Introduction

Severe limb ischemia, due to advanced peripheral artery disease (PAD), is associated with a high risk of cardiovascular events and all-cause mortality^1–4 and a poor prognosis with respect to limb preservation.^5,6 In nonrevascularizable, so-called no-option patients with severe limb ischemia, 6-month major amputation rates have been reported to range from 10% to 40%.^3,4 Severe limb ischemia is associated with poor quality of life (QoL)⁷ and high treatment costs,⁸ especially when amputation is inevitable.^8,9 With an estimated annual incidence of 500 to 1000 new cases per million individuals in Western society,³ which is ever increasing in concert with the increase in cardiovascular risk factors,^10–12 severe limb ischemia poses a substantial burden on patients, healthcare providers, and resources; hence, new treatment modalities are urgently needed.

Clinical Perspective on p 860

Cell-based regenerative therapies aiming at enhanced neovascularization and improved limb perfusion have been proposed as novel treatment strategies. In 2002, a small first-in-man clinical trial reported safety and promising effects of autologous bone marrow (BM)–derived cell therapy in patients with critical limb ischemia.¹³ Since then, several studies have suggested benefit of BM-derived cell therapy for advanced PAD. However, studies were small, lacked appropriate controls, often did not consider clinically relevant outcomes, and thus did not provide definite proof on clinical effectiveness.^14,15 Our recent meta-analysis on 12 randomized clinical trials (RCTs) that studied BM-derived cell therapy in a total of 510 patients with critical limb ischemia underlined the promising potential of this therapy but also showed divergent results between placebo-controlled and non–placebo-controlled RCTs, stressing the need for a large, well-designed, placebo-controlled RCT with clinically relevant outcomes.¹⁵

We designed the Rejuvenating Endothelial Progenitor Cells via Transcutaneous Intra-arterial Supplementation (JUVENTAS) trial, a double-blind, placebo-controlled RCT (NCT00371371) to investigate whether repetitive intra-arterial infusion of BM mononuclear cells (BMMNCs) reduces amputation rates in a large cohort of patients with severe, nonrevascularizable limb ischemia.¹⁶

JUVENTAS was developed and initiated in 2006, and its design was based on the available literature at that time. The rationale has been published elsewhere.¹⁶ In short, we chose the intra-arterial administration route based on preclinical data suggesting that this route allows the migration to different zones of ischemic tissue, which is particularly important in multilevel disease, and facilitates homing to ischemic tissue with preserved nutrient blood supply, leading to improved local cell survival,^17,18 and to the extrapolation of clinical data, as well, suggesting the benefit of intracoronary injection of BMMNC in myocardial infarction.^19,20 We chose to administer BMMNC obtained from 100 mL of BM, which can be obtained under local anesthesia in an outpatient setting with a low risk of complications. We adopted a repetitive infusion scheme assuming that repeated administration would enhance cell retainment based on observations that only limited numbers of cells are retained in injured tissue after injection.

Methods

Detailed methods can be found in the online-only Data Supplement.

Trial Design and Study Population

In this single-center, double-blind, placebo-controlled RCT, we investigated the clinical effects of repetitive infusion of BMMNC into the common femoral artery in 160 patients with severe, nonrevascularizable PAD included from September 2006 through June 2012 (clinicaltrials.gov NCT00371371; Figure 1). Study design and detailed inclusion and exclusion criteria are described in the Methods in the online-only Data Supplement and have been reported previously.¹⁶

• Download figure
• Download PowerPoint

The institutional review board of the University Medical Center Utrecht approved the study protocol, the study was conducted according to the Declaration of Helsinki, and all patients provided written informed consent before the study interventions.

Patients were randomly assigned by means of computerized block randomization with variable block sizes to receive either 3 repetitive intra-arterial infusions of BMMNC or placebo (3-week intervals) into the common femoral artery of the affected limb. BM aspirates (100 mL) were obtained from the right iliac crest in all patients. BMMNCs were isolated by density gradient centrifugation (Lymphoprep, Axis-Shield Inc, Oslo, Norway). For the placebo group, a placebo was prepared by using autologous peripheral blood erythrocytes to match the color of the BMMNC product. Of both products, one-third was prepared for direct infusion, and two-thirds were cryopreserved and stored for subsequent infusions. For cryopreservation of the BMMNC product, 10% dimethyl sulfoxide was added. Because dimethyl sulfoxide releases a specific odor at the time of infusion, the same amount of dimethyl sulfoxide was added to the cryopreserved placebo product to guarantee blinding of the trial staff. Syringes without information about the product were provided to the clinical staff at the time of infusion. At infusion, the product was slowly administered into the common femoral artery of the affected limb by hand injection.

Numbers of white blood cells, CD34⁺ hematopoietic progenitor cells, colony-forming unit granulocytes and monocytes, and burst-forming unit-erythroid capacity of the BM product were assessed.

Surveillance Protocol and Outcome Assessment

Primary outcome was major amputation, defined as amputation through or above the ankle joint, within 6 months after randomization. The combined safety outcome was all-cause mortality, occurrence of malignancy, or hospitalization due to infection. Secondary outcomes were the combined occurrence of major amputation or death, minor amputations, changes in clinical status, ulcer size, rest pain, pain-free walking distance, ankle-brachial index (ABI), transcutaneous oxygen pressure measurements (tco₂; Radiometer Medical ApS, Copenhagen, Denmark), and health-related QoL (Short Form 36 [SF-36] and EuroQoL 5D [EQ-5D]). SF-36 scores were converted to a norm-based score and the EQ-5D was converted to the Preference-Based EuroQoL Tariff as reported previously.^7,21,22 To allow comparisons with recently published clinical trials, we included previously published composite outcomes for success and failure of cell-based therapy.^23,24 Clinical evaluation by the same investigator, treadmill exercise testing, and tco₂ measurement were performed at baseline and at 2 and 6 months follow-up.

The assessment of renal and liver function–related laboratory measurements was performed at inclusion and before and after each intra-arterial infusion. During conduct of the trial, all adverse events (AEs) and serious adverse events (SAEs) were recorded and processed according to national guidelines.

Interim Analyses and Data Monitoring

An independent Data and Safety Monitoring Board evaluated the results of sequential interim analyses. Every 6 months, or after every fourth major amputation, or occurrence of a safety outcome, unblinded analyses were performed on the cumulative database. Based on the results of these analyses, the Data and Safety Monitoring Board informed the trial’s executive committee on whether the data provided sufficient evidence for either efficiency or harm of the intervention with respect to the primary outcome and on safety and advised the executive committee whether to (dis)continue the trial (see Methods in the online-only Data Supplement). During the conduct of the trial, no recommendations were made to discontinue the trial.

Sample-Size Calculation and Statistical Analyses

Estimation of the sample size of the JUVENTAS trial was based on the best available evidence at trial initiation in 2006, which reported a 6-month risk of major amputation in patients with nonrevascularizable chronic critical limb ischemia of 42%²⁵ and a 50% risk reduction of major amputation by BMMNC infusion.^26,27 Based on these assumptions, it was estimated that with a 2-sided α of 0.05 and a power of 80%, ≈110 to 160 patients should be included in the trial. For a sequential design, no fixed sample size estimate can be provided.²⁸

Continuous variables are expressed as means±standard deviation or as medians and interquartile ranges for nonnormal data. Depending on data characteristics, group differences for continuous variables were tested with the independent t test or the Mann-Whitney U test. Dichotomous variables were analyzed with the Fisher exact test. Analyses were performed in accordance with the intention-to-treat principle. Sequential analyses²⁸ of the primary outcome, ie, major amputation, and of the safety outcome were performed by using the PEST program (version 4.4),²⁹ with adjustment of the point and interval estimates for cumulative testing. Additionally, the risks for major amputation and combined major amputation and death were analyzed by the Kaplan-Meier method with the use of the log-rank test. Primary analyses were performed at 6 months of follow-up. Major amputation, mortality, and combined major amputation and death were also analyzed beyond 6 months. Data for patients were censored at the date of death, last visit, or last known to be alive. Because of the lack of power, no subgroup analyses were performed. With the exception of the sequential analysis, no adjustments for multiple statistical testing were made.³⁰

Furthermore, the results with respect to major amputation were added to our recently published meta-analysis, with an updated literature review, and reanalyzed in Review Manager (The Cochrane Collaboration, version 5.3) with the use of a random-effects model according to the methods published previously.¹⁵ Other statistical analyses were performed by using SPSS for Windows version 20.0 (SPSS Inc., Chicago, IL).

Role of Funding Sources

The sponsors of the study had no role in the study design, data collection, analysis, interpretation, or writing of the report. All authors had full access to all data in the study and had final responsibility for the decision to submit for publication.

Results

Enrollment and Baseline Patient Characteristics

One hundred sixty patients with severe, nonrevascularizable PAD were randomly assigned to repetitive intra-arterial BMMNC (n=81) or placebo (n=79) infusions. A detailed overview of the trial flow is depicted in Figure 1. The 2 trial arms were well comparable with respect to baseline characteristics and concomitant pharmacological therapy during the study (Table 1). The majority of patients had tissue loss (Rutherford stage 5 or 6; 63% in both groups). Until primary trial follow-up at 6 months, no patients were lost to follow-up. Major amputation, mortality, and combined major amputation and death were also analyzed beyond the prespecified primary trial follow-up at 6 months; because this was not a prespecified outcome, follow-up beyond 6 months was not available for all patients (n=127; 79% of the total study population, median follow-up of 9 [6–20] months, not different between groups).

Intervention

Beside a slightly higher number of CD34⁺ hematopoietic progenitor cells in the BMMNC group, the characteristics of the BM aspirate did not differ between the 2 groups (Tables 2 and 3). Additionally, the colony-forming capacity of patients’ BM was compared with that of healthy controls (n=32; median age, 32 [21–36]), which showed no significant differences for both the colony-forming unit granulocytes and monocytes and burst-forming unit-erythroid (colony-forming unit granulocytes and monocytes 250 [174–355] versus 282 [197–394], P=0.23; burst-forming unit-erythroid 149 [108–204] versus 174 [139–242], P=0.07, for patients and healthy controls, respectively).

A total of 15 patients, 6 in the BMMNC and 9 in the placebo group, did not complete the scheduled 3 infusions owing to the occurrence of a major amputation or mortality before the completion of the scheduled infusions (Figure 1). In patients who completed 3 BMMNC infusions, total numbers of 500×10⁶ (313–717) BMMNC and 8.4×10⁶ (5.8–11.9) CD34⁺ hematopoietic progenitor cells were infused.

Major Amputation, All-Cause Mortality, and Composite Outcomes

The primary outcome was not different between the 2 groups with a major amputation rate of 19% and 13% at 6 months in the BMMNC and control group (adjusted relative risk BMMNC versus placebo, 1.46; 95% confidence interval [CI], 0.62–3.42; Table 4 and Figure 2A). The combined safety outcome was not different between the groups (15% and 10%, respectively; relative risk, 1.46; 95% CI, 0.63–3.38). All-cause mortality at 6 months was also not different between the groups with 5% versus 6% (relative risk, 0.78; 95% CI, 0.22–2.80), nor was the combined risk for major amputation and death, with a 6-month risk of 23% in the BMMNC and 16% in the placebo group (relative risk, 1.43; 95% CI, 0.76–2.69; Table 4 and Figure 2B). In the BMMNC group, the number of infused cells did not relate with any of the outcome measures (data not shown). When follow-up beyond 6 months was considered, no differences with respect to major amputation, mortality, and combined major amputation and death were observed between groups (Table 4).

• Download figure
• Download PowerPoint

Meta-Analysis

When adding the results of the JUVENTAS trial and newly published RCTs^31–34 to our recently published meta-analysis,¹⁵ a subtle, but significant difference was observed for major amputation (Figure 3A) in favor of the cell therapy–treated group. However, when only the blinded placebo-controlled trials were included, no significant difference between the cell- and placebo-treated groups was observed (Figure 3B).

• Download figure
• Download PowerPoint

Hemodynamic Measurements and Ulcer Healing

Baseline measurements were comparable for ABI, tco₂, and ulcer area (Table 1). For the analyses of ABI, only patients with a reliable index at baseline were included (n=53 BMMNC group; n=51 placebo group). ABI increased in both groups during follow-up (0.11; 95% CI, 0.06–0.15 and 0.08; 95% CI, 0.02–0.13 for the BMMNC and placebo group; Table 5), without difference in ΔABI between the groups (0.03; 95% CI, –0.04 to 0.10). tco₂ also improved in the BMMNC (10.4 mm Hg; 95% CI, 4.2–16.6) and placebo group (6.7 mm Hg; 95% CI, 1.3–12.1), without difference between the groups (3.7; 95% CI, –4.4 to 11.9). In patients with tissue loss, the ulcer area tended to decrease in both groups without a difference between the groups (difference in Δulcer area at 6 months –0.1; 95% CI, –6.1 to 5.9). Complete ulcer healing at 6 months was observed in 37% of the BMMNC in comparison with 28% of the placebo group (Table I in the online-only Data Supplement). Improvement of Rutherford stage, defined as at least a decrease of 1 Rutherford stage, was observed in 36% of the BMMNC and 33% in the placebo group.

Quality of Life and Rest Pain

Throughout the study, the proportion of QoL forms completed to such an extent that total scores could be obtained ranged from 89% and 97% at inclusion to 80% and 85% at 6 months follow-up (of those alive without amputation) for the SF-36 and EQ-5D groups, respectively, with no difference between the groups at any time point. Baseline QoL scores were low on the physical components of the SF-36, ie, physical functioning, role physical, bodily pain, and general health, reflected by the low Physical Composite Score of 29.5±7.1 in the BMMNC and 30.7±7.3 in the placebo group. During follow-up, QoL scores improved over a wide range of SF-36 components, the physical components in particular, resulting in a significant increase of the Physical Composite Score in both the BMMNC, and in the placebo group, as well. No difference was observed in ΔPhysical Composite Score at 6 months between the 2 groups (–0.1; 95% CI, –3.3 to 3.1; Table 5).

The EQ-5D scores 6 months after the first infusion improved significantly in the BMMNC and placebo groups (0.21; 95% CI, 0.12–0.31 and 0.19; 95% CI, 0.08–0.31), without difference between the 2 groups (Table II in the online-only Data Supplement). The visual analog scale pain score decreased in both groups at 6 months, with the most pronounced change in visual analog scale score at 6 months in the placebo group (–2.4; 95% CI, –3.2 to –1.5; Table II in the online-only Data Supplement).

Safety

During follow-up, no effects on liver or kidney function were observed in either group. The total number of SAEs and AEs observed during the study was 213, with 108 (S)AEs and 105 (S)AEs in the BMMNC and placebo groups (difference of mean number of [S]AEs, 0.0; 95% CI, –0.31 to 0.31). The number of patients experiencing a SAE in the BMMNC and placebo groups was 38 and 34 (difference of mean number of SAE, 0.0; 95% CI, –0.30 to 0.32). All SAEs were identified as not related to the study interventions, with the exception of one, an inguinal hematoma as a result of intra-arterial infusion.

Discussion

The JUVENTAS trial is the largest double-blind, placebo-controlled RCT to study the effects of BM-derived cell administration in patients with severe PAD to date. The study shows that repetitive intra-arterial autologous BMMNC administration was not effective in reducing the incidence of the primary outcome, ie, major amputation at 6 months, and even leaves room for a potential negative effect of the intervention. During the 6-month follow-up period, improvement of secondary outcomes, including QoL, ABI, tco₂, and ulcer healing, was observed in both treatment groups without any significant differences between the groups.

Strategies for BM-derived cell therapy in patients with severe limb ischemia differ between studies in many aspects, including administration route, cell source and cell-subtype, cell dose, and single or repetitive treatments. These variables may influence study results. For the JUVENTAS trial, initiated in 2006, we chose a strategy of infusing BMMNCs obtained from a fixed amount of 100 mL of BM into the common femoral artery on 3 separate occasions. It is unlikely that the lack of benefit in our study is due to differences in cell administration or dose. Previously, potential beneficial effects have been reported by studies that used intra-arterial administration and by studies using intramuscular administration, as well. Studies that directly compared the intra-arterial versus intramuscular route did not show superiority of either route, but lacked the proper design to draw definite conclusions.^35,36 Furthermore, beneficial results have been reported after the administration of highly variable doses of BMMNC with volumes of aspirated BM ranging from 50 to 1000 mL, from which varying amounts of BMMNC and CD34⁺ cells were retrieved.¹⁵ Although some studies have suggested a dose-response relation between numbers of administered CD34⁺ cells or BMMNCs, this has not been a consistent finding.^31,37,38 In the JUVENTAS trial, no relation was found between cell numbers or cell characteristics of the administered BM product and clinical improvement. The results of the Intraarterial Progenitor Cell Transplantation of Bone Marrow Mononuclear Cells for Induction of Neovascularization in Patients With Peripheral Arterial Occlusive Disease (PROVASA) trial suggested benefit of repeated intra-arterial BMMNC administration.³⁸ This could not be confirmed in our study. It has also been suggested that assessing potential clinical benefits of cell therapy in patients with severe limb ischemia requires longer follow-up.³⁸ However, also with longer follow-up (median, 9 months), we observed no differences between groups in any of the outcome measures.

Cell-manufacturing procedures have been reported to influence BMMNC potency.^39–41 We used Lymphoprep for BMMNC isolation and included heparin in our isolation protocol, which is similar to previous studies that showed potential clinical effects.^13,38,42 Heparin has been reported to negatively affect cellular homing by impairment of the SDF-1α/CXCR4 axis.⁴⁰ We cannot exclude that the use of heparin in our isolation procedure could have reduced the potency of our product. Contamination of erythrocytes in the BMMNC product may also impair its effectiveness.³⁹ Assmus et al showed substantial impairment of the BMMNC product when >0.2×10⁹ erythrocytes/mL are present. Erythrocytes in our product were far below this threshold (0.08×10⁹/mL±0.06×10⁹/mL) and therefore unlikely to have influenced our product.

The JUVENTAS trial included mainly elderly (median age, 67 years) white patients with a high systemic atherosclerotic burden and prevalence of cardiovascular disease, a typical Western population with advanced PAD. In contrast, several initial positive studies were conducted in Asian populations, often including relatively young patients with thromboangiitis obliterans.¹⁴ Aging and comorbidities have been shown to induce functional impairment of BMMNCs which may limit the therapeutic potential of autologous BM-derived cell therapy. The lack of benefit in our patients could be partly due to BM cell dysfunction in patients with high atherosclerotic burden.⁴³

The major amputation rate in our study was lower than expected based on the available literature at the time of the initiation of the trial.²⁵ More recent data suggest that amputation rates in no-option patients with advanced PAD have decreased over the past decades, in line with the observation in our study.⁴⁴ The low number of major amputations decreased the statistical power of the study with respect to major amputation, but the general lack of benefit on secondary outcomes in comparison with placebo, support a lack of benefit of autologous BMMNCs as applied in our protocol. Moreover, our observations are in line with our recently published meta-analysis on cell therapy in critical limb ischemia.¹⁵ The inclusion of the JUVENTAS trial (and other new trial) results in the meta-analysis of all RCTs shows a minor significant beneficial effect on major amputation, whereas analysis of only the blinded placebo-controlled RCTs shows a clear lack of benefit.

The fact that, in our trial, where rigorous blinding of participants and investigators was applied, a general improvement in both BMMNC- and placebo-treated patients was observed emphasizes the essential role of accurate blinding, which, together with the large population size, is a major strength of our study. Similar observations have previously been reported for cell therapy trials for heart disease. Several recent well-designed and larger RCTs that studied BMMNC administration in both acute myocardial infarction, and in chronic heart failure, as well, failed to prove the beneficial effects of BMMNC administration on cardiac function in comparison with placebo intervention.^45,46 A recent meta-analysis showed that the observed improvement in left ventricular ejection fraction observed after BMMNC therapy in myocardial infarction disappeared when only sham-controlled studies were analyzed.⁴⁷

In conclusion, this double-blind, placebo-controlled RCT of autologous BM-derived cell therapy in patients with severe, nonrevascularizable PAD shows no effect of repetitive intra-arterial infusion of autologous BMMNC on prespecified outcomes. The lack of benefit of autologous BMMNCs in our study does not exclude the potential benefit of other cell sources or subpopulations, different administration locations and routes, or in patients with milder disease. Further study is warranted to investigate whether cell therapy strategies with selected cell populations, enhanced BM cell function, or different modes of administration can provide therapeutic benefit in patients with advanced PAD. The general improvement observed in clinical outcomes in both groups during follow-up stresses the need for future RCTs to implement a rigorous double-blinded, placebo-controlled design.

Acknowledgments

We thank Dr G.J. de Borst, Dr R.J. Toorop, Dr C.E.V.M. Hazenberg, Dr J.A. van Herwaarden, Dr P. Berger, Dr E. Waasdorp, and Dr A.Kh. Jahrome for patient recruitment, screening, and general trial support. Dr L. te Boome and Dr P.E. Westerweel are acknowledged for performing part of the bone marrow aspiration procedures. The Cell Therapy Facility of the University Medical Center Utrecht, in particular K. Westinga, is thanked for processing of the trial products. The nursing staff of the vascular surgery outpatient clinic and surgical wards of the University Medical Center Utrecht are thanked for their medical-site support. Furthermore, all referring vascular surgeons, in particular, from the Sint Antonius Hospital Nieuwegein (Dr J.P.P.M de Vries, Dr H.D.W.M van de Pavoordt, and Dr R.H.W. van de Mortel) and the Meander Medical Center Amersfoort (Dr C.H.P. Arts, Dr M.C. Loubert, Dr A.J.C. Mackaay, and Dr R. Voorhoeve) are kindly acknowledged for their support. Authors’ contributions: Drs Sprengers, Verhaar, van der Graaf, and Moll conceived and designed the study, with advise from Drs Schutgens, Slaper-Cortenbach, Doevendans, and Mali. Dr Teraa and Sprengers enrolled the patients. Drs Teraa, Sprengers, Algra, van der Tweel, van der Graaf, Schutgens, Slaper-Cortenbach, Moll, and Verhaar were involved in data acquisition, analysis, and interpretation. Drs Teraa and Verhaar wrote the first draft of the report. All authors edited and approved the report for final submission.

Footnotes