Objective The relationship between right ventricular (RV) fibrosis and right heart reverse remodelling following pulmonary valve replacement (PVR) has not been well studied in adults with repaired tetralogy of Fallot (rTOF). Our aims were to histologically quantify RV fibrosis and to explore the relationship between fibrosis severity and cardiac remodelling post-PVR.
Methods Adults with rTOF and pre-PVR cardiovascular (CMR) imaging were consented to procurement of RV muscle during PVR. Samples were stained with picrosirius red to quantify collagen volume fraction. Clinical data at baseline and at last follow-up were reviewed. Adverse cardiovascular outcomes included death, sustained arrhythmia and heart failure.
Results Fifty-three patients (male 58%, 38±11 years) were studied. Those with severe fibrosis (collagen volume fraction >11.0%, n=13) had longer aortic cross-clamp times at initial repair compared with the remainder of the population (50 vs 33 min, p=0.018) and increased RV mass:volume ratio pre-PVR (0.20 vs 0.18 g/mL, p=0.028). Post-PVR, the severe fibrosis group had increased indexed RV end-systolic volume index (RVESVi) (74 vs 66 mL/m2, p=0.044), decreased RVESVi change (Δ29 vs Δ45 mL/m2, p=0.005), increased RV mass (34 vs 25 g/m2, p=0.023) and larger right atrial (RA) area (21 vs 17 cm2, p=0.021). A trend towards increased heart failure events was observed in the severe fibrosis group (15% vs 0%, p=0.057).
Conclusions Severe RV fibrosis was associated with increased RVESVi, RV mass and RA area post-PVR in rTOF. Further study is required to define the impact of fibrosis and persistent right heart enlargement on clinical outcomes.
- cardiac magnetic resonance (cmr) imaging
- tetralogy of fallot
- congenital heart disease surgery
Statistics from Altmetric.com
Tetralogy of Fallot (TOF), the most common form of cyanotic congenital heart disease, is a lesion with a unique susceptibility to myocardial fibrosis given the presence of cyanosis in childhood1 and the high prevalence of haemodynamic sequelae as a result of chronic pulmonary valve dysfunction (insufficiency or stenosis).2 Despite successful TOF repair in early life, the adult TOF population remains increasingly vulnerable to the development of adverse outcomes including heart failure, ventricular tachycardia (VT) and/or sudden cardiac death.3 Autopsy studies and surgical specimen review have established an association between the presence of myocardial fibrosis and sudden cardiac death as well as malignant ventricular tachyarrhythmias.4 5
Over the recent decades, cardiovascular magnetic resonance (CMR) imaging has gained increasing prominence for assessment of myocardial fibrosis in patients with repaired TOF (rTOF). Early studies explored late gadolinium enhancement as a marker of replacement fibrosis in rTOF and subsequently delineated an association between hyperenhancement and impaired cardiovascular status, including subnormal exercise capacity.6 More recent literature has concentrated on the application of emerging parametric imaging techniques for estimation of extracellular volume (ECV) quantification and by extension, quantification of diffuse right ventricular (RV) myocardial fibrosis in the TOF population.7 Specifically, RVECV has recently been associated with measures of increased haemodynamic burden in the RV and increased risk of major adverse cardiovascular events late after TOF repair.8
Our understanding of the implications of myocardial fibrosis at the histological level in rTOF has been limited. Literature examining the relationship between histological fibrosis and outcomes late after TOF repair is currently restricted to studies which are few in number, which include only modest numbers of patients with RV specimen collection at the time of pulmonary valve replacement (PVR) and which have only incomplete CMR data analysed.9 The clinical impact of RV fibrosis on heart size and function in rTOF remains incompletely understood and the implications of histological fibrosis on the extent of right heart change following PVR is largely undefined. Therefore, the objectives of this study were to evaluate RV histological fibrosis in adults after TOF repair and to evaluate the impact of severity on the extent of cardiac reverse remodelling following PVR as determined by comparison of preoperative and postoperative CMR studies.
Among adults previously enrolled in the Canadian Outcomes Registry Late After Tetralogy of Fallot Repair (CORRELATE) study,10 those scheduled for surgical PVR at Toronto General Hospital from 2013 to 2017 were approached for inclusion and were prospectively enrolled following consent to intraoperative myocardial biopsy and completion of a preoperative CMR study for assessment of ventricular volumes/mass, ventricular systolic function and atrial dimensions. Patients with contraindications to preoperative CMR (including pregnancy, claustrophobia or implanted devices (pacemakers or automatic implanted cardioverter defibrillator (AICD)) or those with an inadequate myocardial biopsy specimen were excluded from the study. The study complies with the Declaration of Helsinki and written informed consent was obtained from all participants.
Baseline demographic characteristics, surgical history and clinical details are shown in table 1. Additionally, available ECG, echocardiogram and cardiopulmonary exercise test reports within 6 months of CMR completion and at last follow-up were recorded. Major adverse cardiovascular outcomes were defined as death, cardiac transplantation, sustained ventricular or atrial arrhythmia (>30 s) and heart failure requiring admission for intravenous diuretic therapy.
All CMR studies were performed on a 1.5 Tesla imager (Magnetom Avanto Fit; Siemens Healthcare, Erlangen, Germany). Retrospectively gated cine steady-state free precession images were obtained for assessment of ventricular volumes, mass, systolic function and atrial areas. All of the RV and left ventricular (LV) volumes were calculated in the short-axis plane (6–8 mm slice thickness and 0–2 mm interslice gap). Right and left atrial areas were measured in the four chamber view. Free breathing, ECG-gated, cine phase-contrast flow measurements were obtained at the main pulmonary artery for assessment of pulmonary regurgitation (PR). Intravenous contrast injection was included for late gadolinium enhancement imaging ±T1 mapping for ECV quantification in a subset of patients, as previously described by our group.8
Measurements of ventricular volumes, function, mass, atrial size and PR flow were performed using commercially available software (QMASS MR, Medis Medical Imaging Systems, Leiden, the Netherlands). The RV and LV endocardial and epicardial borders were contoured by a single observer blinded to clinical information according to a standardised protocol for measurements of biventricular end-diastolic and end-systolic volumes, biventricular ejection fraction (EF), biventricular mass and biatrial areas, as previously published.10 Atrial dimensions were measured in ventricular end systole and the atrial appendages were excluded. The PR fraction was calculated as the percentage of backward flow over forward flow. The presence of myocardial late gadolinium enhancement was recorded (contiguous slices in short axis and four-chamber views) and ECV was quantified in the RV and LV (base, mid and apical locations in the short-axis view) with comparison to institutional normal values, as previously established.8
All of the RV muscle specimens were obtained immediately after opening the RV outflow tract from the parietal trabeculations located approximately 1–2 cm proximal to the pulmonary valve annulus, with preferential sampling occurring from regions of maximal wall thickness at a depth of 5 mm yielding tissue samples approximately 10 mm in length. Following specimen collection, histological analysis was performed by a single observer blinded to clinical data. Myocardial samples were formalin fixed and embedded in paraffin. For analysis, 4 μm paraffin sections were stained with picrosirius red for quantification of RV fibrosis. Images were taken with an Olympus BX50 microscope (Olympus Canada, Richmond Hill, Canada), 20× magnification, and 3–5 fields were selected from every slide. Automated image analysis was performed using Aperio ImageScope (Leica Biosystems Imaging Vista, California, USA) to quantify collagen burden. A positive pixel count algorithm was used to automatically quantify the area occupied by stain colours within each scanned slide image. Required input parameters for each stain were based on the hue, saturation and intensity as previously established.11 To detect the red colour of collagen, a hue value of 0.9 and a hue width of 0.3 was spedified. Star-shaped microscars (replacement fibrosis) and perivascular fibrosis were excluded from analysis12 (figure 1, panels A and B). Collagen volume fraction was calculated based on the percent of red collagen staining quantified within a tissue section (figure 1, panels C and D). Collagen volume fraction was averaged among all available biopsy specimens analysed from each individual.13
The study population was divided into quartiles according to fibrosis severity; those in the upper quartile were designated as severe histological fibrosis corresponding to collagen volume fraction >11.0% (n=13 patients), in accordance with a previously established threshold of clinically significant fibrosis.2 Demographic and clinical characteristics of subjects were compared according to severity of fibrosis. Continuous variables were tested for normalcy using the Shapiro-Wilk test. Means (with SD) or medians (with IQRs) were reported, as appropriate. To evaluate correlations, the Pearson or Spearman correlation coefficients were used. For comparisons between groups, the Student’s t-test or Wilcoxon rank sum test was used for continuous variables and the Χ2 test or Fisher exact test was applied for categorical variables. Data were analysed using JMP V.13 statistical software package (SAS Institute, Cary, North Carolina, USA).
Fifty-three patients (58% male, 38±11 years) were studied. The median clinical follow-up duration following PVR was 2 years (IQR 1–3). Baseline clinical characteristics of the study population are shown in table 1. Pulmonary valve dysfunction resulting in PVR was attributed to moderate or severe PR in the majority of patients (n=49 or 92%) and mixed valve disease (PR and significant pulmonary stenosis) in the minority (n=4 or 8%).
The median histological collagen volume fraction for the population was 7.1% (IQR 4.1–11.4%). Median fibrosis in the top quartile (n=13 patients) was 17.9% (IQR 12.2–19.8%) as compared with the remainder of the population (n=40 patients) where median fibrosis was 5.3% (IQR 3.5–7.6%). Typical imaging characteristics following staining with picrosirius red are shown in figure 2.
A baseline CMR study was available for analysis in all subjects prior to PVR (median time from CMR to PVR, 12 months (IQR 7–19)). Late gadolinium enhancement imaging was included in six studies; hyperenhancement was observed in expected regions of surgical patch repair and at ventricular insertion points in all, but not within the RV or LV myocardium in any of the patients. Additionally, RVECV and LVECV were measured by T1 mapping in six patients (median RVECV 33% (IQR 27– 37%) and median LVECV 27% (IQR 23–30%). The RVECV was statistically higher in rTOF as compared with institutional controls8 (33% vs 27%, p=0.034), while LVECV did not differ significantly between these groups (27% vs 24%, p=0.35). We were unable to reliably match location of procurement of the RV specimen with a corresponding CMR image (T1 mapping images were limited to the short-axis stack and did not include reliable coverage of the RVOT); therefore, the relationship between quantification of histological fibrosis and RVECV estimation by parametric imaging could not be explored.
Post-PVR CMR analysis was completed in 37 patients at a median of 11 months following surgery (IQR 7–14). Sixteen patients did not undergo post-PVR CMR because of death (n=3), pacemaker/AICD implantation (n=7), relocation out of district or loss to follow-up (n=6). The CMR findings pre-PVR (n=53), post-PVR (n=37) and the delta change post-PVR are summarised in tables 2 and 3.
Associations between fibrosis severity clinical parameters and CMR measurements
Comparing the most severe fibrosis group with the remainder of the population, the aortic cross-clamp time at the time of initial TOF repair was significantly longer in patients with the most severe fibrosis as compared with the remainder of the population (50 min (IQR 40–70) vs 33 min (IQR 27–44), p=0.018). However, total cardiopulmonary bypass time did not differ between the groups. There were no other associations found between measured collagen volume fraction and other clinical characteristics at baseline. Similarly, there were no statistically significant relationships between fibrosis severity and results on ECG, Holter monitor or cardiopulmonary exercise testing, at the time of CMR or at last follow-up.
Pre-PVR, the only right heart CMR measurement to differ significantly in patients with severe RV fibrosis was RV mass:volume ratio, which was higher in patients with severe RV fibrosis (median 0.20 g/mL (IQR 0.15–0.24) vs 0.18 g/mL (IQR 0.12–0.27), p=0.028). The only left heart CMR measurement at baseline to differ according to severity of fibrosis was LV EF, which was significantly lower in the patients with severe fibrosis (51% vs 56%, p=0.041).
Post-PVR, indexed RV end-systolic volume (RVESV) was larger in patients with more severe fibrosis (median 74 mL/m2 vs 66 mL/m2, p=0.044) and accordingly the magnitude of pre-PVR/post-PVR change in indexed RVESV was less pronounced in the severe fibrosis group as compared with the remainder of the population (Δ29 mL/m2 vs Δ45 mL/m2, p=0.005). Similarly, the magnitude of change in indexed RVESV divided by pre-PVR indexed RVESV (this ratio was established in order to account for pre-PVR RV size)14 was smaller in the severe fibrosis group (25% vs 44%, p=0.005, figure 3). The indexed RV mass was larger in patients with severe fibrosis (median 34 g/m2 (IQR 31–38) vs 25 g/m2 (IQR 22–30), p=0.023) as was right atrial area (median 21 cm2 (IQR 18–26) vs 17 cm2 (IQR 15–18), p=0.021).
There was a modest but statistically significant correlation between collagen volume fraction and right atrial area post-PVR CMR (r=0.322, p=0.044). No other statistically significant correlations were identified between collagen volume fraction and CMR measurements.
Over the study period, major adverse cardiovascular events occurred in 9 (17%) of the 53 patients. These mutually exclusive outcomes were death (n=3), sustained VT (n=2), sustained atrial tachycardia (n=3) and heart failure (n=1). One patient died on the day of PVR following rupture of the pulmonary artery patch. A second patient had multiple embolic strokes and died 7 days following PVR. A third patient was known to have liver cirrhosis preoperatively and died following multiorgan failure in the context of overwhelming sepsis 30 days after PVR. It should be noted that this analysis involves a small subset of our total PVR experience (n=468, five deaths; 1%). Perioperative mortality over the past 10 years includes three deaths out of the 260 patients with PVR (1%); all of these three deaths were included in this study thereby distorting the mortality within the current series. There was no association between collagen volume fraction and post-PVR adverse outcomes, as listed above. When major adverse cardiovascular events prior to PVR were included, the number of events increased as follows: sustained VT (n=6), sustained atrial arrhythmia (n=6) and heart failure requiring hospital admission for intensification of intravenous therapy (n=2). The incidence of lifetime heart failure events trended towards statistical significance in the group of patients with severe fibrosis compared with those without (15% vs 0%, p=0.057).
Our data demonstrate for the first time that an association exists between the presence of RV myocardial fibrosis and extent of right heart reverse remodelling following PVR. This study furthers our understanding about the implications of RV myocardial fibrosis in adult patients with rTOF in several ways.
Surgical factors at the time of initial repair may contribute to the development of diffuse myocardial fibrosis which is evident in adult life;.
The presence of severe RV myocardial fibrosis is associated with increased mass:volume ratio pre-PVR suggesting a connection between excessive RV hypertrophy and fibrosis at the tissue level.
Persistence of relatively increased RV end-systolic volumes, RV mass and RA dimensions post-PVR in patients with severe RV fibrosis supports a potential pathophysiological link between myocardial tissue characteristics and imaging outcomes post-PVR.
Histological quantification of myocardial fibrosis late after TOF repair
To the best of our knowledge, this is the largest study of RV myocardial biopsy specimens obtained from adults undergoing PVR for significant PR along with comprehensive CMR measurements of cardiac size and systolic function. Previously, autopsy studies confirmed the presence of diffuse myocardial fibrosis in patients with ventricular arrhythmia resulting in sudden cardiac death late after TOF repair.4 More recently, Pradegan and colleagues quantified the extent of biventricular fibrosis in an autopsy study of 14 adults with TOF (eight repaired, six unrepaired).2 Notable observations from this study included increased degree of RV fibrosis in unrepaired versus repaired hearts (mean RV anterior wall 16% vs 13%, p=0.04; mean RV inferior wall 18% vs 13%, p=0.03), the presence of fibrosis in the LV and an association between severity of RV fibrosis and age at death, although no association was established with age at repair.
The degree of RV fibrosis in our population (median fibrosis severity 7.1%) was somewhat lower than the report by Pradegan and colleagues at the time of autopsy in keeping with better relative myocardial health in a population of patients undergoing PVR. While we did not directly sample the LV myocardium, the presence of LV myocardial fibrosis is a plausible explanation for the lower LV systolic function that we observed in patients with severe RV fibrosis and may represent a manifestation of ventricular–ventricular interactions at the tissue level, as previously demonstrated (LV and RV were positively correlated in a CMR study of T1 measurements for estimation of diffuse myocardial fibrosis).7 In keeping with prior reports, we did not observe any relationship between age at repair and severity of fibrosis; additionally, we did not see any association between age at death and extent of fibrosis, which might be expected given the fact that mortality was attributed to extracardiac causes in all cases.
Myocardial fibrosis and measurements of right heart size late after TOF repair
An important strength of the present study is the comprehensive approach to CMR measurements before and after PVR. A salient finding in our population is the association between severe fibrosis and elevated RV mass:volume ratio. This observation expands on previous studies which have defined RV hypertrophy as a predictor of adverse outcomes in rTOF. Elevated RV mass:volume ratio was identified as a strong independent predictor of death and sustained VT in the large, multicentre, retrospective International Multicenter TOF Registry (INDICATOR) study.15 Similarly, in a recent prospective CMR study of extracellular volume fraction quantification late after TOF repair, increased indexed RV mass was found to be an independent predictor of major adverse cardiovascular events and was also correlated with RVECV.8 Taken together, a mechanistic link appears to exist between RV fibrosis, RV hypertrophy and adverse clinical outcomes underscoring the deleterious effects of poor myocardial health on clinical status in rTOF.
Following PVR, we found that RV mass was persistently elevated in patients with severe fibrosis. This is a histologically plausible observation as hypertrophied myocardial cells have more potential for reversible change as compared with fibrotic tissue.16 Improvement in volume load in the presence of newly imposed valve competence may not necessarily translate into complete reverse remodelling, a process which may be hampered, at least in part, by extent of collagen deposition at the tissue level. The association between elevated RVESV post-PVR in patients and severe fibrosis further implicates status of the myocardium as an important determinant of adaptive change within the heart. Previous literature has identified RVESV, a marker of both of RV size and systolic function, as a higher fidelity discriminator of persistent RV enlargement and clinical outcomes following PVR as compared with RV end-diastolic volume.17–19 In addition to the prognostic relevance of RV volume enlargement pre-PVR on the macroscopic level, our data suggest that diffuse fibrosis is an additional determinant of post-PVR response on the microscopic level. Finally, we observed that right atrial dimensions were relatively enlarged in those with severe fibrosis and that right atrial dimensions were correlated with extent of fibrosis. These intriguing findings may reflect the presence of diastolic dysfunction, although detailed evaluation of diastolic function was beyond the scope of the present study (invasive haemodynamic measurements were only available for a small subset of patients who were referred for cardiac catheterisation prior to PVR).
Clinical implications of fibrosis in the right heart
Given the limited follow-up time post-PVR (median 2 years) and the relatively low number of major adverse events (nine patients, 17%), we were unable to fully explore the prognostic value of RV myocardial fibrosis on outcomes post-PVR. The incidence of death in our population is somewhat higher than previously reported19 but is driven by three cases where death occurred as a result of extracardiac complications following PVR and may be a reflection of our higher risk, quaternary care practice. Worthy of mention is the trend towards increased number of heart failure events in patients with more severe fibrosis pre-PVR which is in keeping with the presence of decreased LV systolic function in the severe fibrosis group (median LVEF 51%) as compared with the remainder of the population (median LVEF 56%). In terms of evolution of myocardial fibrosis, the finding of severe fibrosis in patients with long aortic cross-clamp time during primary repair points to the vulnerability of the myocardium to irreversible damage during cardiopulmonary bypass and is consistent with previous reports.20–22 Although exposure to chronic hypoxia is known to be associated with severe myocardial fibrosis,23 in our population, collagen volume fraction was not associated with prior aortopulmonary shunt or age at primary repair.
Despite the critical contributions of fibrosis to maladaptive remodelling and poor outcomes in rTOF based on findings from our cohort and from others,7 8 effective antifibrotic pharmacological agents for targeted therapy of the RV are conspicuously lacking. Previously published randomised control trials in rTOF did not show any benefit of ramipril on RV volumes after 6 months of therapy or losartan on RV function after 18–24 months of therapy.24 25 Genetic predisposition to development of fibrosis, although studied in children with rTOF1 has not been elucidated in adults, but may be of clinical relevance. Yet, adults with rTOF have been found to be in a profibrotic state with excessive collagen type 1 synthesis along with dysregulated degradation.26 Further definition of which adults are at risk for fibrosis and what belies individual susceptibility to myocardial disease may allow for enhanced precision in development and implementation of antifibrotic therapies. Additional studies focused on diastolic dysfunction and regional function may shed further light on the clinical implications of myocardial fibrosis.27–29 Finally, unlike the LV,12 13 validation of myocardial extracellular volume estimation using CMR parametric imaging techniques against histological fibrosis measurement has yet to occur in the RV but would likely be of tremendous clinical value.
There are several limitations to our study which are worthy of mention. The duration of follow-up is relatively short and adverse cardiovascular events are small in number, thereby precluding detailed study of the association between fibrosis severity and subsequent development of adverse clinical outcomes. Consistent assessment of diastolic function in our patient population is lacking but would be of interest given the presence of RV hypertrophy and heart failure events in patients with severe RV fibrosis, features previously associated with diastolic dysfunction in other patient populations.30 Furthermore, our study population is reflective of those with severe PR who have been referred for PVR; therefore, myocardial fibrosis in patients with predominant stenosis is likely incompletely understood. Patients with indwelling devices/leads (ie, AICD) were excluded from enrolment and/or from post-PVR CMR study; consequently, our evaluation of fibrosis in some high-risk patients is limited. Furthermore, only a subset completed a post-PVR CMR (71% of patients). Finally, the number of patients with intravenous gadolinium injection for CMR assessment of fibrosis was modest and therefore no conclusions could be reliably drawn between histological and imaging assessments of fibrosis severity.
The presence of RV fibrosis is associated with cardiac reverse remodelling following PVR in adults late after TOF repair. Specifically, RV end-systolic volume, RV mass and RA dimensions were relatively increased in patients with severe RV fibrosis. Further study is required to define the impact of fibrosis and persistent right heart enlargement on clinical outcomes following PVR.
What is already known on this subject?
The association between myocardial fibrosis and major adverse outcomes in patients with repaired tetralogy of Fallot (rTOF) was first established following study of autopsy specimens. However, direct quantification of right ventricular (RV) fibrosis at the tissue level in adults referred for pulmonary valve replacement (PVR) and the implications of fibrosis severity for post-PVR right heart remodelling was previously undefined.
What might this study add?
This study supports the hypothesis that severity of histological RV fibrosis is related to the extent of right heart reverse remodelling following PVR. Specifically, adults with severe fibrosis at the time of PVR were found to have larger RV end-systolic volumes, increased RV mass and larger right atrial dimensions post-PVR as compared with those without severe fibrosis.
How might this impact on clinical practice?
Optimal timing of PVR in patients with rTOF is a critical, but as yet unresolved, question in contemporary congenital heart disease practice. Enhanced risk stratification in rTOF may arise from a deeper understanding of the clinical implications of poor myocardial health at the tissue level and may ultimately allow for refinement of PVR referral strategies in the high-risk rTOF population with chronic pulmonary regurgitation.
Contributors KY and RMW conceived and designed the research, acquired data, performed statistical analysis, drafted the manuscript and critically reviewed the manuscript. DY, EJH, XH and KH acquired data, wrote the first draft and critically reviewed the manuscript. RRC, MKF, LG-W, FB and MEF wrote the first draft and critically reviewed the manuscript. All authors interpreted the results and approved the final version of the manuscript.
Funding Canadian Institutes of Health Research Operating Grant (MOP 119353) to RW.
Competing interests None declared.
Patient consent Obtained.
Ethics approval The study protocol was approved by the institutional research ethics board (study number 12–0242).
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.