The effect of 1.5 T cardiac magnetic resonance on human circulating leucocytes

Abstract

Aims

Investigators have proposed that cardiovascular magnetic resonance (CMR) should have restrictions similar to those of ionizing imaging techniques. We aimed to investigate the acute effect of 1.5 T CMR on leucocyte DNA integrity, cell counts, and function in vitro, and in a large cohort of patients in vivo.

Methods and results

In vitro study: peripheral blood mononuclear cells (PBMCs) were isolated from healthy volunteers, and histone H2AX phosphorylation (γ-H2AX) expression, leucocyte counts, and functional parameters were quantified using flow cytometry under the following conditions: (i) immediately following PBMC isolation, (ii) after standing on the benchside as a temperature and time control, (iii) after a standard CMR scan. In vivo study: blood samples were taken from 64 consecutive consenting patients immediately before and after a standard clinical scan. Samples were analysed for γ-H2AX expression and leucocyte counts. CMR was not associated with a significant change in γ-H2AX expression in vitro or in vivo, although there were significant inter-patient variations. In vitro cell integrity and function did not change with CMR. There was a significant reduction in circulating T cells in vivo following CMR.

Conclusion

1.5 T CMR was not associated with DNA damage in vitro or in vivo. Histone H2AX phosphorylation expression varied markedly between individuals; therefore, small studies using γ-H2AX as a marker of DNA damage should be interpreted with caution. Cardiovascular magnetic resonance was not associated with loss of leucocyte viability or function in vitro. Cardiovascular magnetic resonance was associated with a statistically significant reduction in viable leucocytes in vivo.

Introduction

Cardiovascular magnetic resonance (CMR) is a cost-effective cardiac imaging modality that provides important diagnostic and prognostic information, which significantly impacts upon patient management.¹ Integral to the expansion of CMR and magnetic resonance imaging (MRI), in general, is its safety profile. Magnetic resonance imaging is free from ionizing radiation and is considered to be ‘one of the safest medical procedures currently available’.²

Recently, however, the safety of CMR has been questioned. Cardiovascular magnetic resonance has been reported to induce T-cell death via loss of DNA integrity, as evidenced by increased histone H2AX phosphorylation (γ-H2AX) combined with a reduction in circulating T cells post-CMR.^3,4 This has led some to conclude that CMR should be used with caution and should have restrictions similar to those of ionizing imaging techniques.^3–6 Other work has found CMR is not associated with loss of DNA integrity.⁷ However, all of these studies have generally been insufficiently powered, and no contemporary ancillary in vitro studies have been performed.

We investigated the acute effect of CMR on T-cell and monocyte DNA integrity, cell counts, and cell function in vitro, and in a large cohort of patients in vivo.

Methods

Study design

This was a prospective research study. An ethics committee of the UK National Research Ethics Service approved the study, and written informed consent was obtained from all participants. The work was conducted according to the Helsinki Declaration.

The study comprises two parts: (i) an in vitro study and (ii) an in vivo study. Recruitment to both parts was prospective.

In vitro study

Blood samples (4 mL) were obtained from healthy volunteers into ethylenediaminetetraacetic acid (EDTA) vacutainers, and peripheral blood mononuclear cells (PBMCs) were immediately isolated via Ficoll-Paque separation. Cells were resuspended in Roswell Park Memorial Institute (RPMI) 1640 cell culture medium at 1×10⁶ cells/mL and split into three groups: (i) baseline control, (ii) temperature and time control, and (iii) CMR intervention. Group 1 samples were processed and analysed immediately. Group 2 samples were processed and analysed after being placed on the laboratory bench for the duration of the CMR scan (see Group 3). Group 3 samples were processed and analysed after being placed in the CMR scanner bore and receiving a standard clinical ‘viability-type’ CMR scan, which lasted 40 min [1.5 T scanner (Avanto, Siemens Medical Imaging); sequences included localizers, steady-state free precession (SSFP) cine imaging, and late enhancement imaging (no contrast agent)]. Cell analysis was led by J.E.F., W.R.C., and D.M.D., assisted by J.P.S., A.L.B., and T.M., and conducted at the Manchester Collaborative Centre for Inflammation Research (MCCIR), part of the University of Manchester, UK.

All samples were exposed to one of the three conditions described above and were then split into two duplicate aliquots to evaluate (i) DNA damage, cell counts, and apoptosis and (ii) T-cell function. The details of these experiments are described below:

First, in a single tube, 1×10⁶ cells were stained with viability dye (LIVE/DEAD fixable violet dead cell stain kit, ThermoFisher), followed by monoclonal antibodies directed to CD3 (PE-Cy7-conjugated mouse anti-human CD3; clone SK7, BD Biosciences) and CD14 (PerCPCy5.5-conjugated mouse anti-human CD14; clone MφP9, BD Biosciences). The cells were subsequently fixed and permeabilized using the TrueNuclear Transcription Factor Buffer Set (BioLegend). The intracellular antigens γ-H2AX (AlexaFluor647-conjugated mouse anti-human γ-H2AX; clone N1-431, BD Biosciences) and active caspase-3 (FITC-conjugated rabbit anti-human active caspase 3; clone C92-605, BD Biosciences) were then stained before the cells were washed and absolute count beads were added. Samples were analysed by flow cytometry (BD LSR II flow cytometer).

CD3 and CD14 expression combined with forward scatter/side scatter characteristics were used to identify T cells and monocytes, respectively. Mean expression of γ-H2AX per cell (as determine by fluorescence intensity) was used to determine the extent of DNA double-strand breaks (DSBs). To determine the reproducibility of the γ-H2AX assay, one sample was split into 5 and each was concurrently stained with CD14 and γ-H2AX. The reproducibility of the γ-H2AX assay was found to be high, with a coefficient of variation of 3.2%. Mean expression of active caspase-3 per cell was used as a marker of cell apoptosis.

Second, PBMCs from each group (after intervention) were placed into cell culture plates with RPMI 1640 at a concentration of 1×10⁶ cells/mL, stimulated with lipopolysaccharide (100 ng/mL) to activate the cells and cultured for 1 week with a change of media and restimulation on Day 3. On Day 7, cells were retrieved from the plate, and cell function evaluated by flow cytometry using monoclonal antibodies directed to CD3 (PE-Cy7-conjugated mouse anti-human CD3; clone SK7, BD Biosciences), CD69 (PE-CF594-conjugated mouse anti-human CD69; clone FN50, BD Biosciences) and CD107a (PE-conjugated mouse anti-human CD107a; clone H4A3, BD Biosciences). Mean expression of CD69 per cell was used as a marker of T-cell activation. Mean expression of CD107a per cell was used as a marker of T-cell degranulation.

In vivo study

All patients referred for a standard ‘viability-type’ clinical CMR scan at the University Hospital of South Manchester NHS Foundation Trust over a period of 3 months were prospectively approached before attending for scan. Inclusion criteria: (i) adult (i.e. aged 18 or older); (ii) referred for a clinical ‘viability-type’ CMR scan at the University Hospital of South Manchester NHS Foundation Trust (details of the scan are provided below). Patients were excluded if they were due to receive a gadolinium-based contrast agent other than Gadovist, at a dose other than 0.1 mmol/kg, or if they were due to receive any other medications during or after the scan (e.g. adenosine or dobutamine). Blood was obtained from consecutive consenting patients.

All patients underwent a standard ‘viability-type’ clinical CMR scan using a 1.5 T scanner (Avanto, Siemens Medical Imaging; maximum gradient strength 45 mT/m, maximum slew rate 200 mT/m/ms). The CMR scans included localizers, SSFP cine imaging, and late enhancement imaging per standard clinical protocols. All patients received 0.1 mmol/kg Gadovist during the scan per standard clinical protocol. Patients did not receive any other medications before, during or after the scan, that is no patients underwent stress (perfusion or wall motion) imaging.

Blood samples (4 mL) were obtained into EDTA vacutainers immediately prior to CMR scanning and immediately post-CMR scanning. A representative 100 µL blood sample was processed immediately. Samples were analysed by flow cytometry, with monoclonal antibodies directed to CD3 (PE-Cy7-conjugated mouse anti-human CD3; clone SK7, BD Biosciences), CD14 (PE-conjugated mouse anti-human CD14; clone MφP9, BD Biosciences), and γ-H2AX (AlexaFluor647-conjugated mouse anti-human γ-H2AX; clone N1-431, BD Biosciences), as described above.

Statistical analysis

The in vitro experiments evaluating the effect of CMR on γ-H2AX and cell counts were initially conducted in blood samples from five healthy volunteers. This provided data for the sample size calculation for the in vitro work as follows: In the pilot data, mean difference in T-cell count between Group 2 and Group 3 was 69 000 cells/mL, with an SD of 58 000 cells/mL. Therefore, blood samples from 16 healthy volunteers were required to detect a conservative minimum difference between Group 2 and Group 3 of 50 000 T cells/mL with 80% power using a paired t-test with the conventional two-sided 5% significance level, assuming a conservative SD of 65 000 cells/mL. To account for 10% of healthy volunteers not attending, the number invited to take part was inflated to 18.

There were limited data on which to base a power calculation for the in vivo work at the time of designing this study. The SD of the change in DNA DSBs as determined using flow cytometry in the study by Fiechter et al.³ was estimated to be ∼880, assuming a correlation between DNA DSBs detected before and after CMR of 0.5. The change in DNA DSBs before and after CMR in the study by Fiechter et al.³ was 470. Therefore, 60 patients were required to detect a conservative minimum change in DNA DSBs of 325, with 80% power, using a paired t-test with the conventional two-sided 5% significance level, assuming an SD of 880. To account for 20% of patients not completing the study, we aimed to recruit 75 patients.

Data were analysed in a blinded fashion. Blood samples were labelled using random codes by the ‘CMR team’. The CMR team sent the codes, together with the patient identifier and sample collection timing to which the codes referred, to an independent statistician (J.M.). Blood sample processing and analysis were performed by the ‘blood analysis team’, who were independent of the CMR team, using the codes as sample identifiers. The blood analysis team sent the results of the analysis to the independent statistician. The independent statistician then assigned the data to each patient and sample collection time according to the code and performed the statistical analysis.

Data distribution was determined by the Shapiro–Wilk test. Normally distributed variables are summarized with means and standard deviations and were analysed using paired t-tests. In vitro data were log transformed due to non-normality and are expressed as geometric mean (95% confidence intervals). Comparisons of the three groups in the in vitro studies were made within subjects using linear regression analysis with generalized estimating equations and an exchangeable correlation structure accounting for within-subject clustering. Bonferroni adjustments were applied for specific comparisons between pairs of groups. All data were used for analysis, that is possible outlying data were not excluded from the analyses. Analyses were performed using IBM SPSS version 22 and GraphPad Prism v7.00.

Results

In vitro effects of cardiovascular magnetic resonance

Study population

Blood samples were obtained from 17 healthy volunteers in addition to the blood samples from the 5 healthy volunteers that formed the pilot work. Histone H2AX phosphorylation and cell count data from all 22 healthy volunteers were pooled. The other parameters were assessed in the n = 17 group. Mean age of the 22 healthy volunteers was 31 ± 10 years. Eleven (50%) were male.

DNA damage

There was a significant difference in γ-H2AX expression in T cells between groups [390 (95% confidence interval (CI) 280–550) vs. 710 (95% CI 510–970) vs. 750 (95% CI 530–1100), for Groups 1, 2, and 3, respectively; P < 0.001; Figure 1A]. Similarly, there was a significant difference in γ-H2AX expression in monocytes between groups [580 (95% CI 440–750) vs. 1300 (95% CI 1070–1500) vs. 1160 (95% CI 940–1430), for Groups 1, 2, and 3, respectively; P < 0.001; Figure 1B]. Post hoc analysis demonstrated that the differences in γ-H2AX expression were driven by a significant increase in γ-H2AX expression in Groups 2 and 3 compared with Group 1, in both T cells and monocytes. There was no difference in γ-H2AX expression between Groups 2 and 3, in either T cells or monocytes (Figure 1A and B). Representative histograms are displayed in Supplementary material online, Figure S1.

Figure 1.

Effect of CMR on leucocyte DNA integrity in vitro: (A) T-cell histone H2AX phosphorylation expression and (B) monocyte histone H2AX phosphorylation expression. Group 1: baseline control. Samples were processed and analysed immediately. Group 2: temperature and time control. Samples were processed and analysed after being placed on the laboratory bench for the duration of the CMR scan (see Group 3). Group 3: CMR intervention. Samples were processed and analysed after being placed in the CMR scanner bore and receiving a standard clinical 1.5 T CMR scan, lasting 40 min.

Cell counts and apoptosis

There was a significant difference in T-cell count between groups [cells/mL: 282 000 (95% CI 194 000–410 000) vs. 68 600 (95% CI 50 400–90 700) vs. 59 300 (95% CI 46 300–76 000), for Groups 1, 2, and 3, respectively; P < 0.001; Figure 2A]. Similarly, there was a significant difference in monocyte count between groups [82 100 (95% CI 65 200–103 000) vs. 26 600 (95% CI 20 300–34 900) vs. 24 100 (95% CI 18 000–32 100), for Groups 1, 2, and 3, respectively; P < 0.001; Figure 2B]. Post hoc analysis demonstrated that the differences in T-cell and monocyte counts were driven by significant reductions in cell counts in Groups 2 and 3 compared with Group 1. There were no differences in T-cell or monocyte counts between Groups 2 and 3 (Figure 2A and B).

Figure 2.

Effect of CMR on leucocyte integrity in vitro: (A) T-cell count, (B) monocyte count, (C) T-cell expression of caspase-3 (a marker of apoptosis), and (D) monocyte expression of caspase-3. Groups as defined in Figure 1.

There was a significant difference in T-cell active caspase-3 expression between Groups [51 (95% CI 47–56) vs. 100 (95% CI 87–110) vs. 104 (95% CI 92–120), for Groups 1, 2 and 3 respectively; P < 0.001; Figure 2C]. Similarly, there was a significant difference in monocyte active caspase-3 expression between groups [84 (95% CI 75–95) vs. 200 (95% CI 180–240) vs. 230 (95% CI 200–270), for Groups 1, 2, and 3, respectively; P < 0.001; Figure 2D]. Post hoc analysis demonstrated that the differences in active caspase-3 expression were driven by a significant increase in active caspase-3 expression in Groups 2 and 3 compared with Group 1, in both T cells and monocytes. There was no difference in active caspase-3 expression between Groups 2 and 3, in either T cells or monocytes (Figure 2C and D). Representative histograms are displayed in Supplementary material online, Figure S2.

Cell function

There was no significant difference in T-cell CD69 expression between groups [220 (95% CI 180–270) vs. 240 (95% CI 190–300) vs. 240 (95% CI 200–300), for Groups 1, 2, and 3, respectively; P = 0.15; Figure 3A]. There was also no significant difference in T-cell CD107a expression between groups [710 (95% CI 530–960) vs. 540 (95% CI 470–620) vs. 520 (95% CI 450–600), for Groups 1, 2, and 3, respectively; P = 0.098; Figure 3B]. Representative histograms are displayed in Supplementary material online, Figure S3.

Figure 3.

Effect of CMR on T-cell function in vitro: (A) T-cell expression of CD69 (a marker of T-cell activation) and (B) T-cell expression of CD107a (a marker of T-cell degranulation). Groups as defined in Figure 1.

In vivo effects of cardiovascular magnetic resonance

Study population

Seventy-one patients consented to take part in the study; however, four patients were excluded because they did not undergo CMR scanning (three patients were claustrophobic and one patient’s size precluded scanning), and three patients were excluded because they did not undergo blood sampling for logistical reasons (two patients had neither the pre- or post-CMR blood sample taken and one patient did not have a pre-CMR sample taken).

The analysis therefore included 64 patients. Thirty-seven (58%) were male, aged 51 ± 16 years, body surface area 2.0 ± 0.3 m², and estimated glomerular filtration rate 79 ± 13 mL/min/1.73 m². Scan duration was 42 ± 11 min. Scan diagnoses (some diagnoses coexisted): normal 23 (36%), cardiomyopathy or possible cardiomyopathy 26 (40%), ischaemic heart disease 10 (16%), valvular heart disease 5 (8%), myocarditis 2 (3%), hypertensive heart disease 2 (3%), and other 5 (8%).

DNA damage

Cardiovascular magnetic resonance was not associated with a significant change in the expression of γ-H2AX in T cells in vivo (pre-CMR 8680 ± 3090, post-CMR 8410 ± 2730; difference −270 ± 2950; P = 0.47; Figure 4A). Similarly, CMR was not associated with a significant change in the expression of γ-H2AX in monocytes in vivo (pre-CMR 3470 ± 1350, post-CMR 3340 ± 990; difference −130 ± 1480; P = 0.50; Figure 4B). However, there were significant inter-patient variations in γ-H2AX expression, with both large increases and large decreases seen in γ-H2AX expression following CMR in both T cells and monocytes (Figure 4A and B).

Figure 4.

Effect of CMR on leucocyte DNA integrity and leucocyte integrity in vivo: (A) T-cell histone H2AX phosphorylation expression before and after CMR, with a representative histogram; (B) monocyte histone H2AX phosphorylation expression before and after CMR, with a representative histogram; (C) T-cell count before and after CMR; and (D) monocyte cell count before and after CMR.

Cell counts

There was a significant reduction in the number of circulating T cells following CMR (pre-CMR 215 000 ± 158 000 vs. post-CMR 146 000 ± 133 000; P = 0.002; difference −69 000 ± 173 000; Figure 4C). There was also a reduction in the number of circulating monocytes following CMR, although the difference was not statistically significant (pre-CMR 89 300 ± 79 400 vs. post-CMR 72 200 ± 65 400; difference −17 100 ± 87 400; P = 0.12; Figure 4D).

Discussion

In the largest in vitro and in vivo studies to date, CMR was not associated with an increase in DNA damage in leucocytes, as determined by γ-H2AX expression, in vitro or in vivo. In vitro, CMR was also not associated with a change in leucocyte viability or function. However, CMR was associated with a statistically significant reduction in T cells, and a non-significant reduction in monocytes, in vivo.

The published in vitro data regarding DNA damage are inconsistent. Szerencsi et al.⁸ found 3-T MRI scanning (clinical brain protocol, scan duration 0–82 min) was not associated with single-strand DNA breaks or micronuclei in human lymphocytes immediately post-scanning. Similarly, Schwenzer et al.⁹ found static magnetic field (3 T) and MRI scanning (scan duration up to 2 h) was not associated with DNA DSBs in cultured human cancer cells at 0, 1, and 24 h post-imaging, and Reddig et al.¹⁰ found that static magnetic field (7 T) and MRI scanning (1 h) were not associated with DNA DSBs in human lymphocytes at 0, 1, and 20 h post-imaging. In all 3 studies, DNA damage was found in cells exposed to positive control environments (gamma rays or X-rays). In contrast, Lee et al.¹¹ observed a significant increase in single strand DNA breaks, micronuclei and chromosomal aberrations in human lymphocytes following 3 T MRI, the frequency of which increased in a scan duration-dependent manner (0–89 min). Similarly Simi et al.¹² reported a significant increase in the frequency of micronuclei in human lymphocytes immediately following 1.5 T CMR. In the only in vitro study prior to our study to assess the impact of MRI on cell viability, Reddig et al.¹⁰ found that MRI was not associated with loss of cell viability.

Our study is, by some distance, the largest and most comprehensive in vitro study to date to evaluate the effect of MRI on DNA integrity. Importantly, our study included assessments of cell viability and function to evaluate the impact of DNA damage, if this had occurred. The in vitro findings are consistent; demonstrating no significant change in DNA integrity, cell viability, cell apoptosis, or cell function with CMR.

A large increase in DNA damage and loss of cell integrity was observed in vitro in both the group exposed to CMR and in the temperature and time control group, emphasizing the importance of the control group to the study, and suggesting that these changes occurred as a result of removing the cells from their physiological environment, which is associated with, in particular, cell hypoxia,^13,14 rather than an effect of CMR.

The published in vivo data are also inconsistent. Fiechter et al.³ reported a statistically significant increase in DNA DSBs in human lymphocytes immediately following contrast agent-enhanced CMR (n = 20), with a similar scanning protocol to our study, albeit longer duration (68 ± 22 min). In contrast, Brand et al.⁷ found contrast agent-enhanced CMR (scan duration range 30–60 min), again with a similar scanning protocol to our study, was not associated with an immediate increase in DNA DSBs in human lymphocytes (n = 45). Neither study examined the impact of CMR on cell numbers. Lancellotti et al.⁴ found non-contrast CMR (scan duration range 35–40 min) was not associated with an increase in DNA DSBs in human blood lymphocytes in the early-phase post-scanning (1 and 2 h; n = 20) but was associated with an increase in DNA DSBs at 2 days (n = 19) and 1 months (n = 15) post-scan. There was a statistically significant reduction in total lymphocyte count at 1 h post-scan, due to a decrease in natural killer and natural killer T cells, which recovered to prescan levels at 2 h post-scan. Lancellotti et al. also found a transient expression of activation markers in neutrophils and monocytes, which was interpreted as a brief, CMR-induced, inflammatory response. Yildiz et al.¹⁵ (n = 28) found non-contrast MRI was not associated with an immediate increase in single-strand DNA breaks in human lymphocytes, but contrast agent-enhanced MRI was associated with an immediate increase in single-strand DNA breaks.

The finding that CMR was not associated with immediate DNA damage in leucocytes in vivo in our study is therefore in keeping with the findings of Brand et al.⁷ and Lancellotti et al.⁴, albeit in a larger population. The reduction in viable leucocytes that we observed is in keeping with the findings of Lancellotti et al.⁴

Taking the results of our study together with those from the other studies discussed, it may be that contrast-enhanced CMR does have a biological effect. However, the precise nature of this effect and, crucially, its subsequent biological significance, if any, remain unknown. For example, the biological relevance of the statistically significant reduction in viable leucocytes observed in vivo in our study and in the study by Lancellotti et al.,⁴ the magnitude of which is small in comparison with natural variation,^16,17 is unclear. Magnetic resonance imaging has been widely used for two decades, with more than 60 million MRI scans currently performed annually, with no apparent excess adverse outcome, although epidemiological studies to investigate this in more detail, and to determine the accurate long-term impact of imaging modalities associated with ionizing radiation, would be highly informative.^18,19

The mechanisms for the potential biological effect of MRI are unclear. The electromagnetic fields associated with MRI are non-ionizing and are therefore not able to generate free electrons with sufficient energy to produce complex DNA damage, which may be of most biological importance (see below).²⁰ It has been proposed that the delayed but persistent DNA damage observed by Lancellotti et al. [seen at 2 days and 1 month post-CMR but not in the early period post-scan; the 1 month γ-H2AX expression correlated strongly with the specific absorption rate (r = 0.79)] may suggest that impaired repair mechanisms may mediate the potential biological effect, although the findings are inconsistent with the DNA damage occurring during imaging.^6,20 Others have proposed that MRI may lead to an increase in oxidative stress, which may in turn increase the rate of production of DNA damage by endogenous processes.^20–22

Similar to the other studies in this field, our in vivo work did not include control data, data to contextualize the findings or investigation of the subsequent biological impact of the findings. To this end, the results of this study, and those of other studies discussed, can really only be considered preliminary exploratory work, from which clinically relevant conclusions should not be drawn. Comprehensive, definitive studies are urgently required.

Finally, the validity of γ-H2AX expression as a marker of DNA damage is worthy of discussion. Similar to Fiechter et al.³ and Lancellotti et al.⁴, we observed γ-H2AX expression to vary markedly between individuals, with both large increases and large decreases in γ-H2AX expression seen following CMR. As such, the results of studies using γ-H2AX as a marker of DNA damage in small cohorts, such as the study by Fiechter et al., should be interpreted with caution. Also, γ-H2AX is not specific for DNA DSBs; it is a marker of stalled replication forks and DNA transcriptional activity.^23–27 Furthermore, it may be the type of DNA damage, rather than the number of lesions, that is most important. The number of lesions produced by ionizing radiation is significantly less than the number produced by endogenous processes (∼50 000 per cell per day); however, the complexity of the damage is far greater, which is of crucial importance, because the ability to repair these complex lesions, and maintain genome integrity, is reduced.²⁰ As such, it may be more relevant to assess downstream effects, such as chromosomal aberrations, for which there is more evidence of an association with cancer risk.

Limitations

The lack of in vivo control data, data to put the findings into context and investigation into the biological significance of the in vivo findings, has been discussed. Only one magnetic field strength (1.5 T) was investigated. In view of the in vivo CMR scans being clinical scans, the additional time required to determine specific absorption rate for each sequence meant that it was not possible to collect this data.

Conclusion

In the largest study to date, 1.5 T CMR of ∼40 min duration was not associated with DNA DSBs in vitro or in vivo. Histone H2AX phosphorylation expression varied markedly between individuals; therefore, small studies using γ-H2AX as a marker of DNA damage should be interpreted with caution. Cardiovascular magnetic resonance was also not associated with loss of leucocyte viability or function in vitro. Cardiovascular magnetic resonance was associated with a statistically significant reduction in viable leucocytes in vivo, although the clinical relevance of the magnitude is unclear.

Supplementary material

Supplementary material is available at European Heart Journal online.

Funding

C.A.M is funded by a Clinician Scientist Award (CS-2015-15-003) from the National Institute for Health Research, UK. The views expressed in this publication are those of the authors and not necessarily those of the NHS, the National Institute for Health Research or the Department of Health. DMD was supported by the Medical Research Council (Award G1001044), a Wellcome Trust Investigator Award (110091) and the Manchester Collaborative Centre for Inflammation Research (funded by a pre-competitive open-innovation award from GSK, AstraZeneca and The University of Manchester, UK).

Conflict of interest: none declared.