Author + information
- Received April 2, 2013
- Revision received June 21, 2013
- Accepted July 15, 2013
- Published online March 4, 2014.
- Christopher A. Miller, BSc, MBChB∗,†,‡∗ (, )
- Jaydeep Sarma, MA, MB BChir, PhD∗,‡,
- Josephine H. Naish, PhD†,
- Nizar Yonan, MD∗,‡,
- Simon G. Williams, MD∗,‡,
- Steven M. Shaw, PhD∗,‡,
- David Clark, BSc§,
- Keith Pearce, BSc∗,
- Martin Stout, PhD∗,
- Rahul Potluri, MBChB∗,†,
- Alex Borg, MD∗,
- Glyn Coutts, PhD‖,
- Saqib Chowdhary, PhD∗,‡,
- Gerry P. McCann, MD¶,
- Geoffrey J.M. Parker, PhD†,
- Simon G. Ray, MD∗,‡ and
- Matthias Schmitt, MD, PhD∗,‡
- ∗North West Heart Centre and Transplant Centre, University Hospital of South Manchester, Wythenshawe Hospital, Manchester, United Kingdom
- †Centre for Imaging Sciences and Biomedical Imaging Institute, University of Manchester, Manchester, United Kingdom
- ‡Institute of Cardiovascular Sciences, University of Manchester, Manchester, United Kingdom
- §Alliance Medical Cardiac MRI Unit, Wythenshawe Hospital, Manchester, United Kingdom
- ‖Christie Medical Physics and Engineering, Christie Hospital, Manchester, United Kingdom
- ¶NIHR Leicester Cardiovascular Biomedical Research Unit and Department of Cardiovascular Sciences, University of Leicester, Leicester, United Kingdom
- ↵∗Reprint requests and correspondence:
Dr. Christopher Miller, North West Heart Centre and The Transplant Centre, University Hospital of South Manchester, Wythenshawe, Manchester, Manchester M23 9LT, United Kingdom.
Objectives This study sought to evaluate the diagnostic performance of multiparametric cardiovascular magnetic resonance (CMR) for detecting cardiac allograft vasculopathy (CAV) using contemporary invasive epicardial artery and microvascular assessment techniques as reference standards, and to compare the performance of CMR with that of angiography.
Background CAV continues to limit the long-term survival of heart transplant recipients. Coronary angiography has a Class I recommendation for CAV surveillance and annual or biannual surveillance angiography is performed routinely in most centers.
Methods All transplant recipients referred for surveillance angiography at a single UK center over a 2-year period were prospectively screened for study eligibility. Patients prospectively underwent coronary angiography followed by coronary intravascular ultrasound, fractional flow reserve, and index of microcirculatory resistance. Within 1 month, patients underwent multiparametric CMR, including assessment of regional and global ventricular function, absolute myocardial blood flow quantification, and myocardial tissue characterization. In addition, 10 healthy volunteers underwent CMR.
Results Forty-eight patients were recruited, median 7.1 years (interquartile range: 4.6 to 10.3 years) since transplantation. The CMR myocardial perfusion reserve was the only independent predictor of both epicardial (β = −0.57, p < 0.001) and microvascular disease (β = −0.60, p < 0.001) on stepwise multivariable regression. The CMR myocardial perfusion reserve significantly outperformed angiography for detecting moderate CAV (area under the curve, 0.89 [95% confidence interval (CI): 0.79 to 1.00] vs. 0.59 [95% CI: 0.42 to 0.77], p = 0.01) and severe CAV (area under the curve, 0.88 [95% CI: 0.78 to 0.98] vs. 0.67 [95% CI: 0.52 to 0.82], p = 0.05).
Conclusions CAV, including epicardial and microvascular components, can be detected more accurately using noninvasive CMR-based absolute myocardial blood flow assessment than with invasive coronary angiography, the current clinical surveillance technique.
- cardiac allograft vasculopathy
- cardiovascular magnetic resonance
- microvascular disease
- myocardial blood flow
Cardiac allograft vasculopathy (CAV) continues to represent the major limitation to long-term survival in heart transplant recipients (1). CAV is characterized by diffuse coronary intimal and medial thickening. It affects both the epicardial arteries and the microvessels; however, it does so independently, and epicardial and microvascular disease are both independently predictive of prognosis (2–4). Because of denervation of the transplanted heart, CAV usually does not become clinically apparent until it has progressed to an advanced stage, when sequelae such as myocardial infarction, progressive heart failure, or arrhythmic sudden death ensue. Screening is therefore required for its early detection.
Coronary intravascular ultrasonography (IVUS) is considered the gold standard technique for diagnosing CAV; however, its broad clinical use in this context is limited by cost and lack of widespread expertise, and its evaluation is limited to epicardial vessels (5). Invasive coronary angiography has a low sensitivity for detecting CAV because of the diffuse nature of the disease with a lack of normal reference segments, and relatively late occurring luminal narrowing (6). Furthermore, angiography is associated with significant, albeit uncommon, complications, and repeated studies are costly and carry a considerable cumulative radiation burden. However, despite these shortcomings, angiography has a Class I recommendation for CAV screening, and annual or biannual surveillance angiography is performed routinely in most centers (7).
Noninvasive imaging approaches to the detection of CAV offer a number of theoretical advantages; however, the findings of the majority of studies assessing their diagnostic performance have been unconvincing (8). Importantly, most such studies have been limited by the use of coronary angiography as the reference standard, with stenoses of 50% or 70% as the significance thresholds, despite adverse events frequently occurring well before such advanced disease is reached (9). Furthermore, none has included reference assessment of the microvasculature.
Cardiovascular magnetic resonance (CMR) is a potentially attractive screening modality for CAV due to its lack of ionizing radiation and its multiparametric nature, namely, its ability to assess multiple aspects of pathology in a single examination (including ventricular function, myocardial perfusion, and myocardial tissue characterization). Systematic evaluation of multiparametric CMR for the diagnosis of CAV has not been reported to date. The aim of this study was to evaluate the diagnostic performance of multiparametric CMR in CAV, and to compare it with that of invasive coronary angiography, using contemporary invasive epicardial artery and microvascular assessment techniques as reference standards.
Patients and study design
All heart transplant recipients referred for CAV surveillance coronary angiography at University Hospital of South Manchester NHS Trust, United Kingdom (1 of 6 UK adult heart transplant centers), between November 1, 2010, and November 1, 2012, were prospectively screened for study eligibility (Fig. 1). Patients were excluded if they had a contraindication to CMR or adenosine infusion, an estimated glomerular filtration rate of 35 ml/min/1.73 m2 or less, or current confirmed or suspected acute allograft rejection.
In addition, 10 age- and sex-matched healthy volunteers were recruited. The volunteers were completely asymptomatic with no known risk factors or history of cardiac disease, normal physical examination, and normal electrocardiogram (i.e., they were not patients who had been referred for CMR that was subsequently found to be normal).
Patients underwent coronary angiography, followed immediately by invasive coronary physiological measurements and coronary intravascular ultrasonography (IVUS) as described in the following text. Within 1 month of the invasive investigations, patients underwent multiparametric CMR assessment. The order of the invasive assessment and CMR was determined randomly, and no patient had an interim cardiovascular event. Healthy volunteers underwent CMR only. 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.
Coronary angiography was performed according to the standard Judkin technique with a 6F system. Multiple projections of the coronary arteries were acquired, including at least 2 orthogonal views of the proximal, mid, and distal left anterior descending (LAD) artery.
After acquiring the angiographic images, 70 IU/kg heparin was administered intravenously, and a 6-F guiding catheter was used to engage the left coronary artery. Intracoronary nitroglycerin (200 μg) was given. A 0.014-inch coronary pressure wire (Radi Medical Systems, St. Jude Medical, St. Paul, Minnesota) was calibrated, equalized, and advanced to the distal portion of the LAD. Maximal hyperemia was induced by administration of intravenous adenosine (140 μg/kg/min) through a 18G cannula in a large peripheral vein for 3 min before and during data acquisition. Hyperemic mean transit time was determined by averaging the transit times of 3 × 3 ml boluses of room temperature saline. Mean aortic and distal coronary pressures were recorded. Meticulous attention was paid to guide catheter engagement. Fractional flow reserve (FFR) was calculated by dividing the mean distal coronary pressure by the mean proximal coronary pressure during hyperemia. Index of microcirculatory resistance (IMR) was calculated by multiplying the distal coronary pressure by the hyperemic mean transit time, measured simultaneously (10).
After the physiological measurements, a 40 MHz IVUS catheter (Atlantis SR Pro, Boston Scientific, Natick, Massachusetts) was advanced over the wire, so that the transducer was positioned in the distal artery, as close as possible to the pressure transducer mounted on the pressure wire. Automated pullback at a constant rate of 0.5 mm/s was performed along the length of the vessel.
Cardiovascular magnetic resonance
The CMR was performed using a 1.5-T scanner (Avanto, Siemens Medical Imaging, Erlangen, Germany) equipped with a 32-element phased-array coil.
Global and Regional Left Ventricular Function
Steady-state free precession cine images were acquired in standard long-axis views and in a stack of short-axis slices covering the left ventricle (LV). Short-axis tagged images were acquired at basal, mid, and apical ventricular levels using a segmented k-space fast gradient echo sequence with spatial modulation of magnetization in orthogonal planes.
Using a saturation recovery gradient echo sequence, basal, mid, and apical short-axis images were acquired every heartbeat during pharmacological vasodilation (“stress”) and at rest. For stress imaging, intravenous adenosine (140 μg/kg/min) was administered through a 18G cannula in a large peripheral vein for 3 min before, and during, data acquisition. A 0.05 mmol/kg bolus of gadolinium-based contrast agent (gadopentetate dimeglumine [Gd-DTPA]; Magnevist, Bayer HealthCare, Wuppertal, Germany) was administered intravenously at 5 ml/s followed by a 30-ml saline flush. Rest imaging was performed 10 min after stress imaging with a further 0.05 mmol/kg of contrast agent. After rest perfusion image acquisition, a further 0.1 mmol/kg of contrast agent was administered to bring the total dose to 0.2 mmol/kg.
A single-shot modified Look Locker inversion recovery sequence was acquired in short-axis view at midventricular level before the contrast agent was administered, and 15 min after the final dose of the contrast agent (11). Blood samples were taken at the time of CMR to measure hematocrit. Standard late gadolinium enhancement imaging was performed at least 10 min after the final dose of the contrast agent using spoiled gradient echo segmented inversion recovery and phase-sensitive inversion recovery segmented gradient echo sequences.
Quantitative Coronary Angiography
Angiographic images were analyzed using quantitative coronary angiography, performed on the proximal, mid, and distal LAD with guiding catheter calibration (QAngio XA, Medis Medical Imaging Systems, Leiden, the Netherlands). Reference and minimal lumen diameters for the 3 sites were measured, and the greatest percent diameter stenosis was recorded.
Quantitative analysis was performed using QIVUS (Medis Medical Imaging Systems), according to the method described by Fearon et al. (12). The external elastic lamina and luminal border were traced on images acquired every 0.1 mm. “Plaque area” (also known as intima-media area) was calculated as vessel area minus luminal area. By means of Simpson’s method, vessel volume, luminal volume, and hence, plaque volume were calculated as the sum of the respective areas multiplied by the “segment” length of 0.1 mm. Subsequently the plaque volume index, defined as plaque volume expressed as a percentage of vessel volume, was calculated to normalize for vessel size and length of IVUS pullback.
Cardiovascular magnetic resonance
The LV mass, end-diastolic volume, end-systolic volume, and ejection fraction (EF) were quantified from steady-state free precession images using CMRtools (Cardiovascular Imaging Solutions, London, United Kingdom) (13). Peak systolic circumferential strain (εcc) and strain rate (systolic and early diastolic) were measured from midventricular short-axis tagged images using SinMod (inTag software, version 5.0, CREATIS Laboratory, Lyon, France; and Maastricht University, the Netherlands) (14). Basal and apical short-axis rotation, calculated from basal and apical tagged images using the same software, and epicardial areas, measured on corresponding steady-state free precession images, were incorporated into a custom-written algorithm (Microsoft Excel using Visual Basic). After expression of time coordinates as a percentage of systolic duration (time between peak electrocardiogram R-wave and aortic valve closure) and cubic spline interpolation, twist was calculated by subtracting basal rotation from apical rotation at each time point. Normalized twist was calculated as the twist angle divided by distance between basal and apical slice positions. Torsion (represented by an approximation of the circumferential-longitudinal shear angle) was calculated by multiplying normalized twist by the mean of the basal and apical epicardial radii at each time point.
Endocardial and epicardial contours were drawn on the perfusion images using Osirix Imaging Software, version 4.0 (Pixmeo, Geneva, Switzerland). An additional region of interest (ROI) was drawn in the blood pool on the basal images, avoiding papillary muscles and trabeculae. The ROIs were manually translated on each perfusion image of the same slice to compensate for rigid-body translational motion. Perfusion quantification was performed in MatLab, version R2009a (MathWorks, Natick, Massachusetts) using algorithms written in house. Signal intensity curves were extracted from the average signal in the blood pool, to provide an arterial input function, and on a voxel-wise basis from the myocardial ROI. Signal intensity was converted to contrast agent concentration (15). Data for quantitative perfusion analysis were restricted to the first pass of the contrast agent through the heart, which was automatically detected from the blood pool signal curve. Perfusion values were obtained on a voxel-wise basis using generalized Tikhonov deconvolution with a b-spline representation of the impulse response function (16). Myocardial perfusion reserve (MPR) was calculated by dividing median hyperemic myocardial blood flow (MBF) by median resting MBF.
The LGE images were reported visually by 2 experienced operators, and the presence or absence of LGE, and its distribution pattern, were recorded. Myocardial T1 relaxation time was measured by drawing endocardial and epicardial contours on the modified Look Locker inversion recovery images using Osirix Imaging Software, version 4.0 (Pixmeo). In keeping with Wong et al. (17), myocardium in the vicinity of infarcted myocardium was excluded, but foci of LGE in myocardium free from infarction were not excluded. An additional ROI was drawn in the blood pool for measurement of blood T1. The ROIs were manually translated on each effective inversion time (TIeff) image to compensate for rigid-body translational motion. To obtain voxel-wise T1 relaxation maps, a 3-parameter fit to the signal intensity, S as a function of TIeff was performed according to S(TIeff) = A − Be(-TIeff/T1∗), and T1 was calculated as T1 = T1*((B/A) − 1). Fitting was carried out using MatLab, version R2009a (MathWorks). After applying a heart-rate correction algorithm, mean midventricular pixel T1 relaxation times before and after contrast were then used to calculate myocardial extracellular volume according to the following formula: extracellular volume fraction (ECV) = λ × (1− hematocrit), where the partition coefficient, λ = ΔR1(myocardium) / ΔR1(blood). The ΔR1 is proportional to contrast agent concentration: ΔR1 = R1(post-contrast) − R1(pre-contrast).
All data were analyzed in a blinded fashion, with independent analysis of CMR and invasive data. The IVUS plaque volume index and IMR were used as the reference standards for epicardial and microvascular disease, respectively (see the Discussion section). Statistical analysis was performed using SPSS, version 19 (IBM, Armonk, New York), and STATA, version 11.0 (StataCorp, College Station, Texas). Continuous variables are expressed as mean ± SD unless stated. An independent-samples t test (or Mann-Whitney U test where appropriate) was used to compare data from transplant patients and healthy volunteers. Linear regression was used to investigate possible associations between continuous invasive and CMR data, and stepwise selection methods were used to determine the most important associations. Separate stepwise analyses were performed for the angiographic and CMR data. Because of the relatively small size of the study, the number of univariable associations entered into the multivariable model was limited to 5, and overlapping variables were avoided. Receiver-operating characteristic (ROC) curve analyses were used to determine optimal cut-off points for CMR MPR and coronary angiographic stenosis, respectively, for detecting CAV (epicardial and microvascular disease), and estimates of sensitivity and specificity were derived. Given the absence of well-defined severity thresholds, the 75th centile (“severe disease”) and median values (“moderate disease”) for both plaque volume index and IMR were used to define the presence of epicardial and microvascular disease, respectively. The diagnostic performance of CMR MPR and coronary angiography were compared using a chi-square comparison of the trapezoidal area under the respective ROC curves.
Forty-eight patients were recruited (Fig. 1). In 2 patients the LAD was occluded, meaning that IVUS and invasive coronary physiological assessment were not possible (for the purpose of the subsequent ROC curve analyses, both were considered to have severe epicardial disease). Three patients underwent CMR without gadolinium contrast (estimated glomerular filtration rate deteriorated to <35 ml/min/1.73 m2 between the invasive studies and CMR in 2 patients, and 1 patient terminated the scan early because of claustrophobia). There were no complications. Demographic data are presented in Table 1 and Online Table 1. Median time from transplantation to enrollment was 7.1 years (interquartile range: 4.6 to 10.3 years).
Mean plaque volume index was 22.4 ± 9.8% and mean maximal intima-media thickness was 1.21 ± 0.57 mm. Mean FFR was 0.90 ± 0.06. An FFR <0.80 was observed in 1 patient (2%). Mean IMR was 23.7 ± 12.5. There was a significant correlation between plaque volume index and FFR (r = −0.46, p = 0.001), but there was no correlation between plaque volume index and IMR (r = 0.24, p = 0.103). The FFR was seen to improve as IMR deteriorated (r = 0.32, p = 0.028).
Mean maximum angiographic stenosis was 23.9 ± 16.0%. Maximum angiographic stenosis showed a significant correlation with plaque volume index (r = 0.33, p = 0.024) and FFR (r = −0.38, p = 0.010). There was no correlation between maximum angiographic stenosis and IMR (r = −0.16, p = 0.281).
Cardiovascular magnetic resonance
The CMR data are presented in Table 2 and Online Table 2. In keeping with the significantly higher resting rate pressure product seen in transplant patients (Table 1), resting MBF was significantly higher in transplant patients than in healthy volunteers. Stress MBF was significantly lower in transplant recipients compared with healthy volunteers, as was MPR. Significant differences were also seen between transplant patients and healthy volunteers in indexed LV end-diastolic volume, εcc, and twist, but not in normalized twist, torsion, or time to peak torsion (i.e., time to onset of untwisting). Atypical LGE was seen in almost half of transplant recipients, and 4 of those (9% of the study population) also had infarct-typical LGE. Inferior right ventricular septal insertion point enhancement was the most common type of atypical LGE (16 patients; 36%), but midwall (5 patients; 11%), “punched-out” (4 patients; 9%) and epicardial (2 patients; 4%) patterns were also observed. Pericardial LGE was observed in 3 patients (7%). Myocardial ECV was significantly higher in transplant patients compared with healthy volunteers.
Associations with plaque volume index
On univariable analysis, maximum angiographic stenosis showed a significant association with plaque volume index; however, after correcting for time since transplantation, this relationship was no longer significant (p = 0.295) (Table 3, Online Table 3). Early diastolic strain rate, stress MBF, MPR and infarct LGE were significantly associated with plaque volume index on univariable analyses, but only MPR and early diastolic strain rate remained independently associated with plaque volume index on multivariable analysis (Fig. 2A).
Associations with index of microcirculatory resistance
Maximum angiographic stenosis was not significantly associated with IMR on univariable analysis (Table 4). Patient characteristics including donor age, the presence of hypertension in the donor, and recipient:donor body mass index ratio showed significant associations with IMR on univariable analyses, as did CMR parameters such as EF, εcc, stress MBF, and MPR. On multivariable analysis only donor hypertension, EF, and MPR remained independently associated with IMR (Fig. 2B).
Diagnostic performance of cardiovascular magnetic resonance and angiography
In light of the multivariable regression results, MPR was the only CMR parameter used for ROC curve analysis. The diagnostic performance of angiography and CMR MPR are displayed in Table 5. When epicardial and microvascular disease were considered together, as in vivo, CMR MPR outperformed angiography (chi-square = 6.6, p = 0.01 for detecting moderate epicardial or microvascular disease; chi-square = 3.7, p = 0.05 for detecting severe epicardial or microvascular disease) (Fig. 3).
Associations with myocardial perfusion reserve
There was no association between MPR and maximum angiographic stenosis (β = −0.11, p = 0.469) or between MPR and FFR (β = 0.13, p = 0.415). On univariable analysis, donor age (β = −0.32, p = 0.038), LV mass index (β = −0.34, p = 0.021), εcc (β = −0.32, p = 0.049), infarct LGE (β = −0.29, p = 0.051), and ECV (β = −0.34, p = 0.045) were associated with MPR; but on multivariable stepwise regression only LV mass index (β = −0.39, p = 0.015) and infarct LGE (β = −0.36, p = 0.025) remained independently associated with MPR.
This study provides comprehensive assessment of cardiac structure and function in the medium to long term after heart transplantation. The MPR, on the basis of measurement of absolute stress and rest MBF using CMR, was the only independent predictor of both epicardial and microvascular disease, and its diagnostic performance was significantly greater than that of invasive coronary angiography, the current clinical standard.
CAV is an exemplifier of a disease that affects both the epicardial and microvascular coronary compartments. In keeping with histological work and other invasive coronary physiology studies, the current study serves to confirm that CAV affects the epicardial arteries and the microvasculature independently (2,3,18,19). Comprehensive assessment of CAV, both in terms of evaluating disease severity and evaluating the performance of new diagnostic approaches, therefore requires assessment of both compartments. The current study is the first to evaluate a diagnostic approach to CAV with epicardial and microvascular benchmarks.
In the present study IVUS, rather than FFR, was used as the epicardial reference standard. IVUS is generally regarded as the gold standard technique for epicardial artery assessment in CAV and is considerably more established, and a number of studies have demonstrated that IVUS-derived vessel wall parameters, irrespective of hemodynamic significance, predict outcome in transplant recipients (20–24). In addition, because of the complex interplay between epicardial and microvascular disease, FFR may not provide a good indication of epicardial disease in CAV. Hirohata et al. (25) found that in patients >2 years post-transplant, for a given plaque burden, FFR varied significantly according to IMR; with FFR seen to improve (i.e., increase/become closer to 1.0) as IMR deteriorated (increased). These findings are directly in keeping with the findings of the current study, where FFR and IMR were shown to have a significant positive correlation. These observations reflect that in the setting of microvascular dysfunction, maximal achievable coronary flow is diminished, and thus the impact of an epicardial stenosis is lessened.
Such observations are not unique to CAV. For example, in the case of a given epicardial stenosis, FFR will be lower when the artery subtends viable myocardium (i.e., with minimal microvascular resistance, hence a large “pressure drop” down the artery) compared with if the artery were to supply infarcted myocardium (i.e., substantial microvascular resistance, hence a smaller pressure drop). Nevertheless, given the prevalence and importance of microvascular disease in CAV, these findings led Hirohata et al. (25) to conclude that “FFR may not provide a good representation of epicardial plaque burden late (i.e., more than 2 years) after heart transplantation.” Indeed, over a number of studies, the correlation between FFR and IVUS parameters has varied considerably, from no correlation to a correlation coefficient of −0.58 (19,23).
In keeping with most other contemporary CAV studies, plaque volume index, rather than other IVUS parameters such as intima-media thickness, was used as it better reflects disease burden (12,19). In the absence of established disease severity cut-offs, 75th centile and median values were pragmatically chosen for ROC curve analysis. The IMR was used instead of invasive coronary flow reserve to assess the microvasculature because, unlike coronary flow reserve, IMR is specific for the microcirculation (i.e., independent of epicardial artery function) and largely independent of hemodynamic variations, and is more reproducible (10,26,27). Median and 75th centile IMR values were used for ROC curve analysis in keeping with other studies (28,29). Epicardial and microvascular disease severity in the current study was almost identical to that in the previous largest studies, and hence the disease severity thresholds used here appear reasonable (4,12,19,25).
Myocardial blood flow
The myocardial perfusion findings with regard to epicardial artery disease in the current study are in keeping with the results of 2 smaller studies that assessed MBF in CAV using 13N-ammonia positron emission tomography. Kofoed et al. (30) and Wu et al. (31) found MPR and hyperemic MBF to be inversely related to IVUS parameters of CAV severity, with very similar univariable correlations to those found here (MPR: correlation coefficients up to −0.61 and −0.40, respectively, in the studies by Kofoed et al. and Wu et al. compared to a β value of −0.55 in the current study; hyperemic MBF: −0.49 and −0.46, respectively, compared to −0.51 in the current study). Also in keeping with the current study, in both positron emission tomography studies, resting MBF was higher in transplant patients than in control subjects, at least in part secondary to the higher resting heart rate seen in transplant recipients due to allograft vagal denervation, and was unrelated to the severity of epicardial artery disease.
A number of studies in other cardiovascular diseases have considered microvascular dysfunction to be present when MPR, or hyperemic MBF, are reduced in the presence of normal epicardial arteries (usually assessed with angiography) (32,33). Other studies have demonstrated a correlation between MPR and invasive coronary flow reserve (measured using intracoronary Doppler) in patients without significant epicardial coronary stenoses (34,35). To our knowledge, the current study is the first in any cardiac pathology to demonstrate that MPR, and hyperemic MBF, are independent predictors of microvascular function, when microvascular function itself is independently measured. Highlighting the importance of microvascular disease in CAV, MPR was governed by epicardial and microvascular disease to a similar degree.
The inverse association between MPR and LV mass index seen here has been described in other conditions where microvascular dysfunction is prominent (36–38), but this is the first time it has been demonstrated in transplanted hearts. There was also an association between the presence of infarct-typical LGE and reduced MPR; however, although myocardial extracellular volume was inversely associated with MPR on univariable analysis, the relationship did not remain significant on multivariable analysis. The discordance seen here between MPR and FFR is in keeping with that reported in other conditions and, as described by Johnson et al. (39), is likely to reflect the nature of the coronary pathophysiology involved.
Myocardial tissue characterization
Infarct prevalence in the current study (9%) was in keeping with the study by Butler et al. (11%) (40), although lower than in the study by Steen et al. (37%) (41); however, the cohort studied by Steen et al. represented advanced epicardial disease (including 19% with focal angiographic stenoses of ≥75%). The prevalence of infarct-atypical LGE in the current study (49%) was very similar to that found by Butler et al. (40%) (40) and by Steen et al. (51%) (41), and as in these other studies, was unrelated to epicardial disease. In addition to these other studies, which did not have a microvascular reference, the current study showed that infarct-atypical LGE was unrelated to microvascular disease.
In keeping with the findings of Weis et al. (42), who used CFR to assess microvascular dysfunction in patients with angiographically normal epicardial arteries, microvascular function was independently associated with LV function, which may reflect chronic subendocardial ischemia leading to impaired LV function. Early diastolic strain rate was independently associated with epicardial disease, which is in keeping with the study by Korosoglou et al. (43), but in general, myocardial deformation parameters were not discriminatory for CAV.
Why screen at all?
It has been suggested that routine CAV screening should not be performed at all given the lack of evidence-based treatment options for improving symptoms or prognosis in established disease, other than retransplantation, which is possible in only a minority (44). Certainly significant luminal narrowing on angiography represents advanced vascular disease that is unlikely to be modifiable, and adverse events frequently occur well before this degree of disease is reached (9). This study demonstrates that CAV can potentially be detected noninvasively at a much earlier stage using CMR-based MBF assessment, which may allow earlier intensification of preventative therapy and may help pave the way for the development of disease-modifying therapies. However, although the CMR perfusion sequence used in the current study is in routine clinical use, and thus could be readily applied, CMR MBF quantification is technically challenging, and although considerable effort is being put into developing commercially available quantification software, at present it remains a research tool. Furthermore, it should be noted that 30% of patients screened had contraindications to CMR scanning or gadolinium contrast agent, and, therefore, such an approach would not be appropriate for all patients.
In keeping with the majority of studies involving transplant recipients, the current study is a relatively small, single–center study. The results require confirmation in larger, multicenter studies. The IVUS and invasive physiological assessments were performed on the LAD only; however, that also is in keeping with most other transplant studies (12). The IMR measurements were made without taking collateral flow into account (requiring balloon occlusion of the vessel); however, that is in keeping with all other studies using IMR in transplant recipients as collateralization is not a feature of CAV (12,19). Finally, this is a cross-sectional study and, therefore, the effect of disease progression on CMR MPR, and the prognostic value of CMR MPR, have not been assessed.
In this comprehensive assessment of cardiac structure and function in the medium to long term after transplantation, CMR-based MPR was independently predictive of both epicardial and microvascular components of CAV. Furthermore, the diagnostic performance of CMR MPR was significantly higher than that of coronary angiography, the current clinical screening technique.
The authors thank New Start Transplant Charity, the radiographers and administrative staff of the Alliance Wythenshawe CMR Unit, the secretarial and nursing staff of the Transplant Unit, and the Cardiology Waiting List coordinators in the North West Heart Centre. Finally, this study would not have been possible without the willing cooperation of the patients of the Transplant Unit, Wythenshawe, United Kingdom.
Dr. Miller is supported by a Fellowship from the National Institute for Health Research, United Kingdom (NIHR-DRF-2010-03-98). Drs. Miller, Yonan, Williams, and Schmitt have received research funding from the New Start Transplant Charity, United Kingdom. Dr. McCann is supported by a Fellowship from the National Institute for Health Research, United Kingdom (NIHR-PDF-2011-04-51). Dr. Parker is a director and shareholder of Bioxydyn, Ltd. All the other authors have reported they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- cardiac allograft vasculopathy
- cardiovascular magnetic resonance
- fractional flow reserve
- index of microcirculatory resistance
- intravascular ultrasonography
- left anterior descending coronary artery
- late gadolinium enhancement
- myocardial blood flow
- myocardial perfusion reserve
- receiver-operating characteristic
- region of interest
- Received April 2, 2013.
- Revision received June 21, 2013.
- Accepted July 15, 2013.
- American College of Cardiology Foundation
- Hiemann N.E.,
- Wellnhofer E.,
- Knosalla C.,
- et al.
- Haddad F.,
- Khazanie P.,
- Deuse T.,
- et al.
- Tuzcu E.M.,
- De Franco A.C.,
- Hobbs R.,
- et al.
- St. Goar F.G.,
- Pinto F.J.,
- Alderman E.L.,
- et al.
- Miller C.A.,
- Chowdhary S.,
- Ray S.G.,
- et al.
- Tuzcu E.M.,
- Kapadia S.R.,
- Sachar R.,
- et al.
- Fearon W.F.
- Fearon W.F.
- Wong T.C.,
- Piehler K.,
- Meier C.G.,
- et al.
- Rickenbacher P.R.,
- Pinto F.J.,
- Lewis N.P.,
- et al.
- Kobashigawa J.A.,
- Tobis J.M.,
- Starling R.C.,
- et al.
- Aarnoudse W.
- Ng M.K.C.
- Fearon W.F.,
- Shah M.,
- Ng M.,
- et al.
- McGeoch R.,
- Watkins S.,
- Berry C.,
- et al.
- Kofoed K.F.,
- Czernin J.,
- Johnson J.,
- et al.
- Wu Y.W.,
- Chen Y.H.,
- Wang S.S.,
- et al.
- Shelton M.E.,
- Senneff M.J.,
- Ludbrook P.A.,
- Sobel B.E.,
- Bergmann S.R.
- Knaapen P.,
- Germans T.,
- Camici P.G.,
- et al.
- Steen H.,
- Merten C.,
- Refle S.,
- et al.
- Grant S.C.,
- Brooks N.H.,
- Levy R.D.