always correlate with the location of the cell injection. Scar size was not a pre-specified outcome but was quantified post hoc. Pre-operative scar mass was 27 ± 15 g (20 ± 10% of the total myocardial mass) in the placebo group and 32 ± 16 g (24 ± 10%) in the CD133 group. At 6-month follow-up, scar mass was 27 ± 15 (22 ± 11%) in the placebo group and 30 ± 17 (23 ± 10%) in the CD133 group (inter-group P = 0.5 (mass) and P = 0.7 (%)). On average, scar mass had decreased by 2.2 ± 5 g in CD133 patients (P = 0.05), but was unchanged in the placebo group (0.3 ± 4 g, P = 0.7). In relation to the pre-operative measurements, scar mass decreased by 9.7% (IQR −20 to 2%) in CD133 patients, and by 0.6% (IQR −15 to 12%) in placebo patients (inter-group P = 0.1) (Figure 3). The between-group difference of the change in scar mass at follow-up was 2.0 g (95% CI −1.1 to 5.0), P = 0.2.
(Enlarge Image)
Figure 2.
Left ventricular function by MRI: (A–C) LV function and dimension parameters measured by cardiac magnetic resonance imaging (MRI) before surgery and at 6-month follow-up. (A) Left ventricular ejection fraction (LVEF): time effect P < 0.001; group effect P = 0.5, interaction P = 0.3. (B) Left ventricular end-diastolic diameter (LVEDD): time effect P < 0.001; group effect P = 0.6, interaction P = 0.2; (C) left ventricular end-diastolic volume (LVEDV): time effect P < 0.001; group effect P = 0.9, interaction P = 0.4. (D) Change in LVEF 6 months after CABG surgery and CD133 or placebo injection, compared with pre-operative.
(Enlarge Image)
Figure 3.
Infarct scar size: late gadolinium enhancement images in the short-axis orientation of the heart at baseline (A) and 6 months after injection of CD133 bone marrow stem cells (B) demonstrating a slight reduction in scar size (−4 g or 6.3% of total LV myocardial mass in this patient). Scar tissue appears as areas of hyperintensity (bright signal) whereas viable myocardium appears dark. (C) Change in scar size from between pre-operative and 6-month follow-up expressed in % of the pre-operative scar mass.
Again, most global LV function and dimension parameters improved after surgery irrespective of the group assignment, but there were no significant differences between groups at follow-up (Figure 4). Regarding regional contractility, there were no differences in the average 2D strain between groups, pre-operatively or at follow-up (pre-specified secondary endpoint). When the speckle-tracking echocardiography data were stratified according to the location of the AOI (post hoc subgroup analysis), we observed a tendency towards better improvement of regional LV function in patients with posterior or inferior infarction and CD133 cell injection, which was not apparent in the placebo group (group effect P = 0.1, Figure 5A). When the proportions of segments with improved longitudinal strain were compared between the groups, regional contractile function recovery was better in the CD133 group at 3 months (P = 0.03, OR = 2.2 95% CI 1.1–4.6) but not at 6 months (P = 0.4). Figure 5B summarizes the 2D longitudinal strain data in CD133 patients according to the location of the AOI and illustrates that regional contractile function recovered only when the AOI was posterior or inferior (localization effect P = 0.02). However, there was no correlation between AOI location and LVEF at 6 months or the change in global LV function.
(Enlarge Image)
Figure 4.
Left ventricular function by echocardiography: global left ventricular function and dimension parameters measured by echocardiography before surgery and at 3- and 6-month follow-ups. (A) Left ventricular ejection fraction (LVEF): time effect P < 0.001, group effect P = 0.9, interaction P = 0.8; (B) fractional shortening (LV-FS): time effect P = 0.01, group effect P = 0.7, interaction P = 0.08; (C) end-diastolic diameter (LVEDD): time effect P = 0.001, group effect P = 0.9, interaction P = 0.8; (D) end-systolic diameter (LVESD): time effect P = 0.001; group effect P = 0.7, interaction P = 0.7. (E) End-diastolic volume (LVEDV): time effect P = 0.1, group effect P = 1, interaction P = 0.6; (F) end-systolic volume (LVESV): time effect P = 0.001, group effect P = 0.6, interaction P = 0.9.
(Enlarge Image)
Figure 5.
Regional left ventricular function: (A) change in regional longitudinal strain between pre-operative and 3-month follow-up in the area of interest that received CD133 or placebo injection stratified according to the localization of the area of interest. Recovery of myocardial function was better in the CD133 group when the area of interest was posterior or inferior (time effect P = 0.2; group effect P = 0.1, interaction P = 0.08. (B) Longitudinal strain data measured by speckle tracking echocardiography in CD133 patients before surgery and at 3- and 6-month follow-ups, stratified according to the localization of the area of interest. Time effect P = 0.5; localization effect P = 0.02, interaction P = 1.
Exercise Capacity At 6-month follow-up, 50 patients underwent exercise testing (placebo vs. CD133: 25 vs. 25 patients). By spiroergometry, all tested patients reached the anaerobic threshold of RER > 1.01. Prior to the operation peak oxygen uptake (VO2 max) and ventilation perfusion mismatch (VE/VCO2-slope) were 12.6 ± 4.6 mL/min/kg and 30.3 ± 6 in the placebo group and 14 ± 3.7 mL/min/kg and 33.4 ± 6.3 in the CD133 group. At 6-month follow-up testing VO2 max and VE/VCO2-slope were not significant different between groups (placebo vs. CD133: 14.4 ± 4.3 vs. 14.9 ± 6 mL/min/kg (P = 0.7) and 31.5 ± 6.8 vs. 31.3 ± 5.6 (P = 0.9)). Although the mean and median values of the 6-min walk at follow-up (Figure 6) or their change from pre-operatively (placebo vs. CD133: 28.7 (−15.9 to 73.3) vs. −2.3 (−35.3 to 34.8 m), P = 0.3)) were not significantly different, the data spread was much larger in the CD133 group than in the placebo group (Figure 6). While range was similar (CD133, 120–661 m; placebo, 153–600 m), the inter-quartile range was approximately three times greater, indicating a larger inter-individual variability in the CD133 group.
(Enlarge Image)
Figure 6.
Exercise capacity: exercise capacity assessed my 6-min walking distance before surgery and at 6-month follow-up. Time effect P = 0.3; group effect P = 0.3, interaction P = 0.8.
The majority of patients had less heart failure symptoms and angina at 6-month follow-up (NYHA, P < 0.001; CCS, P = 0.006) (Figure 7). However, inter-group comparison of the proportions of patients in the different NYHA classes revealed between-group differences at 6 months, suggesting that NYHA class had improved more in the placebo group than in the CD133 group (P = 0.004). For CCS class, this was not the case (P = 0.2).
(Enlarge Image)
Figure 7.
Symptoms: (A) severity of heart failure symptoms before surgery and at 6-month follow-up quantified according to the New York Heart Association classification (NYHA). (B) Angina symptoms according to the Canadian Cardiovascular Society (CCS).
CABG surgery led to an improvement of quality of life as assessed by Minnesota Living with Heart Failure Score at 6 months post-operatively, with no significant difference between groups (placebo vs. CD133: 12.7 (3.5–21.9) vs. 2.4 (−5.0 to 9.8); time effect P = 0.001; group effect P = 0.08, interaction P = 0.1).
Results
Patient Characteristics
Fifty-seven patients were male and three were female. The mean age was 62.7 ± 10.6 years in the placebo group and 61.9 ± 7.3 years in the CD133 group, and all other pre-operative characteristics were also similar between groups. The last documented myocardial infarction event was 2.3 months ago (median, range: 21 days to 28.5 years), and the average LVEF was 26.2 ± 5.6% (range: 15–35%) (Table 2). Nine patients in the CD133 group and nine in the placebo group had the last documented myocardial infarction <3 months before surgery. The decision for surgical re-vascularization was made in accordance with the guidelines on the management of stable angina pectoris of the European Society of Cardiology. Fifty-six patients had three-vessel disease and 4 had two-vessel disease. After surgery, re-vascularization was incomplete in five patients. In three of those, an occluded right coronary artery could not be grafted and, in two, the occluded LAD was not graftable. Two of the incompletely re-vascularized patients were in the CD133 group and three were in the placebo group. By gadolinium first-pass imaging, the proportion of LV myocardial segments with perfusion deficit at rest was similar between the groups (placebo, 23%; CD133, 23%).
Primary Outcome
Six-month MRI follow-up was performed in 48 patients, 22 patients in the placebo group and 26 in the CD133 group. Three patients refused the examination, two had died early post-operatively, one developed dyspnoea in the MRI scanner, and two patients had post-operative pacemaker implantation. At 6-month follow-up, ejection fraction at rest as determined by MRI was 31 ± 7% in the CD133 group and 33 ± 8% in the placebo group (P = 0.3), with an average inter-group difference of −2.1% (95% CI −6.3 to 2.1).
Secondary Outcomes
Mortality Thirty-day mortality was 3.3%, with two deaths in the placebo group and none in the CD133 group. In one patient with pre-operative LVEF of 20%, the LV failed despite intra-aortic balloon pump and catecholamine treatment, and implantation of a mechanical assist device was necessary. Later, the patient developed Candida sepsis with and died on POD 8. The other patient developed pneumonia, mediastinitis, and sepsis, and died on POD 22 although LVEF had increased from 25 to 35%. Beyond the follow-up period, two more patients died 31 and 34 months post-operatively, one in the CD133 and one in the placebo group. The causes of death were given as heart failure and/or sudden cardiac death but autopsies were not performed.
Adverse Events The similarity of AE and SAE frequency between the groups demonstrates the safety of the cell treatment Table 3. The overall high incidence of AE reflects the invasiveness of CABG surgery in patients with severely impaired heart function.
Magnetic Resonance Imaging The pre-operative LV function parameters were similar between the groups. Overall, global LV function improved after surgery (Figure 2). However, in line with the LVEF data (see above), there were no significant between-group differences in LVEDD or LVEDV measured at 6-month follow-up (placebo vs. CD133: 59 ± 7 vs. 61 ± 6 mm, P = 0.4; 218 ± 50 vs. 224 ± 57 mL, P = 0.7). Similarly, there was no significant difference in the change of LVEF (2.3% (−2.0 to 6.6), P = 0.3) (Figure 2D). Per qualitative assessment of regional contractile function in the placebo group, akinesia was documented in 97 LV regions pre-operatively and 72 regions post-operatively. It improved in 42 regions, worsened in 17. In the CD133 group, this pattern was similar (106 akinetic regions pre-operatively, 84 post-operative, improvement in 39 (P = 0.8), worsening in 17 (P = 0.9). The proportions of LV myocardial segments with a perfusion deficit that changed after surgery were different between groups. In placebo patients, perfusion deficit at rest was unchanged in 19% of all analysed segments, better in 2% and worse in 7%. In the CD133 group, the perfusion deficit was unchanged in 15% (P = 0.2 vs. placebo), improved in 9% (P = 0.001, OR 4.0, 95% CI 1.6–9.8), and worse in 1% of all analysed segments (P = 0.0007, OR 0.2, 95% CI 0.07–0.5). However, the perfusion improvement in the CD133 group did not always correlate with the location of the cell injection. Scar size was not a pre-specified outcome but was quantified post hoc. Pre-operative scar mass was 27 ± 15 g (20 ± 10% of the total myocardial mass) in the placebo group and 32 ± 16 g (24 ± 10%) in the CD133 group. At 6-month follow-up, scar mass was 27 ± 15 (22 ± 11%) in the placebo group and 30 ± 17 (23 ± 10%) in the CD133 group (inter-group P = 0.5 (mass) and P = 0.7 (%)). On average, scar mass had decreased by 2.2 ± 5 g in CD133 patients (P = 0.05), but was unchanged in the placebo group (0.3 ± 4 g, P = 0.7). In relation to the pre-operative measurements, scar mass decreased by 9.7% (IQR −20 to 2%) in CD133 patients, and by 0.6% (IQR −15 to 12%) in placebo patients (inter-group P = 0.1) (Figure 3). The between-group difference of the change in scar mass at follow-up was 2.0 g (95% CI −1.1 to 5.0), P = 0.2.
(Enlarge Image)
Figure 2.
Left ventricular function by MRI: (A–C) LV function and dimension parameters measured by cardiac magnetic resonance imaging (MRI) before surgery and at 6-month follow-up. (A) Left ventricular ejection fraction (LVEF): time effect P < 0.001; group effect P = 0.5, interaction P = 0.3. (B) Left ventricular end-diastolic diameter (LVEDD): time effect P < 0.001; group effect P = 0.6, interaction P = 0.2; (C) left ventricular end-diastolic volume (LVEDV): time effect P < 0.001; group effect P = 0.9, interaction P = 0.4. (D) Change in LVEF 6 months after CABG surgery and CD133 or placebo injection, compared with pre-operative.
(Enlarge Image)
Figure 3.
Infarct scar size: late gadolinium enhancement images in the short-axis orientation of the heart at baseline (A) and 6 months after injection of CD133 bone marrow stem cells (B) demonstrating a slight reduction in scar size (−4 g or 6.3% of total LV myocardial mass in this patient). Scar tissue appears as areas of hyperintensity (bright signal) whereas viable myocardium appears dark. (C) Change in scar size from between pre-operative and 6-month follow-up expressed in % of the pre-operative scar mass.
Echocardiography
Again, most global LV function and dimension parameters improved after surgery irrespective of the group assignment, but there were no significant differences between groups at follow-up (Figure 4). Regarding regional contractility, there were no differences in the average 2D strain between groups, pre-operatively or at follow-up (pre-specified secondary endpoint). When the speckle-tracking echocardiography data were stratified according to the location of the AOI (post hoc subgroup analysis), we observed a tendency towards better improvement of regional LV function in patients with posterior or inferior infarction and CD133 cell injection, which was not apparent in the placebo group (group effect P = 0.1, Figure 5A). When the proportions of segments with improved longitudinal strain were compared between the groups, regional contractile function recovery was better in the CD133 group at 3 months (P = 0.03, OR = 2.2 95% CI 1.1–4.6) but not at 6 months (P = 0.4). Figure 5B summarizes the 2D longitudinal strain data in CD133 patients according to the location of the AOI and illustrates that regional contractile function recovered only when the AOI was posterior or inferior (localization effect P = 0.02). However, there was no correlation between AOI location and LVEF at 6 months or the change in global LV function.
(Enlarge Image)
Figure 4.
Left ventricular function by echocardiography: global left ventricular function and dimension parameters measured by echocardiography before surgery and at 3- and 6-month follow-ups. (A) Left ventricular ejection fraction (LVEF): time effect P < 0.001, group effect P = 0.9, interaction P = 0.8; (B) fractional shortening (LV-FS): time effect P = 0.01, group effect P = 0.7, interaction P = 0.08; (C) end-diastolic diameter (LVEDD): time effect P = 0.001, group effect P = 0.9, interaction P = 0.8; (D) end-systolic diameter (LVESD): time effect P = 0.001; group effect P = 0.7, interaction P = 0.7. (E) End-diastolic volume (LVEDV): time effect P = 0.1, group effect P = 1, interaction P = 0.6; (F) end-systolic volume (LVESV): time effect P = 0.001, group effect P = 0.6, interaction P = 0.9.
(Enlarge Image)
Figure 5.
Regional left ventricular function: (A) change in regional longitudinal strain between pre-operative and 3-month follow-up in the area of interest that received CD133 or placebo injection stratified according to the localization of the area of interest. Recovery of myocardial function was better in the CD133 group when the area of interest was posterior or inferior (time effect P = 0.2; group effect P = 0.1, interaction P = 0.08. (B) Longitudinal strain data measured by speckle tracking echocardiography in CD133 patients before surgery and at 3- and 6-month follow-ups, stratified according to the localization of the area of interest. Time effect P = 0.5; localization effect P = 0.02, interaction P = 1.
Exercise Capacity At 6-month follow-up, 50 patients underwent exercise testing (placebo vs. CD133: 25 vs. 25 patients). By spiroergometry, all tested patients reached the anaerobic threshold of RER > 1.01. Prior to the operation peak oxygen uptake (VO2 max) and ventilation perfusion mismatch (VE/VCO2-slope) were 12.6 ± 4.6 mL/min/kg and 30.3 ± 6 in the placebo group and 14 ± 3.7 mL/min/kg and 33.4 ± 6.3 in the CD133 group. At 6-month follow-up testing VO2 max and VE/VCO2-slope were not significant different between groups (placebo vs. CD133: 14.4 ± 4.3 vs. 14.9 ± 6 mL/min/kg (P = 0.7) and 31.5 ± 6.8 vs. 31.3 ± 5.6 (P = 0.9)). Although the mean and median values of the 6-min walk at follow-up (Figure 6) or their change from pre-operatively (placebo vs. CD133: 28.7 (−15.9 to 73.3) vs. −2.3 (−35.3 to 34.8 m), P = 0.3)) were not significantly different, the data spread was much larger in the CD133 group than in the placebo group (Figure 6). While range was similar (CD133, 120–661 m; placebo, 153–600 m), the inter-quartile range was approximately three times greater, indicating a larger inter-individual variability in the CD133 group.
(Enlarge Image)
Figure 6.
Exercise capacity: exercise capacity assessed my 6-min walking distance before surgery and at 6-month follow-up. Time effect P = 0.3; group effect P = 0.3, interaction P = 0.8.
New York Heart Association and Canadian Cardiovascular Society class
The majority of patients had less heart failure symptoms and angina at 6-month follow-up (NYHA, P < 0.001; CCS, P = 0.006) (Figure 7). However, inter-group comparison of the proportions of patients in the different NYHA classes revealed between-group differences at 6 months, suggesting that NYHA class had improved more in the placebo group than in the CD133 group (P = 0.004). For CCS class, this was not the case (P = 0.2).
(Enlarge Image)
Figure 7.
Symptoms: (A) severity of heart failure symptoms before surgery and at 6-month follow-up quantified according to the New York Heart Association classification (NYHA). (B) Angina symptoms according to the Canadian Cardiovascular Society (CCS).
Quality of Life
CABG surgery led to an improvement of quality of life as assessed by Minnesota Living with Heart Failure Score at 6 months post-operatively, with no significant difference between groups (placebo vs. CD133: 12.7 (3.5–21.9) vs. 2.4 (−5.0 to 9.8); time effect P = 0.001; group effect P = 0.08, interaction P = 0.1).
