Introduction
With advances in cardiovascular magnetic resonance, a detailed physiologic and anatomic assessment of the cardiovascular system can be performed non-invasively in children with single ventricle lesions. Reference Fogel1,Reference Goldstein, Holzer and Trucco2 Many studies have described the utility of cardiac catheterisation and MRI indices in determining ideal Fontan candidacy and guiding postoperative management and outcomes. Reference Fogel, Pawlowski and Whitehead3–Reference Mohammad Nijres, Murphy, Diab, Awad and Abdulla5 Indices assessed include systemic ventricular size and function, atrioventricular valve regurgitation, aortopulmonary collateral burden, lymphatic abnormalities, pulmonary artery diameter, pulmonary vascular resistance, and venous decompression channels. Reference Fogel, Pawlowski and Whitehead3,Reference Grosse-Wortmann, Drolet and Dragulescu6–Reference Ghosh, Griffis and Glatz8 MRI not only provides a three-dimensional intracardiac and extracardiac structural assessment that is important in Fontan reconstruction, but it also provides accurate volumetric and flow assessment to calculate ejection fraction, valvar regurgitation, pulmonary blood flow and permits quantification of aortopulmonary collateral flow. Reference Brown, Gauvreau and Powell9 Cardiac catheterisation provides pressure and saturation data for calculating pulmonary vascular resistance. Reference Rogers, Ratnayaka and Lederman10 Due to the emerging association of lymphatic abnormalities with early post-Fontan outcomes, T2-weighted imaging by CMR has been used to assess the lymphatic system prior to undergoing Fontan palliation. Reference Biko, DeWitt and Pinto11 Hence, CMR and cardiac catheterisation are now used in conjunction for pre-Fontan evaluation at many centres worldwide. Reference Rogers, Ratnayaka and Lederman10
Performing cardiac magnetic resonance followed by cardiac catheterisation in one setting under the same anaesthesia provides valuable MRI information prior to undergoing catheterisation, allowing for a focused hemodynamic and angiographic assessment, eliminating the need for multiple anaesthesia episodes. Reference Rogers, Ratnayaka and Khan12 Additionally, performing hemodynamic and volumetric data at similar timepoints and loading conditions allows accurate assessment for children prior to Fontan palliation. Reference Arar, Hussain and Abou Zahr13 MRI-derived flow measurements have shown higher test–retest reliability and allow reproducible hemodynamic assessments in CHD with multiple cardiac and extracardiac shunts. Reference Arar, Hussain and Abou Zahr13 Additionally, the use of invasive cardiovascular magnetic resonance wherein cardiac catheterisation is performed in the MRI scanner has been described in children with CHD for essential structural, functional, and hemodynamic assessment. Reference Rogers, Ratnayaka and Lederman10 At our centre, we mostly perform cardiac MRIs and cardiac catheterizations in succession under the same anaesthesia, and it has become increasingly common for the cardiac interventionalist to perform invasive cardiovascular magnetic resonance wherein diagnostic right and left heart catheterisation is performed in the MRI scanner. Reference Arar, Hussain and Abou Zahr13 Invasive cardiac magnetic resonance is advantageous in avoiding ionising radiation and reduces the total amount of anaesthesia exposure. Reference Rogers, Ratnayaka and Lederman10 If these patients require intervention, they are subsequently transported to the fluoroscopy-based laboratory where angiography and interventions are performed under the same anaesthesia. Reference Rogers, Ratnayaka and Khan12
Despite these comprehensive evaluations, a select group of patients is unable to tolerate Fontan circulation and develop unfavourable outcomes soon after surgery. In this study, we describe our centre’s MRI and cardiac catheterisation variables obtained during pre-Fontan evaluation and analyse the association with early adverse outcomes following Fontan palliation.
Materials and method
Data for all cardiac MRIs performed after the superior cavopulmonary anastomosis at Children’s Medical Center, Dallas, between April 2017 and November 2022 were first extracted from our Circle cvi42 MRI database. Cardiac catheterisation data from these patients were then obtained from electronic medical records. Patients who did not undergo Fontan palliation were excluded from the analysis (Figure 1). For patients who had received multiple cardiac MRIs, only data from the most recent MRI obtained before the Fontan operation were recorded. Sequences of CMR included (i) balanced 2D Steady state free precision cine imaging for functional analyses, (ii) phase contrast imaging to estimate pulmonary and systemic flow (Qp:Qs), (iii) T2-weighted turbo-spin echo sequence to assess the lymphatic system, (iv) 3D-SSFP, a non-contrast method for angiography, and (v) 3D-black blood VISTA (Volume Isotropic Turbo Spin Acquisition), a non-contrast method to show dark blood angiography with bright vessel walls. Cardiac catheterisation was performed in either the fluoroscopy-based laboratory or the MRI suite via invasive cardiac magnetic resonance. During the catheterisation, patients were placed under general anaesthesia with oral endotracheal intubation, and hemodynamics were collected under 21% FiO2. Data were then collected and analysed in a retrospective fashion under four categories: demographic, MRI, catheterisation, and pre- and post-operative variables. Demographic data included age, weight, gender, diagnosis of heterotaxy syndrome, dominant morphologic ventricle, and Fontan type. MRI data included lymphatic classification into grades 1–4 as described by Biko et al., end-diastolic volume, end-systolic volume, ejection fraction (%), systemic AV valve regurgitation fraction (calculated by
${{stroke\,volume}-{total\,antegrade\,flow} \over {stroke\,volume}}$
), collateral flow (estimated by aortic flow – [superior caval vein flow + descending aorta flow] or aortic flow – [superior caval vein flow + inferior caval vein flow]), right and left pulmonary artery measurements, aortic arch patency and coarctation index, and differential pulmonary artery blood flow percentage.
Reference Biko, DeWitt and Pinto11
Effective pulmonary blood flow (Qep) was calculated by excluding the aorto-pulmonary collateral flow whereas total pulmonary blood flow (Qp) included the aortopulmonary collateral flow. Systemic blood flow (Qs) was obtained by assessing net antegrade flow across the aortic valve, and effective systemic blood flow (Qes) was calculated by subtracting the aorto-pulmonary collateral flow from the aortic outflow. Subsequently, the Qep:Qes and Qp:Qes ratios were obtained. Catheterisation data included Qp:Qs, pulmonary vascular resistance, transpulmonary gradient, systemic ventricle end-diastolic pressure, left atrial pressure, Glenn/branch pulmonary artery patency, collaterals, and interventions. Pulmonary vascular resistance was calculated using
$PVR=\,{TPG \over Qp}$
, and Qp was calculated using the Fick equation such that
$Qp={VO2 \over Hgb*13.6*\left({pulmonary\,vein\,saturation}-{pulmonary\,artery\,saturation}\right)}$
. Pulmonary vascular resistance was also calculated using combined MR and catheter data such that
$PVR\,=\,{{TPG\,from\,catheter} \over {\,Qep\,from\,MRI}}$
[excluding aortopulmonary collateral flow]. In most cases, VO2 was measured. In cases where the transpulmonary gradient was significantly discrepant between the left and right pulmonary arteries, the “worst case scenario” was likely reported. Peri/post-operative data included type of Fontan, conduit, duration of cardiopulmonary bypass, duration of aortic cross clamp, cardiac arrest, extracorporeal membrane oxygenation support, ventricular assist device implantation, transplant, Fontan takedown, duration of intubation, duration of ICU stay, duration of hospitalisation, presence and duration of chylous effusion, plastic bronchitis, survival to discharge, and date of death. Major adverse cardiac events as well as a few markers for adverse outcomes were grouped together such that composite adverse outcomes were defined as either takedown of Fontan circulation, mechanical circulatory support, heart transplantation, or death. MRI and hemodynamic parameters were compared between groups with adverse and favourable outcomes. Statistical analysis included the Mann–Whitney U Test for non-parametric data, Student’s T-test for parametric continuous variables, and Fisher’s Exact Test for categorical variables.
Flow chart illustrating outcomes of patients who underwent pre-Fontan assessments.
Results
Of the 138 initial MRIs, 19 were excluded due to the repetition of the MRI or the Fontan operation done prior to full data acquisition. Of the total 138 cases, 54 were performed by an invasive cardiac magnetic resonance approach wherein cardiac catheterisation was performed under MRI guidance, 62 evaluations were performed by the combined approach (MRI followed by cardiac catheterisation on the same day under the same anaesthesia), and in 22 cases, the catheterizations and MRIs were performed on separate occasions. Eight were then excluded due to the Fontan operation being deferred or scheduled for a later date. Of the 111 remaining, 30 were then excluded as they did not undergo Fontan at our institution for a variety of reasons. Of these 30 patients, 12 were deemed to have prohibitive risk and were undergoing transplant evaluation, were listed for transplant, or had already undergone transplant. Reasons for avoiding the Fontan pathway included Glenn/pulmonary artery obstruction and elevated Glenn/pulmonary artery pressures, elevated pulmonary vascular resistance, poor systolic or diastolic function, elevated atrioventricular valve regurgitation, high-risk Dori lymphatic classification, pulmonary vein abnormality, and/or reasons unspecified. 4 of the 30 died before they were able to get their Fontan surgery, and the final 14 did not obtain Fontan surgeries as they were awaiting formal evaluation at the time of this article, had their operation done at an outside institution, or planned to undergo biventricular repair. Thus, 81 patients with pre-operative MRI and catheterisation data underwent subsequent Fontan palliation and were included in the study. The outcomes of these Fontan procedures are outlined in Figure 1. Of these 81 patients with a median (interquartile range) age of 4.1 (3.4–6.1), the composite adverse outcome was recorded in 8 (10%): Fontan takedown in 1 (12.5%), heart transplant in 1 (12.5%), mechanical circulatory support in 1 (12.5%), and death in 5 (62.5%) during a median (interquartile range) follow-up of 21 (3–32) months. In all 5 deaths, each appeared to be cardiac-related though autopsies were declined. Pre-Fontan demographic, MRI, and catheter data were then compared between the group with composite adverse outcomes and the group without composite adverse outcomes. The comparison is summarised in Table 1. Of the demographic data collected, there was a significant association between the composite adverse outcomes group and the diagnosis of heterotaxy syndrome (p = 0.04). Of the MRI data collected, there was a significant association between the composite adverse outcomes group and moderate-severe atrioventricular valve regurgitation defined as ≥30% (p = 0.046). Of the catheter data, there was a significant association between elevated indexed pulmonary vascular resistance (median 2.93 versus 2, p = 0.03) and composite adverse outcomes when using indexed pulmonary vascular resistance calculated with transpulmonary gradient from the catheterisation and effective pulmonary blood flow (Qep) from MRI. Notably, there was no significant association found between elevated pulmonary vascular resistance and composite adverse outcome when calculating pulmonary vascular resistance using transpulmonary gradient and pulmonary blood flow from catheterisation data alone. There was no significant association between lymphatic classification, ejection fraction, or end-diastolic volume and adverse outcome.
Demographic, clinical, MRI, and catheterisation variables for patients with composite adverse versus favourable outcomes
Discussion
Nuance of heterotaxy
In our study, three of the eight patients (38%) who underwent Fontan palliation and had a composite adverse outcome had the diagnosis of heterotaxy syndrome. Six of the 73 patients (8.2%) who underwent Fontan palliation and did not have an adverse outcome had the diagnosis of heterotaxy syndrome. There are many potential reasons for this association. A study published by Hsin-Chia Lin et al posits that while the patients with heterotaxy syndrome in their cohort had high peri-operative mortality, the result may have been confounded by high preoperative pulmonary artery pressure, fewer extracardiac conduits used, and more severe atrioventricular valve regurgitation in the heterotaxy group. Reference Lin, Wu and Wang15 In their statistical analysis, they adjusted for confounders and found that the association between heterotaxy syndrome and peri-operative outcomes dissolved. Reference Lin, Wu and Wang15 Thus, in our study, the diagnosis of heterotaxy may be further risk stratified, according to the factors such as hemodynamics, to obtain a more precise risk assessment. Marathe et al in their multicenter study, concluded that the pre-operative and immediate post-operative course for patients with heterotaxy syndrome may be more complicated, but there was no difference in 15-year freedom from Fontan failure for the heterotaxy versus non-heterotaxy group, implying that long-term outcomes are not necessarily worse for heterotaxy patients if peri-operative and early post-operative courses are survived. Reference Marathe, Cao and Celermajer16 This same study, along with other studies before it, reiterates that the diagnosis of heterotaxy places patients at higher risk of arrhythmias, contributing to increased post-operative mortality. Reference Marathe, Cao and Celermajer16 –Reference Niu, Dickerson and Moore18 Son Q. Duong et al. consider the extracardiac manifestations that commonly affect patients with heterotaxy, such as respiratory and renal comorbidities, that may also contribute to poor outcomes. Reference Duong, Zaniletti, Lopez, Sutherland, Shin and Collins19 Within the diagnosis of heterotaxy, there are multiple factors that contribute to increased morbidity and mortality, and further delineation of these factors may provide a more nuanced understanding of the risk associated with the diagnosis.
Pulmonary blood flow (Qp), effective pulmonary blood flow (Qep), and pulmonary vascular resistance (PVR)
A higher pre-Fontan pulmonary vascular resistance was also significantly associated with composite adverse outcomes in our study: the mean indexed pulmonary vascular resistance of the adverse outcome group was 2.93 versus 2 Woods Um2 in the group without adverse outcomes (p = 0.03). Once total cavopulmonary connection is created in the Fontan procedure, pulmonary blood flow then becomes a function of central venous pressure and pulmonary vascular resistance. Reference Becker, Uebing and Hansen20 Thus, with increased pulmonary vascular resistance, pulmonary blood flow decreases, ultimately decreasing cardiac output. Reference Gewillig and Brown21 Additionally, venous congestion increases with increased pulmonary vascular resistance. Reference Gewillig and Brown21 Over time, pulmonary vascular resistance inevitably increases, further exacerbating issues caused by diminished cardiac output and venous congestion. Reference Gewillig and Brown21
An important distinction to make is the calculation of pulmonary vascular resistance in this study. Indexed pulmonary vascular resistance was calculated via two methods. The first method used transpulmonary gradient (mmHg) and pulmonary blood flow (Qp, L/min/m2) from catheterisation such that
$PVRi ={TPG \over Qp}$
. The second method used transpulmonary gradient (mmHg) from catheterisation and effective pulmonary blood flow (Qep, L/min/m2) from cardiac MRI such that
$PVRi = {TPG \over Qep}$
. A significant association with indexed pulmonary vascular resistance and composite adverse outcome was found when using pulmonary vascular resistance derived from combined MRI and catheterisation data rather than catheterisation data alone. We postulate that effective pulmonary blood flow (Qep) from cardiac MRI may be more accurate in describing pulmonary blood flow as it measures the net antegrade forward flow across the branch pulmonary arteries and excludes the aortopulmonary collateral flow. Hence, we found that higher pulmonary vascular resistance calculated using Qep was associated with adverse post-Fontan outcomes. This differs from prior studies, such as studies published by Downing et al and Hart et al, which found in pre-Fontan patients, MRI measured higher Qp compared to cardiac catheterisation measurements of Qp using the Fick principle.
Reference Downing, Whitehead and Dori14,Reference Hart, Whiteside and Yu22
However, these findings were explained by inclusion of collaterals that led to systematically higher Qp measurements by MRI versus catheterisation in single ventricle patients. Our use of Qep excludes collateral flow and does not rely on the oxygen saturations of the branch pulmonary arteries, which are unpredictable in the setting of these collaterals and dependent on sampling of streaming flow. Greater aortopulmonary collateral flow has been associated with adverse early post-operative outcomes, and inclusion of this measured flow, which renders pulmonary vascular resistance calculations to be less accurate, may be falsely reassuring.
Reference Grosse-Wortmann, Drolet and Dragulescu6
It may be that in these cases, Qp was overestimated by catheterisation in comparison to Qep by MRI, thereby underestimating pulmonary vascular resistance. Thus, the inclusion of MRI along with cardiac catheterisation for the pre-Fontan evaluation adds valuable information when assessing crucial variables for the selection of ideal Fontan patients.
Lymphatics and end-diastolic volume
Biko DM et al. demonstrated that patients with type 4 lymphatic abnormalities have significantly higher rates of Fontan failure, need for transplant, and postoperative hospital stay than patients with less severe lymphatic abnormalities. Reference Biko, DeWitt and Pinto11 In this study, there was no significant association with type 4 lymphatics and composite adverse events (p = 0.1). This discrepancy is likely explained by inadequate power, and thus a tendency towards type II error, to comment on associations with lymphatic abnormalities in this study. Only three patients examined in this study were categorised as having type 4 lymphatic abnormalities, and only one of these three patients underwent Fontan palliation. This patient had an adverse outcome. The other two patients with type 4 lymphatic abnormalities were considered to be too high-risk to undergo the Fontan procedure. Similarly, ventricular indexed end-diastolic volume has been previously shown to be a predictor of death or transplant late after the Fontan operation. Reference Rathod, Prakash and Kim23 The lack of significant association between end-diastolic volume and composite adverse outcome (p = 0.22) in this study again may be due to lack of power. Additionally, a shorter median follow-up time in this study may also have contributed to this discrepancy.
Conclusions, limitations, and future directions
Based on our findings, we conclude that diagnosis of heterotaxy syndrome (p = 0.04), higher pre-Fontan indexed pulmonary vascular resistance (p = 0.03), and moderate-severe (≥30%) atrioventricular valve regurgitation (p = 0.046) are significantly associated with adverse outcomes. This study is limited by the fact that it is a single-centre study that examined patients who underwent Fontan operations over a 5-year period. This may explain its inability to demonstrate significant associations between adverse events and factors previously established to predict poor outcomes, such as severe lymphatic abnormalities and end-diastolic volume. Additionally, as a retrospective study, data were gathered in a clinical context, resulting in a high risk of unmeasured confounding. The fact that our analysis only included patients who were discussed and determined to be of lower risk also introduces bias into our results. Future directions include analysing similar variables in a multi-centre study with inclusion of multivariate analysis for further risk stratification, accounting of confounders, and determination of hazard ratio. Additionally, secondary outcomes, such as hospital length of stay and chest tube duration, should be added in future iterations of this study.
The positive findings highlight the usefulness of cardiac MRI in calculating effective pulmonary blood flow and pulmonary vascular resistance to aid in predicting outcome.
Acknowledgements
None.
Financial support
This research received no specific grant from any funding agency, commercial, or not-for-profit sectors.
Competing interests
None.
Ethical standard
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national guidelines on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008, and have been approved by the institutional committees: IRB STU 032016-009.
