Heart 99:791-798 doi:10.1136/heartjnl-2012-302949
  • Coronary artery disease
  • Original article

Endothelial progenitor cells, atheroma burden and clinical outcome in patients with coronary artery disease

  1. Nicholas L Mills1
  1. 1British Heart Foundation Centre for Cardiovascular Science, University of Edinburgh, Edinburgh, UK
  2. 2Medical Research Council Scottish Centre for Regenerative Medicine, University of Edinburgh, United Kingdom
  1. Correspondence to Dr Gareth J Padfield, British Heart Foundation Centre for Cardiovascular Science, The University of Edinburgh, Chancellor's Building, Edinburgh EH16 4SU, UK; gareth.padfield{at}, gareth.padfield{at}
  • Received 24 October 2012
  • Revised 3 January 2013
  • Accepted 5 January 2013
  • Published Online First 6 February 2013


Objective We wished to determine the effect of an acute coronary syndrome (ACS) on putative endothelial progenitor cell (EPC) populations, and define their relationship to coronary artery disease (CAD) severity and clinical outcome, in order to clarify their clinical relevance.

Design and setting A prospective cohort study conducted in a tertiary referral cardiac centre.

Patients Two-hundred-and-one patients undergoing coronary angiography for suspected angina or ACS.

Main outcome measures Putative EPC populations were determined by flow cytometry. CAD was quantified using the Gensini scoring system. Survival free from revascularisation, recurrent myocardial infarction and death were determined at 3 years.

Results Circulating CD34+VEGFR-2+ and CD34+VEGFR-2+CD133+ cells were rare (<0.007% of mononuclear cells), were not increased in patients with ACS, and were unrelated CAD severity or clinical outcome (p>0.1 for all). By contrast, CD34+CD45 cells were increased in patients with CAD compared with those with normal coronary arteries (p=0.008) and correlated with atheroma burden (r=0.44, p<0.001). Increased concentrations of circulating CD34+CD45 cells were associated with a shorter cumulative event-free survival (p<0.02). Proangiogenic monocytes (CD14+VEGFR-2+Tie-2+) and endothelial cell-colony forming units were increased in patients with ACS (p<0.01 for both), however, concentrations reflected myocardial necrosis, and did not predict the extent of CAD or clinical outcome.

Conclusions Traditional EPC populations, CD34+VEGFR-2+ and CD34+VEGFR-2+CD133+ are not related to the extent of CAD or clinical outcome. However, CD34+CD45 cells are increased in patients with CAD and predict future cardiovascular events. It is likely that CD34+CD45 concentrations reflect the extent of vascular injury and atheroma burden.


Cell therapy may accelerate neovascularisation and enhance vascular repair through the differentiation and proliferation of progenitors into mature phenotypes. Progress is limited by an incomplete understanding of the progenitor cell populations involved in vascular repair. Clinical trials have been largely encouraging,1 but cell therapy for coronary heart disease has failed to deliver consistent long-term clinical benefits.

Endothelial progenitor cells (EPC) are traditionally defined by the coexpression of vascular endothelial growth factor receptor-2 (VEGFR-2), a cell surface receptor necessary for the regulation of vascular development and homeostasis, and the stem cell markers, CD342 ,3 and CD133.4 ,5 These putative EPC are mobilised in response to angiogenic stress,6 ,7 accelerate re-endothelialisation and attenuate neointimal hyperplasia,8 ,9 and higher basal concentrations are associated with favourable cardiovascular outcomes.10 However, it should be recognised that there are conflicting data regarding the ability of cell preparations enriched for triple positive CD34+VEGFR-2+CD133+ cells to differentiate into mature endothelial cells,11 ,12 and a robust phenotypic definition of an EPC is yet to be established. Recent studies have identified non-haematopoietic CD34+ subpopulations, negative for the pan-leukocyte marker CD45 (CD34+CD45), that are capable of forming late-outgrowth endothelial colonies. Most notably, these cell types have robust proliferative potential and are morphologically indistinguishable from mature endothelial cells.12 ,13

Monocytes are instrumental in maintaining vascular homeostasis through augmenting the differentiation and proliferation of naïve progenitor cells, and the secretion of proangiogenic factors.14 In particular, angiogenic monocytes expressing the ‘endothelial’ surface receptors VEGFR-2 and Tie-2 accelerate re-endothelialisation and improve endothelial function following experimental vascular injury.14 ,15 Combined with a relative abundance in the peripheral circulation, such properties make monocytic subpopulations attractive therapeutic targets in cardiovascular medicine. Composed predominantly of monocytes, the traditional early outgrowth EPC colony described by Hill et al,15 is now recognised to be unable to form mature endothelial cells, although still considered to be a relevant measure of cardiovascular repair on account of its paracrine proangiogenic activity.14

We examined the behaviour of circulating CD34+ and CD14+ subpopulations following acute coronary syndrome (ACS), and their relationship to the severity of coronary artery disease (CAD) and clinical outcome in order to determine their clinical relevance.



The study was performed with the approval of the local research ethics committee in accordance with the Declaration of Helsinki, and written informed consent was obtained. Patient recruitment followed referral for investigation of suspected stable angina or ACS, defined as an increase in the plasma troponin concentration to ≥0.20 ng/ml or ≥0.5 mm ST-segment deviation, in the context of anginal symptoms. We excluded patients with significant comorbid illnesses including haematological or internal malignancy, hepatic or renal failure or concurrent infection.

Coronary angiography

All patients underwent diagnostic coronary angiography using conventional arterial catheters. Patients with a ≥50% stenosis of a major epicardial arterial segement determined by off-line visual inspection were classified as having CAD. CAD severity was graded using the Gensini scoring system.16

Flow cytometric analyses

Putative EPC and angiogenic monocytes were identified by flow cytometry and analysed as described previously17 using the following preconjugated antihuman monoclonal antibodies: anti-CD45-PercP (Becton Dickinson, UK), anti-CD34-FITC, anti-VEGFR-2-PE, anti-Tie-2-APC (R&D systems, USA), anti-CD-133-PE (Miltenyi Biotec, UK) and anti-CD14-FITC (Caltag Systems, UK) (figure 1). A total of 500 000 events were acquired in the leukocyte gate for each sample using a FACS-Calibur flow-cytometer (Becton Dickinson, UK). Unstained samples were used as negative controls. Compensation was performed using commercially available flow cytometry compensation beads (Miltenyi Biotec, UK).

Figure 1

Flow cytometric analysis of putative progenitor cells. Leukocytes were identified on the basis of their characteristic forward and side scatter profile (A). Leukocytes expressing CD34+-FITC, CD133+-PE, CD45+-PercP, and VEGFR-2+-APC were identified (B–E). Coexpression was determined using Boolean principles. CD34+ events are shown with respect to CD45 positivity (F). The proportion of CD14+-FITC events expressing Tie-2-APC and or VEGFR-2-PE were determined using quadrant analysis (G–I). Single or unstained stained negative controls were used where appropriate. This figure is only reproduced in colour in the online version.

Endothelial cell-colony forming units

Endothelial cell-colony forming units (EC-CFU)18 were generated from circulating mononuclear cells cultured on fibronectin-coated plates (Becton Dickinson, UK) using a commercially available assay.17

Clinical outcomes

Clinical outcomes were determined by review of medical records and using TrakCare (InterSystems Corporation, USA); an electronic patient record system used by the Acute Hospitals Division of Lothian National Health Service Health Board, UK. Major adverse clinical events (MACE) included any of the following: myocardial infarction (MI), as defined as the detection of an elevated troponin (≥0.20 ng/ml) in the context of anginal symptoms or ST-segment deviation of ≥0.5 mm19; coronary revascularisation, including all percutaneous and surgical coronary artery procedures and hospitalisation, including any unplanned hospitalisation for ACS, heart failure, stroke or uncontrolled arrhythmia. Events were adjudicated by a clinician blinded to EPC concentrations.

Data analysis and statistics

At a significance level of 5% and based on power calculations derived from previous studies,17 ,20 we estimated that a sample size of n=90 would give 80% power of detecting a putative clinically meaningful 15% difference in phenotypic EPC (CD34+VEGFR-2+) between patients with stable angina and ACS. A sample size of n=90 per group will give 95% power, at a 5% significance level, to detect meaningful correlations (r2>0.12) between measures of EPC number and clinical variables including Gensini score. Flow cytometric analyses, EC-CFU enumeration, and Gensini scoring were performed by an observer blinded to the patient's clinical profiles. Statistical analyses were performed with SPSS V.17 (SPSS Inc, Chicago, USA). Continuous variables are reported as mean±SD or median (interquartile range) where appropriate. Student t test, Mann–Whitney test and Pearson's χ2 test were used for comparisons between groups where appropriate. Spearman's test was used to test for correlation analysis between variables. For partial correlation analyses controlling for covariates, data were normalised where appropriate. Cell populations were categorised into tertiles after natural logarithmic transformation in order to evaluate associations between circulating progenitors and clinical endpoints. Multivariate Cox-regression analysis was performed to determine associations between progenitor cells and event-free survival with adjustment for the diagnosis of ACS on enrolment and cardiovascular risk factors. Hazard Ratios represent the predicted change in the hazard between the lowest and the highest tertiles. Statistical significance was taken at a two-sided p value <0.05.


Two-hundred-and-one patients undergoing diagnostic coronary angiography for the investigation of suspected stable angina (n=90) or ACS (n=111) participated in the study (table 1). Both groups were well matched for standard clinical variables (p>0.1 for all). Patients with ACS were less likely to be treated with antihypertensive drugs (see online supplementary table S1). The Gensini score was marginally greater in patients with ACS than those with suspected stable angina (21 (11–48) vs 17 (5–35); p=0.052). Patients with ACS were enrolled at a median of 3 (1–5) days following hospitalisation and had a median plasma troponin concentration of 0.42 (0.20–3.3) ng/ml.

Table 1

Demographic, clinical and angiographic characteristics of the study population

Seventy patients undergoing coronary angiography for the investigation of suspected stable angina, were classified as having obstructive coronary heart disease. Thirty-nine patients presenting with an ACS, were classified as having unstable angina, and 72 with acute MI. Patients with mild, or no CAD, were generally younger, female, and had a lower incidence of diabetes. (see online supplementary table S2).

Angiogenic cell populations

CD34+ progenitor cells

CD34+ cells were marginally increased in patients following ACS compared with those with suspected stable angina (3.44 (2.54–4.85) vs 2.84 (2.15–4.18)×106 cells/l; p=0.04; see online  supplementary table S2). However, the majority of circulating CD34+ cells were CD45+, and the concentrations of CD34+CD45+ cells were similar between both groups (p>0.11 for all). The less abundant CD34+CD45 population was increased in patients with ACS compared with patients with suspected stable angina (1.05 (0.70–1.64) vs 0.90 (0.55–1.30)×106 cells/l; p=0.02) (table 2).

Table 2

Putative progenitor cell population frequencies in the study population

Most notably, CD34+CD45 cell concentrations were not influenced by myocyte necrosis (p=0.71; figure 2) and the concentration of CD34+CD45 cells was increased in patients with both stable and unstable coronary disease compared with patients with suspected stable angina but normal coronary arteries at angiography (ANOVA, p=0.008). In patients with ACS, the CD34+CD45 cell concentration was similar regardless of the time from the onset of symptoms to cell quantification (p=0.72; figure 3).

Figure 2

Putative progenitor cell frequencies according to clinical presentation and the presence of coronary artery disease or myocyte necrosis. Patients with suspected stable angina but normal coronary arteries had significantly lower CD34+CD45 concentrations. In those patients with at least 1 ≥50% stenosis, the CD34+CD45 concentration was similar regardless of presentation or troponin concentration (A). CD34+VEGFR-2+CD133+ cells were extremely rare and were similar in all groups (B). CD14+VEGFR-2+Tie-2+ and endothelial cell-colony forming units were increased in patients with myocardial infarction (C, D respectively). Data are the median concentration and interquartile distance. Error bars define the 10th and 90th centiles. This figure is only reproduced in colour in the online version.

Figure 3

CD34+CD45 and CD14+VEGFR-2+Tie-2+ cell concentrations in patients with an acute coronary syndrome (ACS) stratified by time from symptom onset. CD34+CD45 cell concentration did not differ when patients with ACS were stratified by time from the onset of symptoms to quantification (ANOVA, p=0.72). CD14+VEGFR-2+Tie-2+ cell concentration was highest in patients with ACS where quantification was performed on the same day as the onset of symptoms and fell progressively until concentrations were comparable with patients with suspected stable angina after approximately 5 days (ANOVA, p=0.025). Data are the median concentration and interquartile distance. Error bars define the 10th and 90th centiles. This figure is only reproduced in colour in the online version.

Regardless of CD45 expression, CD34+VEGFR-2+ and CD34+VEGFR-2+CD133+ cells were rare (<0.007% of mononuclear cells), frequently undetectable, and were present at similar concentrations in patients with ACS compared with those undergoing angiography for suspected stable angina (0.13 (0.07–0.27) vs 0.10 (0.06–0.18)×106 cells/l; p=0.06, and 0.05 (0.02–0.12) vs 0.05 (0.03–0.09)×106 cells/l; p=0.33 respectively; figure 2 and table 2).

CD14+ proangiogenic monocytes

CD14+ subpopulations were determined in a subgroup of 119 patients (see online supplementary table S3), 69 patients with ACS and 50 patients with suspected stable angina. The concentration of circulating CD14+ cells was similar in both groups (474 (384–591) vs 457 (356–547)×106 cells/l; p=0.29), however, there was an increase in CD14+VEGFR-2+ (74.8 (47.1–136.9) vs 53.9 (30.6–103.1)×106 cells/l; p=0.02) and CD14+VEGFR-2+Tie-2+ cells (9.0 (3.9–20.5) vs 4.5 (2.1–8.8)×106 cells/l; p=0.003) following ACS (table 2). Concentrations were highest in patients with acute MI compared with patients with stable angina or normal coronary arteries (ANOVA, p=0.015; figure 2). CD14+VEGFR-2+Tie-2+ concentrations were highest in the first day following the onset of symptoms, and fell progressively until concentrations were comparable to those of  patients with suspected stable angina after approximately 5 days (p=0.025; figure 3).

EC-CFU colonies

EC-CFU concentrations were increased following ACS compared with patients with suspected stable angina (12 (4–27) vs 6 (2–17) EC-CFU colonies per well; p=0.005; table 2). EC-CFU concentrations were highest in patients with MI compared with patients with stable angina or normal coronary arteries (ANOVA, p=0.018; figure 2).

Clinical correlates

In patients with suspected stable angina, there was a positive correlation between CD34+CD45 cells and the Gensini score (r=0.44; p<0.0001; table 3). By contrast, CD34+CD45+ cells expressing VEGFR-2+, CD133+ and VEGFR-2+CD133+ did not correlate with CAD severity (p>0.1 for all). The correlation between CD34+CD45 cells and CAD severity persisted after controlling for cardiovascular risk factors and use of cardiac medication (r=0.35; p=0.04; table 3). Similarly, when patients were stratified into quartiles according to Gensini score, CD34+CD45 concentrations increased in a step-wise fashion with each increment of severity of CAD (median CD34+CD45 concentrations were 0.75 (0.45–1.13), 0.91 (0.58–1.52), 1.01 (0.71–1.53) and 1.34 (0.76–1.59)×106 cells/l, given Gensini scores of <9, 9–20, 21–42 and >43, respectively; ANOVA=0.013; figure 4). CD14+ concentrations did not correlate with CAD severity (r=0.22; p=0.08), and although there was a weak positive association between CD14+VEGFR-2+ cells and CAD severity (r=0.29; p=0.03), this was not so following adjustment for cardiovascular risk factors and use of cardiac medication (r=0.15; p=0.39).

Table 3

Bivariate and partial correlation analysis between putative progenitor cell concentration and coronary artery disease severity in patients with suspected stable angina

Figure 4

CD34+CD45 concentration according to Gensini score. Patients were stratified into quartiles according to Gensini score. Median CD34+CD45 concentrations fell in a step-wise fashion with each decrement in the severity of coronary artery disease (ANOVA=0.013).

Clinical events

Patients were followed-up for approximately 3 years: median 1068 (970–1145) days. No patients were lost to follow-up. During this period, 123 patients (64%) underwent coronary revascularisation; 28 patients underwent coronary artery bypass grafting (CABG) (15%), and 95 underwent percutaneous coronary intervention (49%). This included 57 patients (30%) who underwent ad hoc revascularisation following the initial angiogram. Six patients died (3%), five from cardiovascular causes and one from lung cancer. Fifty-seven patients were hospitalised for a cardiovascular event (30%), and 15 patients had MI (7%). A MACE occurred in 74 patients (35%) (table 4).

Table 4

Clinical outcome at 3 years stratified into tertiles by CD34+CD45 cell concentration

The CD34+CD45 population alone was predictive of clinical events. CD34+CD45 concentrations were significantly higher in those patients with recurrent MI (1.40 (0.88–1.64) vs 0.96 (0.62–1.44)×106 cells/l; p=0.023), undergoing coronary revascularisation (1.05 (0.75–1.53) vs 0.77 (0.47–1.27)×106 cells/l; p=0.003) or with any MACE (1.04 (0.7–1.53) vs 0.87 (0.45–1.27)×106 cells/l; p=0.010) during the follow-up period. CD34+CD45 concentrations did not differ significantly with respect to death or hospitalisation alone (p>0.05 for all). There were no differences in event-free survival according to concentrations of circulating CD45+CD34+, CD34+VEGFR-2+, CD34+VEGFR-2+133+, CD14+VEGFR-2+ or CD14+VEGFR-2+Tie-2+ cells (p>0.3 for all).

Patients were stratified into tertiles according to CD34+CD45 concentration. Patients with the highest CD34+CD45 concentrations had higher Gensini scores and creatinine concentrations, and a greater prevalence of peripheral vascular disease and clopidogrel use (p<0.05 for all; see online supplementary table S4). Event rates were lower in patients in the lowest tertile compared with the highest tertile: for MI (HR 0.20 (95% CI 0.06 to 0.67, p=0.009)); for revascularisation (HR 0.47 (95% CI 0.29 to 0.877, p=0.003)); and for any MACE (HR 0.50 (95% CI 0.27 to 0.92, p=0.026)). In patients with higher circulating CD34+CD45 cell concentrations, the cumulative event-free survival was lower for coronary revascularisation (p=0.01), recurrent MI (p=0.02), or any MACE (p=0.02). In a Cox regression analysis correcting for covariates, these associations remained significant for all covariates with the exception of the Gensini score (table 4).


We have evaluated a comprehensive range of putative angiogenic cells and EPC in 201 patients undergoing coronary angiography. Traditional EPC populations (CD34+VEGFR-2+ and CD34+VEGFR-2+CD133+) were not increased by an ACS and were unrelated to either the extent of CAD or clinical outcome. However, CD34+CD45 cells were increased relative to the extent of CAD and predicted future cardiovascular events, suggesting that they represent the extent of vascular injury and are a measure of atheroma burden. Angiogenic monocytes derived by phenotype (CD14+VEGFR-2+Tie-2+) or cell culture (EC-CFU) were increased in patients who had sustained an MI, but did not relate to the extent of CAD, indicating that they are part of an acute-phase response to cardiovascular stress.

Consistent with previous studies we observed increased CD34+ concentrations in patients with ACS, however we found this to be as a result of increased CD34+CD45 concentrations. However, the difference in CD34+CD45 concentration between patients with stable and unstable disease was due to low concentrations of CD34+CD45 cells in stable patients with normal coronary arteries at angiography rather than ACS per se. Among patients with significant CAD, the CD34+CD45 concentration was similar regardless of stable or unstable disease. Furthermore, in patients with ACS, CD34+CD45 concentrations were similar regardless of the time at which they were measured following the onset of symptoms, indicating that they had not been mobilised acutely following ACS. Finally, we identified a strong and independent association between CD34+CD45 concentration and the severity of CAD in patients with stable angina. We therefore suggest that circulating CD34+CD45 concentrations do not represent a dynamic response to discrete plaque rupture or myocardial ischaemia, but provide a measure of the extent of vascular injury and an index of atherosclerotic burden.

Similarly, Guven et al,23 found that late outgrowth colonies were increased in patients with more severe CAD. Although CD34+CD45 cells are capable of forming late outgrowth colonies,12 their origin and function in vivo is unknown. A small proportion of mature endothelial cells are CD34+CD45,24 and CD34+CD45 cells may therefore simply be endothelial cells released into the circulation as a consequence of endovascular injury, rather than EPC mobilised from the bone marrow. Although it is plausible that CD34+CD45 cells possess reparatory function, given that we observed that patients with the highest CD34+CD45 concentrations were most likely to be readmitted with MI, or require revascularisation, it is more likely that the presence of CD34+CD45 in the circulation reflects vascular denudation, therefore serving as a surrogate biomarker of CAD severity, predictive of the occurrence of adverse events.

CD34+ populations expressing CD133 or VEGFR-2 were not increased in patients with ACS, and concentrations did not reflect the severity of CAD or predict clinical outcome. Mobilisation of CD133+ cells following extensive tissue injury, such as following CABG or burns, is recognised,7 and Friedrich et al25 describe a CD133+CD34VEGFR-2+ population thought to be a precursor of CD34+ cells that is present in increased concentrations in unstable atheromatous plaques. Timmermans et al,12 have found that circulating CD133+ cells do not directly differentiate to endothelial cells, and that CD133+ cells are not responsible for late outgrowth colonies. Instead late outgrowth colonies, indistinguishable from endothelial cells, were isolated exclusively from the CD34+CD45CD133 cell fraction. Although CD34+VEGFR-2+ and CD34+VEGFR-2+CD133+ concentrations have been used widely in clinical studies, we found that both populations are exceedingly rare, and frequently undetectable. Therefore, even in a carefully conducted, large clinical study such as this, these populations are difficult to quantify accurately. We believe that VEGFR-2+ or CD133+ CD34+ subpopulations are unlikely to play a significant role in the acute response to vascular injury or AMI.

EC-CFU are largely composed of activated monocytes and lymphocytes, and are mobilised acutely following angiogenic stress.20 EC-CFU probably potentiate re-endothelialisation and neovascularisation by enhancing proliferation and migration of endothelial cells and EPC via the secretion of angiogenic cytokines.14 Consistent with this hypothesis, angiogenic monocytes coexpressing VEGFR-2 and Tie-2 increased acutely following MI. Mobilisation of CD14+VEGFR-2+ monocytes following MI has been reported previously in a small clinical study by Bruno et al,26 who also demonstrated that intramyocardial injection of CD14+VEGFR-2+ cells improved ventricular function in a murine model of MI. By contrast, Hu et al27 found no difference in CD14+VEGFR-2+ cells between patients with stable or unstable CAD and healthy controls, however, this study assessed only 25 patients, 2 weeks post-ACS, and troponin concentrations were not reported making direct comparisons with our study difficult. We believe proangiogenic monocytes are mobilised in the acute response to cardiovascular stress, and may be a measure of myocardial rather than vascular or endothelial injury.

Study limitations

Conventional angiography will generally underestimate atheroma burden, and we exercise caution in interpreting correlations between putative EPC concentrations and atheroma burden quantified using angiography. However, it is more likely that under these circumstances we would have encountered a type I error rather than produce a false positive result, and we are therefore confident that for the purposes of this study, the use of conventional angiography as a surrogate of coronary atheroma is appropriate and provides meaningful information. Drawing definitive conclusions about clinical outcomes in a study of this size is challenging, however, the increased MACE rate in patients with high circulating CD34+CD45 concentrations is entirely consistent with the fact that these patients had more severe CAD. Finally, patients presenting with ACS were less likely to be on antihypertensive medication, and this may have affected circulating progenitor concentrations.28 However, we feel this is unlikely to have significantly affected our observations as the association between CD34+CD45 cells and CAD severity was independent of traditional cardiovascular risk factors including hypertension and the use of cardiac medication.


Proangiogenic monocytes and EC-CFU are increased in patients with acute MI, but do not predict the extent of CAD or clinical outcome. Traditional CD34+VEGFR-2+CD133+ EPC populations are impractical to measure and are unlikely to mediate cardiovascular regeneration. CD34+CD45 cells are not mobilised following an ACS, but increased concentrations reflect a greater severity of CAD, and predict adverse cardiovascular events. CD34+CD45 concentrations probably reflect the extent of vascular injury and atherosclerotic burden.


The BHF supported Dr Padfield (SS/CH/92010), Dr Mills (FS/10/024) and Professor Newby (CH/09/002) to undertake the work, and the research was supported by a BHF Project Grant (PG/07/017/22405). The Wellcome Trust Clinical Research Facility is supported by NHS Research Scotland through NHS Lothian.


  • Contributors All authors contributed significantly to the work. GJP and NLM devised and conducted the research and wrote the manuscript along with DEN. EF assisted with data acquisition and analyses. OT, GRB and MT assisted with data analysis and manuscript preparation. GJP and NLM take overall responsibility for the manuscript. The authors have no conflict of interests. Technical appendix, statistical code, and dataset are available on request. Explicit informed consent for data sharing was not given but the risk of identification is low.

  • Funding British Heart Foundation.

  • Competing interests None.

  • Ethics approval NREC.

  • Provenance and peer review Not commissioned; externally peer reviewed.


Free sample
This recent issue is free to all users to allow everyone the opportunity to see the full scope and typical content of Heart.
View free sample issue >>

Don't forget to sign up for content alerts so you keep up to date with all the articles as they are published.