Abstract
Background:
Atrial right to left (aRL) shunting is often identified on echocardiograms in the early postoperative period following repair of tetralogy of Fallot (TOF) and thought to reflect poor right ventricular (RV) compliance, but to be possibly beneficial in serving as a “pop off” for the RV. We aimed to investigate the relationship between aRL shunting to echocardiographic diastolic function and early postoperative outcomes, hypothesizing that aRL would be associated with worse diastolic function, and with post-operative length of stay (LOS).
Methods:
Single center cohort study of patients who underwent repair of TOF. Echocardiograms were obtained 2–5 days after repair. Patients were grouped as “elective” if repaired after 30 days of age without prior palliation, “staged” if they had a neonatal palliation prior to repair, or as “neonatal” repair if repaired <30 days age. aRL shunting was compared to all others: bidirectional, left to right, and no atrial shunt detected. Linear regression tested the relationship of aRL with right atrial volumes and right atrial emptying fraction (RAEF), RV inflow/tissue Doppler velocities, and RA peak longitudinal strain and early strain rate. Multivariable negative binomial regression tested the association between aRL with LOS, stratified by repair group.
Results:
There were 197 TOF patients (60% male, 74% White), most (127, 64%) had elective, 41 (21%) staged, and 29 (15%) neonatal repair. aRL was present in 68 patients (35%). In the overall cohort, aRL shunting was associated with lower RA end diastolic volume, higher RAEF, higher A wave peak velocity, and higher RA peak longitudinal strain. In the subgroup analysis, aRL was associated with higher RAEF and peak longitudinal strain in the elective repair group only, where aRL was also associated with shorter LOS.
Conclusions:
aRL after TOF repair is associated with better atrial function, and possibly with a combination of robust atrial function in the presence of RV noncompliance, and shorter LOS in patients undergoing elective rTOF, but not in those undergoing a neonatal intervention.
Keywords: atrial strain, diastolic function, tetralogy of Fallot, speckle tracking echocardiography, atrial level shunt
Introduction:
Tetralogy of Fallot (TOF) is the most common cyanotic congenital heart defect, occurring in approximately 1 in 3,500 births and accounting for almost 10% of all congenital cardiac malformations [1]. Surgical repair is most often electively performed at 3–6 months of age, however some patients undergo a neonatal intervention either via complete surgical repair or palliative procedure (i.e., right ventricular outflow tract stent, patent ductus arteriosus stent, Blalock-Taussig-Thomas Shunt) to augment pulmonary blood flow [2, 3]. Despite excellent operative results, with greater than 95% survival, patients with repaired TOF are at risk for postoperative ventricular diastolic dysfunction [4–6]. It is important to identify and treat diastolic dysfunction promptly, as it can result in a low cardiac output state, and significantly impact postoperative clinical course and risk for future interventions, even late into adulthood [4, 5, 7, 8]
Right to left shunting (aRL) at the atrial level results from an end-diastolic pressure gradient and is commonly interpreted as evidence of increased RV end-diastolic pressure, thus serving as a measure of diastolic function (i.e., RV non-compliance) in patients with TOF early after repair. Clinicians use aRL shunting intuitively to help guide patient management and interpret oxygen saturation trends. A common surgical management strategy is to leave the foramen ovale patent as a “pop-off” valve, intended to assist patients through the postoperative period. The theory behind this strategy being that, in the noncompliant RV, it may be beneficial to allow for aRL shunting and therefore preserving cardiac output at the expense of systemic oxygen saturation in the event of significant RV noncompliance. However, the relationship of aRL shunting at the atrial level to other echocardiographic indicators of ventricular and atrial diastolic dysfunction and its association with postoperative outcomes in this population remains understudied.
Our understanding and assessment of diastolic function has evolved significantly with the use of speckle tracking echocardiography (STE), and STE-derived strain metrics of the right atrium may prove useful in echocardiographic assessment of repaired TOF[9]. Atrial function is important to consider in the assessment of diastolic function because it both directly augments ventricular filling and it indirectly reflects the diastolic properties of the ventricle [9]. Atrial function is characterized by three phases: a reservoir phase during atrial filling (ventricular systole), a conduit phase during passive emptying into the right ventricle, and an active contractile phase during atrial systole [10–12]. STE allows for evaluation of RA deformation throughout these phases [13]. Our group has previously demonstrated a significant reduction in RA (and RV) strain in the early postoperative period after repair of TOF compared to paired preoperative evaluations [9]. These abnormalities are also important to identify early given that they can persist beyond the postoperative period even into adulthood [4, 11, 14]. To date, few studies examined RA function in the early postoperative period after repair and no studies investigated the association of RA strain parameters with early clinical outcomes, particularly considering the variation in surgical/treatment repair strategies.
In this study we sought to evaluate the association of early post-operative aRL shunting with echocardiographic indices of diastolic function and early clinical outcomes in neonates and infants undergoing repair of TOF. We hypothesized that aRL shunting in the early postoperative period would be associated with impaired diastolic function and shorter hospital length of stay (LOS).
Material and methods:
Study design
We conducted a single-center retrospective analysis of a prospectively recruited cohort of patients younger than 2 years of age undergoing surgical repair for TOF between April 5, 2012 and May 12, 2023. Patients had postoperative research protocol echocardiograms obtained 2–5 days after surgery, prior to hospital discharge. The study cohort was previously published [15]. Patients with TOF and pulmonary stenosis, pulmonary atresia, and absent pulmonary valve were included. Patients with discontinuous pulmonary arteries and multiple aorto-pulmonary collateral vessels supplying pulmonary blood flow were excluded due to the heterogeneous phenotype and outcome associated with these anatomic sub-types.
Due to possible differences in RV compliance related to age at repair and TOF phenotype, the cohort was subdivided based on repair strategy: elective (i.e., elective, or non-neonatal repair) (≥ 30 days of age at surgery, no prior palliation), staged repair (≥30 days of age with prior neonatal palliation), and neonatal repair of TOF. Atrial level shunt was interrogated on subcostal windows with color doppler with low color scale and with pulse waved spectral doppler confirmation. Shunting direction was categorized as exclusive or pure aRL shunting, bidirectional, exclusive left to right, or no atrial shunting present. We compared aRL shunting to a composite group of all others and conducted subgroup analysis by repair strategy. For all comparisons to the exposure of interest (pure aRL shunting), those with bidirectional, left to right, or no atrial level shunting were grouped together given the similar theoretical lack of the “pop-off” effect hypothesized in this study. Given the prospective and observational study nature of this study, and the goal to assess aRL shunting in relation to hospital length of stay with adequately powered sample size, patients with no evidence of atrial-level shunting in the postoperative echocardiogram were included in the analysis, and we performed sensitivity analysis in the group limited to patients with evidence of atrial level shunting. In our institution, the decision to close or create an atrial communication during repair is at the discretion of the surgeon, and usually based upon degree of RV hypertrophy and outflow tract obstruction. In most cases a PFO is present and left open, or it is partially closed if it is an atrial septal defect to avoid leaving a hemodynamically significant residual lesion.
The study protocol was approved by the Institutional Review Board for the Protection of Human Subjects at Children’s Hospital of Philadelphia, and parents of subjects gave informed consent to participate in the study.
Echocardiographic and clinical outcomes:
Echocardiographic parameters included outcome variables of diastolic function, as well as parameters that were deemed as potential confounders or effect modifiers of the association between aRL shunting and outcomes, including: significant tricuspid regurgitation (defined as greater than mild), early postoperative pulmonary insufficiency (PI), graded as moderate or greater if flow reversal was present in the branch pulmonary arteries, and left ventricular (LV) function as expressed by shortening fraction (LVSF; M-mode). Echocardiographic outcome variables included: RA volumetric parameters obtained from tracing of the RA endocardium included: end systolic and end diastolic volume (ESV and EDV, respectively), and RA emptying fraction (RAEF) % calculated as (EDV-ESV)/EDV. Inflow spectral Doppler and tissue Doppler parameters included tricuspid valve early inflow velocity (E), late inflow velocity (A), tissue Doppler lateral RV free wall peak early velocity (e’), and the calculated E/e’ ratio. Speckle tracking echocardiography was used to obtain strain values including RA peak longitudinal strain, RA early strain rate, RV global longitudinal strain (RVGLS), and peak RV longitudinal strain rate. RV diastolic strain rate was not included in analysis owing to inconsistent data availability, with early and late diastolic strain curves often fused in this population.
RA peak longitudinal strain was obtained and measured offline per standard protocols previously used by our group [9] and using iE33 (Philips Medical Systems, Andover, MA) and Epiq (Philips Medical Systems, Andover, MA) machines and 2D speckle tracking TomTec software (Cardiac Performance Analysis software; TomTec Imaging, Unterschleissheim, Germany). Clinical data were obtained from review of electronic medical records and included relevant demographic and genetic data, operative characteristics, and disposition status including the study outcome total hospital length of stay (LOS), defined as date of TOF repair to date of hospital discharge. Documentation from postoperative cardiac catheterization was reviewed in this study, but too few patients had complete invasive hemodynamic data for inclusion in analysis.
Descriptive Statistics
Descriptive statistics are reported as n (%) for categorical and median (IQR) for continuous variables. Differences in characteristics between shunting groups were compared using the Wilcoxon rank sum or Kruskal-Wallis rank sum tests for continuous variables, and Chi-squared or Fisher’s exact test (for tabulations with small cell sizes < 5) for categorical variables.
Primary Outcome analyses
Linear regression and generalized linear regression models, unadjusted and adjusted for confounders, including presence of 22q11.2 deletion syndrome, were used to test the relationship between aRL shunting and each outcome. For the subgroup analysis of echocardiographic outcomes by repair strategy (elective, staged, or neonatal), we used linear regression to test the relationship of aRL shunting to mean difference in right atrial emptying fraction (RAEF) and RA peak longitudinal strain. For the analyses of hospital LOS, multivariable negative binomial regression was used to obtain the incidence rate ratio (IRR) for the effect of aRL shunting on LOS, among the 193 subjects who survived to hospital discharge. We conducted the LOS model using the SAS %margins macro, with the %RunBY macro used to process by imputation, to calculate post-estimation predictive mean values of LOS [16, 17].
We used the same set of adjustment characteristics in all multivariable models run on the full sample. Analyses for the full cohort were adjusted for gestational age, preoperative discharge-status (i.e., discharged home after birth prior to admission for TOF repair), preoperative oxygen saturation, operative weight, cardiopulmonary bypass (CBP) time, surgical repair access (transatrial-transpulmonary including those that required extension of a transannular patch into the RV outflow tract, or ventriculotomy approach), and delayed sternal closure, which is not standard of care at our institution. Analyses run on subgroups were adjusted for operative weight and CPB time.
Missing data was minimal and believed to be missing at random, thus multiple imputation was performed for missing adjustment variables (Supplemental Table 1) using SAS PROC MI. Analyses were performed on all imputed datasets and combined with SAS PROC MIANALYZE. Linear regression with variance inflation factors (VIF) was used to check for collinearity between covariates included in the model. All VIF values observed were less than 2, indicating little concern for collinearity. Due to the exploratory nature of this study, no nonlinear associations were explored, and no adjustments were made for multiple hypothesis testing. An α of < 0.05 was considered significant. All analyses were performed using SAS Enterprise Guide 8.3 (SAS Institute, Cary, NC), and R software (R Studio 2023.06.1 Build 524, Posit Software).
Results:
Clinical and Operative Characteristics:
In total, 197 subjects (60% male, 74% White race) who underwent repair of TOF were included for analysis. Preoperative pulmonary valve anatomy included 82% of patients with stenosis, 15% with pulmonary atresia, and 3% with absent pulmonary valve. Most patients were non-syndromic (76%) and 6.3% had a 22q11.2 deletion. Surgical access for closure of the ventricular septal defect and relief of RV outflow tract obstruction included 28% atriotomy only, 14% ventriculotomy only, and 57% both atriotomy and ventriculotomy. Transannular patch augmentation of the pulmonary valve was performed in 74% of cases.
Demographic characteristics, TOF anatomy, co-morbidities, surgical approach, and mortality were similar between aRL shunting vs. non-aRL shunting groups (Table 1), apart from a later median gestational age in the RL group by one week. Most patients (127, 64%) had elective complete TOF surgical repair, while 41 patients (21%) had staged, and 29 patients (15%) had neonatal TOF repair. Staged palliation included various strategies to augment pulmonary blood flow, including patent ductus arteriosus (PDA) stent in 11/41 (27%), RV outflow tract (RVOT) stent in 8/41 (20%), Blalock-Taussig-Thomas (BTT) shunt in 21/41 (51%). Of 197 patients, a postoperative atrial communication was detected in 85%. By surgical operative report, a pre-existing atrial communication was created in 14.4% (N=19), left open in 72.0% (N=95), and closed (including partial/complete) in 13.6% (N=18). On postoperative echocardiogram, aRL shunting (exposure of interest) was present in 68 patients (35%), left to right in 9%, and bidirectional in 42%. There was not a significant presence of arrhythmia requiring treatment in the study cohort (11/197 patients, 6%). There were minimal patients in the cohort who were on positive pressure ventilation at the time of the postoperative echocardiogram (non-invasive: 3.8%, N=7; invasive: 10.2%, N=19).
Table 1: Baseline cohort characteristics and distribution by atrial level shunting.
Clinical and operative characteristics seen in the overall cohort, as well as by atrial-level shunting (aRL vs all others).
| Direction of Shunting | |||||
|---|---|---|---|---|---|
| Characteristic | N | Total N = 1971 | R to L N = 681 | All others N = 1291 | p-value2 |
| Sex | 196 | 0.2 | |||
| Male | 118 (60%) | 37 (54%) | 81 (63%) | ||
| Female | 78 (40%) | 31 (46%) | 47 (37%) | ||
| White Race | 196 | 145 (74%) | 48 (72%) | 97 (75%) | 0.6 |
| Weight (kg): | 197 | 5.45 (4.30, 6.23) | 5.61 (4.32, 6.21) | 5.44 (4.30, 6.23) | 0.6 |
| Gestational Age at birth (weeks): | 195 | 38.3 (37.0, 39.1) | 39.0 (37.6, 39.4) | 38.0 (36.4, 39.0) | 0.016 |
| O2 saturation before surgery (%): | 192 | 93 (86, 97) | 93 (88, 98) | 92 (85, 97) | 0.3 |
| Air Type | 196 | 0.11 | |||
| Room Air | 168 (86%) | 62 (91%) | 106 (83%) | ||
| On O2 | 28 (14%) | 6 (8.8%) | 22 (17%) | ||
| Genetic Syndrome | 197 | 0.4 | |||
| Absent/Unknown | 149 (76%) | 54 (79%) | 95 (74%) | ||
| Genetic Syndrome | 48 (24%) | 14 (21%) | 34 (26%) | ||
| 22q11.2 Deletion Status | 175 | 0.7 | |||
| Deleted | 11 (6.3%) | 3 (4.9%) | 8 (7.0%) | ||
| Not Deleted | 164 (94%) | 58 (95%) | 106 (93%) | ||
| Pulmonary Valve Anatomy | 196 | 0.5 | |||
| Stenosis | 161 (82%) | 54 (79%) | 107 (84%) | ||
| Atresia | 30 (15%) | 13 (19%) | 17 (13%) | ||
| Absent | 5 (2.6%) | 1 (1.5%) | 4 (3.1%) | ||
| Surgical Characteristics | |||||
| Ventriculotomy | 193 | 138 (72%) | 48 (72%) | 90 (71%) | >0.9 |
| Transannular patch | 197 | 146 (74%) | 51 (75%) | 95 (74%) | 0.8 |
| Total CPB time (min): | 197 | 67 (47, 93) | 66 (42, 93) | 67 (48, 93) | 0.6 |
| Total aortic cross-clamp (min): | 197 | 50 (28, 68) | 47 (27, 62) | 50 (29, 69) | 0.3 |
| Residual VSD (any size) | 197 | 71 (36%) | 21 (31%) | 50 (39%) | 0.3 |
| Delayed Sternal Closure | 197 | 12 (6.1%) | 4 (5.9%) | 8 (6.2%) | >0.9 |
| Preoperative discharge to home | 197 | 0.8 | |||
| No | 61 (31%) | 21 (31%) | 40 (31%) | ||
| Yes | 134 (68%) | 47 (69%) | 87 (67%) | ||
| Unknown | 2 (1.0%) | 0 (0%) | 2 (1.6%) | ||
| Preoperative Readmission | 195 | 33 (17%) | 9 (13%) | 24 (19%) | 0.3 |
| Postoperative Disposition | 197 | 0.9 | |||
| Death | 4 (2.0%) | 1 (1.5%) | 3 (2.3%) | ||
| Home | 188 (95%) | 66 (97%) | 122 (95%) | ||
| Other hospital/facility | 5 (2.5%) | 1 (1.5%) | 4 (3.1%) | ||
| Transplant | 0 (0%) | 0 (0%) | 0 (0%) | ||
n (%); Median (IQR)
Fisher’s exact test; Pearson’s Chi-squared test; Wilcoxon rank sum test
Echocardiographic parameters of diastolic function, irrespective of the atrial level shunting, were similarly distributed among the repair subgroups, with the exception of a lower RA peak longitudinal strain in the elective repair group compared to all other repair strategies (Supplemental Table 2). Hemodynamic parameters relevant to diastolic processes and atrial-level shunting directionality, including pulmonary insufficiency, tricuspid regurgitation, end-diastolic anterograde pulmonary blood flow, and left ventricular dysfunction, were analyzed for a relationship to aRL shunting in all patients inclusive of all treatment strategies (Table 2A). There was no relationship of aRL shunting with early postoperative pulmonary insufficiency. The presence of significant tricuspid regurgitation (moderate or greater) was also not associated with aRL shunting. There was a trend towards a lesser proportion of patients with end-diastolic forward flow in the group of patients with right to left atrial level shunting as compared to all others, but this was a small effect size and was not statistically significant (p=0.0077, Table 2A). Diminished LV function was not prevalent in the cohort and the median LV shortening fraction was normal in both comparison groups, albeit minimally higher in those with aRL shunting (median 36.5% compared to 35.0%).
Table 2: Echocardiographic parameters associated with aRL Shunt.
Echocardiographic hemodynamic factors relevant to diastole and atrial-level shunting (A), along with diastolic parameters by 2-dimensional, Doppler (B), and speckle-tracking (C) echocardiography shows pertinent associations to aRL shunting.
| Echocardiography Characteristic | Variable Distribution1 | Statistical Analysis2 | ||||
|---|---|---|---|---|---|---|
| aRL Shunting | All Others | X2 | P-value | |||
| A | > Mild TR | 46/68 (67.7%) | 90/128 (70.3%) | 0.149 | 0.700 | |
| > Mild Pulmonary Insufficiency | 24/73 (32.9%) | 40/131 (30.5%) | 0.120 | 0.729 | ||
| End-Diastolic Anterograde Pulmonary Flow | 29/57 (50.9%) | 70/119 (58.8%) | 5.129 | 0.077 | ||
| Wilcoxon rank sum | P-value | |||||
| LV Shortening Fraction (%) | 36.5 (33.0–40.5) | 35.0 (30.0–38.0%) | Z = 2.17 | 0.032 | * | |
| aRL Shunting | All Others | Estimated Effect (β) | P-value | |||
| B | RA Emptying Fraction (%) | 40 (26, 49) | 34 (27, 40) | 5.51 (1.46, 9.55) | 0.008 | * |
| RA End Diastolic Volume (mL) | 2.80 (1.76, 4.19) | 3.55 (2.26, 4.70) | −0.58 (−1.15, −0.01) | 0.047 | * | |
| RA End Systolic Volume (mL) | 4.59 (2.89, 6.77) | 5.39 (3.53, 6.91) | −0.64 (−1.44, 0.16) | 0.119 | ||
| Tricuspid E Velocity (m/s) | 0.78 (0.67, 0.98) | 0.79 (0.63, 0.92) | −0.06 (−0.04,0.16) | 0.256 | ||
| Tricuspid A Velocity (m/s) | 0.69 (0.59, 0.88) | 0.65 (0.49, 0.83) | 0.12 (0.02, 0.21) | 0.015 | * | |
| E/A Ratio | 1.09 (0.87, 1.22) | 1.19 (1.01, 1.35) | −0.2 (−0.44, 0.03) | 0.094 | ||
| RV Lateral e’ Velocity (cm/s) | 5.00 (3.39, 6.50) | 5.00 (4.00, 6.79) | −0.31 (−1.07, 0.44) | 0.418 | ||
| E/e’ Ratio | 17 (12, 25) | 14 (11, 22) | 4.15 (−0.22, 8.53) | 0.063 | ||
| C | RA Peak Longitudinal Strain | 22 (15, 30) | 19 (16, 27) | 4.21 (0.76, 7.67) | 0.017 | * |
| RA Early Strain Rate | −1.35 (−1.82, −0.87) | −1.20 (−1.55, −0.91) | 0.02 (−0.26, 0.30) | 0.890 | ||
| RV Global Longitudinal Strain | −13.4 (−16.7, −11.1) | −14.8 (−16.8, −9.6) | 0.23 (−1.64, 2.09) | 0.811 | ||
| RV Longitudinal Strain Rate | −1.20 (−1.40, −1.00) | −1.20 (−1.52, −0.96) | 0.08 (−0.09, 0.25) | 0.353 | ||
Distribution of characteristics displayed as “subjects with characteristic / total with data available (%)” for categorical variables, and “median (interquartile (IQR) range)” for continuous variables.
Statistical analyses used: A - chi-square and Wilcoxon rank sum with chi-square (χ2) and Z statistic displayed, respectively; B-C: adjusted linear regression with mean difference (β) and IQR displayed. P-values also displayed from each respective analysis and
denotes significance (α < 0.05)
The assessment of echocardiographic parameters of diastolic function in the full cohort inclusive of all treatment strategies demonstrated an association of aRL shunting with lower RA end-diastolic volume and higher RA emptying fraction (Table 2B) and no significant relationship to RA end-systolic volume. Regarding tissue and inflow Doppler data, there was an association of aRL shunting with higher tricuspid inflow peak A wave velocity, but no significant associations were seen between aRL shunting and E velocity, e’ velocity, and E/A or E/e’ ratios (Table 2B). Regarding STE-derived strain parameters, aRL shunting was associated with a higher average RA peak longitudinal strain compared to other atrial shunting patterns, but there were no significant associations between type of shunting and RA early strain rate, or RV global longitudinal strain or strain rate (Table 2C).
When the analysis was stratified by treatment strategy, there were significant associations only in the elective repair strategy group, including associations between aRL shunting and: higher RAEF (Central Illustration), higher average RA peak longitudinal strain (Central Illustration) and a higher E/e’ ratio (β 5.74 [0.81–10.67], p = 0.026).
Central Illustration: Atrial right to left shunting after repair of tetralogy of Fallot is associated with better RA emptying fraction and peak longitudinal strain, especially in the elective repair strategy.
Atrial right to left shunting (aRL) is a common finding after repair of tetralogy of Fallot (C, color and spectral doppler illustration of predominantly right to left atrial shunting), but is poorly understood. In this study, we aimed to investigate the associations of this finding to echocardiographic measures of diastolic function, such as right atrial peak longitudinal strain, a metric that is dramatically changed from pre to postoperative states (D, right atrial ejection fraction and strain change from preoperative (top) and postoperative (bottom) states as assessed by speckle tracking echocardiography (STE)). In our study, aRL was associated with higher RA emptying fraction (A, global function) and peak longitudinal strain (B, conduit function), both in the overall cohort and in the elective repair group (B: Mean differences (β) of aRL shunting group to reference group of all other shunting patterns, with 95% confidence intervals and stratification by repair strategy).
While aRL shunting was not associated with LOS within the overall cohort (Figure 1), aRL shunting was associated with a 28% shorter LOS (mean adjusted 8.6 vs. 12.1 days, IRR 0.72, IQR 0.52–0.98) for the elective repair strategy group. This relationship was also seen when analyzing pooled infant repairs (elective and staged repairs), where aRL shunting was associated with a 29% shorter LOS (mean adjusted 8.5 vs 11.9 days, IRR 0.71, IQR 0.54–0.93). There was no significant association between aRL shunting and LOS in the neonatal or staged repair groups (Figure 1). A sensitivity analysis excluding those with no atrial level shunting (Supplemental Table 3) and adjusting for the presence of 22q11.2 deletion syndrome did not significantly change results (Supplemental Table 4).
Figure 1. aRL shunting in all infants, and more specifically in infants with no prior palliation, is associated with shorter hospital length of stay.
Adjusted multivariable regression incidence risk ratio for aRL shunting group to reference group of all other shunting patterns for exposure of hospital length of stay (LOS), stratified by repair strategy. IRR=1: both groups equal LOS, IRR < 1: aRL shunting group lower LOS, IRR > 1: aRL shunting group higher LOS.
Discussion:
In this large for its kind single-center cohort study, we aimed to investigate the association of atrial-level right-to-left shunting with echocardiographic parameters of diastolic function and clinical outcome following TOF surgical repair. We hypothesized that aRL shunting would be associated with worse RV diastolic function compared to those with other shunting patterns, but that this “pop-off” mechanism would be associated with a more expedient progression to hospital discharge. We found that aRL shunting is associated with better right atrial global and reservoir function (RAEF and RA peak longitudinal strain, respectively) in patients undergoing elective repair of TOF. In the elective group, aRL shunting is also associated with worse RV compliance (higher E/e’) but with shorter LOS.
Our main findings in this study highlight the important role of atrial function in association with aRL shunting. We showed an association of aRL shunting to higher RA peak longitudinal strain, highlighting atrial reservoir function where the RA is filling against a closed tricuspid valve. Similarly, we showed a higher A wave peak inflow Doppler velocity in the overall cohort, indicating better contractile (pump) function of the RA. Both findings are corroborated by the association of aRL shunting with RAEF, indicative of more global/overall atrial function, seen in both the overall cohort and subgroup analyses. One could also argue that the trend towards a lesser proportion of patients with end-diastolic anterograde pulmonary blood flow in the aRL shunting group also speaks in support of this, however, we acknowledge that this is a complex phenomenon and the difference not significant. With the caveat that echocardiographic markers of diastolic function are also inherently reflective of ventricular properties and that our study design cannot establish cause and effect, our data suggest that aRL shunting is also reflective of atrial function, rather than purely RV noncompliance. In this context, understanding whether or not leaving an atrial communication at surgery facilitates better RA function could be an important question for further investigation.
While we identified associations between aRL and atrial function, we showed mixed results in relation to aRL shunting and its association with impaired ventricular relaxation. We identified an association with higher peak A wave velocity in the overall cohort, and higher E/e’ ratio in the elective repair group. However, we did not identify a significant association between aRL and end diastolic anterograde pulmonary blood flow in the overall cohort, and we did not identify significant associations with other parameters of diastolic function that one might expect if RV noncompliance was the most important factor associated with aRL shunting. Ultimately, our study design cannot definitively discern the relative contribution of RV impaired relaxation vs. atrial function in the resulting aRL and this phenomenon may result from a contribution from both. Future prospective study is warranted, especially with regards to aRL and clinical outcomes, which may provide data to inform institutional preferences on creating/preserving an atrial communication at surgery.
In this study we found that patients with aRL shunting had lower right atrial end diastolic volume. RA size measured as end diastolic area or volume has been shown to be associated with elevated ventricular filling pressures on invasive hemodynamics in has been shown to be associated with elevated ventricular filling pressures on invasive hemodynamics in infants and children with repaired TOF (6). In this study, it is possible that end diastolic volume is lower in patients with aRL shunting given the volume escaping across the atrial communication, but one could also argue that this indicates better RV compliance and a lower filling pressure when aRL shunting is present.
Subgroup analysis of associations between aRL shunting and RA function showed that this relationship was most pronounced in the elective treatment group and least relevant to the clearly distinct neonatal repair group. The elective group is also the only subpopulation where we identified an association of aRL shunting with higher E/e’ ratio, a finding that has not been demonstrated previously and is the strongest indicator in this study in support of aRL shunting being associated with higher RV filling pressures in this subgroup. Thus, we postulate that, in older patients at the time of TOF repair, the RV may have had more time to become noncompliant under preoperative pressure-load physiology, given that older age at repair is associated with increased RV hypertrophy and fibrosis [18]. However, it could be conversely hypothesized that there are some patients who required neonatal repair who have prohibitively low pulmonary blood flow from significant hypertrophied muscular obstruction, and that these patients should have also demonstrated echocardiographic evidence of RV noncompliance.
Little is known overall about how atrial function evolves after TOF repair and how it varies across timing of surgical repair. Prior studies have shown reduced RA function in repaired TOF patients in the early postoperative period, including impaired reservoir, conduit, and contractile function [19–21]. Past reports by our group showed that there is a decline in RA emptying fraction and peak longitudinal strain between preoperative and postoperative echocardiographic assessments [9]. Despite this decline, our results suggest that RA function is important in the postoperative period.
We believe this is the first study to examine the relationship of strain to atrial level shunting in the postoperative period after surgical repair. While it has been previously shown that infants undergoing TOF repair experience a significant drop in RA and RV strain from pre to postoperative states, there are few studies examining strain in the early postoperative period following surgical repair of TOF [9, 22]. The impact of early RA function on late outcomes is yet to be determined and should be investigated in future studies, considering that in adults with repaired TOF, diminished RA strain is associated with greater risk of new-onset atrial arrhythmias [23]. Furthermore, atrial strain is emerging as an imaging biomarker in other disease processes, such as in pulmonary hypertension where measures of reduced RA longitudinal strain correlate with higher N-terminal-pro-brain natriuretic peptide levels, indicating a worse prognosis [24].
We believe our study supports the importance of routine interrogation of atrial level shunting early after repair of TOF, particularly in light of the difficulty in assessing diastolic function by echocardiography in infants and children [25–28]. Our data show that interrogation of atrial shunting could serve clinicians in prognostication with regards to LOS. Additionally, if the aRL “pop-off” for cardiac output is what is driving the association with improved LOS, then these data could inform institutional surgical policies regarding PFO creation or preservation at repair of TOF.
Limitations:
We acknowledge limitations to our study. We focused on echocardiographic findings in the early postoperative period only, however postoperative echocardiograms were performed as part of a research protocol that mandated performing postoperative echocardiograms between 2 and 5 days after surgery, therefore we cannot extrapolate our findings to later echocardiographic findings. We also acknowledge that there are often physiological differences that may be significant between 2–5 postoperative days. While most of these echocardiograms served as a pre-discharge study and obtaining a research grade echocardiogram can be regarded as a strength, these factors could also constitute limitations of our study. We are also unable to comment on how aRL evolves later in the postoperative period. Several operative factors are also important to consider in the interpretation of our results. The first is surgical approach, where right atriotomy was universally used across the cohort. This exposure was equally distributed across the comparison groups and therefore is unlikely to have affected analysis but should be considered given that the loss of pericardial integrity may also contribute to diastolic dysfunction and may have an impact on atrial level shunting that is yet to be fully understood. Our study design attempts to adjust for other operative factors (i.e. use of a ventriculotomy and type of outflow tract repair), but the observational nature of our study limited our ability to test the association between operative factors and atrial shunting. The study population included a smaller sample size in the neonatal and staged repair strategy groups, however this is reflective of the general population of patients with TOF. Therefore, the lack of significant findings in the neonatal or staged sub-group analyses may result from lack of statistical power to identify differences, which could be identified in larger TOF cohorts. While we considered other hemodynamic factors such as significant tricuspid regurgitation, pulmonary insufficiency, and left ventricular dysfunction, postoperative TR and LV dysfunction were either nonexistent or mild. Future analyses with more representation of these variables would allow for the evaluation of their contribution to right to left shunting. We do not report on invasive hemodynamic data from cardiac catheterizations because there were only a handful of postoperative cardiac catheterizations in this cohort. Finally, while this study’s cohort size can be regarded as a strength for a single-center study, it was conducted at a quaternary referral center and may not be fully generalizable to all patients presenting for repair of TOF.
Conclusions:
Interrogation of atrial shunting is a simple, accessible, and clinically relevant echocardiographic parameter which can inform the care of these patients in the early postoperative period. aRL shunting early after repair could be more reflective of atrial function rather than RV noncompliance (or a combination of these factors), and may be beneficial for patients undergoing elective TOF repair. The relevance of aRL shunting in patients with TOF undergoing neonatal interventions requires further study.
Supplementary Material
Highlights:
Atrial right to left shunt (aRL) after repair of TOF is common, but not well understood
The relationship of aRL to outcomes and echo metrics of diastole are unclear
Large single center cohort inclusive of neonatal, staged, and elective repair
aRL is associated with better atrial function, rather than RV noncompliance alone
In elective TOF repair, aRL is also associated with shorter length of stay
Acknowledgements:
We thank Yan Wang, Anysia Fedec, Valerie Capone, Devon Ash and the inpatient sonographers of the Children’s Hospital of Philadelphia Echocardiography Laboratory for their support performing research echocardiograms. The authors would also like to thank the Children’s Hospital of Philadelphia Cardiac Center Clinical Research Core for statistics support.
Funding Sources:
Laura Mercer-Rosa’s research was funded by National Institutes of Health National Heart, Lung and Blood Institute grant K01-HL125521 and Pulmonary Hypertension Association supplement to K01-HL-125521. Research reported in this publication was also supported by the National Center for Advancing Translational Sciences of the National Institutes of Health under award number UL1TR001878. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Abbreviations:
- aRL
atrial right to left shunting
- TOF
tetralogy of Fallot
- RA
right atrium
- EF
ejection/emptying fraction
- RV
right ventricle/ventricular
- LOS
length of stay
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Declaration of Competing Interests:
The authors of this manuscript have no competing interests or disclosures to report.
References
- [1].Villafañe J, Feinstein JA, Jenkins KJ, et al. Hot topics in tetralogy of Fallot. Journal of the American College of Cardiology. 2013;62:2155–66. [DOI] [PubMed] [Google Scholar]
- [2].Miller JR, Stephens EH, Goldstone AB, et al. The American Association for Thoracic Surgery (AATS) 2022 Expert Consensus Document: Management of infants and neonates with tetralogy of Fallot. J Thorac Cardiovasc Surg. 2023;165:221–50. [DOI] [PubMed] [Google Scholar]
- [3].Bailey J, Elci OU, Mascio CE, et al. Staged Versus Complete Repair in the Symptomatic Neonate With Tetralogy of Fallot. The Annals of thoracic surgery. 2020;109:802–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Aboulhosn JA, Lluri G, Gurvitz MZ, et al. Left and right ventricular diastolic function in adults with surgically repaired tetralogy of Fallot: a multi-institutional study. Can J Cardiol. 2013;29:866–72. [DOI] [PubMed] [Google Scholar]
- [5].Maskatia SA, Morris SA, Spinner JA, et al. Echocardiographic Parameters of Right Ventricular Diastolic Function in Repaired Tetralogy of Fallot Are Associated with Important Findings on Magnetic Resonance Imaging. Congenit Heart Dis. 2015;10:E113–22. [DOI] [PubMed] [Google Scholar]
- [6].DiLorenzo M, Hwang WT, Goldmuntz E, et al. Diastolic dysfunction in tetralogy of Fallot: Comparison of echocardiography with catheterization. Echocardiography. 2018;35:1641–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [7].Shiina Y, Funabashi N, Lee K, et al. Right atrium contractility and right ventricular diastolic function assessed by pulsed tissue Doppler imaging can predict brain natriuretic peptide in adults with acquired pulmonary hypertension. Int J Cardiol. 2009;135:53–9. [DOI] [PubMed] [Google Scholar]
- [8].Cullen S, Shore D, Redington A. Characterization of right ventricular diastolic performance after complete repair of tetralogy of Fallot. Restrictive physiology predicts slow postoperative recovery. Circulation. 1995;91:1782–9. [DOI] [PubMed] [Google Scholar]
- [9].Ekhomu O, Faerber JA, Wang Y, et al. Right atrial function early after tetralogy of Fallot repair. The international journal of cardiovascular imaging. 2022;38:1961–72. [DOI] [PubMed] [Google Scholar]
- [10].Richter MJ, Zedler D, Berliner D, et al. Clinical Relevance of Right Atrial Functional Response to Treatment in Pulmonary Arterial Hypertension. Front Cardiovasc Med. 2021;8:775039. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].Alenezi F, Rajagopal S. The right atrium, more than a storehouse. Int J Cardiol. 2021;331:329–30. [DOI] [PubMed] [Google Scholar]
- [12].Rai AB, Lima E, Munir F, et al. Speckle Tracking Echocardiography of the Right Atrium: The Neglected Chamber. Clin Cardiol. 2015;38:692–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [13].Hope KD, Calderón Anyosa RJC, Wang Y, et al. Right atrial mechanics provide useful insight in pediatric pulmonary hypertension. Pulm Circ. 2018;8:2045893218754852. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].Monti CB, Secchi F, Capra D, et al. Right ventricular strain in repaired Tetralogy of Fallot with regards to pulmonary valve replacement. Eur J Radiol. 2020;131:109235. [DOI] [PubMed] [Google Scholar]
- [15].Mercer-Rosa L, Elci OU, DeCost G, et al. Predictors of Length of Hospital Stay After Complete Repair for Tetralogy of Fallot: A Prospective Cohort Study. J Am Heart Assoc. 2018;7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [16].SAS Knowledge Base 63038: Predictive margins and average marginal effects. SAS Institute Inc. [Google Scholar]
- [17].SAS Knowledge Base 66249: RunBY macro: Repeatedly run or add BY processing to any macro, procedure, or special code. SAS Institute Inc. [Google Scholar]
- [18].Peck D, Tretter J, Possner M, et al. Timing of Repair in Tetralogy of Fallot: Effects on Outcomes and Myocardial Health. Cardiol Rev. 2021;29:62–7. [DOI] [PubMed] [Google Scholar]
- [19].Hou J, Yu HK, Wong SJ, et al. Atrial mechanics after surgical repair of tetralogy of Fallot. Echocardiography. 2015;32:126–34. [DOI] [PubMed] [Google Scholar]
- [20].Riesenkampff E, Luining W, Seed M, et al. Increased left ventricular myocardial extracellular volume is associated with longer cardiopulmonary bypass times, biventricular enlargement and reduced exercise tolerance in children after repair of Tetralogy of Fallot. Journal of cardiovascular magnetic resonance : official journal of the Society for Cardiovascular Magnetic Resonance. 2016;18:75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [21].Kutty S, Shang Q, Joseph N, et al. Abnormal right atrial performance in repaired tetralogy of Fallot: A CMR feature tracking analysis. Int J Cardiol. 2017;248:136–42. [DOI] [PubMed] [Google Scholar]
- [22].Sun Z-Y, Li Q, Li J, et al. Echocardiographic evaluation of the right atrial size and function: Relevance for clinical practice. American Heart Journal Plus: Cardiology Research and Practice. 2023:100274. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [23].Egbe AC, Miranda WR, Madhavan M, et al. Right atrial dysfunction is associated with atrial arrhythmias in adults with repaired tetralogy of fallot. Am Heart J. 2023;263:141–50. [DOI] [PubMed] [Google Scholar]
- [24].Querejeta Roca G, Campbell P, Claggett B, et al. Right Atrial Function in Pulmonary Arterial Hypertension. Circ Cardiovasc Imaging. 2015;8:e003521; discussion e. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [25].Lai WW, Mertens L, Cohen M, et al. Echocardiography in pediatric and congenital heart disease : from fetus to adult. Third edition. ed. Hoboken, NJ: Wiley-Blackwell; 2021. [Google Scholar]
- [26].Ho SY, Nihoyannopoulos P. Anatomy, echocardiography, and normal right ventricular dimensions. Heart. 2006;92 Suppl 1:i2–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [27].DiLorenzo MP, Bhatt SM, Mercer-Rosa L. How best to assess right ventricular function by echocardiography. Cardiol Young. 2015;25:1473–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [28].Camarda JA, Patel A, Carr MR, et al. Practice Variations in Pediatric Echocardiography Laboratories. Pediatr Cardiol. 2019;40:537–45. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.