Trasplante intramuscular local de células mononucleares autólogas de médula ósea para la isquemia crítica de los miembros inferiores
Resumen
Antecedentes
La enfermedad arterial periférica es un problema importante de salud y, en alrededor del 1% al 2% de los pacientes, la enfermedad evoluciona a isquemia crítica de los miembros inferiores (ICMI). En un número considerable de personas con ICMI no se dispone de opciones terapéuticas eficaces aparte de la amputación, por lo que alrededor del 25% de estos pacientes requerirá una amputación mayor durante el siguiente año. Esta es la segunda actualización de una revisión publicada por primera vez en 2011.
Objetivos
Evaluar los efectos beneficiosos y perjudiciales del trasplante intramuscular local de células mononucleares de médula ósea (CMMO) autólogas adultas como tratamiento para la ICMI.
Métodos de búsqueda
Se utilizaron los métodos exhaustivos estándares de búsqueda de Cochrane. La última fecha de búsqueda fue el 8 de noviembre de 2021.
Criterios de selección
Se incluyeron todos los ensayos controlados aleatorizados (ECA) sobre ICMI en los que los participantes se asignaron al azar a la administración intramuscular de CMMO autólogas adultas (o ninguna intervención, tratamiento conservador convencional o placebo).
Obtención y análisis de los datos
Se utilizaron los métodos Cochrane estándares. Los desenlaces principales de interés fueron la mortalidad por todas las causas, el dolor y la amputación. Los desenlaces secundarios de la revisión fueron el análisis angiográfico, el índice tobillo‐brazo (ITB), la distancia caminada sin dolor, los efectos secundarios y las complicaciones. La certeza de la evidencia se evaluó mediante el método GRADE.
Resultados principales
Se incluyeron cuatro ECA con un total de 176 participantes con diagnóstico clínico de ICMI. Los participantes fueron asignados al azar para recibir la implantación intramuscular de CMMO o el control. Los grupos control variaron entre los estudios e incluyeron el tratamiento convencional, la sangre periférica autóloga diluida y el suero fisiológico. No hubo evidencia clara de un efecto sobre la mortalidad relacionada con la administración de CMMO en comparación con el control (razón de riesgos [RR] 1,00; intervalo de confianza [IC] del 95%: 0,15 a 6,63; tres estudios, 123 participantes; evidencia de certeza muy baja). Todos los ensayos evaluaron los cambios en la intensidad del dolor, pero los ensayos utilizaron diferentes herramientas de evaluación del dolor, por lo que no se pudieron agrupar los datos. Tres estudios afirmaron que no se observaron diferencias en la reducción del dolor entre los grupos de CMMO y control. Un estudio informó de que la reducción del dolor en reposo fue mayor en el grupo de CMMO en comparación con el grupo control (evidencia de certeza muy baja). Los cuatro ensayos informaron sobre la tasa de amputación al final del periodo de estudio. No se sabe con certeza si las amputaciones se redujeron en el grupo de CMMO en comparación con el grupo control, ya que un posible efecto pequeño (RR 0,52; IC del 95%: 0,27 a 0,99; cuatro estudios, 176 participantes; evidencia de certeza muy baja) se perdió después del análisis de sensibilidad (RR 0,52; IC del 95%: 0,19 a 1,39; dos estudios, 89 participantes). Ninguno de los estudios incluidos informó sobre un análisis angiográfico. El índice tobillo‐brazo se presentó de forma diferente en cada estudio, por lo que no fue posible agrupar los datos. Tres estudios informaron que no se produjeron cambios entre los grupos y un estudio informó sobre una mejoría mayor en el ITB (como mejoría hemodinámica) en el grupo de CMMO en comparación con el grupo control (evidencia de certeza muy baja). Un estudio informó sobre la distancia caminada sin dolor y no encontró diferencias claras entre los grupos de CMMO y control (evidencia de certeza baja). Se agruparon los datos de los efectos secundarios informados durante el seguimiento y no se observaron diferencias claras entre los grupos de CMMO y control (RR 2,13; IC del 95%: 0,50 a 8,97; cuatro estudios, 176 participantes; evidencia de certeza muy baja). La certeza de la evidencia se disminuyó debido a las preocupaciones sobre el riesgo de sesgo, la imprecisión y la inconsistencia.
Conclusiones de los autores
Se identificó un pequeño número de estudios que cumplieron los criterios de inclusión, los cuales diferían en los controles que utilizaron y en la forma de medir los desenlaces importantes. Los datos limitados de estos ensayos proporcionan evidencia de certeza muy baja a baja, por lo que no es posible establecer conclusiones que apoyen el uso del trasplante intramuscular local de CMMO para mejorar los desenlaces clínicos en personas con ICMI. Hace falta evidencia de ECA más grandes para proporcionar una potencia estadística suficiente para evaluar la función de esta intervención.
PICO
Resumen en términos sencillos
¿Tratar el flujo sanguíneo reducido en las piernas con células extraídas de la médula ósea ayuda a mejorar los síntomas?
¿Por qué es importante esta pregunta?
La isquemia crítica de los miembros inferiores ocurre cuando se reduce el flujo sanguíneo a las piernas por el empeoramiento de la enfermedad arterial periférica. Al principio, los pacientes presentan dolor y calambres en las piernas que les impiden caminar (llamado claudicación intermitente), pero con el tiempo algunos presentarán síntomas más graves que incluyen dolor en reposo, ulceración de las piernas y gangrena. Existen muy pocas opciones de tratamiento disponibles cuando la enfermedad alcanza esta etapa, especialmente cuando la revascularización quirúrgica o con catéter (intervención para restablecer el flujo sanguíneo en las arterias o venas obstruidas) no es una opción. A muchos de estos pacientes se les amputará la extremidad afectada. El uso de la terapia con células mononucleares (con células extraídas del propio paciente) ofrece la posibilidad de un tratamiento alternativo, al proporcionar células que podrían estimular la formación de vasos capilares estables que mejoren el flujo sanguíneo de la extremidad afectada. Estas células pueden obtenerse de la médula ósea. Posteriormente se purifican en un laboratorio y se inyectan en el músculo grande de la parte posterior de la pierna.
¿Qué se encontró?
Se buscaron ensayos controlados aleatorizados (un tipo de estudio en el que los participantes se asignan al azar a uno de dos o más grupos de tratamiento) que compararan el tratamiento con células de médula ósea seleccionadas con un control (ya sea ninguna intervención, tratamiento conservador convencional o placebo [tratamiento falso]). Se encontraron cuatro estudios con un total combinado de 176 participantes que analizaron la seguridad y la efectividad de este tratamiento. Los estudios compararon la terapia celular con diferentes controles. El análisis mostró que no existe un efecto claro de la terapia celular sobre la muerte por cualquier causa. Los estudios midieron el dolor de diferentes maneras y no se proporcionó toda la información, por lo que no fue posible agrupar los datos. Por separado, tres estudios no encontraron una diferencia en la reducción del dolor entre los grupos de tratamiento y control. Un estudio informó que el dolor se redujo más en el grupo de terapia celular que en el grupo control. La agrupación de los datos de los cuatro estudios mostró que la terapia celular podría reducir las amputaciones, pero no se sabe con certeza, porque el posible beneficio se perdió al rehacer el análisis excluyendo los datos de los estudios que suscitaban dudas. Tres estudios comunicaron que no hubo una mejoría en el índice tobillo‐brazo (una forma de medir el flujo sanguíneo en la pierna), mientras que un estudio informó una mejoría más grande en el índice tobillo‐brazo del grupo de terapia celular en comparación con el grupo control. No se observaron mejorías en la distancia caminada sin dolor entre los grupos. No se observaron diferencias claras en los efectos secundarios entre los grupos.
¿Qué confianza se tiene en la evidencia?
La confianza en la evidencia fue muy baja debido a las dudas sobre cómo se llevaron a cabo los estudios. Los cuatro estudios incluidos difieren entre sí en la forma de medir los efectos en diferentes puntos temporales; utilizaron diferentes controles; hubo pocos episodios y participantes en general; y hubo diferencias en los resultados de los estudios por separado.
Authors' conclusions
Summary of findings
| Intramuscular injection of bone marrow mononuclear cells (BMMNC) compared to control for treatment of critical lower limb ischaemia (CLI) | ||||||
| Patient or population: people with CLI who were not suitable for revascularisation and showed no improvement in response to the best standard therapy | ||||||
| Outcomes | Anticipated absolute effects (95% CI)* | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with control | Risk with BMMNC | |||||
| All‐cause mortality (follow‐up: 6 or 12 months) | Study population | RR 1.00 (0.15 to 6.63) | 123 (3 RCTs) | ⊕⊝⊝⊝ LOWc | No deaths were reported for 2 studies (Barc 2006; Pignon 2017). Li 2013 reported 4 deaths (2 in each group) during 6 months follow‐up; these deaths were not considered to be related to treatment. Information regarding mortality was not available in 1 study (Lindeman 2018). | |
| 32 per 1000 | 32 per 1000 (5 to 210) | |||||
| Reduction in pain Various scales including VAS, PISQ; scales ranged from 0 to 4/10/100, where 0 = no pain (follow‐up: 3 to 12 months) | All studies reported on pain, but due to different measurements and incomplete information we were unable to pool the data. See comment | 176 | ⊕⊝⊝⊝ LOWd | Barc 2006 assessed pain with VAS, where 0 was no pain at all and 10 was the most severe pain experienced. Pain levels decreased in both BMMNC and control groups. Li 2013 assessed pain with VAS, but reported the improvement of pain defined as a > 50% decrease in VAS during study follow‐up. Study authors reported a greater reduction in pain in the BMMNC group (P = 0.045). Lindeman 2018 used the PISQ, and reported that no differences were observed in average pain reduction between BMMNC and control group (P = 0.23). Pignon 2017 assessed pain with VAS, where 0 was no pain at all and 100 was the most severe pain experienced. Study authors reported that no differences in pain reduction were observed between the BMMNC and control group. | ||
| Incidence of amputation (follow‐up: 6 or 12 months) | Study population | RR 0.52 (0.27 to 0.99) | 176 | ⊕⊝⊝⊝ | All studies reported the incidence of amputation. A possible small benefit was no longer seen after removal of 2 studies at high risk of bias in sensitivity analysis (RR 0.52, 95% CI 0.19 to 1.39; 89 participants, 2 studies) (Barc 2006; Li 2013). | |
| 250 per 1000 | 130 per 1000 (68 to 248) | |||||
| Angiographic analysis | See comment | ‐ | ‐ | None of the studies reported angiographic analysis. | ||
| Increase in ABI An ABI ratio of 0.9 or less indicates PAD. Values between 0.9 and 1.0 are borderline, and above 1.0 is considered normal (follow‐up: range 1 month to 12 months) | All studies measured ABI, but due to incomplete information we were unable to pool the data. See comment | 176 | ⊕⊝⊝⊝ | Barc 2006 reported that ABI did not change from baseline during study follow‐up in both BMMNC and control groups. Li 2013 used an absolute increase of > 15% ABI to quantify haemodynamic improvement. Study authors reported that this was greater in the BMMNC group compared to control (P = 0.002). Lindeman 2018 reported no difference in mean ABI between the BMMNC and control groups after 12 months follow‐up (P = 0.50). Pignon 2017 did not show any changes in median ABI value between groups. | ||
| Increase in PFWD (m) (follow‐up: 12 months) | See comment | 53 | ⊕⊕⊝⊝ | Only Lindeman 2018 reported PFWD, finding no clear difference between groups. The mean PFWD for the treatment and control groups at 12 months follow‐up was 128 ± 71 m vs 160 ± 11 m, P = 0.87. | ||
| Side effects and complications (follow‐up: 6 or 12 months) | Study population | RR 2.13 (0.50 to 8.97) | 176 (4 RCTs) | ⊕⊝⊝⊝ | All trials reported adverse events. 2 studies reported that no identifiable treatment‐related adverse events were observed, and that the therapy was well‐tolerated (Barc 2006; Pignon 2017). Li 2013 reported adverse events in both BMMNC and control groups (3 vs 1 fever, 1 vs 0 MI, 0 vs 1 stroke). There were no differences in the incidence of adverse events between groups, and the therapy was well‐tolerated Lindeman 2018 reported that leukaemia occurred in 1 participant during the follow‐up period. It is unclear if this was related to the procedure. After applying the sensitivity analysis by removing 2 studies at high risk of bias (Barc 2006; Li 2013), the significance of the results did not change (RR 2.69, 95% CI 0.11 to 63.18). | |
| 23 per 1000 | 48 per 1000 (11 to 204) | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). ABI: ankle‐brachial index;BMMNC: bone marrow mononuclear cells; CI: confidence interval; CLI: critical limb ischaemia;MI: myocardial infarction; PAD: peripheral arterial disease; PFWD: pain‐free walking distance;PISQ: pain inventory score questionnaire; RCT: randomised controlled trial; RR: risk ratio; VAS: visual analogue scale | ||||||
| GRADE Working Group grades of evidence | ||||||
| aThe stem cells used in the included RCTs originate from bone marrow. All four trials used mononuclear cells collected during the harvesting procedure from bone marrow and implanted into affected limbs (Barc 2006; Li 2013; Lindeman 2018; Pignon 2017). | ||||||
Background
Description of the condition
Peripheral arterial disease is a major health problem with a total disease prevalence of 3% to 10% that increases to 15% to 20% in individuals over the age of 70 years (Hirsch 2006). As the disease progresses, about 1% to 2% of patients develop critical limb ischaemia (CLI), also known as critical limb‐threatening ischaemia, which is characterised by chronic ischaemic rest pain, ischaemic ulcers, or gangrene (Norgren 2007).
The current mainstay treatment of CLI has been surgical or catheter‐based revascularisation. However, for a substantial number of patients revascularisation is not be feasible because of the involvement of distal vessels. For individuals with CLI who are not candidates for revascularisation, around a quarter will require a major amputation during the following year (Norgren 2007).
Description of the intervention
Pre‐clinical studies of the use of mononuclear cells from the bone marrow has shown promising results in animal models with CLI. These data demonstrated an improvement in capillary density in hindlimb ischaemia and promoted collateral vessel formation (Shintani 2001). On the basis of these results in animals, clinical trials were started in order to test cell therapy with autologous bone marrow mononuclear cells (BMMNCs) in people with ischaemic lower limbs.
At present, the procedure is as follows. Mononuclear cells are derived either directly from bone marrow aspiration (Higashi 2004; Tateishi‐Yuyama 2002), or through mobilisation into the peripheral blood using granulocyte colony‐stimulating factor (G‐CSF) (Huang 2004; Huang 2005b). In the first procedure, bone marrow cells are usually collected (mostly under general anaesthesia) from the iliac crest. Thereafter, the BMMNCs are enriched away from other bone marrow cells in sterile conditions in a laboratory. In the second procedure, mononuclear cells mobilised into the peripheral blood following G‐CSF administration for four to five days are collected from a peripheral blood sample and then enriched from other blood cells in sterile conditions. In both procedures, the enriched mononuclear cells are directly implanted into the gastrocnemius muscle of the ischaemic leg.
How the intervention might work
Although the exact mechanism of action of mononuclear cell implantation has not been fully elucidated, BMMNCs have been shown to secrete a number of angiogenic cytokines (Shintani 2001; Takahashi 2006). It therefore seems that implantation into ischaemic limbs could enhance angiogenesis through paracrine mechanisms by supplying endothelial progenitor cells and providing multiple angiogenic factors or cytokines. These combined mechanisms may subsequently lead to the formation of stable capillary vessels and so reverse the ischaemic status of the affected limb.
Why it is important to do this review
Since mononuclear cell implantation has emerged as a novel intervention in clinical practice for CLI, it is important that a systematic review is undertaken in order to assess the safety and efficacy of this intervention. We have therefore updated the systematic review first published in 2014, Moazzami 2014, in order to identify recent evidence regarding the potential therapeutic benefits and possible harms of local intramuscular BMMNC implantation for people with CLI.
Objectives
To evaluate the benefits and harms of local intramuscular transplantation of autologous adult bone marrow mononuclear cells (BMMNCs) as a treatment for critical limb ischaemia (CLI).
Methods
Criteria for considering studies for this review
Types of studies
We included randomised controlled trials (RCTs).
Types of participants
We included participants with a clinical diagnosis of critical limb ischaemia (CLI) who had been admitted to hospital for treatment. We included participants who were not candidates for open or endovascular revascularisation (Rutherford category 5/6) according to angiographic evidence of superficial femoral artery or infrapopliteal disease in the affected limb, as well as those who did not show any evidence of improvement in response to best standard therapy in the previous four weeks. There was no age restriction.
Types of interventions
We included studies involving the administration of autologous adult BMMNCs by direct implantation into the gastrocnemius muscle of ischaemic legs of participants as a treatment for CLI, compared to either no intervention, conventional conservative therapy (e.g. pharmacological treatment of pain, wound care, or bed rest), or administration of an inert placebo such as isotonic saline.
Types of outcome measures
Primary outcomes
-
All‐cause mortality
-
Reduction in pain, as assessed by analgesic requirements or a pain analogue scale
-
Incidence of amputation (minor or major, where major amputation is above the ankle, and minor amputation is below the ankle or part of the foot)
Secondary outcomes
-
Angiographic analysis (changes in the number of visible vessels or any change in the intensity or apparent size of previously visible vessels)
-
Increase in ankle‐brachial index (ABI)
-
Increase in pain‐free walking distance (PFWD)
-
Side effects and complications, such as local or systemic inflammation, cardiovascular abnormalities, and thromboembolic complications
Search methods for identification of studies
Electronic searches
The Cochrane Vascular Information Specialist conducted systematic searches of the following databases for RCTs and controlled clinical trials without language, publication year, or publication status restrictions:
-
Cochrane Vascular Specialised Register via the Cochrane Register of Studies (CRS Web searched from 17 February 2014 to 8 November 2021);
-
Cochrane Central Register of Controlled Trials (CENTRAL); Cochrane Register of Studies Online (CRSO 2021, Issue 10) (searched on 8 November 2021);
-
MEDLINE (Ovid MEDLINE Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations, Ovid MEDLINE Daily and Ovid MEDLINE) (searched from 1 January 2017 to 8 November 2021);
-
Embase Ovid (searched from 1 January 2017 to 8 November 2021);
-
CINAHL EBSCO (Cumulative Index to Nursing and Allied Health Literature) (searched from 1 January 2017 to 8 November 2021);
-
AMED Ovid (Allied and Complementary Medicine) (searched from 1 January 2017 to 8 November 2021).
The Information Specialist modelled search strategies for other databases on the search strategy designed for CENTRAL. Where appropriate, search strategies were combined with adaptations of the Highly Sensitive Search Strategy designed by Cochrane for identifying RCTs and controlled clinical trials (as described in Chapter 6 of the Cochrane Handbook for Systematic Reviews of Interventions) (Lefebvre 2021). Search strategies for major databases are provided in Appendix 1.
The Information Specialist searched the following trials registries on 8 November 2021:
-
World Health Organization International Clinical Trials Registry Platform (who.int/trialsearch);
-
US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (clinicaltrials.gov).
Searching other resources
We checked the reference lists of papers and reports retrieved from the electronic searches to identify additional potentially relevant studies.
Data collection and analysis
Selection of studies
Three review authors (KM, EF, and ZM) screened the titles and abstracts of references identified by the search, and two review authors (BM and ZM or ZZ) independently assessed their eligibility for inclusion in the review. Any disagreements were resolved by consensus or by a discussion with the third review author (KM). We obtained full versions of articles deemed potentially relevant based on title or abstract, and three review authors (BM, AR, and ED) assessed these independently against the inclusion criteria. We recorded the reasons for exclusion of any study excluded at the full‐text stage in the Characteristics of excluded studies table.
Data extraction and management
Three review authors (BM, AR, and ZM) independently extracted and recorded the dichotomous and continuous data for the prespecified outcomes on forms developed by Cochrane Vascular. We recorded details on study design, setting, participant numbers, inclusion and exclusion criteria, cell type, route of delivery, control used, outcomes, funding and declarations of interest made by the study authors. Any disagreements were resolved by another review author (ED); if necessary, we sought additional information from the study authors.
Assessment of risk of bias in included studies
We used the Cochrane risk of bias 1 (RoB 1) tool as described in the Cochrane Handbook for Systematic Reviews of Interventions to evaluate risk of bias of the included studies (Higgins 2017). We assessed the following risk of bias domains: randomisation and allocation (selection bias), blinding (performance bias and detection bias), incomplete outcome data (attrition bias), selective reporting (reporting bias), and other potential sources of bias, judging each domain to be either low, high, or unclear risk of bias according to the guidance in Higgins 2017.
Measures of treatment effect
We expressed the measure of effect as risk ratios (RRs) with 95% confidence intervals (CI) for each dichotomous outcome. Where continuous scales of measurement were used to assess the effects of treatment, we used mean difference (MDs) with 95% CI. When different scales were used in the different studies, we standardised the results where possible and combined them using standardised mean difference (SMD) with 95% CI.
Unit of analysis issues
The unit of analysis was the individual participant randomised to either BMMNC therapy or a control group.
Dealing with missing data
Where necessary, we sought missing data and data regarding participant demographics and outcome measures by contacting the corresponding study author. We planned that if some outcome data remained missing despite our attempts at retrieval, we would exclude the trials from the analyses where there were more than 10% incomplete or missing entries for each variable.
Assessment of heterogeneity
We explored and assessed heterogeneity using the I2 and Q statistics, and by the subjective judgement of the comparability of participants, interventions, and outcomes. We planned to assess statistical heterogeneity of trial data by using the Mantel‐Haenszel Chi2 test of heterogeneity and the I2 statistic of heterogeneity (Higgins 2021). We considered data to be heterogeneous if P was less than 0.10 for the first method. For the I2 method, we used the guidelines on interpretation described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2021), which suggest that an I2 statistic of 0% to 40% might not be important; 30% to 60% may represent moderate heterogeneity; 50% to 90% may represent substantial heterogeneity; and 75% to 100% considerable heterogeneity.
Assessment of reporting biases
We planned that if sufficient trials were available, we would use funnel plots to assess publication bias. However, as only four studies were included in the review, this was not possible.
Data synthesis
We only undertook meta‐analyses where the treatments, participants, and underlying clinical questions were similar enough for pooling to be meaningful. The overall treatment effect was estimated by the pooled RR with 95% CI. Since there were differences in the methods of the studies (sources of error were both within‐study and between‐study variance), we used a random‐effects model to perform the analysis. Each test for significance was two‐tailed.
Subgroup analysis and investigation of heterogeneity
We planned to perform subgroup analyses by the type of disease (atherosclerosis versus thromboangitis obliterans) if sufficient trials were available. Performing subgroup analyses was not feasible owing to inadequate information regarding participants with or without significant comorbidity, gender, ethnicity, and different age groups.
Sensitivity analysis
We planned to undertake sensitivity analyses to examine the robustness of the observed findings in relation to a number of factors including study quality and patient type if sufficient studies were identified. We performed sensitivity analysis for the outcomes amputation and side effects and complications by removing two studies at high risk of bias (Barc 2006; Li 2013). We judged studies to be at high risk of bias if any of the following domains were at high risk of bias or if all of the following domains were at unclear risk of bias: random sequence generation and allocation concealment (selection bias), blinding of participants and personnel (performance bias), and blinding of outcome assessment (detection bias).
Summary of findings and assessment of the certainty of the evidence
We created a summary of findings table to present the main findings of this review using GRADEpro GDT software (GRADEpro GDT), and the guidelines provided in Chapter 14 of the Cochrane Handbook for Systematic Reviews of Interventions (Schünemann 2021). We used the five GRADE criteria (risk of bias, inconsistency, imprecision, indirectness, and publication bias) to assess the certainty of the body of evidence (GRADE 2004). We justified all decisions to downgrade the certainty of the evidence in footnotes. We included the outcomes considered to be of most clinical relevance in summary of findings Table 1, as follows.
-
All‐cause mortality
-
Reduction in pain
-
Incidence of amputation
-
Angiographic analysis
-
ABI
-
PFWD
-
Side effects and complications
Results
Description of studies
Search results are presented in Figure 1.
Study flow diagram.
Results of the search
We included three new studies in this update (Li 2013; Lindeman 2018; Pignon 2017). We excluded one previously included study following clarification of our inclusion criteria (Huang 2005a). In total, we excluded 38 new studies (Benoit 2011; BONMOT 2008; Burt 2010; Dong 2018; Du 2017; Flugelman 2017; Frogel 2017; Horie 2018; Huang 2005a; Iafrati 2016; JUVENTAS 2008; Korymasov 2009; Majumdar 2015; Molavi 2016; Murphy 2017; NCT01584986; NCT02336646; NCT03174522; NCT03214887; NCT03304821; NCT03339973; Niven 2017; Ohtake 2017; Perin 2017; Poole 2013; Prochazka 2010; PROVASA 2011; RESTORE‐CLI 2012; Sharma 2021; Skóra 2015; Teraa 2015; Tournois 2015; Wang 2014; Wang 2017; Wang 2018; Wijnand 2018; Zhou 2017a; Zhou 2017b). We identified three new ongoing studies (NCT00753025; NCT01446055; NCT02454231). See Characteristics of included studies, Characteristics of excluded studies, and Characteristics of ongoing studies.
Included studies
For details, see Characteristics of included studies.
We identified four included studies involving a total of 176 participants for this review (Barc 2006; Li 2013; Lindeman 2018; Pignon 2017). All studies were RCTs comparing the effect of intramuscular injections of BMMNCs versus control. Barc 2006 compared the effect of intramuscular injections of BMMNCs versus control in people with CLI. Control therapy was described by the authors as standard conservative therapy, which was also given to the treatment group. Li 2013 investigated the effect of administration of BMMNCs in comparison with 0.9% sodium chloride (also known as saline or saline solution). Lindeman 2018 compared administration of BMMNCs to placebo (diluted autologous peripheral blood). In Pignon 2017, participants who received BMMNCs were compared with participants who received placebo (30 mL saline with 4 mL autologous peripheral blood). One trial was a multicentre study conducted in seven academic centres in France (Pignon 2017); the other three studies were single centre (Barc 2006; Li 2013; Lindeman 2018).
Trial duration
The duration of follow‐up varied from six to 12 months. The treatment duration in two studies lasted six months (Barc 2006; Li 2013), and in two studies participants were followed up for 12 months (Lindeman 2018; Pignon 2017). All studies used the end of the treatment as the final follow‐up time for the treatment phase.
Sample sizes
All of the included studies had small sample sizes, ranging from 29 participants in Barc 2006 to 58 participants in Li 2013. Overall, 176 participants were included. Of these, 88 participants were randomised to receive BMMNCs, and 88 participants were randomised to control. In Barc 2006, 14 participants were randomised to the BMMNC treatment group and 15 to the control group (conventional therapy). Li 2013 randomised 58 participants to BMMNC and 29 to control (saline). In Lindeman 2018, 53 participants were randomised, 28 to BMMNCs and 25 to placebo (diluted autologous peripheral blood). In Pignon 2017, 17 participants received BMMNCs and 19 participants received placebo (30 mL saline with 4 mL autologous peripheral blood).
Participants
All included studies involved participants with a diagnosis of CLI. The most common aetiology of CLI amongst the included studies was atherosclerosis obliterans. Data regarding participant age, sex, the severity of limb ischaemia, and baseline comorbidities (diabetes mellitus, hypertension, and chronic obstructive pulmonary disease (COPD)) were assessed and compared between groups by the individual studies. Barc 2006 included participants with CLI at risk of amputation who showed no progress after eight weeks of conventional therapy and with no option for operative therapy. Li 2013 included participants with CLI without improvement after a minimum of four weeks on conventional therapy. Lindeman 2018 included participants with end‐stage peripheral arterial disease and CLI without improvement after six months of optimal therapy and with no revascularisation options. Pignon 2017 included participants with CLI and no improvement after medical therapy after a duration of 12 months.
Characteristics of bone marrow cell therapy
The autologous mononuclear cells were obtained from bone marrow in all of the included studies (Barc 2006; Li 2013; Lindeman 2018; Pignon 2017). In Barc 2006, the isolation of mononuclear cells from collected marrow was performed with Baxter Fenwal CS 3000 plus blood cell separator. In Li 2013, BMMNCs were isolated from the bone marrow by density gradient centrifugation with lymphocyte separating fluid. In Lindeman 2018, the collected bone marrow suspension was filtered and concentrated to a final volume of 40 mL on the COBE Spectra Apheresis System (Gambro, Stockholm, Sweden) without further manipulation. In Pignon 2017, two different approaches were performed for BMMNC separation across centres, including blood‐cell separator using COBE Spectra, version 4, Bone Marrow Processing Program (Gambro BCT, Lakewood, CO, USA), and blood‐cell separator requiring a Ficoll density‐gradient for isolation of the BMMNC (COBE 2991, Gambro BCT).
The median cell dosage of mononuclear cell count varied from 1.3 x 109, in Pignon 2017, to 1.7 x 109, inLindeman 2018. The median number of CD34(+) cells were 1 x 107, 33.5 x 106, and 48.8 x 106 in Li 2013, Pignon 2017, and Lindeman 2018, respectively. In Barc 2006, the exact number of mononuclear cells was not stated. There was no standard definition for high versus low cell amounts across all studies. Intramuscular cell implantation was the route of administration in all studies. The standard procedure for cell implantation was performed following multiple intramuscular injections into the gastrocnemius muscle. The cells were administered multiple times, ranging from four times in Barc 2006 to 40 times in Lindeman 2018.
None of the included studies provided additional information regarding the teams responsible for the administration of the BMMNC or deciding on the indication of open or endovascular revascularisation, or details of the angiographic assessments.
Outcomes
The outcomes reported in the individual trials are reported in the Characteristics of included studies tables and include:
-
all‐cause mortality (Barc 2006; Li 2013; Pignon 2017);
-
increase in ankle‐brachial index (ABI) (Barc 2006; Lindeman 2018; Pignon 2017);
-
reduction in pain assessed by analgesic requirements or a pain analogue scale (Barc 2006; Li 2013; Lindeman 2018; Pignon 2017);
-
progression of disease in terms of the incidence of amputation (Barc 2006; Li 2013; Lindeman 2018; Pignon 2017);
-
increase in pain‐free walking distance (PFWD) (Lindeman 2018);
-
measurement of transcutaneous pressure of oxygen (TcPO2) (Pignon 2017);
-
side effects (Barc 2006; Li 2013; Lindeman 2018; Pignon 2017).
Excluded studies
We excluded 38 additional studies in this update (Benoit 2011; BONMOT 2008; Burt 2010; Dong 2018; Du 2017; Flugelman 2017; Frogel 2017; Horie 2018; Huang 2005a; Iafrati 2016; JUVENTAS 2008; Korymasov 2009; Majumdar 2015; Molavi 2016; Murphy 2017; NCT01584986; NCT02336646; NCT03174522; NCT03214887; NCT03304821; NCT03339973; Niven 2017; Ohtake 2017; Perin 2017; Poole 2013; Prochazka 2010; PROVASA 2011; RESTORE‐CLI 2012; Sharma 2021; Skóra 2015; Teraa 2015; Tournois 2015; Wang 2014; Wang 2017; Wang 2018; Wijnand 2018; Zhou 2017a; Zhou 2017b), for a total of 73 excluded studies (see Characteristics of excluded studies).
The main reasons for exclusion were as follows.
-
Fifty studies investigated the effect of preparations of cells other than mononuclear cells. These included bone marrow cells (NCT00498069), CD34+ cells (Dong 2013; Dong 2018; Madaric 2011; NCT00616980), peripheral blood‐derived stem cell therapy (Capiod 2009; Huang 2005a; NCT00922389; Niven 2017; Ohtake 2017; Szabo 2013), CD133+ cells (Burt 2010; Korymasov 2009; NCT00913900), bone marrow concentrate (BONMOT 2008; Gurunathan 2009; Iafrati 2011; Iafrati 2016; Murphy 2017; NCT00595257; NCT01049919; NCT01245335; Prochazka 2010; Wang 2017), allogeneic mesenchymal stem cells (MSCs) (Du 2017; Gupta 2013; Lu 2011; Majumdar 2015; Mohamed 2020; NCT02336646; Pawan 2012; Wang 2018; Wijnand 2018), bone marrow mesenchymal stem cells (BMMSCs) (Chen 2009; Debin 2008; NCT00955669), adult stem cells (Lasala 2011), bone marrow‐derived aldehyde dehydrogenase bright cells (Perin 2011), expanded autologous bone marrow‐derived tissue, Ixmyelocel‐T and vascular repair cells (RESTORE‐CLI 2012), autologous angiogenic cell precursors (NCT01584986), Rexmyelocel‐T (REX‐001) (NCT03174522), granulocyte‐macrophage colony stimulating factor (GM‐CSF) (NCT03304821; Poole 2013; Zafarghandi 2010), allogeneic ABCB5‐positive stem cells (NCT03339973), venous endothelial cells (ECs) (Flugelman 2017), aldehyde dehydrogenase (ALDH) bright cells (Perin 2017), vascular endothelial growth factor (VEGF) 165 gene‐modified PB CD34+ cells (Zhou 2017b), circulating blood‐derived progenitor cells (Frogel 2017), and autologous cell therapy products (Tournois 2015).
-
Ten studies did not include people with CLI (Holzinger 1994; Horie 2018; Kirana 2007; Kirana 2012; NCT00306085; NCT00539266; NCT03214887; Subramaniyam 2009; Zhang 2010; Zhou 2017a).
-
In six studies, cells were not administered intramuscularly, and the intra‐arterial route of administration was applied (JUVENTAS 2008; NCT00282646; PROVASA 2011; Sharma 2021; Teraa 2015; Walter 2011).
-
In one study, the active intervention arm consisted of BMMNCs co‐administered with peripheral blood mononuclear cells (PBMNCs) (Zhao 2008).
-
In one study, BMMNCs were co‐administered with a selective 5‐HT(2A) antagonist in the treatment arm (Higashi 2010).
-
In one study, bone marrow was processed into 40 mL of concentrate (Benoit 2011).
-
In one study, non‐cell‐based products were compared to cell treatment (Wang 2014).
-
In one study, the therapeutic effects of intramuscular and intra‐arterial delivery of bone marrow cells were compared (Klepanec 2012).
-
One study evaluated the safety and efficacy of repeated bone marrow mononuclear cell injections compared with a single bone marrow mononuclear cell injection (Molavi 2016).
-
One study used BMMNCs in combination with gene therapy (Skóra 2015).
Ongoing studies
We identified three new ongoing studies (NCT00753025; NCT01446055; NCT02454231). For details, see Characteristics of ongoing studies.
Risk of bias in included studies
Please refer to Figure 2 and Figure 3. Details and justification for the risk of bias ratings for each included study are presented in the risk of bias tables in Characteristics of included studies.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Allocation
All included trials explicitly stated that randomisation occurred; however, in two studies the method used to generate the random sequence was not described, and we judged them to be at unclear risk of selection bias (Barc 2006; Li 2013). The randomisation methods and allocation were adequately performed in two trials, and these were judged as being at low risk of selection bias (Lindeman 2018; Pignon 2017).
Blinding
In two trials, both outcome assessors and participants were blinded, and these were judged as being at low risk of bias (Lindeman 2018; Pignon 2017). One study was single‐blinded, and was thus judged as being at unclear risk of performance and detection bias (Li 2013). One study provided insufficient information concerning blinding and was judged to be at unclear risk of performance and detection bias (Barc 2006).
Incomplete outcome data
We assessed three studies as having a low risk of attrition bias (Barc 2006; Lindeman 2018; Pignon 2017). We assessed the remaining study as having a high risk of attrition bias, as substantial numbers of participants were not evaluated for some outcomes (Li 2013).
Selective reporting
We did not identify any reporting bias in two studies (Li 2013; Lindeman 2018). We judged two studies as being at high risk of selection bias, as ABI and pain were not reported in sufficient detail to use in the analysis (Barc 2006; Pignon 2017).
Other potential sources of bias
We identified no other potential sources of bias in the four included studies (Barc 2006; Li 2013; Lindeman 2018; Pignon 2017).
Effects of interventions
See summary of findings Table 1.
All‐cause mortality
No deaths were reported during the study period in two included trials (Barc 2006; Pignon 2017). Information regarding mortality was not available in one study (Lindeman 2018). One study reported four deaths during study follow‐up (two in the therapy group and two in the control group); the study authors stated that none of the deaths were related to treatment (Li 2013). Pooled data of three studies showed no clear effect of BMMNC therapy on all‐cause mortality (risk ratio (RR) 1.00, 95% confidence interval (CI) 0.15 to 6.63; 3 studies, 123 participants; P = 1.0; very low‐certainty evidence; Analysis 1.1). We downgraded the certainty of the evidence a total of three levels due to risk of bias (two studies with unclear risk of selection, performance, and detection bias); imprecision (few participants and events); and inconsistency (wide CIs and clinical heterogeneity) (summary of findings Table 1).
Reduction in pain, as assessed by analgesic requirements or a pain analogue scale
We were unable to pool data, as different measures were used to assess pain, so we have provided a narrative report.
In Barc 2006, ischaemic pain was assessed with a visual analogue pain scale (VAS) with 10 levels, where 0 was no pain at all and 10 was the most severe pain experienced. The study reported that pain decreased in both groups. Data were not reported in sufficient detail to use.
Li 2013 used the same assessment method (VAS), but reported improvement of pain defined as a > 50% decrease in pain scores during study follow‐up. Study authors reported a significant difference for pain reduction between the BMMNC and control groups (42% versus 12%, P = 0.045).
Lindeman 2018 assessed both the severity of pain and the impact of pain on daily activities by a Pain Inventory Score Questionnaire. The mean pain score reduction at 12 months in the BMMNC group was 4.6 ± 2.6, compared to 4.8 ± 1.9 in the control group (P = 0.23). Overall, no differences were observed in average pain reduction between the BMMNC group and the control group (diluted autologous peripheral blood).
Pignon 2017 evaluated pain severity using a VAS scoring system ranging from 0 to 100. Study authors reported that no differences in pain reduction were observed between participants who received BMMNC and those who received control (diluted autologous peripheral blood), but data were not reported in sufficient detail to include in the analysis.
Overall, we downgraded the certainty of the evidence for pain to very low due to risk of bias, imprecision (few participants and events), and inconsistency (clinical heterogeneity) (summary of findings Table 1).
Incidence of amputation
All four studies examined the rate of amputations during their study period (Barc 2006; Li 2013; Lindeman 2018; Pignon 2017). The pooled findings showed that compared with control (including conventional therapy, diluted autologous peripheral blood, and saline), intramuscular mononuclear cell implantation was possibly associated with a small reduction in risk of amputation (RR 0.52, 95% CI 0.27 to 0.99; 4 studies, 176 participants; P = 0.05; very low‐certainty evidence; Analysis 1.2). This possible slight benefit was lost after applying sensitivity analysis by removing two studies at a high risk of bias (RR 0.52, 95% CI 0.19 to 1.39; 2 studies, 89 participants; P = 0.19) (Analysis 1.3) (Barc 2006; Li 2013). We downgraded the certainty of the evidence due to risk of bias (two studies with unclear risk of selection, performance, and detection bias); imprecision (few participants and events); and inconsistency (wide CIs) (summary of findings Table 1).
Angiographic analysis
None of the included studies reported data on angiographic assessments.
Increase in ankle‐brachial index (ABI)
All four included studies measured ABI, but incomplete information precluded the pooling of data, therefore we have provided a narrative report. Barc 2006 reported that ABI did not change from baseline during study follow‐up in either the BMMNC or the control group. Li 2013 used an absolute increase of > 15% ABI to quantify haemodynamic improvement, and reported that this was greater in the therapy group compared to the control group (52% versus 5%, P = 0.002). Lindeman 2018 reported no difference in mean ABI between BMNC treatment and control groups after 12 months follow‐up (mean ABI: 0.68 ± 0.32 versus 0.50 ± 2.0, respectively; P = 0.50). Pignon 2017 reported the median ABI in a figure which did not show any changes in ABI value between groups; we were unable to obtain the specific data. We downgraded the certainty of the evidence to very low due to concerns related to risk of bias, imprecision (few participants and events), and inconsistency (clinical heterogeneity) (summary of findings Table 1).
Increase in pain‐free walking distance (PFWD)
Only one study evaluated the effect of intramuscular BMMNC implantation on PFWD (Lindeman 2018). There was no clear difference in mean PFWD between participants who received BMMNC and those who received control (diluted autologous peripheral blood). The mean (± standard deviation) PFWD for the treatment and control groups at 12 months follow‐up was 128 ± 71 m compared to 160 ± 11 m, respectively; P = 0.87. We downgraded the certainty of the evidence to low due to imprecision (few participants and events) and inconsistency (clinical heterogeneity) (summary of findings Table 1).
Side effects and complications
All of the included trials reported adverse events. Two studies reported that no identifiable treatment‐related adverse events were observed, and that therapy was well‐tolerated (Barc 2006; Pignon 2017). Lindeman 2018 reported that leukaemia occurred in one participant during the follow‐up period. However, no further information was provided to permit an assessment of whether this adverse event was thought to be directly related to the procedure. Whilst Li 2013 reported some adverse events in both the BMMNC and the control group (3 versus 1 fever, 1 versus 0 myocardial infarction, 0 versus 1 stroke, respectively), they reported there were no significant differences in the incidence of adverse events between groups and that therapy was well‐tolerated. The pooled data showed no clear difference between groups in the incidence of side effects and complications (RR 2.13, 95% CI 0.50 to 8.97; 4 studies, 176 participants; P = 0.30; Analysis 1.4). The overall effect was unchanged after sensitivity analysis in which two studies at high risk of bias were removed (RR 2.69, 95% CI 0.11 to 63.18; 2 studies, 89 participants; P = 0.54; Analysis 1.5) (Barc 2006; Li 2013). We downgraded the certainty of the evidence to very low due to risk of bias, imprecision (few participants and events), and inconsistency (wide CIs and clinical heterogeneity) (summary of findings Table 1).
Subgroup and sensitivity analysis
Performing subgroup analysis was precluded by inadequate information regarding participants with or without significant comorbidity, gender, ethnicity, and different age groups.
We performed sensitivity analysis for the outcomes amputation (Analysis 1.3) and side effects and complications (Analysis 1.5) by removing the two studies at high risk of bias (Barc 2006; Li 2013). Results are reported above.
Discussion
Summary of main results
See summary of findings Table 1.
A large body of evidence exists regarding the use of bone marrow mononuclear cells (BMMNCs) for the treatment of patients with various haematological malignancies, but their role in the treatment of other diseases, including critical limb ischaemia (CLI), has not been exclusively addressed. Moreover, the findings of this review suggest that there is limited evidence to support this practice at present. In the current update, we included four studies with a total of 176 participants who were randomised to receive either intramuscular BMMNC implantation or control for the treatment of CLI. There was heterogeneity between study control arms, with groups receiving either conventional therapy, diluted autologous peripheral blood, or 0.9% sodium chloride (saline) solution.
There were no mortality events related to the administration of BMMNCs, and analysis did not show any clear difference between the BMMNC and control groups with respect to mortality risk. We deemed the certainty of this evidence to be very low due to risk of bias, imprecision, and inconsistency. All trials assessed changes in pain severity with different forms of pain assessment tools, and so we were unable to pool data. Three studies individually reported that no differences in pain reduction were observed between participants who received BMMNC and those who received control. Li 2013 reported that reduction in rest pain was greater in the BMMNCs group compared to the control group. We downgraded the certainty of the evidence for this outcome to very low due to risk of bias, imprecision, and inconsistency.
All four trials reported on the rate of amputation at the end of the study period. Analysis showed that BMMNC treatment possibly reduced the risk of amputation slightly compared to control treatment (very low‐certainty evidence). This possible benefit was lost after removal of two studies at high risk of bias in a sensitivity analysis. We downgraded the certainty of the evidence to very low due to risk of bias, imprecision, and inconsistency. The evidence for risk of amputation following BMMNC treatment is uncertain.
None of the included studies reported angiographic analysis.
All of the included studies measured ankle‐brachial index (ABI) index. Incomplete information precluded the pooling of data. Three studies reported no change to ABI between groups (Barc 2006; Lindeman 2018; Pignon 2017). Li 2013 reported greater improvement in ABI in response to BMMNCs compared to placebo (P = 0.002). We downgraded the certainty of the evidence to very low due to risk of bias, imprecision, and inconsistency.
Only Lindeman 2018 reported pain‐free walking distance (PFWD), and results did not differ between the BMMNC and control groups. We downgraded the certainty of the evidence for this outcome to low due to imprecision and inconsistency.
All studies reported on side effects, and our analysis showed no clear difference in the numbers of side effects between BMMNC and control groups. We downgraded the certainty of the evidence to very low due to risk of bias, imprecision, and inconsistency.
The results of our review are limited due to the small number of randomised controlled trials (RCTs) meeting our inclusion criteria and differences in the control arms. The certainty of the evidence presented is very low to low, and we cannot draw any strong conclusions on the efficacy of BMMNC‐based therapy for CLI patients.
Overall completeness and applicability of evidence
The four included trials involved a very small number of participants, and there were substantial variations in the treatment strategies, control groups, and follow‐up duration, which limited our ability to combine the findings of these trials. Moreover, the use of different control groups in all studies may not be applicable to a much broader spectrum of CLI patients and their clinicians, and resulted in heterogeneity across the included studies. Given that only four studies with low numbers of participants were eventually included, significant concern exists on the completeness and applicability of the evidence presented in this review. Further studies with larger sample sizes are needed to permit a definitive conclusion. Future trials need to address important outcomes such as the role of treatment in the relief of pain, which can be beneficial in improving a patient's quality of life, as well as understanding the possible adverse effects of treatment. Furthermore, focusing on the comparison of different routes of administration, types of cells, and supportive bioengineered matrices could provide fundamental information to establish the effectiveness and safety assessment of the method. Moreover, specific studies about the implantation of stem cells in ischaemic tissues at an earlier stage of peripheral arterial disease are urgently needed to clarify how it can affect the course of this disabling disease.
Quality of the evidence
Overall, we judged the certainty of the evidence as very low (summary of findings Table 1). Major reasons for downgrading the certainty of the evidence were concerns related to risk of bias, imprecision, and inconsistency. Methodological limitations included unclear randomisation methods and allocation concealment, as well as lack of blinding of participants and insufficient outcome data. Based on the available data, two studies did not adequately report random sequence generation or allocation concealment, putting them at risk of selection bias, and provided insufficient information regarding the blinding of participants (Barc 2006; Li 2013). We assessed two studies as at an overall low risk of bias (Lindeman 2018; Pignon 2017). We also downgraded the certainty of the evidence for imprecision of effect estimates due to the small numbers of studies with limited participants that contributed to all outcomes (the total number of participants from four trials was fewer than 200 participants), and for inconsistency of the results (due to wide confidence intervals across analyses and differences in the control groups). Statistical heterogeneity greatly affected our ability to answer our review question, and has impacted the validity of our conclusion. There is a need for high‐quality studies with larger sample sizes to assess the effects of the treatment regimen of cell‐based therapy for CLI in clinical practice. The data obtained from these studies are currently not of sufficient certainty to draw robust conclusions for the outcomes evaluated in this review regarding the use of BMMNC for the treatment of CLI.
Potential biases in the review process
An extensive literature search was performed by Cochrane Vascular. Two review authors independently determined study eligibility, and two review authors independently extracted data and performed quality assessment in order to reduce bias and subjectivity. There were no significant disagreements during the review process. We included only randomised clinical trials in our review. We are confident that all potential sources of data to be included in this review were carefully vetted.
Agreements and disagreements with other studies or reviews
To date, limited published RCTs have directly evaluated the effectiveness of intramuscular implantation of cell‐based products derived from bone marrow and compared it with conventional treatment options for the management of CLI. Moreover, these studies have yielded inconsistent results in terms of effectiveness. There are non‐RCT studies that have investigated the role of intramuscular implantation of various cell types (Franz 2009; Iso 2010). These studies have shown promising results: some studies have reported improved outcomes with intramuscular implantations, whilst few studies have not shown this approach to be beneficial for people with CLI (Dong 2013; Lara‐Hernandez 2010). A systematic review reported that although non‐RCTs suggest a benefit of cell treatment (lower the risk of amputation by 37%, increased amputation‐free survival by 18%, and improved wound healing by 59%, without affecting mortality), controlled trial studies with a low risk of bias do not support the promising results of these studies thus far (Rigato 2017). Rigato 2017 also reported that cell therapy with peripheral blood mononuclear cells (PBMNCs), but not other cell types, was associated with a significant improvement in amputation and amputation‐free survival, whereas only BMMNCs significantly improved wound healing. On the other hand, both BMMNCs and PBMNCs significantly improved ABI, transcutaneous pressure of oxygen (TcPO2), and rest pain scores.
This update includes three new trials, bringing the total number of participants to 176. Although additional data are included, the findings presented in this update are consistent with the results of the two previous versions of this systematic review (Moazzami 2011; Moazzami 2014).
Previous meta‐analyses have shown promising results of stem cell‐based therapies versus conventional treatments in the management of CLI patients (Liu 2012; Teraa 2013; Wen 2011). In contrast to our review, these studies pooled and analysed outcome data of patients treated with cells from different sources and different types of cell administration techniques (BMMNCs, bone marrow mesenchymal stem cells, mobilised peripheral blood stem cells, Ixmyelocel‐T) and collectively reported them as bone marrow‐derived cell therapy. Another systematic review and meta‐analysis concluded that cell‐based therapy significantly reduced the amputation rate and improved ABI, as well as ulcer healing, when compared to non‐cell treatment groups (Liu 2015). Contrary to our review, Liu 2015 combined and analysed non‐randomised studies in their meta‐analysis. Consequently, in these meta‐analyses, the use of various types of cell products for implantation, as well as differences in routes of administration, may have had a profound impact on clinical outcomes and the generalisability of the conclusion.
A recent Cochrane Review involving seven RCTs with a total of 359 participants compared the efficacy and safety of autologous cells derived from different sources, prepared using different protocols, administered at different doses, and delivered via different routes for the treatment of 'no‐option' CLI patients (Abdul Wahid 2018). Similarly, pooled analyses did not show a clear difference in clinical outcomes between different stem cell sources and different treatment regimens of autologous cell implantation, and no strong conclusions were possible (Abdul Wahid 2018).
Study flow diagram.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Comparison 1: BMMNC versus control, Outcome 1: All‐cause mortality
Comparison 1: BMMNC versus control, Outcome 2: Amputation
Comparison 1: BMMNC versus control, Outcome 3: Amputation ‐ sensitivity analysis
Comparison 1: BMMNC versus control, Outcome 4: Side effects and complications
Comparison 1: BMMNC versus control, Outcome 5: Side effects and complications ‐ sensitivity analysis
| Intramuscular injection of bone marrow mononuclear cells (BMMNC) compared to control for treatment of critical lower limb ischaemia (CLI) | ||||||
| Patient or population: people with CLI who were not suitable for revascularisation and showed no improvement in response to the best standard therapy | ||||||
| Outcomes | Anticipated absolute effects (95% CI)* | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with control | Risk with BMMNC | |||||
| All‐cause mortality (follow‐up: 6 or 12 months) | Study population | RR 1.00 (0.15 to 6.63) | 123 (3 RCTs) | ⊕⊝⊝⊝ LOWc | No deaths were reported for 2 studies (Barc 2006; Pignon 2017). Li 2013 reported 4 deaths (2 in each group) during 6 months follow‐up; these deaths were not considered to be related to treatment. Information regarding mortality was not available in 1 study (Lindeman 2018). | |
| 32 per 1000 | 32 per 1000 (5 to 210) | |||||
| Reduction in pain Various scales including VAS, PISQ; scales ranged from 0 to 4/10/100, where 0 = no pain (follow‐up: 3 to 12 months) | All studies reported on pain, but due to different measurements and incomplete information we were unable to pool the data. See comment | 176 | ⊕⊝⊝⊝ LOWd | Barc 2006 assessed pain with VAS, where 0 was no pain at all and 10 was the most severe pain experienced. Pain levels decreased in both BMMNC and control groups. Li 2013 assessed pain with VAS, but reported the improvement of pain defined as a > 50% decrease in VAS during study follow‐up. Study authors reported a greater reduction in pain in the BMMNC group (P = 0.045). Lindeman 2018 used the PISQ, and reported that no differences were observed in average pain reduction between BMMNC and control group (P = 0.23). Pignon 2017 assessed pain with VAS, where 0 was no pain at all and 100 was the most severe pain experienced. Study authors reported that no differences in pain reduction were observed between the BMMNC and control group. | ||
| Incidence of amputation (follow‐up: 6 or 12 months) | Study population | RR 0.52 (0.27 to 0.99) | 176 | ⊕⊝⊝⊝ | All studies reported the incidence of amputation. A possible small benefit was no longer seen after removal of 2 studies at high risk of bias in sensitivity analysis (RR 0.52, 95% CI 0.19 to 1.39; 89 participants, 2 studies) (Barc 2006; Li 2013). | |
| 250 per 1000 | 130 per 1000 (68 to 248) | |||||
| Angiographic analysis | See comment | ‐ | ‐ | None of the studies reported angiographic analysis. | ||
| Increase in ABI An ABI ratio of 0.9 or less indicates PAD. Values between 0.9 and 1.0 are borderline, and above 1.0 is considered normal (follow‐up: range 1 month to 12 months) | All studies measured ABI, but due to incomplete information we were unable to pool the data. See comment | 176 | ⊕⊝⊝⊝ | Barc 2006 reported that ABI did not change from baseline during study follow‐up in both BMMNC and control groups. Li 2013 used an absolute increase of > 15% ABI to quantify haemodynamic improvement. Study authors reported that this was greater in the BMMNC group compared to control (P = 0.002). Lindeman 2018 reported no difference in mean ABI between the BMMNC and control groups after 12 months follow‐up (P = 0.50). Pignon 2017 did not show any changes in median ABI value between groups. | ||
| Increase in PFWD (m) (follow‐up: 12 months) | See comment | 53 | ⊕⊕⊝⊝ | Only Lindeman 2018 reported PFWD, finding no clear difference between groups. The mean PFWD for the treatment and control groups at 12 months follow‐up was 128 ± 71 m vs 160 ± 11 m, P = 0.87. | ||
| Side effects and complications (follow‐up: 6 or 12 months) | Study population | RR 2.13 (0.50 to 8.97) | 176 (4 RCTs) | ⊕⊝⊝⊝ | All trials reported adverse events. 2 studies reported that no identifiable treatment‐related adverse events were observed, and that the therapy was well‐tolerated (Barc 2006; Pignon 2017). Li 2013 reported adverse events in both BMMNC and control groups (3 vs 1 fever, 1 vs 0 MI, 0 vs 1 stroke). There were no differences in the incidence of adverse events between groups, and the therapy was well‐tolerated Lindeman 2018 reported that leukaemia occurred in 1 participant during the follow‐up period. It is unclear if this was related to the procedure. After applying the sensitivity analysis by removing 2 studies at high risk of bias (Barc 2006; Li 2013), the significance of the results did not change (RR 2.69, 95% CI 0.11 to 63.18). | |
| 23 per 1000 | 48 per 1000 (11 to 204) | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). ABI: ankle‐brachial index;BMMNC: bone marrow mononuclear cells; CI: confidence interval; CLI: critical limb ischaemia;MI: myocardial infarction; PAD: peripheral arterial disease; PFWD: pain‐free walking distance;PISQ: pain inventory score questionnaire; RCT: randomised controlled trial; RR: risk ratio; VAS: visual analogue scale | ||||||
| GRADE Working Group grades of evidence | ||||||
| aThe stem cells used in the included RCTs originate from bone marrow. All four trials used mononuclear cells collected during the harvesting procedure from bone marrow and implanted into affected limbs (Barc 2006; Li 2013; Lindeman 2018; Pignon 2017). | ||||||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1.1 All‐cause mortality Show forest plot | 3 | 123 | Risk Ratio (M‐H, Random, 95% CI) | 1.00 [0.15, 6.63] |
| 1.2 Amputation Show forest plot | 4 | 176 | Risk Ratio (M‐H, Random, 95% CI) | 0.52 [0.27, 0.99] |
| 1.3 Amputation ‐ sensitivity analysis Show forest plot | 2 | 89 | Risk Ratio (M‐H, Random, 95% CI) | 0.52 [0.19, 1.39] |
| 1.4 Side effects and complications Show forest plot | 4 | 176 | Risk Ratio (M‐H, Random, 95% CI) | 2.13 [0.50, 8.97] |
| 1.5 Side effects and complications ‐ sensitivity analysis Show forest plot | 2 | 89 | Risk Ratio (M‐H, Random, 95% CI) | 2.69 [0.11, 63.18] |
