Articles. For personal use. Only reproduce with permission from The Lancet. Summary

Transcription

1 Articles Effects of intracoronary infusion of peripheral blood stem-cells mobilised with granulocyte-colony stimulating factor on left ventricular systolic function and restenosis after coronary stenting in myocardial infarction: the MAGIC cell randomised clinical trial Hyun-Jae Kang, Hyo-Soo Kim, Shu-Ying Zhang, Kyung-Woo Park, Hyun-Jai Cho, Bon-Kwon Koo, Yong-Jin Kim, Dong Soo Lee, Dae-Won Sohn, Kyou-Sup Han, Byung-Hee Oh, Myoung-Mook Lee, Young-Bae Park Summary Background Bone-marrow stem-cell transplantation has been shown to improve cardiac function in patients with myocardial infarction. We examined the feasibility and efficacy of granulocyte-colony stimulating factor (G-CSF) therapy and subsequent intracoronary infusion of collected peripheral blood stem-cells (PBSCs) in such patients. Methods We prospectively randomised 27 patients with myocardial infarction who underwent coronary stenting for the culprit lesion of infarction into three groups; cell infusion (n=10), G-CSF alone (n=10), and control group (n=7). Changes in left ventricular systolic function and perfusion were assessed after 6 months. By December, 2003, seven patients from the cell infusion group, three from the G-CSF group, and one from the control group had been assessed. Findings G-CSF injection and intracoronary infusion of the mobilised PBSC did not aggravate inflammation and ischaemia during the periprocedural period. Exercise capacity (mean treadmill exercise time: 450 s [SD 178] at baseline vs 578 s [168] at 6 months follow-up, p=0 004), myocardial perfusion (perfusion defect 11 6% [9 6] vs 5 3% [5 0], p=0 020) and systolic function (left ventricular ejection fraction 48 7% [8 3] vs 55 1% [7 4], p=0 005) improved significantly in patients who received cell infusion. However, we noted an unexpectedly high rate of in-stent restenosis at culprit lesion in patients who received G-CSF, and therefore we stopped enrolment. Interpretation G-CSF therapy with intracoronary infusion of PBSC showed improved cardiac function, and promoted angiogenesis in patients with myocardial infarction. However, aggravation of restenosis could be a serious problem. In future studies with G-CSF based stem-cell therapy, patients should be carefully monitored for unexpected effects. Published online March 2, See Commentary Cardiovascular Laboratory, Clinical Research Institute (H-J Kang MD, H-S Kim MD, S-Y Zhang MD, K-W Park MD, H-J Cho MD, B-K Koo MD, Y-J Kim MD, D-W Sohn MD, B-H Oh MD, M-M Lee MD, Y-B Park MD), Cardiovascular Centre (H-J Kang, H-S Kim, K-W Park, H-J Cho, B-K Koo, Y-J Kim, D-W Sohn, B-H Oh, M-M Lee, Y-B Park), and Departments of Internal Medicine (H-J Kang, H-S Kim, K-W Park, H-J Cho, B-K Koo, Y-J Kim, D-W Sohn, B-H Oh, M-M Lee, Y-B Park), Nuclear Medicine (D S Lee MD), and Laboratory Medicine (K-S Han MD), Seoul National University Hospital, Seoul, Korea Correspondence to: Dr Hyo-Soo Kim, Department of Internal Medicine, Seoul National University Hospital, 28 Yongon-dong, Chongno-gu, Seoul , Republic of Korea ( hyosoo@snu.ac.kr) Bone-marrow stem-cell transplantation has been shown to improve myocardial perfusion and systolic function in patients with myocardial infarction, 1 4 but the invasiveness of bone-marrow cell collection limits its clinical application. Mobilisation of stem cells with granulocytecolony stimulating factor (G-CSF) and stem-cell factor have been studied as alternative, less invasive approaches in mice, 5 and favourable results have brought attention to the need for such treatments to be assessed in man. In patients with haematological diseases, G-CSF is a well established stem-cell mobiliser for peripheral-blood stem-cell (PBSC) transplantation, in which CD34 is used as a marker of haemopoietic stem-cells. Likewise, most studies 1 4 of stem-cell transplantation in patients with myocardial infarction have used CD34 as a marker of stem cells for transplantation. The proven efficacy and safety of G-CSF both in healthy donors 6,7 and patients with haematological disease, 8 along with favourable results from studies of CD34+ cell transplantation in patients with myocardial infarction or ischaemia, 1 4 suggest that G-CSF based PBSC transplantation should be assessed in patients with myocardial infarction. However, the feasibility and safety of G-CSF and intracoronary infusion of PBSC mobilised by G-CSF in patients with acute and old myocardial infarction have not been studied. Additionally, no study has compared outcomes with G-CSF alone and G-CSF with additional intracoronary PBSC infusion with regard to cardiac function in patients with myocardial infarction. We aimed to test the feasibility and safety of these treatments in patients with myocardial infarction, and to assess their effectiveness with regard to improvement of cardiac function. Methods Study protocol The study was a randomised, controlled phase II clinical trial. The institutional review board of Seoul National University Hospital approved the study protocol. We obtained informed written consent from patients after explaining the procedure and risk. The overall trial profile is shown in figure 1. At least 48 h after onset of acute myocardial infarction, patients underwent coronary angiography. Patients in whom the culprit lesion of the infarct-related artery was eligible for percutaneous coronary intervention (PCI), who were free from chest pain, and showed stable vital signs for at least 24 h, were randomised into one of three groups; cell infusion, G-CSF alone, or control group, by use of a randomisation table. Randomisation was done by a blinded independent co-ordinator; after randomisation, study processes were not blinded. In accordance with the recommendations of the institutional review board, the first three patients were not randomised and were assigned to the cell infusion THE LANCET Published online March 2,

2 AMI patients with >48 h after onset of AMI and clinically stable for >24 h, OMI patients Coronary angiography Cell infusion group (n=10) G-CSF group (n=10) Control group (n=8) n=3 Apheresis Cell infusion Ongoing follow-up G-CSF injection for 4 days Thallium/sestamibi SPECT PCI (stent implantation), coronary flow reserve Dobutamine stress echcardiography, treadmill test group, to verify the safety of intracoronary cell infusion and the adequacy of target-cell number. In the groups given cell infusion and G-CSF alone, PBSC were mobilised by daily subcutaneous injections of G-CSF (Dong-A Pharmaceutical, Seoul, Korea) at 10 g/kg body weight for 4 days before PCI. After completion of G-CSF injection, all patients underwent PCI and implantation of stents for the culprit lesion of the infarct-related artery. Immediately after PCI, patients in the cell infusion group received intracoronary infusion of PBSC, which had been mobilised and collected before PCI. No placebo was given to the control group. We assessed the safety of G-CSF-based PBSC therapy on the basis of: development of major adverse cardiac events (death, new myocardial infarction, revascularisation, admittance to hospital due to aggravation of ischaemia, or heart failure); clinical status, including G-CSF related pain, dyspnoea, and chest pain; biochemical tests including creatine kinase-mb and C-reactive protein; electrocardiogram (ECG); and coronary flow reserve. The primary endpoint was change in left ventricular ejection fraction, which was measured by thallium/sestamibi gated myocardial single photon emission computed tomography (SPECT). These measurements were validated by echocardiography. The secondary endpoints were changes in myocardial perfusion, exercise capacity, and the development of major adverse cardiac events. The change in myocardial perfusion was assessed with SPECT, dobutamine stress echocardiography, and coronary flow reserve. Exercise capacity was measured by treadmill test. The sample size was determined to assess the efficacy of G-CSF and intracoronary infusion of collected PBSC with regard to improvement of left ventricular ejection fraction on the basis of results from a previous study. 1 Exclusion (n=1) n=7 n=7 n=3 n=6 n=1 Ongoing follow-up 6 months' follow-up assessment: coronary angiography, coronary flow reserve, thallium/sestamibi SPECT, echocardiography, treadmill test Figure 1: Study protocol AMI=acute myocardial infarction. OMI=old myocardial infarction. Ongoing follow-up Detection of a difference of 7% in left ventricular ejection fraction, with an 80% power and an error of 5 %, would require 48 patients (16 patients for each study group). We adjusted the sample size for an estimated 25% loss of follow-up, which resulted in 20 patients in each group and a total sample size of 60 patients. Participants Our study was initiated in February, 2003, at Seoul National University. By December, 2003, we had prospectively enrolled 28 patients with acute and old myocardial infarction who underwent elective PCI for the culprit lesion of infarction. One patient in the control group was excluded from the study because he underwent elective coronary artery bypass graft surgery due to unsuccessful PCI in the culprit lesion. At present, 11 patients have completed the 6 months follow-up assessment; seven from the cell infusion group, three from the G-CSF group, and one from the control group. We stopped enrolment of additional patients due to the unexpectedly high rate of in-stent restenosis at the culprit lesion in patients treated with G-CSF and only followed up the patients already enrolled. Exclusion criteria were: persistent severe heart failure (greater than Killip class II or left ventricular ejection fraction <25%); uncontrolled myocardial ischaemia or ventricular tachycardia; culprit lesion of infarct related artery not feasible for PCI; age older than 75 years; malignant disease; serious current infection or haematological disease; and life expectancy less than 1 year. Procedures On the day of PCI, we obtained PBSC from patients in the cell infusion group by use of COBE spectra apheresis system (COBE BCT Inc, Lakewood, CO, USA) with mononuclear cell collection methods as recommended in the manual. 9 PBSCs were preserved at 37 C in a shaking chamber before infusion. We infused PBSC by over-the-wire angioplasty balloon catheter after PCI. Additional cell preparation was not done, to minimise risk of contamination before cell infusion. The lowest CD34+ cell yield from the first three patients was 0 7% of whole leucocytes obtained by apheresis. The infusion cell doses in the first three patients were leucocytes per patient. To guarantee the minimum target cell dose of CD34+ cells for infusion, we used mononuclear cells as the infusion cell dose in the rest of the cell infusion group. Every infusion was done within 1 hour of PBSC collection. Before intracoronary infusion, we administered nicorandil and nitroglycerin by coronary guiding catheter, and achieved activated clotting time of more than 250 s with intravenous bolus infusion of heparin to minimise the risk of no reflow phenomenon (reduction in antegrade epicardial bloodflow despite the absence of mechanical obstruction). 201 We did resting Tl-dipyridamole and stress 99m Tc-sestamibi gated SPECT before PCI at baseline and before coronary angiography at 6 months follow-up. Gated sestamibi projection images were reconstructed 2 THE LANCET Published online March 2,

3 Cell infusion G-CSF Controls (n=10) (n=10) (n=7) Age (years)* 59 4 (9 6) 54 4 (11 7) 52 1 (12 7) Male:female 9: 1 9: 1 7: 0 AMI:OMI 6: 4 4: 6 5: 2 Infarct-related artery (LAD:LCx:RCA) 4: 3: 3 4: 3: 3 3: 1: 3 LVEF* (%) 48 9 (9 0) 53 0 (14 3) 40 3 (8 7) Risk factors Hypertension Diabetes mellitus Hypercholesterolaemia Current cigarette smoking Medication Aspirin+clopidogrel Statin blocker inhibitor/at-ii receptor blocker Data are number of patients unless otherwise indicated. *Mean (SD). AMI=acute myocardial infarction. OMI=old myocardial infarction. =angiotensin-converting enzyme. AT-II=angiotensin-II. LAD=left anterior descending artery. LCx=left circumflex artery. RCA=right coronary artery. LVEF=left ventricular ejection fraction. Table 1: Baseline characteristics of patients enrolled and successfully revascularised percutaneously and ventricular volumes and ejection fraction were calculated with Cedars quantitative gated single photon emission CT software (Philips Medical Systems, Milpitas, CA, USA). 10,11 We measured coronary flow reserve of the infarctrelated artery with the use of coronary Doppler wire, positioned just distal to the implanted stent in the culprit lesion. The patients in G-CSF and control groups underwent measurement of coronary flow reserve immediately after PCI. In the cell infusion group, coronary flow reserve of the infarct-related artery was measured immediately before and after intracoronary cell infusion. Intracoronary bolus injections of 18 g adenosine for left coronary artery, and 12 g for right coronary artery were used to induce hyperaemia. Low-dose dobutamine stress echocardiography was done the day after PCI, to assess viable and hibernating myocardium. We assessed regional wall motion at rest and during infusion of dobutamine at doses of 5, 10, and 20 g/kg per min with echocardiography (Sequoia, Siemens, Mountain View, CA, USA). We assessed regional left ventricular wall motion as described by the committee on standards of the American Society of Echocardiography. 12 The wall motion score index was calculated as the average sum of wall motion scores at rest and low dose dobutamine echocardiography. At baseline echocardiography, left ventricular volume and left ventricular ejection fraction which were calculated by biplane modified Simpson s method were measured and verified by two experienced observers. Symptom-limited treadmill test was done with a modified Bruce protocol. Anti-anginal medications were not discontinued during tests. Duration of exercise, symptoms, ECG changes, and presence of substantial arrhythmia were recorded. None of the patients had noncardiac illness that limited treadmill exercise capacity. Quantitative coronary angiography was done at baseline, immediately after PCI, and at 6 months follow-up by an independent specialist with Quantcor QCA V4.0 program (Pie medical imaging, Maastricht, Netherlands). Binary Sex Age Days Risk factors IRA Drugs WBC CRP Flow cytometry (%) LVEF LVESV WMSI WMSI At 6 months follow-up from (cells (mg/ of collected PBSC (%) (ml) at with HT DM Chol Cig LVEF LVESV WMSI Reste- AMI per L) CD 34 AC 133 CD 31 base- low- (%) (ml) nosis ml) line dose dobutamine Cell infusion Patient 1 M LCx APA, BB, Patient 2 F Ex LAD APA, BB, < Patient 3 M RCA APA, BB, < Patient 4 M Ex LCx APA, BB, NA NA NA Patient 5 M LCx APA, BB, < Patient 6 M LAD APA, BB, Patient 7 M RCA APA, BB, G-CSF only Patient 8 M LAD APA, BB, Patient 9 M LAD APA, BB, Patient 10 M Ex LAD APA, BB, NA + Control Patient 11 M LAD APA, BB, M=male. F=female. AMI=acute myocardial infarction. HT=hypertension. DM=diabetes mellitus. Chol=hypercholesterolaemia. Cig=current cigarette smoker. Ex=ex-smoker. IRA=infarct-related coronary artery. LCx=left circumflex artery. LAD=left anterior descending artery. RCA=right coronary artery. APA=aspirin+clopidogrel. BB= blocker. =angiotensin-converting enzyme inhibitor or angiotensin receptor blocker. = Co A inhibitor. WBC=baseline peripheral leucocyte counts. CRP=baseline C-reactive protein. NA=not available. LVEF=left ventricular ejection fraction. LVESV=left ventricular end systolic volume. WMSI=wall motion score index. Table 2: Clinical characteristics at baseline and 6 months follow-up THE LANCET Published online March 2,

4 restenosis was defined as more than 50% loss of diameter. All patients underwent follow-up visits at 1, 2, and 4 months after enrolment and were assessed for clinical status, safety, adverse reactions, and medication. At 6 months after enrolment, we did follow-up coronary angiography and measurements of coronary flow reserve in infarct-related arteries. If the culprit lesion was complicated with restenosis at follow-up coronary angiography, elective PCI was done at the investigators decision. SPECT was done before follow-up coronary angiography. Stress echocardiography and the treadmill test were also repeated at follow-up visits. The planned duration of follow-up was 1 year. Statistical analysis Analyses for safety were based on the intention-to-treat principle (27 patients) and those of efficacy were based on the per-protocol principle (11 patients). Data are presented as mean (SD). We used a paired t test for comparison of continuous variables. The correlation between variables was analysed with the Pearson test. Statistical significance was assumed at a value of p<0 05. Statistical analysis was done with SPSS version A A 70 p=0 005 Stress 60 Rest Stent Restenosis 50 LVEF (%) At baseline At 6 month follow-up 20 B B Baseline p=0 050 At 6 months Proportional change in LVEF (%) y= x r=0 75 p= LVESV (%) Increases in plaque volume ( L) Baseline At 6 months Figure 2: Changes in (A) left ventricular ejection fraction (LVEF), and (B) left ventricular end systolic volume (LVESV) between baseline and 6 month follow-up in cell infusion group LVEF and LVESV improved in the cell infusion group but not in the group given G-CSF alone. WMSI=wall motion score index. Figure 3: Effects of PBSCs on myocardial perfusion and restenosis (A) and correlation between changes in left ventricular systolic function and plaque volume (B) (A) Coronary flow reserve at infarcte region (upper row) improved from 1 4 at baseline (left) to 2 4, 6 months after cell infusion (right), showing improvement of microvascular dysfunction. Large fixed perfusion defect of inferior left ventricular wall at baseline (middle row left; thallium/sestamibi SPECT showed improvement at rest 6 months after cell infusion (middle row right). Mild reversible perfusion defect remained in stress images at 6 months (middle row right), related to restenosis. Coronary angiography at baseline (bottom row left) showed good immediate results after stenting of culprit lesion of infarction (arrow) at proximal right coronary artery. Coronary angiography (bottom row right) at 6 months after cell infusion, however, showed diffuse in-stent restenosis (arrow). (B) In the cell infusion group, proportional changes in left ventricular ejection fraction closely correlated with increases in neointimal volume at stented segment. 4 THE LANCET Published online March 2,

5 Role of the funding sources The sponsors of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. Results The characteristics of the enrolled and percutaneously revascularised patients are shown in table 1. Because enrolment was prematurely terminated, some baseline parameters were not similar among treatment groups. The specific details of the patients who completed 6 months of follow-up are presented in table 2. Periprocedural safety results were assessed with the results of the patients who underwent PCI successfully. No serious adverse reactions related to G-CSF administration arose during the periprocedural period. Only three patients complained of headache, which was easily controlled with analgesics and completely relieved with discontinuation of G-CSF. During G-CSF injection and hospital stay, we did not observe aggravation of angina or any ECG changes suggesting ischaemia or substantial arrhythmia. G-CSF successfully mobilised PBSC without any significant changes in inflammation as indicated by C-reactive protein (baseline mean 21 mg/l [SD 27] vs 15 mg/l [13] after G- CSF, p=0 553). No thrombotic complications were noted during G-CSF injection and the periprocedural period. Intracoronary cell infusion resulted in a mild elevation of cardiac enzyme (creatine kinase-mb before PCI 3 4 IU/L [3 0] vs 5 6 [4 4] IU/L 12 h after infusion, p=0 012), but there were no symptoms or signs of significant ischaemia or arrhythmia during and after cell infusion. Furthermore, cell infusion did not disturb coronary microcirculation assessed with coronary angiography and coronary flow reserve (before infusion 1 5 [0 2] vs after infusion 1 7 [0 5], p=0 225). We collected 9 0 [3 0] 10 9 leucocytes (mononuclear cells: 75 1% [12 5], volume: 53 0 ml [18 7]) during apheresis. We infused 1 5 [0 5] 10 9 collected leucocytes (volume: 9 0 ml [3 0]) by over-the-wire balloon angioplasty catheter after PCI to patients in the cell infusion group. The infused cells contained 8 3% [10 2] CD34+ cells. The remains of the collected PBSC were returned to patients by intravenous continuous infusion. No deaths, aggravation of heart failure or angina, or substantial arrhythmias occurred during follow-up. At 6 months follow-up, the patients in the cell infusion group showed improved functional capacity and cardiac function. Duration of symptom-limited treadmill exercise increased from 450 s (178) to 578 s (168; p=0 004). At baseline, after successful revascularisation of the culprit lesion, three of seven patients showed asymptomatic significant ST depressions during the treadmill test. At follow-up, however, there were no chest pains, significant ST changes, or substantial arrhythmias during the treadmill test. Since medications were not changed during the follow-up, this improvement suggested that intracoronary cell infusion relieved silent ischaemia that was probably due to microvascular insufficiency. At 6 months follow-up in the cell infusion group we noted improvements in left ventricular ejection fraction (p=0 005, figure 2A) and reductions of left ventricular end systolic volume (p=0 050, figure 2B), as measured by SPECT. In the G-CSF alone group, however, improvements of left ventricular ejection fraction were not evident (45 4% [14 4] at baseline vs 44 0% [16 0] at 6 months follow-up, p=0 707). This trend was validated with echocardiography, and the measurements from both methods were well correlated (r=0 933, p=0 002). We assessed wall motion score index during rest and lowdose dobutamine infusion at baseline stress echocardiography. We noted no significant improvements in this index during low-dose dobutamine infusion from resting values in the cell infusion group (p=0 356, table 2), which suggested that there were no substantial portions of hibernating myocardium. Thus, these results suggested that the improvement of left ventricular systolic function at 6 months was beyond the effect of recovery from hibernating or stunned myocardium due to successful revascularisation, and that it might be due partly to angiogenesis and myocardial regeneration induced by stem cells. In the cell infusion group, regions of hypoperfused myocardium measured by SPECT were reduced at 6 months follow-up from 11 6% (9 6) to 5 3% (5 0; p=0 020, figure 3A). Coronary flow reserve also improved from the baseline value of 1 5 (0 2) to 2 6 (1 0) at 6 months (p=0 072). However, these improvements in perfusion were not observed in the G-CSF group (11 7% [18 5] hypoperfused myocardium at baseline vs 10 7% [17 6] at 6 months, p=0 423). Improvements in perfusion could be attributable to angiogenesis by infused stem cells. In addition to these positive results, we also noted an unexpectedly high rate of in-stent restenosis at the culprit lesion of patients who were treated with G-CSF (five of seven patients in the cell infusion group, two of three patients in the G-CSF group, figure 3A). Stents in the culprit lesion had a diameter of 3 0 mm (0 6), and a length of 22 8 mm (6 3). Minimum luminal diameter and length of in-stent restenosis lesion were 1 0 mm (0 6) and 21 1 mm (3 9), respectively. Furthermore, we found a close correlation between gain of neointimal volume and improvements of systolic function in the cell infusion group (figure 3B), suggesting that stem-cell therapy accelerated neointimal growth in proportion to the improvement of systolic function. Discussion Our findings showed good periprocedural results and shortterm safety in patients with myocardial infarction given G-CSF alone or with intracoronary infusion of PBSC. G-CSF did not aggravate inflammation nor ischaemia. By contrast, Hill and colleagues 13 reported that, in 12 patients with intractable angina, the administration of G-CSF was associated with two cases of acute myocardial infarction and one cardiac death. Experience in the use of G-CSF in patients with myocardial infarction is limited, but another study 14 did not report any serious adverse effects of G-CSF therapy. The reason for the discrepancy between our results and those of Hill and colleagues is unclear. However, one major difference between the two studies is the characteristics of the patients enrolled in the study. Hill and colleagues enrolled high-risk patients with intractable angina and coronary lesions inappropriate for further revascularisation, whereas we enrolled patients with myocardial infarction but with relatively stable symptoms, and arterial lesions well capable of being revascularised. Therefore, all of the patients in our study received complete revascularisation, unlike those in the study by Hill and colleagues. This differences in revascularisation and in the underlying coronary bed could have resulted in the difference between the results of the two studies. We infused collected leucocytes without additional preparation by the intracoronary route. Infused leucocytes mainly consisted of mononuclear cells, which contained a mean of 1 4 (2 0) 10 8 CD34+ cells, but also many neutrophils, and platelets. The effectiveness of non-selected bone-marrow mononuclear-cells in improving cardiac function has been shown in previous studies, 1 3 but the efficacy of apheresis products rich in mononuclear cells had THE LANCET Published online March 2,

6 not been previously reported. The possibility that infused inflammatory cells and platelets could aggregate and disturb the microcirculation was a major concern before the start of this trial. We noted mild elevations of creatine kinase-mb, within normal reference values for our institution, with intracoronary cell infusion. However, we were unable to detect any symptom or sign of aggravation of ischaemia. Coronary flow reserve and coronary angiography did not show any deterioration of coronary circulation. From these results, we judged that intracoronary infusion of nonselected apheresis products was a feasible and safe method to deliver PBSC to the target myocardium. Another possible concern with mobilised PBSC infusion was that composition of infused PBSC is different from that of bone-marrow stem-cells. Because we do not know which stem-cell population is the true effector-cell portion that improves cardiac function, we used CD34 as a stemcell marker (as have most other studies) and used whole leucocytes for infusion. Despite these concerns before the start of this trial, improvement of left ventricular ejection fraction in the PBSC infusion group was similar to those of previous studies that used autologous bone-marrow mononuclear cells. 1 In addition to the favourable results, we also noted an unexpectedly high rate of restenosis. Although PBSC infusion improved cardiac function and perfusion, aggravation of restenosis can be a serious problem. One explanation for this restenosis might be the differentiation of progenitor cells into smooth muscle cells within the stented segment. 18 As Sata and colleagues 18 showed, stem cells might contribute to the pathological arterial remodelling by differentiating into smooth muscle cells. Another explanation might be the angiogenesis induced within the atherosclerotic lesion and the aggregation of mobilised inflammatory cells within the plaque, as we found in an atherectomy specimen from a restenosis lesion. In view of the high incidence of in-stent restenosis that we noted, the improvement of cardiac function with stem-cell therapy might be greater if such restenosis could be prevented. We believe that, in order to proceed with further studies using G-CSF based stem-cell therapy, aggressive anti-restenosis measures such as use of drug-eluting stents should be considered. Although safety and outcome of this intervention in acute myocardial infarction and restenosis related to G-CSF need further assessment, adoption of drug-eluting stents seems justifiable. These stents have been shown to exert potent antiproliferative effects on smooth muscle proliferation resulting in a favourable outcome against restenosis. This effect might be extended to G-CSF related restenosis that had a neointima mainly consisting of cellular proliferation and migration. Furthermore, although G-CSF might accelerate growth of neointima with bare metal stents, it might be beneficial with drug-eluting stents in promoting re-endothelialisation, resulting in significant reduction of inflammation and restenosis, as we have shown in intracoronary radiation therapy. 19 The possibility of an unexpected effect of stem-cell mobilisation and therapy should be monitored even more carefully in future studies. In conclusion, the short-term safety and the absence of periprocedural adverse events suggest that G-CSF-based PBSC therapy is a feasible treatment for myocardial infarction. Despite the favourable effects on cardiac function, however, our data warrant a more cautious approach to stem-cell therapy, in view of the possible aggravation of restenosis. Contributors H-J Kang contributed to study design, data analysis, and writing of the paper. H-S Kim, B-K Koo, and M-M Lee enrolled patients, contributed to the study design and draft of the paper, and performed invasive procedures. S-Y Zhang and K-S Han contributed to the collection, preparation, and analysis of peripheral-blood stem-cells. Y-J Kim, D S Lee, and D-W Sohn did noninvasive assessment of patients. K-W Park, H-J Cho, B-H Oh, and Y-B Park contributed to the study design, interpretation of data, enrolment of patients, and clinical follow-up. Conflict of interest statement None declared. Acknowledgments This study was supported by a grant from the Korea Health 21 R&D project, Ministry of Health and Welfare, Republic of Korea (02-PJ10-PG8- EC ; H-S Kim), a grant from KOSEF (Korea Science and Engineering Foundation) through the Ageing and Apoptosis Research Centre at Seoul National University, Seoul, Republic of Korea (H-S Kim) and a grant from Stem Cell Research Centre, Republic of Korea (M102KL K ; Y-B Park). References 1 Assmus B, Schächinger V, Teupe C, et al. Transplantation of progenitor cells and regeneration enhancement in acute myocardial infarction (TOPCARE-AMI). Circulation 2002; 106: Strauer BE, Brehm M, Zeus T, et al. Repair of infarcted myocardium by autologous intracoronary mononuclear bone marrow cell transplantation in humans. Circulation 2002; 106: Tse HF, Kwong YL, Chan JKF, Lo G, Ho CL, Lau CP. Angiogenesis in ischaemic myocardium by intramyocardial autologous bone marrow mononuclear cell implantation. Lancet 2003; 361: Perin EC, Dohmann HFR, Borojevic R, et al. Transendocardial, autologous bone marrow cell transplantation for severe, chronic ischemic heart failure. Circulation 2003; 107: Orlic D, Kajstura J, Chimenti S, et al. Mobilized bone marrow cells repair the infarcted heart, improving function and survival. Proc Natl Acad Sci USA 2001; 98: Anderlini P, Przepiorka D, Champlin R, Korbling M. Biologic and clinical effects of granulocyte colony-stimulating factor in normal individuals. Blood 1996; 88: Cavallaro AM, Lilleby K, Majolino I, et al. Three to six year follow-up of normal donors who received recombinant human granulocyte colonystimulating factor. Bone Marrow Transplant 2000; 25: Bensinger W, Singer J, Appelbaum F, et al. Autologous transplantation with peripheral blood mononuclear cells collected after administration of recombinant granulocyte stimulating factor. Blood 1993; 81: The COBE Spectra Apheresis System Operator s Manual. Lakewood: COBE BCT, Germano G, Kiat H, Kavanagh PB, et al. Automatic quantification of ejection fraction from gated myocardial perfusion SPECT. J Nucl Med 1995; 36: Lee DS, Cheon GJ, Ahn JY, Chung JK, Lee MC. Reproducibility of assessment of myocardial function using gated 99Tc(m)-MIBI SPECT and quantitative software. Nucl Med Commun 2000; 21: Schiller NB, Shah PM, Crawford M, et al. Recommendations for quantitation of the left ventricle by two-dimensional echocardiography: American Society of Echocardiography Committee on Standards, Subcommittee on Quantitation of Two-Dimensional Echocardiograms. J Am Soc Echocardiogr 1989; 2: Hill JM, Paul JD, Powell TM, et al. Efficacy and risk of granulocyte colony stimulating factor administration in patients with severe coronary artery disease. Circulation 2003; 108 (suppl IV): 478 (abstr). 14 Ellis SG. Mobilizing stem cells to treat acute myocardial infarction: The first results from RECOVER pilot. Hearing on transcatheter cardiovascular therapeutics 2003, The Molecular Cardiology Symposium 1: Myogenesis, (September 15, 2003). 15 Wexler SA, Donaldson C, Denning-Kendall P, Rice C, Bradley B, Hows JM. Adult bone marrow is a rich source of human mesenchymal stem cells but umbilical cord and mobilized adult blood are not. Br J Haematol 2003; 121: Theilgaard-Monch K, Raaschou-Jensen K, Palm H, et al. Flow cytometric assessment of lymphocyte subsets, lymphoid progenitors, and hematopoietic stem cells in allogeneic stem cell grafts. Bone Marrow Transplant 2001; 28: Fritsch G, Stimpfl M, Kurz M, et al. The composition of CD34 subpopulations differs between bone marrow, blood and cord blood. Bone Marrow Transplant 1996; 17: Sata M, Saiura A, Kunisato A, et al. Hematopoietic stem cells differentiate into vascular cells that participate in the pathogenesis of atherosclerosis. Nat Med 2002; 8: Cho HJ, Kim HS, Lee MM, et al. Mobilized endothelial progenitor cells by granulocyte-macrophage colony-stimulating factor accelerate reendothelialization and reduce vascular inflammation after intravascular radiation. Circulation 2003; 108: THE LANCET Published online March 2,