Factores de crecimiento para la angiogénesis en la enfermedad arterial periférica
Resumen
Antecedentes
La arteriopatía periférica (AP) se asocia con una carga clínica y socioeconómica alta. Los tratamientos para aliviar los síntomas de la AP y reducir los riesgos de amputación y de muerte son una prioridad alta para la sociedad. Varios factores de crecimiento han mostrado potencial para estimular la angiogénesis. Los factores de crecimiento administrados de forma directa (como proteínas recombinantes), o indirecta (p.ej. por medio de vectores virales o plásmidos de ADN que codifican estos factores), han surgido como una estrategia alentadora para el tratamiento de los pacientes con AP.
Objetivos
Evaluar los efectos de los factores de crecimiento que promueven la angiogénesis para el tratamiento de los pacientes con AP de las extremidades inferiores.
Métodos de búsqueda
El especialista en información del Grupo Cochrane Vascular (Cochrane Vascular Group) buscó en el registro especializado (junio 2016) y en CENTRAL (2016, número 5). Se realizaron búsquedas en registros de ensayos para obtener detalles de estudios en curso o no publicados. También se verificaron las listas de referencias de publicaciones relevantes y, si fue necesario, se intentó contactar con los autores de los ensayos para obtener detalles de los estudios.
Criterios de selección
Se incluyeron ensayos controlados aleatorizados que comparaban los factores de crecimiento (administrados de forma directa o indirecta) con ninguna intervención, placebo u otra intervención no basada en la acción del factor de crecimiento en pacientes con AP de las extremidades inferiores. Los resultados primarios fueron la amputación del miembro, la muerte y los eventos adversos. Los resultados secundarios comprendieron la capacidad de caminata, las medidas hemodinámicas, la ulceración y el dolor en reposo.
Obtención y análisis de los datos
Dos autores de la revisión, de forma independiente, seleccionaron los ensayos y evaluaron el riesgo de sesgo. Se utilizaron los resultados de los estudios en riesgo bajo de sesgo para el análisis principal y de todos los estudios en los análisis de sensibilidad. Se calcularon los odds ratios (OR) para los datos dicotómicos y las diferencias de medias (DM) para los datos continuos, con intervalos de confianza (IC) del 95%. Se evaluó la heterogeneidad estadística mediante la estadística I2 y la prueba Q de Cochrane. Se realizó el metanálisis para el efecto general y para cada factor de crecimiento como un análisis de subgrupos mediante el OR en un modelo de efectos fijos. Se evaluó la solidez de los resultados en un análisis de sensibilidad mediante el riesgo relativo (RR) o un modelo de efectos aleatorios. También se evaluó la calidad de la evidencia para cada resultado.
Resultados principales
Se incluyeron 20 ensayos en la revisión y se utilizaron 14 estudios (con aproximadamente 1400 participantes) con resultados publicados en los análisis. Seis estudios publicados compararon los factores de crecimiento de fibroblastos (FCF), cuatro estudios los factores de crecimiento de hepatocitos (FCH) y otros cuatro estudios los factores de crecimiento endotelial vascular (FCEV), versus placebo o ningún tratamiento. Seis de estos estudios investigaron exclusivamente o principalmente a participantes con claudicación intermitente y ocho estudios exclusivamente a participantes con isquemia crítica del miembro. El seguimiento varió generalmente de tres meses a un año. Dos estudios pequeños proporcionaron algunos datos a los dos años y uno de ellos también a los diez años.
La dirección y el tamaño de los efectos de los factores de crecimiento en las amputaciones de miembros principales (OR 0,99, IC del 95%: 0,71 a 1,38; 10 estudios, N = 1075) y la muerte (OR 0,99, IC del 95%: 0,69 a 1,41; 12 estudios, N = 1371) hasta los dos años son inciertos. La calidad de la evidencia es baja debido al riesgo de sesgo y la imprecisión (un año más tarde, evidencia de calidad moderada debido a la imprecisión). Sin embargo, los factores de crecimiento pueden disminuir la tasa de cualquier amputación de extremidades (OR 0,56, IC del 95%: 0,31 a 0,99; 6 estudios, N = 415). La calidad de la evidencia es baja debido al riesgo de sesgo y al informe selectivo.
La dirección y el tamaño de los efectos de los factores de crecimiento sobre los eventos adversos graves (OR 1,09, IC del 95%: 0,79 a 1,50; 13 estudios, N = 1411) y sobre cualquier evento adverso (OR 1,10, IC del 95%: 0,73 a 1,64; 4 estudios, N = 709) hasta los dos años también son inciertos. La calidad de la evidencia es baja debido al riesgo de sesgo y la imprecisión (para los eventos adversos graves un año más tarde, evidencia de calidad moderada debido a la imprecisión).
Los factores de crecimiento pueden mejorar las medidas hemodinámicas (evidencia de baja calidad), la ulceración (evidencia de muy baja calidad) y el dolor en reposo (evidencia de muy baja calidad) hasta el año, aunque tienen poco o ningún efecto sobre la capacidad de caminata (evidencia de baja calidad). No se identificaron diferencias relevantes en los efectos entre los factores de crecimiento (FCF, FCH y FCEV).
Conclusiones de los autores
Los resultados de esta revisión no apoyan la administración del tratamiento con los factores de crecimiento FCF, FCH ni FCEV en los pacientes con AP de las extremidades inferiores para evitar la muerte o la amputación mayor del miembro ni para mejorar la capacidad de caminata. Sin embargo, la utilización de estos factores de crecimiento puede mejorar las medidas hemodinámicas y disminuir la tasa de cualquier amputación de miembros (probablemente debido a la prevención de amputaciones menores) con un efecto incierto sobre los eventos adversos; una mejora de la ulceración y el dolor en reposo es muy incierta. Se necesitan ensayos nuevos con riesgo bajo de sesgo para generar evidencia de mayor certeza.
PICO
Resumen en términos sencillos
Tratamiento con sustancias que estimulan la formación de vasos sanguíneos nuevos (factores de crecimiento) en pacientes con reducción del suministro de sangre a las piernas
Antecedentes y pregunta de la revisión
Existen diferentes enfermedades, como la aterosclerosis, que pueden causar una reducción en el suministro de sangre a las piernas. Según la gravedad de la enfermedad la misma puede asociarse con síntomas como dolor en las piernas al caminar o al descansar, ulceración (heridas abiertas) y gangrena de la pierna. La reducción del suministro de sangre a las piernas afecta a un 3% a un 10% de todas las personas y a un 15% a un 20% de las personas de más de 70 años de edad. Es una causa principal de reducción de la movilidad y la calidad de vida y de un aumento del riesgo de amputación o muerte.
La estrategia de tratamiento en general comprende cambios conductuales (p.ej. abandono del hábito de fumar, ejercicio y dieta), fármacos (p.ej. antiplaquetarios, estatinas) e intervenciones quirúrgicas o con catéter. Sin embargo, para algunos pacientes la única opción es la amputación de la pierna.
Algunas sustancias producidas naturalmente en el cuerpo, denominadas factores de crecimiento, pueden estimular la formación de nuevos vasos sanguíneos. Actualmente, estas sustancias son producidas por laboratorios con la intención de tratar a los pacientes con un suministro reducido de sangre a las piernas. Por lo tanto, se evaluó la evidencia de los estudios clínicos sobre los efectos de los factores de crecimiento en dichos pacientes.
Resultados clave e implicaciones
Se identificaron 20 estudios y se analizaron los resultados de 14 estudios publicados que incluían a aproximadamente 1400 pacientes y que evaluaban tres tipos de factores de crecimiento (evidencia actual hasta junio de 2016).
La revisión muestra que no se conocen los efectos de los factores de crecimiento en los parámetros clínicos más importantes que comprenden las amputaciones del miembro por encima del tobillo, la muerte y los eventos adversos (evidencia de baja calidad hasta los dos años, pero evidencia de calidad moderada hasta el año). Sin embargo, la tasa de todas las amputaciones del miembro puede disminuir (evidencia de baja calidad). Además, los factores de crecimiento pueden mejorar los parámetros del flujo sanguíneo (evidencia de baja calidad), la ulceración (evidencia de muy baja calidad) y el dolor en reposo (evidencia de muy baja calidad) hasta un año, aunque tienen poco o ningún efecto sobre la capacidad de caminata (evidencia de baja calidad). La calidad de la evidencia se disminuyó principalmente debido al poder estadístico bajo y a la calidad deficiente del estudio. No se identificaron diferencias relevantes en los efectos entre los factores de crecimiento.
Esta revisión no apoya el tratamiento con factor de crecimiento en los pacientes con una reducción en el suministro de sangre a las piernas para prevenir las amputaciones del miembro por encima del tobillo o la muerte ni para mejorar la capacidad de caminata. Sin embargo, la utilización de factores de crecimiento puede mejorar los parámetros del flujo sanguíneo y evitar las amputaciones de miembros por debajo del tobillo con un efecto incierto sobre los efectos adversos; una mejora de la ulceración y el dolor en reposo es muy incierta. Se necesitan estudios nuevos de alta calidad para generar evidencia de mayor certeza.
Authors' conclusions
Summary of findings
| Growth factor therapy versus placebo or no therapy for the treatment of peripheral arterial disease | |||||
| Patient or population: people who have been diagnosed with peripheral arterial disease Setting: all settings Intervention: growth factor therapy Comparison: placebo or no therapy | |||||
| Outcomes | № of participants | Quality of the evidence | Relative effect | Anticipated absolute effects (95% CI)* | |
| Risk with placebo | Risk difference | ||||
| Major limb amputation | 1075 (10 RCTs) | ⊕⊕⊝⊝ lowa | OR 0.99 (0.71 to 1.38) | 19 per 100 (86/452) | 0 per 100 |
| Any limb amputation | 415 (6 RCTs) | ⊕⊕⊝⊝ Lowb | OR 0.56 (0.31 to 0.99) | 28 per 100 (86/452) | 10 fewer per 100 |
| Death | 1371 (12 RCTs) | ⊕⊕⊝⊝ Lowa (⊕⊕⊕⊝ moderate for 1 year) | OR 0.99 (0.69 to 1.41) | 12 per 100 (65/546) | 0 per 100 |
| Serious adverse events | 1411 (13 RCTs) | ⊕⊕⊝⊝ Lowa (⊕⊕⊕⊝ moderate for 1 year) | OR 1.09 (0.79 to 1.50) | 20 per 100 (113/555) | 1 more per 100 |
| Any adverse event | 1411 (13 RCTs) | ⊕⊕⊝⊝ Lowa (⊕⊕⊕⊝ moderate for 1 year) | OR 1.10 (0.73 to 1.64)c | 82 per 100 (338/555)c | 1 more per 100 |
| *The basis for the anticipated risk in the study population was the average risk in the control group (i.e. the total number of participants with events divided by the total number of participants in the control group for all studies reporting the outcome). The estimation of the risk difference (and its 95% confidence interval) was based on the anticipated risk in the control group and the relative effect of the intervention (and its 95% CI). | |||||
| GRADE Working Group grades of evidence | |||||
| aDowngraded by two levels due to imprecision and risk of bias; for 1 year only by one level due to imprecision. | |||||
Background
Description of the condition
Peripheral artery disease (PAD) commonly refers to stenosis (narrowing) and occlusion (blocking) of the peripheral arteries. While the most important cause of PAD is atherosclerosis, other non‐acute or acute causes such as vasculitis, thrombosis and embolic disease are possible. Major risk factors for PAD are smoking, diabetes, dyslipidaemia and hypertension (Fowkes 2013; Tendera 2011). PAD may affect different vascular sites such as lower and upper extremities, the carotid and mesenteric vessels (Hirsch 2006; Tendera 2011). Due to clinical importance, we have limited the scope of this review to PAD of the lower extremity arteries.
While many PAD patients are asymptomatic, symptoms of varying severity may occur depending on the degree of stenosis and insufficiency of blood (i.e. oxygen) supply to the distal tissues. At a low degree of stenosis, PAD is characterised by leg pain induced during exercise or walking (intermittent claudication), and at a higher degree also by rest pain of the affected leg, ulceration and gangrene of the foot (critical limb ischaemia). Common classification systems for the clinical severity of PAD are the Fontaine and Rutherford classifications (Gardner 2008). PAD is a major cause of decreased mobility, functional capacity, quality of life and increases the risks of amputation and/or death (Hirsch 2006; Ouma 2012; Tendera 2011).
PAD affects 3% to 10% of the general population and 15% to 20% of people over 70 years of age and is slightly more prevalent in men than in women (Norgren 2007). Globally, 202 million people were living with PAD in 2010, 69.7% of them in low‐ or middle‐income countries (Fowkes 2013). The annual incidence of critical limb ischaemia ranges from 500 to 1000 and of major amputations from 120 to 500 new cases per million population (Tendera 2011). Since atherosclerosis is a systemic disease, PAD is associated with an increased risk of coronary artery and cerebrovascular diseases, which are the major causes of death in people with PAD (Tendera 2011).
According to current guidelines, the treatment strategy for people with PAD comprises behavioural modifications (e.g. smoking cessation, daily exercise, diet), drugs that decrease the risk of cardiovascular events (e.g. antiplatelet drugs, statins), and/or drugs that improve PAD symptoms (Hirsch 2006; Rooke 2011; Tendera 2011). People with critical limb ischaemia are treated additionally either with catheter‐based revascularisation or bypass surgery. A number of up‐to‐date Cochrane reviews have evaluated some of these treatments (Bedenis 2014; De Backer 2012; Lane 2014; Salhiyyah 2015). For people who fail to respond to conservative therapies and are not candidates for invasive procedures, there is no option to relieve symptoms other than amputation (Hirsch 2006; Tendera 2011).
Description of the intervention
With respect to PAD, the term 'growth factors' comprises all physiologically produced substances with the potential to stimulate new vessel formation (angiogenesis). The following growth factors have already shown the potential to promote angiogenesis and are of particular interest in people with PAD (Ouma 2012).
-
Vascular endothelial growth factor (VEGF, isoforms: VEGF‐A through ‐E).
-
Placental growth factors (PLGF, isoforms: PLGF‐1 and ‐2).
-
Fibroblast growth factor (FGF, isoforms: 23 classes FGF‐1 to FGF‐23).
-
Hepatocyte growth factor (HGF, isoforms HGF/NK1 and HGF/NK2).
-
Platelet‐derived growth factor (PDGF, isoforms: PDGF‐A through ‐D; 5 homo/heterodimers: PDGF‐AA, ‐AB, ‐BB, ‐CC, and ‐DD).
-
Angiopoietin (Ang, isoforms: Ang‐1, ‐2, ‐3, and ‐4).
-
Erythropoietin (EPO).
Growth factors may be delivered to the target tissues directly in the form of recombinant proteins (also using carriers like nanoparticles). The local increase in growth factors can also be achieved indirectly by substances promoting their gene expression (e.g. transcription factors), as well as by viral vectors, DNA plasmids encoding or by cells synthesising these factors (De Haro 2009; Ouma 2012). There are also different routes of application such as intramuscular (IM), intra‐arterial and subcutaneous (Ouma 2012).
How the intervention might work
The exact mechanisms of growth factors' action in PAD are still unclear. Growth factors are important for regulating a variety of cellular processes. They bind to specific receptors on the surface of the target cells and may act by stimulating, for example, cellular growth, proliferation, differentiation and maturation. This promotes therapeutic angiogenesis and blood supply to the distal tissues. Therefore, growth factors may ameliorate the symptoms and prevent amputation in people with PAD (Ouma 2012).
Why it is important to do this review
Given the high clinical and socioeconomic burden of PAD of the lower extremities, treatments for its alleviation are a high societal priority. This is especially important for people with critical limb ischaemia whose option for approved therapy is only amputation.
Several growth factors have already shown the potential to stimulate angiogenesis in animal models. Therefore, their use has emerged as a promising strategy to treat patients with PAD (Ouma 2012). This systematic review of randomised clinical trials will help to determine benefits and harms of growth factor use in people with PAD.
Objectives
To assess the effects of growth factors that promote angiogenesis for treating people with PAD of the lower extremities.
Methods
Criteria for considering studies for this review
Types of studies
Randomised controlled trials (RCTs).
Types of participants
People who have been diagnosed with PAD of the lower extremities (on clinical and investigative assessment). We did not place any restrictions on setting, age, sex or severity of PAD.
Types of interventions
Intervention: growth factors delivered directly (as recombinant proteins) or indirectly (by substances promoting their gene expression, by viral vectors or DNA plasmids encoding or by cells synthesising these factors).
Comparison: no intervention, placebo or any other intervention not based on growth factor's action.
Types of outcome measures
Primary outcomes
-
Limb amputation
-
Death
-
Severe complications/adverse events (e.g. stroke, myocardial infarction, neoplasia)
Secondary outcomes
-
Walking ability (peak walking time/distance, claudication onset time/distance)
-
Haemodynamic measures (e.g. ankle brachial index (ABI), toe brachial index (TBI))
-
Ulceration (e.g. surface area, ulcer healing)
-
Rest pain (e.g. visual analogue scale (VAS), pain improvement)
Search methods for identification of studies
We did not apply any restrictions on date or language of publications.
Electronic searches
The Cochrane Vascular Information Specialist (CIS) searched the following databases for relevant trials.
-
The Cochrane Vascular Specialised Register (9 June 2016).
-
The Cochrane Central Register of Controlled Trials (CENTRAL; 2016, Issue 5) via the Cochrane Register of Studies Online.
See Appendix 1 for details of the search strategy used for CENTRAL.
The Cochrane Vascular Specialised Register is maintained by the CIS and is constructed from weekly electronic searches of MEDLINE Ovid, Embase Ovid, CINAHL and AMED as well as through handsearching relevant journals. The full list of the databases, journals and conference proceedings searched, as well as the search strategies used, are described in the Specialised Register section of the Cochrane Vascular module in the Cochrane Library (www.cochranelibrary.com).
In addition, the CIS searched the following trial databases for details of ongoing and unpublished studies (9 June 2016).
-
World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP) Search Portal. (apps.who.int/trialsearch).
-
ClinicalTrials.gov (clinicaltrials.gov).
-
International Standard Randomised Controlled Trial Number (ISRCTN) registry (www.isrctn.com).
The terms used to search were:
-
'claudication' and 'growth factor';
-
'ischaemia' and 'growth factor';
-
'growth factor' and 'peripheral arterial'.
Searching other resources
We checked reference lists of relevant publications retrieved from the electronic searches to identify additional studies. If necessary, we also contacted the authors of publications and trialists of published, unpublished or ongoing trials for details of the studies.
Data collection and analysis
Selection of studies
Two review authors (VG, AH) independently assessed the eligibility of the retrieved publications on studies for inclusion in the review; we resolved disagreements by consensus or by discussion with a third author (MUB). We performed the selection process in three stages, assessing titles, abstracts and full texts. We recorded the reasons for excluding any full‐text publication in the Characteristics of excluded studies table.
Data extraction and management
One review author (VG) extracted information on trial participants and interventions using a standard data collection form developed by Cochrane Vascular, while a second author (AH) checked the extracted data. The two review authors independently recorded data concerning the outcome measures on forms developed by Cochrane Vascular, resolving disagreements by consensus. When necessary, we requested additional information from the study authors.
Assessment of risk of bias in included studies
We used Cochrane's tool for assessing risk of bias, provided by Cochrane Vascular, to assess allocation (selection bias); blinding of participants, clinical personnel and outcome assessors (performance bias and detection bias); incomplete outcome data (attrition bias); selective reporting (reporting bias); and other potential sources of bias (Higgins 2011). Two review authors (VG, AH) independently assessed the risk of bias of each trial, resolving any disagreements by consensus. We judged each item of bias and produced a summary score of low, unclear or high risk of bias for each outcome of the study. We used the assessment of risk of bias to select the results of the studies at low risk of bias (judgments for all items relevant for this outcome at low risk of bias) for the main analyses, and for interpreting the quality (i.e. certainty) of the determined evidence.
Measures of treatment effect
We chose the effect measures for statistical analysis based on guidance from the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We used the odds ratio (OR) as the measure of effect for dichotomous outcomes and mean difference (MD) for continuous outcomes. In case of different scales in the different studies, we standardised the results and used standardised mean differences (SMDs) where possible. We calculated the effect estimates with the corresponding 95% confidence intervals (CIs) and present the results graphically as forest plots.
Unit of analysis issues
We used the participant as the unit of analysis in each trial and the trial as the unit for data synthesis. We performed separate analyses according to different periods of follow‐up. We did not include studies with non‐standard designs.
Dealing with missing data
We requested missing data regarding participant demographics, description of intervention, details of study design and outcome measures from study authors, if necessary. We considered incomplete outcome data (e.g. due to dropouts or non‐response on questionnaires) to influence the risk of bias for these outcomes.
Assessment of heterogeneity
We evaluated clinical heterogeneity based on data for participants, interventions and outcomes of the studies. We also assessed the results of the studies with respect to statistical heterogeneity using the I2 statistic and Cochrane's Q test according to the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We considered an I2 greater than 30% or a P value of the Chi2 statistic of less than 0.1 to be indicative of statistical heterogeneity among the studies; an I2 greater than 50% indicated substantial statistical heterogeneity. In case of statistical heterogeneity we tried to explore possible causes.
Assessment of reporting biases
To minimise reporting bias, we checked registers of prospective trial registration. When appropriate, we applied funnel plots to assess publication bias.
Data synthesis
We performed data synthesis using Review Manager 5 software (RevMan 5) according to Cochrane Vascular guidelines (Review Manager 2014). We included studies at low risk of bias in the main analysis according to section 8.8.3.1 of Higgins 2011 and all studies in the sensitivity analysis. We combined data using different effect measures (OR, RR) and models (fixed‐effect or random‐effects). We present results of the studies and the meta‐analyses (using OR with a fixed‐effect model) in forest plots. We combined the results of the studies for all growth factors in the meta‐analyses to increase precision of the effect estimates under the assumption of similar effect‐sizes of different growth factors and their isoforms.
Subgroup analysis and investigation of heterogeneity
In case of sufficient data we planned to conduct the following subgroup analyses.
-
Growth factors.
-
Delivery approaches of growth factors.
-
Routes of administration.
-
PAD severity.
-
Sex.
-
Age groups.
Sensitivity analysis
We undertook a sensitivity analysis to examine the robustness of the observed findings by including the studies with a higher risk of bias in the meta‐analysis. We assessed the robustness of the combined results in the meta‐analysis using different statistical measures (OR, RR) and models (fixed‐effect or random‐effects).
Summary of findings table
We summarised the results of the analysis on primary outcomes in a 'Summary of findings' table. The 'Summary of findings' table also provides information concerning the quality of evidence for all comparisons. Two review authors (VG, AH) independently assessed the quality (i.e. certainty) of the evidence for each outcome using the system developed by the GRADE working group by assessing risk of bias, indirectness, inconsistency, imprecision and publication bias (Balshem 2011). We resolved disagreements by consensus. As the context of the Cochrane Review users may differ (subsets of population, settings and modifications of the intervention), we made separate judgments about applicability of the evidence and the quality of evidence.
Results
Description of studies
Results of the search
See Figure 1.
Study flow diagram.
Included studies
Our review included 20 trials presented in 40 records (Cooper 2001; Deev 2015; HGF‐0205; HGF‐STAT; Kibbe 2016; Kusumanto 2006; Lazarous 2000; Makinen 2002; Matyas 2005; NCT00080392; NCT00304837; Rauh 1999; RAVE; Shigematsu 2010; TALISMAN‐201; TALISMAN‐202; TALISMAN‐211; TAMARIS; TRAFFIC; VM202‐China), and we used 14 trials with published results in the qualitative or quantitative analyses (see Table 1, Characteristics of included studies). Six of these studies exclusively or mainly investigated participants with intermittent claudication (Cooper 2001; Deev 2015; Lazarous 2000; Makinen 2002; RAVE; TRAFFIC), and eight studies included only participants with critical limb ischaemia (HGF‐0205; HGF‐STAT; Kibbe 2016; Kusumanto 2006; Matyas 2005; Shigematsu 2010; TALISMAN‐201; TAMARIS). We requested additional information with respect to methodological issues and/or results of the studies from the authors of all studies included in the analyses; however, we did not receive any responses.
| Trial | N participants | Indication | Intervention | Placebo |
| Fibroblast growth factor | ||||
| 19 | Intermittent claudication | 1 × 10 µg/kg b‐FGF, n = 4 | n = 6 | |
| 1 × 30 µg/kg b‐FGF, n = 5 | ||||
| 2 × 30 µg/kg b‐FGF, n = 4 | ||||
| 24 (plan 108) | Intermittent claudication | 6 × 2 µg/kg b‐FGF, n = 16 | n = 8 | |
| 190 | Intermittent claudication | 1 × 30 µg/kg FGF‐2, n = 66 | n = 63 | |
| 2 × 30 µg/kg FGF‐2, n = 61 | ||||
| 125 | Critical limb ischaemia | 16 mg (4 × 4 mg) NV1FGF, n = 59 | n = 66 | |
| 71 | Critical limb ischaemia | 2‐16 mg [5 groups] NV1FGF, n = ? | n = ? | |
| 36 | Critical limb ischaemia | 16 mg (4 × 4 mg) NV1FGF, n = ? | n = ? | |
| 32 mg (4 × 8 mg) NV1FGF, n = ? | ||||
| 525 | Critical limb ischaemia | 16 mg (4 × 4 mg) NV1FGF, n = 266 | n = 259 | |
| 11 + 2b (plan 28) | Critical limb ischaemia | 2.87 × 108 Ad5FGF‐4, n = 3 + 2b | n = 3 | |
| 2.87 × 109 Ad5FGF‐4, n = 3 | ||||
| 2.87 × 1010 Ad5FGF‐4, n = 2 | ||||
| Hepatocyte growth factors | ||||
| 104 | Critical limb ischaemia | 3 × 0.4 mg AMG0001, n = 26 | n = 26 | |
| 2 × 4.0 mg AMG0001, n = 25 | ||||
| 3 × 4.0 mg AMG0001, n = 27 | ||||
| 27 (plan 48) | Critical limb ischaemia | 3 × 4.0 mg AMG0001, n = 21 | n = 6 | |
| 40 (plan 46) | Critical limb ischaemia | 2 × 4.0 mg AMG0001, n = 27 | n = 13 | |
| 52 | Critical limb ischaemia | 2 × 4.0 mg VM202 (HGF‐X7), n = 21 | n = 11 | |
| 4 × 4.0 mg VM202 (HGF‐X7), n = 20 | ||||
| 200 | Critical limb ischaemia | 3 × 4.0 mg NL003 (pCK‐HGF‐X7), n = ? | n = ? | |
| 3 × 6.0 mg NL003 (pCK‐HGF‐X7), n = ? | ||||
| 3 × 8.0 mg NL003 (pCK‐HGF‐X7), n = ? | ||||
| Vascular endothelial growth factor | ||||
| 13 | Critical limb ischaemia | VEGF‐C, n = ? | n = ? | |
| 54 (plan 60) | Intermittent claudicationc Critical limb ischaemia | VEGF‐Ad (2 × 1010 VEGF165), n = 18 | n = 19 | |
| VEGF‐P/L (2 × 2.0 mg VEGF165), n = 17 | ||||
| 105 (plan 105) | Intermittent claudication | 4 × 109 PU AdVEGF121, n = 32 | n = 33 | |
| 4 × 1010 PU AdVEGF121, n = 40 | ||||
| 54 | Critical limb ischaemia | 2 × 2.0 mg phVEGF165, n = 27 | n = 27 | |
| 100 | Intermittent claudicationc, Critical limb ischaemia | 2 × 1.2 mg pCMV‐VEGF165, n = 75 | n = 25d | |
| 10 | Intermittent claudication | VEGF‐A, n = ? | n = ? | |
| ? | Critical limb ischaemia | VEGF‐C, n = ? | n = ? | |
aUnpublished trial.
b2 participants not randomised.
cMainly.
dNo therapy.
Fibroblast growth factor (FGF)
Eight studies assessed FGF: three focused on fibroblast growth factor‐2 ('basic fibroblast growth factor') for intermittent claudication (Cooper 2001; Lazarous 2000; TRAFFIC); four studies evaluated plasmid‐delivered (non‐viral) fibroblast growth factor‐1 (NV1FGF) (TALISMAN‐201; TALISMAN‐202; TALISMAN‐211; TAMARIS), and one adenovirus 5‐delivered fibroblast growth factor‐4 (Ad5FGF‐4) (Matyas 2005) for critical limb ischaemia.
The first of the three studies on FGF‐2 was a small dose‐escalation phase I trial including only 19 participants (Lazarous 2000), followed by a larger phase II multicentre trial including 190 participants (TRAFFIC). Both trials provided results for one or two doses of 30 µg/kg FGF‐2, although the second dose was given on day 2 in Lazarous 2000 and on day 30 in TRAFFIC. The third trial investigated a 6‐week infusion of 2 µg/kg per week (12 µg/kg FGF‐2 in total) (Cooper 2001). Investigators did not finally implement the planned dose escalation in this study because early data from 24 participants (of the 108 who were to be enrolled) showed development of proteinuria in the treatment arm, prompting the premature termination of the study.
Of the four studies on NV1FGF, two had not been published at the time of writing the review, although both were completed in 2005: TALISMAN‐202 (N = 71) was a dose‐escalating trial on five treatment regimens of NV1FGF at doses 2 mg to 16 mg, while TALISMAN‐211 (N = 36) assessed two doses of NV1FGF, 16 mg and 32 mg. Both published trials investigated the total doses of NV1FGF of 16 mg (four sessions of eight injections) and included 125 or 525 participants, respectively (TALISMAN‐201; TAMARIS).
There was also one small trial on Ad5FGF‐4, which investigated escalating doses of 2.87 × 108 to 2.87 × 1010 viral particles (Matyas 2005). The trial presented data for only 11 participants from a protocol with randomisation, as well as for 2 additional participants from a separate but comparable protocol (we do not include results for these participants in this review).
Hepatocyte growth factors (HGF)
Five trials investigated HGF, all studies delivered via plasmid, in participants with critical limb ischaemia. Three studies used the HGF plasmid product AMG0001 (HGF‐STAT; HGF‐0205; Shigematsu 2010). Two studies used the plasmid DNA pCK‐HGF‐X7, which expresses two isoforms of HGF (VM202 in Kibbe 2016 and NL003 in VM202‐China).
Of three studies on AMG0001, the earliest was a dose‐response study with 104 participants that investigated the total doses of 1.2 mg, 8 mg and 12 mg AMG0001 (in three sets of IM injections) (HGF‐STAT). The second trial used only a total dose of 12 mg AMG0001 (in three sets of IM injections) and included 27 participants (HGF‐0205). The third trial investigated only a total dose of 8 mg AMG0001 (in two sets of IM injections); the study was terminated prematurely because of slow recruitment after an interim analysis based on data from the first 40 of the 120 planned participants (Shigematsu 2010).
One trial investigated HGF plasmid VM202 in total doses of 8 mg and 16 mg (in two or four sessions of IM injections, respectively) and included 52 participants with critical limb ischaemia (Kibbe 2016). The second trial was designed to use HGF plasmid NL003 in the total doses of 12 mg, 18 mg and 24 mg (each in three sessions of IM injections) and included 200 participants (VM202‐China); this study had not yet been published at the time of writing, although it was set to be completed in 2014.
Vascular endothelial growth factor (VEGF)
Seven trials assessed the effects of VEGF: five used genes encoding isoform VEGF‐A (Deev 2015; Kusumanto 2006; Makinen 2002; NCT00080392; RAVE), and two isoform VEGF‐C (NCT00304837; Rauh 1999). The results of both studies on isoform VEGF‐C (in participants with critical limb ischaemia) and of one small study on isoform VEGF‐A (in 10 participants with intermittent claudication) have not been published yet (NCT00304837; NCT00080392; Rauh 1999).
Of four published trials on isoform VEGF‐A, one used variant VEGF121 (RAVE). This study included 105 participants with intermittent claudication and investigated the effect of two doses (4 × 109 or 4 × 1010 particle units) of recombinant adenovirus encoding VEGF121, which were applied through IM injections.
Three other published trials on isoform VEGF‐A used the variant VEGF165. Makinen 2002 investigated its effect separately for adenovirus (2 × 1010 particle units) and plasmid (2000 µg) gene delivery through intra‐arterial infusion over balloon catheter after percutaneous transluminal angioplasty; the study included 54 participants with PAD, 14 of them with critical limb ischaemia. One other trial on variant VEGF165, Kusumanto 2006, included only 54 diabetic participants with critical limb ischaemia and delivered VEGF165 as plasmid (2000 µg) in two series of IM injections. Deev 2015 included 100 participants with PAD, 13 of them with critical limb ischaemia, investigating smaller doses of VEGF165 plasmid (1200 µg), also administered in two series of IM injections.
Excluded studies
We excluded 40 studies following full‐text assessment. The reasons for exclusion were: participants without PAD disease, no separate analysis for participants with PAD, irrelevant therapy or comparator, therapy effect based not only on growth factor action, or published study is not RCT (see Characteristics of excluded studies).
Risk of bias in included studies
We present the review authors' judgments and justification for each risk of bias item for each included study in the Characteristics of included studies table. Figure 2 summarises the judgments about each risk of bias item for all studies.
Risk of bias summary: review authors' judgments about each risk of bias item for each included study.
Allocation
Only five trials report appropriate details on random sequence generation (Kusumanto 2006; Makinen 2002; Shigematsu 2010; TAMARIS; TRAFFIC), and just four trials described allocation concealment (Kusumanto 2006; Shigematsu 2010; TAMARIS; TRAFFIC). It is unclear whether many trials retrospectively excluded randomised participants who did not receive any intervention. Therefore, we assigned the judgment of 'low risk' of selection bias to only four trials (Kusumanto 2006; Shigematsu 2010; TAMARIS; TRAFFIC). Moreover, the randomisation in Deev 2015 was probably not truly performed, suggesting a high risk of selection bias for this trial. We judged the remaining trials to be at unclear risk of selection bias.
Blinding
Authors describe all trials but Deev 2015 as being double‐blind. However, just four studies described the blinding procedure of participants and personnel in enough detail to permit judgment of 'low risk' of performance bias (Makinen 2002; Matyas 2005; Shigematsu 2010; TAMARIS). Authors of Deev 2015 describe it as an open‐label study, suggesting a high risk of performance bias. We judged the remaining trials to be at unclear risk of performance bias.
We assume that the risk of bias due to subjective assessment of some outcomes is negligent (death, limb amputation, adverse events/severe complications, haemodynamic parameters and ulceration). Moreover, some trials performed additional steps to mask certain measurements and data analysis. Therefore, we judged all studies to be at low risk of detection bias for objective outcomes. However, only three studies explicitly provided sufficient data on the blinding of participants to permit a judgment of low risk of detection bias for highly subjective outcomes (rest pain and walking ability) (Makinen 2002; Matyas 2005; TAMARIS). Deev 2015 was an open‐label study, suggesting a high risk of detection bias for subjective outcomes. We considered the remaining studies to be at unclear risk of detection bias for subjective outcomes.
Incomplete outcome data
We deemed the risk of attrition bias to be unclear for study outcomes that were not measured or published (Figure 2). Data for adverse events and for death are complete in all published trials as well as data for amputations in eight trials (Deev 2015; Kibbe 2016; Kusumanto 2006; Makinen 2002; Matyas 2005; RAVE; Shigematsu 2010; TAMARIS); we judged these outcomes to be at low risk of bias for these trials. For efficacy analysis, two trials used the so‐called modified intention‐to‐treat (MITT) population, which included only participants who received at least two (in TALISMAN‐201) or all three (in HGF‐0205) treatment doses; therefore, results for all measured efficacy outcomes in these studies may be biased. The risk of attrition bias is also high for the outcomes of rest pain (Kusumanto 2006; Makinen 2002; Shigematsu 2010), ulceration (HGF‐STAT; Kibbe 2016; Kusumanto 2006; Makinen 2002; Shigematsu 2010), and haemodynamic parameters (HGF‐STAT; Kusumanto 2006), which were based on less than 70% of randomised participants. Due to a low proportion of missing data, the risk of attrition bias is low for the outcome ulceration in one trial (Matyas 2005), for rest pain in two (Kibbe 2016; Matyas 2005), for walking ability in three (Lazarous 2000; RAVE; TRAFFIC), and for haemodynamic parameters in three (Lazarous 2000; Matyas 2005; TRAFFIC). For the remaining outcomes, the risk of attrition bias is unclear because the studies do not report the number of participants with corresponding measurements.
Selective reporting
Some trials were completed but their results unpublished (NCT00080392; NCT00304837; Rauh 1999; TALISMAN‐202; TALISMAN‐211; VM202‐China), suggesting a risk of publication bias. One study measured but didn't report results for a number of secondary outcomes: ulceration, rest pain and haemodynamic parameters (because "the primary outcome was non‐significant") (TAMARIS). Three trials did not explicitly provide data for limb amputations (Cooper 2001; Lazarous 2000; TRAFFIC), one for deaths (Lazarous 2000), and many trials for secondary outcomes. In HGF‐STAT, the presented results for death, limb amputations, haemodynamic parameters and rest pain are only descriptive and thus insufficient. Therefore, the risk of selective reporting bias for these studies is unclear. For the other trials, the risk of selective reporting bias is low.
Other potential sources of bias
The studies with published results appear to be free of other potential sources of bias. For other trials, there was insufficient information available to assess whether an important risk of bias exists (NCT00080392; NCT00304837; Rauh 1999; TALISMAN‐202; TALISMAN‐211; VM202‐China).
Effects of interventions
Table 2 shows the reported parameters of each study. The results of meta‐analyses are presented in the section Data and analyses. See also summary of findings Table for the main comparison
| Trial | Amputation | Death | Adverse events/serious adverse events (SAE) | Walking | Haemodynamic parameters | Ulceration | Rest pain |
| Fibroblast growth factor | |||||||
| NR | NR | Data for single events | Improvement | Plethysmography | — | — | |
| NR | All | Data for single events | PWT, COT | ABI* | — | — | |
| NR | All | Data for single events | PWT, COT | ABI* | — | — | |
| Major, Any | All | Aggregate | — | ABI, TBI | Complete healing | VAS | |
| Unpublished trial | |||||||
| Unpublished trial | |||||||
| Major | All | Aggregate | — | NR | NR | NR | |
| Not defined | All | Data for single eventsa | — | ABI | Complete healing | VAS Improvement | |
| Hepatocyte growth factors | |||||||
| Alla | All | Aggregate (SAEs) | — | ABIa, TBIa | Areaa Complete healinga | Improvementa | |
| Major | All | Aggregate | — | ABI, TBI | Area Complete healing | VAS | |
| Major, Any | NR | Aggregate | — | ABIa | Improvement Complete healing | Improvement | |
| Major, Any | All | Aggregate (SAEs) | — | ABI | Improvement Complete healing | VAS Improvement | |
| Unpublished trial | |||||||
| Vascular endothelial growth factor | |||||||
| Unpublished trial | |||||||
| Major | All | Data for single events | — | ABI | Complete healing | Improvement | |
| Not defined | All | Data for single events | PWT, COT | ABIa | — | — | |
| Major | All | Data for single events | — | Improvement | Improvement | Improvement | |
| Not defined | All | Data for single events | PWD | ABI | — | — | |
| Unpublished trial | |||||||
| Unpublished trial | |||||||
ABI: ankle brachial index; COT: claudication onset time; NR: not reported; PWD: peak walking distance; PWT: peak walking time; TBI: toe brachial index; VAS: visual analogue scale.
aData not appropriate for meta‐analyses.
Limb amputation
Seven studies provide results for major amputations (HGF‐0205; Kibbe 2016; Kusumanto 2006; Makinen 2002; TALISMAN‐201; TAMARIS; Shigematsu 2010); three of these studies also reported results for all amputations (Kibbe 2016; TALISMAN‐201; Shigematsu 2010). Three other studies did not specify the types of reported amputations (Deev 2015; Matyas 2005; RAVE). Most reported follow‐up periods ranged from three months to one year. Two small trials provided data for two years (Deev 2015; Makinen 2002), and one of these trials also reported data at 10 years (Makinen 2002).
Size of the effect
In the main analysis, we could only derive an OR estimate at low risk of bias for major limb amputations from a single trial with data at one year (TAMARIS); the study observed similar rates of major limb amputations in the treatment (growth factors therapy) and control groups (25% versus 21%; OR 1.25, 95% CI 0.83 to 1.88; 525 participants).
In the sensitivity analysis, the conducted meta‐analysis found similar rates of major limb amputations in the treatment and control groups using either the data for one year (18% versus 21%; OR 1.00, 95% CI 0.71 to 1.41; 6 studies; 916 participants; Analysis 1.1) or the latest available data up to two years (16% versus 19%; OR 0.99, 95% CI 0.71 to 1.38; 10 studies; 1075 participants; Analysis 1.2). The meta‐analysis revealed decreased rates of any limb amputation in the treatment group using both the data for one year (14% versus 30%; OR 0.54, 95% CI 0.29 to 1.00; 4 studies; 364 participants; Analysis 1.3) and the latest available data for up to two years (14% versus 28%; OR 0.56, 95% CI 0.31 to 0.99; 6 studies; 415 participants; Analysis 1.4). There was no or low heterogeneity between the trials and between the subgroups of growth factors in the conducted analyses (besides in the FGF subgroup for major limb amputations). We found similar results in the corresponding meta‐analyses using risk ratio and/or the random‐effects model. One trial observed similar rates of limb amputations in the treatment and control groups at 10 years (9% versus 5%; 54 participants; Makinen 2002).
Quality of evidence
The evidence on limb amputations was direct. A potential publication bias and selective reporting (relevant for any amputations) may mask results not showing efficacy of growth factors. The evidence for major limb amputations at one year was at low risk of bias; the results were generally consistent between the main and the sensitivity analyses. However, due to imprecision (based on 95% CI, we cannot rule out a clinically relevant effect) the quality of evidence for major limb amputations at one year is moderate. The quality of evidence for major limb amputations for other time points is low due to risk of bias and imprecision. The quality of evidence for any limb amputation is low due to risk of bias and selective reporting (only three studies explicitly report results).
Summary judgment
The direction and magnitude of the effect for growth factors on major limb amputations in people with PAD of the lower extremities is uncertain (low‐quality evidence, but moderate‐quality evidence at one year). However, growth factors may decrease the rate of any amputations (low‐quality evidence).
Death
Twelve trials explicitly provide results for death (Cooper 2001; Deev 2015; HGF‐STAT; HGF‐0205; Kibbe 2016; Kusumanto 2006; Makinen 2002; Matyas 2005; RAVE; TALISMAN‐201; TAMARIS; TRAFFIC). Most reported follow‐up periods ranged from three months to one year. Two small trials provided data for two years (Deev 2015; Makinen 2002), and one of these trials also reported data at 10 years (Makinen 2002).
Size of the effect
In the main analysis, we could only derive an OR estimate at low risk of bias for death from a single trial with data at one year (TAMARIS); the study observed similar rates of death in the treatment and control groups (17% versus 15%; OR 1.18, 95% CI 0.74 to 1.88; 525 participants).
In the sensitivity analysis, the meta‐analysis showed similar rates of death in the treatment and control groups using either the data for one year (12% versus 14%; OR 0.98, 95% CI 0.67 to 1.44; 7 studies; 1038 participants; Analysis 1.5) or using the latest available data up to two years (9% versus 12%; OR 0.99, 95% CI 0.69 to 1.41; 12 studies; 1371 participants; Analysis 1.6). We found no heterogeneity in the meta‐analyses between the trials or between the subgroups of growth factors. We found similar results when using risk ratio and/or the random‐effects model in the meta‐analyses. We also observed similar rates of death in the treatment and control groups in one trial at 10 years (46% versus 53%; 54 participants; Makinen 2002).
Quality of evidence
The evidence on death was direct. The risk of publication bias and of selective reporting for death is low. The evidence for one year was at low risk of bias; the results were consistent between the main and the sensitivity analyses. However, due to imprecision (based on 95% CI, we cannot rule out a clinically relevant effect), the quality of evidence for one year is moderate. The quality of evidence for other time points is low due to risk of bias and imprecision of the data.
Summary judgment
The direction and magnitude of the effect of growth factors on death in people with PAD of the lower extremities is uncertain (low‐quality evidence, but moderate‐quality evidence at one year).
Severe complications/adverse events
All studies investigated adverse events. Many trials reported aggregate data on serious adverse events, some trials on all adverse events or on severe adverse events. Several trials did not provide aggregate data but only data for a single or group of adverse events. Most reported follow‐up periods ranged from three months to one year. Two small trials provided data for up to two years (Deev 2015; Makinen 2002), and one of these also reported data at 10 years (Makinen 2002).
Size of the effect
In the main analysis, we could only derive an OR estimate at low risk of bias for adverse events from a single trial with data for one year (TAMARIS). The study observed similar rates of serious adverse events (9% versus 6%; OR 1.60, 95% CI 0.82 to 3.12; 525 participants) and of any adverse events (80% versus 80%; OR 1.04, 95% CI 0.68 to 1.60; 525 participants) at one year in the treatment and control groups.
In the sensitivity analyses, the meta‐analyses using available data up to two years showed similar rates of serious adverse events in both groups based either only on studies with aggregate data (31% versus 24%; OR 1.11, 95% CI 0.74 to 1.65; 6 studies; 865 participants; Analysis 1.7) or on all studies (i.e. including self‐calculated overall event numbers for trials reporting only data for single events) (23% versus 20%; OR 1.09, 95% CI 0.79 to 1.50; 13 studies; 1411 participants; Analysis 1.8). The results of the trials showed low or no heterogeneity (besides in the FGF subgroup). However, the non‐heterogeneous ORs of the trials with reported aggregate data on any adverse events became moderately heterogeneous when we included self‐calculated overall event numbers from trials providing only data for single events in the analysis (I2 = 36%, P = 0.10). The meta‐analysis for any adverse events also showed similar rates in both groups based either only on the reported aggregate data (84% versus 82%; OR 1.10, 95% CI 0.73 to 1.64; 4 studies; 709 participants; Analysis 1.9) or on all studies (i.e. including self‐calculated event numbers) (61% versus 61%; OR 1.52, 95% CI 1.15 to 2.02; 13 studies; 1411 participants; Analysis 1.10). We found no or low heterogeneity in the meta‐analyses between the subgroups of growth factors. There were similar results when using only data for one year and when using risk ratio and/or the random‐effects model in the meta‐analyses. Likewise, we observed similar rates of adverse events in the treatment and control groups in one trial at 10 years (44% versus 43%; 54 participants; Makinen 2002).
Quality of evidence
The evidence on adverse events was direct. We do not expect any relevant publication bias or selective reporting for safety results. The evidence for one year was at low risk of bias; the results were generally consistent between the main and the sensitivity analyses as well as between analyses based on only studies with aggregate data and on all studies (although the self‐calculated event numbers may cause bias due to multiple counting). However, there was some imprecision of the effect estimate (based on 95% CI, we cannot rule out a relevant increase of serious adverse events). Therefore, the quality of evidence for one year is moderate. The quality of evidence for other time points is low due to risk of bias and imprecision of the data.
Summary judgment
The direction and magnitude of the effects for growth factors on serious and on any adverse events in people with PAD of the lower extremities are uncertain (low‐quality evidence, but moderate‐quality evidence at one year).
Walking ability
Results for walking ability were available from five of six trials including participants with intermittent claudication (Cooper 2001; Deev 2015; Lazarous 2000; RAVE; TRAFFIC). Three studies reported results for the peak walking time and for the claudication onset time up to six months (Cooper 2001; RAVE; TRAFFIC). One small study reported results for the pain‐free walking distance for six months, 12 months and two years (Deev 2015).
Size of the effect
We could not perform the main analysis, as the only trial that could provide results at low risk of bias did not report these parameters (TAMARIS).
The sensitivity meta‐analyses derived from the latest available data up to six months revealed no clinically important difference between the treatment and control groups in peak walking time (MD −0.17 min, 95% CI −0.69 to 0.36; 3 studies; 279 participants; Analysis 1.11) or in claudication onset time (MD −0.07 min, 95% CI −0.31 to 0.17; 3 studies; 279 participants; Analysis 1.12). We found no or low heterogeneity between the trials and between the subgroups of growth factors, although there was moderate heterogeneity between the subgroups for claudication onset time (I2 = 56%; P = 0.13). We found similar results in the meta‐analyses using the random‐effects model. However, the mean pain‐free walking distance investigated in one small study increased in the treatment group compared to the control group at six months, 12 months and two years (151 m, 231 m and 251 m; 100 participants).
Quality of evidence
The evidence on walking ability was direct, and the effect estimates of the meta‐analyses were precise, precluding a clinically relevant effect. We do not expect a relevant publication bias or selective reporting for walking ability. However, the evidence came only from the study results at unclear or high risk of bias. In addition, there was some inconsistency between the results for different parameters. Therefore, we judged the quality of evidence to be low.
Summary judgment
Growth factors have little or no effect on walking ability in people with PAD of the lower extremities (low‐quality evidence).
Haemodynamic measures
Each trial measured at least one haemodynamic parameter (mostly ABI or TBI), but studies often presented the results only graphically, descriptively or without data for the number of participants with measurements or for variability. The reported follow‐up ranged from three months to nine months. Two studies provided results for ABI at one year (Deev 2015; Kibbe 2016), and one provided data at two years (Deev 2015). One small study reported data on haemodynamic improvement (defined as absolute increase more than 15% in the ABI or TBI values) at three months (Kusumanto 2006).
Size of the effect
We could not perform the main analysis, as the only trial that could provide results at low risk of bias did not report these parameters (TAMARIS).
In the sensitivity meta‐analyses, we could include only results from six trials for ABI (Deev 2015; HGF‐0205; Kibbe 2016; Makinen 2002; Matyas 2005; TALISMAN‐201) and from two trials for TBI (HGF‐0205; TALISMAN‐201), derived from the latest available data up to six months. The meta‐analysis showed a small increase in mean ABI in the treatment group (MD 0.04, 95% CI −0.02 to 0.10; 6 studies; 341 participants; Analysis 1.13). We found no heterogeneity for ABI between the trials or between the subgroups of growth factors. However, the effect estimates between the trials and between the subgroups for change in TBI from baseline were substantially heterogeneous (I2 = 89%; P = 0.003); the meta‐analysis showed a small increase in mean TBI in the treatment group (MD 0.04, 95% CI −0.01 to 0.09; 2 studies; 128 participants; Analysis 1.14). We found similar results in the meta‐analyses using the random‐effects model. Results from trials for mean change in ABI within one year and within two years showed similar differences between groups. However, one small study showed an increased rate of participants with haemodynamic improvement at three months in the treatment group (33% versus 6%; 54 participants).
Quality of evidence
The evidence on haemodynamic measures was direct and was generally consistent between results for different parameters. However, the effect estimates of the meta‐analyses were imprecise, and we cannot rule out negative effects for growth factors on haemodynamic measures. Moreover, we could only derive the effect estimates from studies at unclear or high risk of bias. In addition, the potential publication bias and particularly selective reporting may mask results not showing efficacy of growth factors. Therefore, we judged the quality of evidence as low.
Summary judgment
Growth factors may improve haemodynamic measures up to six months in people with PAD of the lower extremities (low‐quality evidence).
Ulceration
Overall, 9 of 10 trials including participants with critical limb ischaemia measured ulceration, and 8 presented results (HGF‐0205; HGF‐STAT; Kibbe 2016; Kusumanto 2006; Makinen 2002; Matyas 2005; Shigematsu 2010; TALISMAN‐201). Two trials presented data with respect to change in the ulceration area within six months (HGF‐0205; HGF‐STAT; we could not pool results, as no data on standard deviations (SD) were available in the HGF‐STAT study). Kibbe 2016 and Shigematsu 2010 presented results for rates of participants with more than 50% (and Kusumanto 2006 with more than 60%) improvement in ulcer size at 3 months to 12 months' follow‐up. Seven studies provided results for rates of participants with complete ulcer healing at up to one year (HGF‐0205; HGF‐STAT; Kibbe 2016; Makinen 2002; Matyas 2005; TALISMAN‐201; Shigematsu 2010).
Size of the effect
We could not perform the main analysis, as the only trial that could provide results at low risk of bias did not report these parameters (TAMARIS).
The sensitivity analysis using last available data up to one year revealed an increased rate of participants with improvement in ulcer size (63% versus 11%; OR 17.57, 95% CI 3.37 to 91.65; 3 studies; 79 participants; Analysis 1.16) and an increased rate of participants with complete ulcer healing (33% versus 16%; OR 1.88, 95% CI 0.89 to 3.97; 6 studies; 189 participants; Analysis 1.15) in the treatment group. There was no or low heterogeneity between the trials and between the subgroups of growth factors for both parameters. We found similar results in the meta‐analyses using risk ratio and/or the random‐effects model.
Quality of evidence
The evidence on ulceration was direct and was generally consistent between results for both parameters. However, the effect estimates of the meta‐analyses were imprecise, and we cannot rule out negative effects of growth factors for complete ulcer healing. Moreover, we could only derive the effect estimates from the study results at unclear or high risk of bias. In addition, the probable publication bias and selective reporting may mask results not showing efficacy of growth factors. Therefore, we judged the quality of evidence as very low.
Summary judgment
Growth factors may improve ulceration up to one year in people with PAD of the lower extremities (very low‐quality evidence).
Rest pain
Overall, 9 of 10 trials including participants with critical limb ischaemia measured rest pain, and 8 presented results (HGF‐0205; HGF‐STAT; Kibbe 2016; Kusumanto 2006; Makinen 2002; Matyas 2005; Shigematsu 2010; TALISMAN‐201). Four trials used a visual analogue scale (VAS) to measure changes in level of rest pain from baseline up to one year (HGF‐0205; Kibbe 2016; Matyas 2005; TALISMAN‐201). Six trials reported results for rates of participants with improvement in rest pain up to nine months (HGF‐STAT; Kibbe 2016; Kusumanto 2006; Makinen 2002; Matyas 2005; Shigematsu 2010).
Size of the effect
We could not perform the main analysis, as the only trial that could provide results at low risk of bias did not report these parameters (TAMARIS).
In the sensitivity analysis, the conducted meta‐analysis revealed a decrease in the mean VAS score up to one year (MD at 10 cm scale −1.09 cm, 95% CI −1.83 to −0.35; 4 studies; 191 participants; Analysis 1.17) and an increased rate of participants with improvement in rest pain up to nine months (44% versus 27%; OR 1.89, 95% CI 0.80 to 4.42; 5 studies; 133 participants; Analysis 1.18) in the treatment group. There was no or low heterogeneity between the trials and between the subgroups of growth factors for both parameters. We found similar results in the meta‐analyses using the random‐effects model for both parameters and/or calculating the risk ratio for the improvement in rest pain.
Quality of evidence
The evidence on rest pain was direct and consistent between results for both parameters. However, the effect estimates of the meta‐analyses were imprecise, not precluding negative effects of growth factors for improvement in rest pain. Moreover, the effect estimates could be derived only from the study results at unclear or high risk of bias. In addition, the probable publication bias and selective reporting may mask results not showing efficacy of growth factors. Therefore, the quality of evidence is very low.
Summary judgment
Growth factors may improve rest pain up to one year in people with PAD of the lower extremities (very low‐quality evidence).
Discussion
Summary of main results
Our systematic review identified 20 randomised trials assessing three growth factors (FGF, HGF and VEGF; delivered directly or via substances promoting their gene expression) in people with different clinical stages of PAD of the lower extremities. Since six of the conducted trials had not been published at the time of writing this review, we could use the results of only 14 of these trials in the analyses. Most reported follow‐up periods ranged from three months to one year. Two small studies provided some data for 2 years, and one of them also for 10 years.
Our systematic review shows, that although growth factors may improve haemodynamic measures as well as ulceration and rest pain in people with PAD of the lower extremities up to one year, they have little or no effect on walking ability. Moreover, as the rate of any amputations was decreased but the rate of major amputations was similar, growth factors may decrease the rate of minor limb amputations. However, their effects on major limb amputations and on death are uncertain. The effect of growth factors on serious adverse events and on any adverse events is also uncertain. We did not identify any relevant differences in effects between growth factors (FGF, HGF and VEGF).
Overall completeness and applicability of evidence
The body of evidence is based on a broad spectrum of participants with respect to age, sex and severity of PAD (different clinical stages of intermittent claudication or of critical limb ischaemia), but participants were mostly white people living in the West. Sufficient data were not available for subgroup analyses with respect to age, sex, PAD severity, delivery approaches of growth factors and routes of application. Moreover, as these studies investigated only FGF, HGF and VEGF, the effect of other growth factors on people with PAD remains unclear. Since only one small study reported data over two years, the long‐term effects of growth factors in people with PAD are uncertain.
Although the review results are based on approximately 1400 participants, the total number of randomised participants and of obtained events, especially from the studies at low risk of bias, was low. Therefore, for most outcomes our analysis lacked the statistical power to provide precise effect estimates of treatment effects.
Although we considered many aspects of applicability such as biologic variability, variability in context, culture, adherence, values and preferences, we think that there are no major concerns about the applicability of the determined evidence of this review. We identified a number of ongoing trials that will probably be available for future updates of this review and should support the completeness of the evidence (Characteristics of ongoing studies).
Quality of the evidence
We identified six randomised trials with more than 300 participants (mostly in people with critical limb ischaemia), which were completed before 2015 but had not been published at the time of writing. Since trials that fail to show efficacy are less likely to be published, publication bias could influence the results for some outcomes showing an effect of growth factors. There may be a similar effect if the studies do not report results for all measured outcomes. Some publications reported the results of the trials insufficiently: either only descriptively, without data on variability (SD or SE levels) or without the number of participants with measurements in the follow‐up, which hampers the use of these results in the meta‐analyses. To minimise publication and reporting bias, we requested all missing data from the authors but unfortunately did not receive any response.
Although analyses for some evaluated parameters showed effects for growth factors, these effects may be due to chance or (more likely) bias, not only to true effect. Bias may be particularly responsible for effect estimates favouring therapy with growth factors for the primary outcome 'any limb amputation' as well as for the secondary outcomes 'haemodynamic measures', 'ulceration' and 'rest pain'. The observed heterogeneity in some meta‐analyses, especially in the FGF subgroup, may also reflect these biases.
Readers may view the evidence in our review as direct (head‐to‐head comparisons generally applicable for populations and interventions of interest), since only haemodynamic parameters represent somewhat indirect (but validated) measures of blood flow.
The results were generally consistent between the main analyses from studies at low risk of bias and the sensitivity analyses from all studies, between different parameters of outcomes as well as between results based on different effect measures and statistical models. However, inconsistency was relevant for judgment of the results for walking ability.
Imprecision was the major problem for most outcomes since we could not rule out a clinically relevant effect for many primary outcomes or negative effect for some secondary outcomes. Only effect estimates for walking ability were precise enough to exclude a clinically relevant effect.
Potential biases in the review process
We used a number of strategies to avoid potential biases in the review process. These included an extensive search for relevant published and unpublished trials, the duplication of the trial selection process, 'Risk of bias' assessment and extraction of outcome measures, requests for relevant information from the study authors, resolution of disagreements by consensus and application of different statistical methods in the sensitivity analyses. Despite these efforts, some concern may arise due to multiple counting of any adverse events as a result of their calculation from single events (where aggregate data were lacking), the use of the numbers of participants at baseline for follow‐up analyses (if data for the number of participants with measurements were lacking in the publication) and inclusion of the measure 'improvement' (e.g. improvement in ulceration does not consider ulcer development or worsening) in the analysis. In all these cases, our analyses may have overestimated the true effect (increased risk of bias). In addition, as the review is based on only randomised studies, it does not allow for detection of rare adverse events.
Agreements and disagreements with other studies or reviews
To our knowledge, there is no published systematic review or meta‐analysis on the use of growth factors in PAD. Systematic reviews focusing on gene therapy included some of the trials analysed in our review (De Haro 2009; Ghosh 2008; Hammer 2013). In accordance with our review, these analyses did not show a clear benefit of growth factors delivered via substances promoting their gene expression in people with PAD. Nevertheless, some relatively recent narrative reviews on therapeutic angiogenesis suggest growth factors are a promising option for people with PAD (Ouma 2012; Pacilli 2010; Powell 2012). However, these reviews did not use rigorous systematic assessment methods to generate conclusive evidence.
Study flow diagram.
Risk of bias summary: review authors' judgments about each risk of bias item for each included study.
Comparison 1 Growth factors versus placebo (or no therapy), Outcome 1 Limb amputation (major or not specified; at 1 year).
Comparison 1 Growth factors versus placebo (or no therapy), Outcome 2 Limb amputation (major or not specified; last data to 2 years).
Comparison 1 Growth factors versus placebo (or no therapy), Outcome 3 Limb amputation (any or not specified; at 1 year).
Comparison 1 Growth factors versus placebo (or no therapy), Outcome 4 Limb amputation (any or not specified; last data to 2 years).
Comparison 1 Growth factors versus placebo (or no therapy), Outcome 5 Death (of any cause; at 1 year).
Comparison 1 Growth factors versus placebo (or no therapy), Outcome 6 Death (from any cause; last data to 2 years).
Comparison 1 Growth factors versus placebo (or no therapy), Outcome 7 Adverse events (serious; only aggregate data, up to 2 years).
Comparison 1 Growth factors versus placebo (or no therapy), Outcome 8 Adverse events (serious; incl. data for single events, up to 2 years).
Comparison 1 Growth factors versus placebo (or no therapy), Outcome 9 Adverse events (any; only aggregated data, up to 2 years).
Comparison 1 Growth factors versus placebo (or no therapy), Outcome 10 Adverse events (any; incl. data for single events, up to 2 years).
![Comparison 1 Growth factors versus placebo (or no therapy), Outcome 11 Walking ability (change in peak walking time [min]; last data to 6 months).](/es/cdsr/doi/10.1002/14651858.CD011741.pub2/media/CDSR/CD011741/image_n/nCD011741-CMP-001-11.png)
Comparison 1 Growth factors versus placebo (or no therapy), Outcome 11 Walking ability (change in peak walking time [min]; last data to 6 months).
![Comparison 1 Growth factors versus placebo (or no therapy), Outcome 12 Walking ability (change in claudication onset time [min]; last data to 6 months).](/es/cdsr/doi/10.1002/14651858.CD011741.pub2/media/CDSR/CD011741/image_n/nCD011741-CMP-001-12.png)
Comparison 1 Growth factors versus placebo (or no therapy), Outcome 12 Walking ability (change in claudication onset time [min]; last data to 6 months).
Comparison 1 Growth factors versus placebo (or no therapy), Outcome 13 Haemodynamic measures (change in ABI; last data to 6 months).
Comparison 1 Growth factors versus placebo (or no therapy), Outcome 14 Haemodynamic measures (change in TBI; last data to 6 months).
Comparison 1 Growth factors versus placebo (or no therapy), Outcome 15 Ulceration (complete ulcer healing; last data to 1 year).
Comparison 1 Growth factors versus placebo (or no therapy), Outcome 16 Ulceration (improvement in ulcer size; last data to 1 year).
Comparison 1 Growth factors versus placebo (or no therapy), Outcome 17 Rest pain (change on 10 cm VAS; last data to1 year).
Comparison 1 Growth factors versus placebo (or no therapy), Outcome 18 Rest pain (improvement; last data to 1 year).
| Growth factor therapy versus placebo or no therapy for the treatment of peripheral arterial disease | |||||
| Patient or population: people who have been diagnosed with peripheral arterial disease Setting: all settings Intervention: growth factor therapy Comparison: placebo or no therapy | |||||
| Outcomes | № of participants | Quality of the evidence | Relative effect | Anticipated absolute effects (95% CI)* | |
| Risk with placebo | Risk difference | ||||
| Major limb amputation | 1075 (10 RCTs) | ⊕⊕⊝⊝ lowa | OR 0.99 (0.71 to 1.38) | 19 per 100 (86/452) | 0 per 100 |
| Any limb amputation | 415 (6 RCTs) | ⊕⊕⊝⊝ Lowb | OR 0.56 (0.31 to 0.99) | 28 per 100 (86/452) | 10 fewer per 100 |
| Death | 1371 (12 RCTs) | ⊕⊕⊝⊝ Lowa (⊕⊕⊕⊝ moderate for 1 year) | OR 0.99 (0.69 to 1.41) | 12 per 100 (65/546) | 0 per 100 |
| Serious adverse events | 1411 (13 RCTs) | ⊕⊕⊝⊝ Lowa (⊕⊕⊕⊝ moderate for 1 year) | OR 1.09 (0.79 to 1.50) | 20 per 100 (113/555) | 1 more per 100 |
| Any adverse event | 1411 (13 RCTs) | ⊕⊕⊝⊝ Lowa (⊕⊕⊕⊝ moderate for 1 year) | OR 1.10 (0.73 to 1.64)c | 82 per 100 (338/555)c | 1 more per 100 |
| *The basis for the anticipated risk in the study population was the average risk in the control group (i.e. the total number of participants with events divided by the total number of participants in the control group for all studies reporting the outcome). The estimation of the risk difference (and its 95% confidence interval) was based on the anticipated risk in the control group and the relative effect of the intervention (and its 95% CI). | |||||
| GRADE Working Group grades of evidence | |||||
| aDowngraded by two levels due to imprecision and risk of bias; for 1 year only by one level due to imprecision. | |||||
| Trial | N participants | Indication | Intervention | Placebo |
| Fibroblast growth factor | ||||
| 19 | Intermittent claudication | 1 × 10 µg/kg b‐FGF, n = 4 | n = 6 | |
| 1 × 30 µg/kg b‐FGF, n = 5 | ||||
| 2 × 30 µg/kg b‐FGF, n = 4 | ||||
| 24 (plan 108) | Intermittent claudication | 6 × 2 µg/kg b‐FGF, n = 16 | n = 8 | |
| 190 | Intermittent claudication | 1 × 30 µg/kg FGF‐2, n = 66 | n = 63 | |
| 2 × 30 µg/kg FGF‐2, n = 61 | ||||
| 125 | Critical limb ischaemia | 16 mg (4 × 4 mg) NV1FGF, n = 59 | n = 66 | |
| 71 | Critical limb ischaemia | 2‐16 mg [5 groups] NV1FGF, n = ? | n = ? | |
| 36 | Critical limb ischaemia | 16 mg (4 × 4 mg) NV1FGF, n = ? | n = ? | |
| 32 mg (4 × 8 mg) NV1FGF, n = ? | ||||
| 525 | Critical limb ischaemia | 16 mg (4 × 4 mg) NV1FGF, n = 266 | n = 259 | |
| 11 + 2b (plan 28) | Critical limb ischaemia | 2.87 × 108 Ad5FGF‐4, n = 3 + 2b | n = 3 | |
| 2.87 × 109 Ad5FGF‐4, n = 3 | ||||
| 2.87 × 1010 Ad5FGF‐4, n = 2 | ||||
| Hepatocyte growth factors | ||||
| 104 | Critical limb ischaemia | 3 × 0.4 mg AMG0001, n = 26 | n = 26 | |
| 2 × 4.0 mg AMG0001, n = 25 | ||||
| 3 × 4.0 mg AMG0001, n = 27 | ||||
| 27 (plan 48) | Critical limb ischaemia | 3 × 4.0 mg AMG0001, n = 21 | n = 6 | |
| 40 (plan 46) | Critical limb ischaemia | 2 × 4.0 mg AMG0001, n = 27 | n = 13 | |
| 52 | Critical limb ischaemia | 2 × 4.0 mg VM202 (HGF‐X7), n = 21 | n = 11 | |
| 4 × 4.0 mg VM202 (HGF‐X7), n = 20 | ||||
| 200 | Critical limb ischaemia | 3 × 4.0 mg NL003 (pCK‐HGF‐X7), n = ? | n = ? | |
| 3 × 6.0 mg NL003 (pCK‐HGF‐X7), n = ? | ||||
| 3 × 8.0 mg NL003 (pCK‐HGF‐X7), n = ? | ||||
| Vascular endothelial growth factor | ||||
| 13 | Critical limb ischaemia | VEGF‐C, n = ? | n = ? | |
| 54 (plan 60) | Intermittent claudicationc Critical limb ischaemia | VEGF‐Ad (2 × 1010 VEGF165), n = 18 | n = 19 | |
| VEGF‐P/L (2 × 2.0 mg VEGF165), n = 17 | ||||
| 105 (plan 105) | Intermittent claudication | 4 × 109 PU AdVEGF121, n = 32 | n = 33 | |
| 4 × 1010 PU AdVEGF121, n = 40 | ||||
| 54 | Critical limb ischaemia | 2 × 2.0 mg phVEGF165, n = 27 | n = 27 | |
| 100 | Intermittent claudicationc, Critical limb ischaemia | 2 × 1.2 mg pCMV‐VEGF165, n = 75 | n = 25d | |
| 10 | Intermittent claudication | VEGF‐A, n = ? | n = ? | |
| ? | Critical limb ischaemia | VEGF‐C, n = ? | n = ? | |
| aUnpublished trial. | ||||
| Trial | Amputation | Death | Adverse events/serious adverse events (SAE) | Walking | Haemodynamic parameters | Ulceration | Rest pain |
| Fibroblast growth factor | |||||||
| NR | NR | Data for single events | Improvement | Plethysmography | — | — | |
| NR | All | Data for single events | PWT, COT | ABI* | — | — | |
| NR | All | Data for single events | PWT, COT | ABI* | — | — | |
| Major, Any | All | Aggregate | — | ABI, TBI | Complete healing | VAS | |
| Unpublished trial | |||||||
| Unpublished trial | |||||||
| Major | All | Aggregate | — | NR | NR | NR | |
| Not defined | All | Data for single eventsa | — | ABI | Complete healing | VAS Improvement | |
| Hepatocyte growth factors | |||||||
| Alla | All | Aggregate (SAEs) | — | ABIa, TBIa | Areaa Complete healinga | Improvementa | |
| Major | All | Aggregate | — | ABI, TBI | Area Complete healing | VAS | |
| Major, Any | NR | Aggregate | — | ABIa | Improvement Complete healing | Improvement | |
| Major, Any | All | Aggregate (SAEs) | — | ABI | Improvement Complete healing | VAS Improvement | |
| Unpublished trial | |||||||
| Vascular endothelial growth factor | |||||||
| Unpublished trial | |||||||
| Major | All | Data for single events | — | ABI | Complete healing | Improvement | |
| Not defined | All | Data for single events | PWT, COT | ABIa | — | — | |
| Major | All | Data for single events | — | Improvement | Improvement | Improvement | |
| Not defined | All | Data for single events | PWD | ABI | — | — | |
| Unpublished trial | |||||||
| Unpublished trial | |||||||
| ABI: ankle brachial index; COT: claudication onset time; NR: not reported; PWD: peak walking distance; PWT: peak walking time; TBI: toe brachial index; VAS: visual analogue scale. | |||||||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Limb amputation (major or not specified; at 1 year) Show forest plot | 6 | 916 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.00 [0.71, 1.41] |
| 1.1 FGF | 2 | 632 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.01 [0.70, 1.46] |
| 1.2 HGF | 2 | 79 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.14 [0.27, 4.81] |
| 1.3 VEGF | 2 | 205 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.70 [0.17, 2.96] |
| 2 Limb amputation (major or not specified; last data to 2 years) Show forest plot | 10 | 1075 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.99 [0.71, 1.38] |
| 2.1 FGF | 3 | 643 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.05 [0.73, 1.50] |
| 2.2 HGF | 3 | 119 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.14 [0.27, 4.81] |
| 2.3 VEGF | 4 | 313 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.62 [0.24, 1.63] |
| 3 Limb amputation (any or not specified; at 1 year) Show forest plot | 4 | 364 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.54 [0.29, 1.00] |
| 3.1 FGF | 1 | 107 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.48 [0.22, 1.04] |
| 3.2 HGF | 1 | 52 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.65 [0.14, 3.00] |
| 3.3 VEGF | 2 | 205 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.70 [0.17, 2.96] |
| 4 Limb amputation (any or not specified; last data to 2 years) Show forest plot | 6 | 415 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.56 [0.31, 0.99] |
| 4.1 FGF | 2 | 118 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.60 [0.29, 1.24] |
| 4.2 HGF | 2 | 92 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.48 [0.13, 1.81] |
| 4.3 VEGF | 2 | 205 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.51 [0.13, 1.91] |
| 5 Death (of any cause; at 1 year) Show forest plot | 7 | 1038 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.98 [0.67, 1.44] |
| 5.1 FGF | 2 | 650 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.03 [0.68, 1.55] |
| 5.2 HGF | 3 | 183 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.74 [0.22, 2.54] |
| 5.3 VEGF | 2 | 205 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.79 [0.17, 3.73] |
| 6 Death (from any cause; last data to 2 years) Show forest plot | 12 | 1371 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.99 [0.69, 1.41] |
| 6.1 FGF | 5 | 875 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.03 [0.69, 1.53] |
| 6.2 HGF | 3 | 183 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.74 [0.22, 2.54] |
| 6.3 VEGF | 4 | 313 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.91 [0.30, 2.74] |
| 7 Adverse events (serious; only aggregate data, up to 2 years) Show forest plot | 6 | 865 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.11 [0.74, 1.65] |
| 7.1 FGF | 2 | 641 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.08 [0.65, 1.81] |
| 7.2 HGF | 4 | 224 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.15 [0.61, 2.16] |
| 7.3 VEGF | 0 | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 8 Adverse events (serious; incl. data for single events, up to 2 years) Show forest plot | 13 | 1411 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.09 [0.79, 1.50] |
| 8.1 FGF | 5 | 874 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.20 [0.76, 1.90] |
| 8.2 HGF | 4 | 224 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.15 [0.61, 2.16] |
| 8.3 VEGF | 4 | 313 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.84 [0.43, 1.62] |
| 9 Adverse events (any; only aggregated data, up to 2 years) Show forest plot | 4 | 709 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.10 [0.73, 1.64] |
| 9.1 FGF | 2 | 641 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.06 [0.70, 1.59] |
| 9.2 HGF | 2 | 68 | Odds Ratio (M‐H, Fixed, 95% CI) | 2.93 [0.38, 22.70] |
| 9.3 VEGF | 0 | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 10 Adverse events (any; incl. data for single events, up to 2 years) Show forest plot | 13 | 1411 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.52 [1.15, 2.02] |
| 10.1 FGF | 5 | 874 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.45 [1.02, 2.06] |
| 10.2 HGF | 4 | 224 | Odds Ratio (M‐H, Fixed, 95% CI) | 2.77 [1.24, 6.22] |
| 10.3 VEGF | 4 | 313 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.34 [0.75, 2.38] |
| 11 Walking ability (change in peak walking time [min]; last data to 6 months) Show forest plot | 3 | 279 | Mean Difference (IV, Fixed, 95% CI) | ‐0.17 [‐0.69, 0.36] |
| 11.1 FGF | 2 | 192 | Mean Difference (IV, Fixed, 95% CI) | ‐0.13 [‐0.67, 0.41] |
| 11.2 HGF | 0 | 0 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 11.3 VEGF | 1 | 87 | Mean Difference (IV, Fixed, 95% CI) | ‐0.70 [‐2.78, 1.38] |
| 12 Walking ability (change in claudication onset time [min]; last data to 6 months) Show forest plot | 3 | 279 | Mean Difference (IV, Fixed, 95% CI) | ‐0.07 [‐0.31, 0.17] |
| 12.1 FGF | 2 | 192 | Mean Difference (IV, Fixed, 95% CI) | ‐0.05 [‐0.29, 0.19] |
| 12.2 HGF | 0 | 0 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 12.3 VEGF | 1 | 87 | Mean Difference (IV, Fixed, 95% CI) | ‐1.6 [‐3.61, 0.41] |
| 13 Haemodynamic measures (change in ABI; last data to 6 months) Show forest plot | 6 | 341 | Mean Difference (IV, Fixed, 95% CI) | 0.04 [‐0.02, 0.10] |
| 13.1 FGF | 2 | 116 | Mean Difference (IV, Fixed, 95% CI) | 0.04 [‐0.07, 0.15] |
| 13.2 HGF | 2 | 73 | Mean Difference (IV, Fixed, 95% CI) | ‐0.03 [‐0.16, 0.10] |
| 13.3 VEGF | 2 | 152 | Mean Difference (IV, Fixed, 95% CI) | 0.07 [‐0.02, 0.15] |
| 14 Haemodynamic measures (change in TBI; last data to 6 months) Show forest plot | 2 | 128 | Mean Difference (IV, Fixed, 95% CI) | 0.04 [‐0.01, 0.09] |
| 14.1 FGF | 1 | 107 | Mean Difference (IV, Fixed, 95% CI) | 0.01 [‐0.04, 0.06] |
| 14.2 HGF | 1 | 21 | Mean Difference (IV, Fixed, 95% CI) | 0.22 [0.09, 0.35] |
| 14.3 VEGF | 0 | 0 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 15 Ulceration (complete ulcer healing; last data to 1 year) Show forest plot | 6 | 189 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.88 [0.89, 3.97] |
| 15.1 FGF | 2 | 113 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.42 [0.53, 3.81] |
| 15.2 HGF | 3 | 62 | Odds Ratio (M‐H, Fixed, 95% CI) | 4.54 [1.02, 20.21] |
| 15.3 VEGF | 1 | 14 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.67 [0.06, 6.87] |
| 16 Ulceration (improvement in ulcer size; last data to 1 year) Show forest plot | 3 | 79 | Odds Ratio (M‐H, Fixed, 95% CI) | 17.57 [3.37, 91.65] |
| 16.1 FGF | 0 | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 16.2 HGF | 2 | 41 | Odds Ratio (M‐H, Fixed, 95% CI) | 17.21 [2.52, 117.38] |
| 16.3 VEGF | 1 | 38 | Odds Ratio (M‐H, Fixed, 95% CI) | 18.10 [0.95, 344.54] |
| 17 Rest pain (change on 10 cm VAS; last data to1 year) Show forest plot | 4 | 191 | Mean Difference (IV, Fixed, 95% CI) | ‐1.09 [‐1.83, ‐0.35] |
| 17.1 FGF | 2 | 118 | Mean Difference (IV, Fixed, 95% CI) | ‐1.07 [‐1.91, ‐0.24] |
| 17.2 HGF | 2 | 73 | Mean Difference (IV, Fixed, 95% CI) | ‐1.16 [‐2.76, 0.43] |
| 17.3 VEGF | 0 | 0 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 18 Rest pain (improvement; last data to 1 year) Show forest plot | 5 | 133 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.89 [0.80, 4.42] |
| 18.1 FGF | 1 | 11 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.2 [0.07, 19.63] |
| 18.2 HGF | 2 | 76 | Odds Ratio (M‐H, Fixed, 95% CI) | 2.83 [0.93, 8.59] |
| 18.3 VEGF | 2 | 46 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.99 [0.21, 4.57] |
