Clinical Status and Reintervention in Neonates With Symptomatic Tetralogy of Fallot

Abstract

Background

In symptomatic neonates with tetralogy of Fallot (sTOF), the initial treatment strategy significantly affects early outcomes, but its long-term impact remains less well defined.

Objectives

The aim of the study was to compare primary (PR) vs staged repair (SR) in sTOF with respect to reintervention (RI) rates and types, clinical and echocardiographic outcomes, and medication use.

Methods

Neonates with sTOF undergoing PR or SR and with >1 year of follow-up after complete repair were included. The primary outcome was cumulative RI incidence; secondary outcomes included mortality and late echocardiographic and clinical findings. Propensity scoring adjusted for baseline differences. Landmark analysis assessed RI risk at yearly intervals following complete repair.

Results

Of 441 neonates, 182 (41%) underwent PR, and 259 (59%) underwent SR. Groups differed in gestational age, intubation, and 22q11 status. Median follow-up postrepair was 5.26 (2.91, 8.21) years. RI burden was high in both groups, with a small, consistent but nonsignificant advantage to PR. The type of RI varied over time. PR was associated with greater pulmonary insufficiency and larger pulmonary arteries. RV pressure was ≤half systemic in 80%; 10% had ≥moderate tricuspid regurgitation, without between-group difference. Elevated RV pressure was associated with ≥moderate tricuspid regurgitation.

Conclusions

Among sTOF survivors beyond the early perioperative period, late RI burden and residual hemodynamic lesions are common and largely unrelated to initial strategy. PR is associated with increased pulmonary insufficiency and pulmonary artery size.

Central Illustration

Neonates with tetralogy of Fallot and symptomatic cyanosis (sTOF) require early intervention to establish a stable circulation. Management strategies involve either primary repair (PR) or staged repair (SR), with initial palliation (IP) followed by complete repair (CR). Identifying the optimal approach has proven challenging, with early studies limited by methodological flaws and evolving care standards.1, 2, 3 Recent data fail to show a mortality difference between strategies but consistently favor PR in terms of early reintervention (RI) rates and cumulative risk exposure.4, 5, 6, 7, 8, 9, 10 Early advantages may, however, come at a long-term cost. To more fully understand the implications of initial treatment strategies, it is essential to assess outcomes beyond the early high-risk period. In this study, we leveraged a large, multicenter cohort of sTOF patients to evaluate long-term outcomes and RI rates among infants who survived beyond the first year of life, stratified by initial management strategy.

Methods

We previously conducted a large, multicenter study analyzing data from 9 hospitals within the Congenital Cardiac Research Collaborative (CCRC). This study included all consecutive newborns diagnosed with tetralogy of Fallot (TOF) who underwent their intervention, either IP or PR, before 30 days of life, between January 1, 2005, and November 30, 2017. Initial procedural and short-term outcomes for this entire group have already been reported.⁴ The current investigation delves deeper into a specific subset of these patients, aiming to better understand how their initial treatment strategy influences longer-term clinical outcomes and the likelihood of requiring RI beyond the high-risk interventional period of early infancy. As such, only patients alive with follow-up beyond the first year after CR were included in this analysis. Having met these criteria, analysis of this subgroup began at the time of the index procedure. Patients were excluded if they had nonconfluent branch pulmonary arteries, TOF with atrioventricular canal, TOF with absent pulmonary valve syndrome, TOF with major aortopulmonary collateral arteries that underwent unifocalization (or if there was intent for unifocalization), or double-outlet right ventricle. Chart abstraction was performed at the center level, with deidentified data entered into an electronic database hosted at Children’s Healthcare of Atlanta, which served as the data coordinating center for the CCRC. Data were rigorously audited through a previously described process.¹¹ This study was approved by the Institutional Review Board at Cincinnati Children’s Hospital in Cincinnati, Ohio, which acted as the single institutional review board, with a waiver of the need for informed consent. A data use agreement was in place among all participating centers and the data coordinating center. The primary exposure was the index procedure, which was defined as the first intervention performed for the purpose of IP or PR of sTOF. The SR strategy included 2 components: IP and subsequent CR. The SR group included any patient undergoing IP, even if the patient did not ultimately undergo CR. Definitive repair was defined as complete sTOF repair with relief of right ventricular outflow tract obstruction and intent for complete ventricular septal defect (VSD) closure and could be either PR or CR.

The primary outcome was defined a priori as cumulative RI burden, defined as any surgical or transcatheter cardiac RI following the index procedure. In the SR group, CR was considered an obligate event and thus was not included as an RI. The at-risk period for death, transplantation, and RI began with the index procedure and extended through the last follow-up. RI categories were based upon type of RI (surgical or transcatheter), with subcategorization based upon the most common RI within a given category. Clinical status, echocardiographic findings, and medication use were extracted from the medical record from the date of the most recent evaluation prior to the study end date of January 31, 2019.

Statistical analysis

Statistical analyses were performed by the CCRC data coordinating center using SAS software version 9.4 (SAS Institute), and landmark analysis figures were generated by a package in R version 4.2.2 (R Foundation for Statistical Computing).¹² Statistical significance was assessed at the 0.05 level. Normality of continuous variables was assessed using the histogram, normal probability plots, and Anderson-Darling test for normality. Descriptive statistics are presented as counts and percentages for categorical variables and as median (25th and 75th percentile) for continuous data with skewed distributions. Continuous data were compared between SR- and PR-treated patients by using Wilcoxon rank-sum tests, and comparisons between categorical variables were performed using chi-square tests or the Fisher exact test when expected cell counts were <5. Because patient characteristics differed between groups at the index procedure (Table 1), inverse probability of treatment weighing using propensity scores was used to weight the regression coefficients in all comparisons of primary and secondary outcomes to adjust for potential confounders between groups. This approach has been used extensively in this cohort and is described in detail.⁴ Briefly, propensity weights were derived using logistic regression models and accounted for 5 potential confounders, including center, preintervention invasive ventilation, prematurity, DiGeorge syndrome, and presence of antegrade pulmonary blood flow. Weights were stabilized and truncated at the 1st and 99th percentiles.¹³ The standardized mean difference was used to quantify the relative imbalance in a covariate between the 2 treatment groups. All adjusted models included the main effect of treatment and were weighted by stabilized propensity score to achieve balance between treatment groups (Supplemental Table 1). Elements with adjusted standardized mean difference <0.10 were considered to have achieved satisfactory balance.¹⁴ For the primary outcome of cumulative RI burden, the total number of RIs was tallied following a patient’s initial procedure. Cumulative RI burden was evaluated as a rate by totaling the number of RIs and dividing by the duration of follow-up. The resultant value was the number of RIs per person-year of follow-up. Rates of RI were compared between treatment strategies using negative binomial regression models and are presented as rate ratios with 95% CIs. In addition to evaluating total RI burden, we utilized unweighted landmark analysis to compare the risk of RI at distinct intervals of follow-up.^15,16 Landmark points were chosen at 2, 3, and 4 years. For continuous outcomes (eg, pulmonary artery [PA] diameter), residual errors were gauged for normality with histograms and quantile-quantile plots. Failing to meet the normality assumption, continuous outcomes were ranked before analysis, and modeling was carried out on the rank-transformed data. Unadjusted and adjusted estimates are presented as unweighted and weighted medians (25th and 75th percentile), and adjusted P values were derived from the propensity score-weighted 2-sample Student’s t-test on the ranked data.

Table 1.

Observed Baseline Patient Demographics by Initial Treatment Strategy

	n	Overall (N = 441)	Primary Repair (n = 182)	Staged Repair (n = 259)	P Value
Center 1	43	43 (9.8%)	13 (7.1%)	30 (11.6%)	<0.001
Center 2	114	114 (25.9%)	75 (41.2%)	39 (15.1%)
Center 3	19	19 (4.3%)	5 (2.8%)	14 (5.4%)
Center 4	55	55 (12.5%)	8 (4.4%)	47 (18.2%)
Center 5	23	23 (5.2%)	14 (7.7%)	9 (3.5%)
Center 6	47	47 (10.7%)	12 (6.6%)	35 (13.5%)
Center 7	62	62 (14.1%)	17 (9.3%)	45 (17.4%)
Center 8	26	26 (5.9%)	14 (7.7%)	12 (4.6%)
Center 9	52	52 (11.8%)	24 (13.2%)	28 (10.8%)
Time from initial repair to last follow-up	441	5.59 (3.22, 8.68)	5.77 (2.96, 8.39)	5.51 (3.32, 8.87)	0.786
Time from complete repair to last follow-up	441	5.26 (2.91, 8.21)	5.77 (2.96, 8.39)	5.05 (2.74, 8.11)	0.135
Age at initial intervention, d	441	7 (5, 14)	7 (5, 17)	7 (5, 13)	0.112
Weight at initial intervention, kg	432	3 (2.5, 3.4)	3.1 (2.7, 3.5)	2.8 (2.5, 3.3)	0.003
Inotrope preintervention	441	45 (10.2%)	15 (8.2%)	30 (11.6%)	0.254
Invasive ventilation preintervention	441	99 (22.5%)	26 (14.3%)	73 (28.2%)	<0.001
Gestational age, wks	421	38.6 (37.1, 39.4)	39.0 (37.4, 39.7)	38.4 (37.0, 39.1)	0.002
Prematurity (<37 wks gestational age)	441	83 (18.8%)	21 (11.5%)	62 (23.9%)	0.001
Birth weight, kg	433	2.9 (2.5, 3.3)	3.0 (2.6, 3.4)	2.8 (2.4, 3.2)	0.005
Genetic syndrome, any	441	131 (29.7%)	50 (27.5%)	81 (31.3%)	0.39
DiGeorge/VCFS [22q11 deletion]	441	49 (11.1%)	12 (6.6%)	37 (14.3%)	0.011
Antegrade pulmonary blood flow	441	223 (50.6%)	97 (53.3%)	126 (48.7%)	0.336
Right pulmonary artery diameter, mm	433	3.7 (3.2, 4.2)	3.8 (3.2, 4.1)	3.6 (3.1, 4.2)	0.2
Left pulmonary artery Z-score	422	−1.4 (−1.99, −0.84)	−1.4 (−1.99, −0.90)	−1.41 (−2.02, −0.84)	1.00
Left pulmonary artery diameter, mm	430	3.5 (3, 4)	3.5 (3, 4)	3.4 (2.8, 4.0)	0.065
Left pulmonary artery Z-score	419	−1.48 (−2.01, −0.92)	−1.42 (−1.91, −0.96)	−1.51 (−2.13, −0.87)	0.554

Values are median (25th to 75th percentiles) or count (percentage of total).

Results

Of 441 neonates who survived with follow-up beyond 1 year after CR for sTOF, 182 (41%) underwent PR, and 259 (59%) underwent SR. The SR group was more likely to have a younger gestational age (38.4 vs 39 weeks, P = 0.002), higher rate of prematurity (62/259 [23.9%] vs 21/182 [11.5%], P = 0.001), smaller birth weight (2.8 vs 3 kg, P = 0.005), higher likelihood of being intubated before initial intervention (73/259 [28%] vs 26/182 [14%], P < 0.001), and higher prevalence of 22q11 deletion (37/259 [14%] vs 12/182 [7%], P = 0.01). Median follow-up duration after CR for the study population was 5.26 years (range 2.91-8.21 years), with no significant difference between groups (Table 1).

Reintervention

RI was common in both groups. Among patients undergoing an SR, 62% experienced at least 1 RI, compared to 53% undergoing a PR approach (Table 2). Although the cumulative RI rate was higher with SR, 0.28 (0.24, 0.33) per patient-year compared to 0.22 (0.18, 0.27) per patient-year in the PR cohort, the difference did not reach statistical significance (adjusted rate ratio 1.26 [0.97, 1.65] per patient-year, P = 0.09). Landmark analyses further indicated a slow decline in overall risk of RI, with no significant differences between the groups regarding attrition prior to each landmark point (P values: 0.376 at 24 months, 0.648 at 36 months, and 0.400 at 48 months) or RI thereafter (P values: 0.632 at 24 months, 0.399 at 36 months, and 0.516 at 48 months) (Figures 1A to 1C). Further, risk modeling utilizing extrapolation from Figure 1 suggested those who survived to 3 years had approximately a 10% risk of RI in the following year in the PR group, compared to 12.5% in the SR group. By 4 years, the risk was 8.3% in the PR group and 10% in the SR group.

Table 2.

Cumulative Incidence Rate of Reintervention, by Group

	Total Number of Reinterventions Median (25th-75th) Total	Person-Years of Follow-Up Median (25th-75th) Total	Rate of Reintervention per patient year (95% CI)	Rate Ratio-Adjusteda
Staged repair n = 259	1 (0, 2) 390	5.5 (3.3, 8.9) 1,593	0.27 (0.23, 0.32)	0.28 (0.24, 0.33)
Primary repair n = 182	1 (0, 2) 211	5.8 (3.0, 8.4) 1,113	0.22 (0.18, 0.28)	0.22 (0.18, 0.27)
Rate ratiob	--	--	1.23 (0.93, 1.61) P = 0.141	1.26 (0.97, 1.65) P = 0.089

^aAdjusted for center, preintervention invasive ventilation, prematurity, DiGeorge syndrome, and presence of antegrade pulmonary blood flow.

^bPrimary repair is the reference group.

Figure 1.

Landmark Analysis of Reintervention Risk by Year

Landmark analyses of reintervention in the second, third, and fourth years of life following complete repair. By design, patients who died before the first year are excluded (flat horizontal line). Attrition from death or loss to follow-up is depicted before each landmark point (dashed vertical line), followed by a conditional probability of reintervention.

Figure 2, Figure 3 and Table 3, Table 4 provide descriptive data on overall RI risk by type and between groups. PA interventions and pulmonary valve replacement were most common. Transcatheter RI was more common than surgical RI across all time frames, and while overall RI rates slowly decreased over the observation period, RI risk remained elevated.

Figure 2.

Trends in Reintervention Type Over Time

Combined bar and line graph displays trends in reintervention type over time by group. Blue bars show the number of patients alive at each interval, gray bars show total reinterventions, and green bars show the number of patients with ≥1 reintervention. Dashed lines (right y-axis) show the proportion of reinterventions that were surgical (red), transcatheter (black), or other (purple). CR = complete repair; PR = primary repair; RI = reintervention.

Figure 3.

Reintervention Types Over Time by Repair Strategy

Comparison of trends in reintervention type between primary and staged repair groups over time. ASD = atrial septal defect; PA = pulmonary artery; RI = reintervention; RVOT = right ventricular outflow tract.

Table 3.

Reintervention Counts and Types, Overall and by Group

	Total				Primary Repair				Staged Repair
Years after CR/PR	1-2	2-3	3-4	4-5	1-2	2-3	3-4	4-5	1-2	2-3	3-4	4-5
Number of Patients	441	368	323	281	182	155	135	120	259	213	188	161
Patients with ≥1 RI in this era	82 (18.6%)	58 (15.8%)	30 (9.3%)	28 (10.0%)	31 (17.0%)	23 (14.8%)	11 (8.2%)	12 (10.0%)	51 (19.7%)	35 (16.4%)	19 (10.1%)	16 (9.9%)
Total number of RI	93	63	34	28	34	26	12	12	59	37	22	16
Transcatheter RI	70 (75.3%)	47 (74.6%)	17 (50.0%)	21 (75.0%)	27 (79.4%)	20 (76.9%)	4 (33.3%)	11 (91.7%)	43 (72.9%)	27 (73.0%)	13 (59.1%)	10 (62.5%)
Surgical RI	23 (24.7%)	15 (23.8%)	16 (47.1%)	7 (25.0%)	7 (20.6%)	5 (19.2%)	8 (66.7%)	1 (8.3%)	16 (27.1%)	10 (27.0%)	8 (36.4%)	6 (37.5%)
Other RI	0 (0.0%)	1 (1.6%)	1 (2.9%)	0 (0.0%)	0 (0.0%)	1 (3.9%)	0 (0.0%)	0 (0.0%)	0 (0.0%)	0 (0.0%)	1 (4.6%)	0 (0.0%)

CR = complete repair; PI = pulmonary insufficiency; RV = right ventricle; TR = tricuspid regurgitation.

Table 4.

Reintervention Counts and Specific Types, Overall and by Group

	Total				Primary Repair				Staged Repair
Years after CR/PR	1-2	2-3	3-4	4-5	1-2	2-3	3-4	4-5	1-2	2-3	3-4	4-5
Total number of RI	93	63	34	28	34	26	12	12	59	37	22	16
PA angioplasty and/or stent	52 (55.9%)	35 (55.6%)	12 (35.3%)	16 (57.1%)	23 (67.7%)	15 (57.7%)	3 (25.0%)	9 (75.0%)	29 (49.2%)	20 (54.1%)	9 (40.9%)	7 (43.8%)
Pulmonary valve implantationa	11 (11.8%)	10 (15.9%)	12 (35.3%)	3 (10.7%)	2 (5.9%)	2 (7.7%)	5 (41.7%)	0 (0.0%)	9 (15.3%)	8 (21.6%)	7 (31.8%)	3 (18.8%)
RVOT Revision	7 (7.5%)	2 (3.2%)	2 (5.9%)	1 (3.6%)	3 (8.8%)	1 (3.9%)	2 (16.7%)	0 (0.0%)	4 (6.8%)	1 (2.7%)	0 (0.0%)	1 (6.3%)
RVOT angioplasty and/or stent	9 (9.7%)	9 (14.3%)	5 (14.7%)	5 (17.9%)	1 (2.9%)	3 (11.5%)	1 (8.3%)	2 (16.7%)	8 (13.6%)	6 (16.2%)	4 (18.2%)	3 (18.8%)
Atrial septal defect closure	2 (2.2%)	5 (7.9%)	1 (2.9%)	1 (3.6%)	0 (0.0%)	3 (11.5%)	0 (0.0%)	1 (8.3%)	2 (3.4%)	2 (5.4%)	1 (4.6%)	0 (0.0%)
Multiple procedures	4 (4.3%)	1 (1.6%)	1 (2.9%)	1 (2.7%)	2 (5.9%)	1 (3.9%)	1 (8.3%)	0 (0.0%)	2 (3.4%)	0 (0.0%)	0 (0.0%)	1 (6.3%)

CR = complete repair; PA= pulmonary artery; PR = primary repair; RVOT = right ventricular outflow tract.

^aIncludes conduit revision.

Clinical status and echocardiographic endpoints

Clinical status was available at a median duration of 5.3 years with a range of 3.1 to 8.6 years in those undergoing an SR strategy and 5.1 years with a range of 2.4 to 8.2 years in those undergoing PR. Overall mortality beyond 1 year was low under both management strategies. At the most recent evaluation, 99% of patients who underwent PR (who had survived beyond 1 year of age) were alive with CR compared to 98% who underwent SR (Supplemental Figure 1). Review of medical records over the same period showed no significant difference in medication use between the groups. The majority of patients in both groups were on no cardiac medications (61% in the SR group vs 69% in the PR group, P = 0.07), Supplemental Table 2. For those prescribed a medication, the most prevalent in both treatment groups were antiplatelet agents and diuretics.

The most recent echocardiogram was available at a median duration of 5.1 years for patients undergoing a PR strategy, with a range of 2.1 to 8.3 years. For those undergoing an SR strategy, the median duration was 4.7 years, ranging from 2.5 to 8.1 years. Across the entire cohort, right ventricular systolic function was generally good, with 96% of patients exhibiting no worse than mild dysfunction. After inverse probability weighting, approximately 18.9% of patients had estimated right ventricular systolic pressures greater than half systemic pressures, while occurrences of RV pressures equal to or greater than systemic pressure were rare (Table 5). Approximately 10% of the cohort exhibited moderate or greater tricuspid regurgitation (TR), and while TR severity was not significantly different between groups, there was a clear association between elevated RV pressure and degree of TR (P = 0.025). In comparing treatment strategies, ≥moderate pulmonary insufficiency (PI) was associated with PR (82% in PR compared to 69% SR; P = 0.005), with a nonsignificant trend toward a greater degree of RV enlargement with PR (P = 0.096). Baseline PA Z-scores were similar between groups. At last available imaging, PR was associated with greater RPA (−0.14 vs −0.66, P ≤ 0.001) and LPA (−0.53 vs −1.27, P = 0.004) Z-scores and Nakata indices (185 vs 148 mm²/m², P = 0.001) (Supplemental Table 3).

Table 5.

Echocardiographic Findings at Time of Last Follow-Up Examination, Overall and by Group

	Overall (N =441)	Primary Repair (n = 182)	Staged Repair (n = 259)	P Value
Age at last echo, y	4.9 (2.3, 8.2)	5.1 (2.1, 8.2)	4.7 (2.5, 8.1)	0.955
Time between CR and last echo	5.3 (2.8, 8.2)	5.8 (3.0, 8.4)	5.1 (2.7, 8.1)	0.156
RV size
Normal - mildly dilated	58.3%	53.4%	61.7%	0.096
≥ Moderate	41.7%	46.6%	38.3%
RV function
Normal - mildly diminished	96.1%	96.2%	96.0%	0.901
≥ Moderately diminished	3.9%	3.8%	4.0%
RV pressure
≤1/2 systemic	80.3%	83.7%	77.7%	0.583
≤1/2 systemic but < systemic	18.9%	16.3%	20.9%
≥ Systemic	0.8%	0.0%	1.5%
Qualitative TR
None- mild	89.8%	87.7%	91.3%	0.236
≥ Moderate	10.2%	12.3%	8.7%
Qualitative PI
None – mild	25.3%	17.7%	30.6%	0.005
≥ Moderate	74.8%	82.3%	69.4%

CR = complete repair; PI = pulmonary insufficiency; RV = right ventricle; TR = tricuspid regurgitation.

Discussion

In this multicenter study comparing late outcomes between survivors of PR and SR strategies in neonates born with sTOF, we found substantial but declining long-term burdens of RI. Overall trends in RI between groups favored PR but were not statistically significant, while RI risk and patterns varied more by time than by management pathway. Clinical status, medication use, and echocardiographic abnormalities were equivalent in both PR and SR groups, with the exception that severity of PI appeared greater with PR. PA dimensions were significantly larger following PR.

The primary goal of this study was to estimate RI risk in this population of patients with sTOF, acknowledging that there are numerous methods to express and quantify risk, each offering distinct insights. The cumulative risk method provides a population-level rate of RI per patient-year, allowing for a comparison of RI rates between groups, and is often an attractive means of summarizing risk in a single ratio reflective of overall risk burden in populations. However, this method does not differentiate between early and late events and instead treats all events equally regardless of when they occur and can obscure important time-dependent relationships. Traditional survival analysis provides insight into the overall risk of an event over time but, again, considers early and late events equally. In contrast, landmark analysis provides a different perspective, the purpose of which is to focus on the outcomes after a specific follow-up time, which might be relevant for understanding long-term effects or reducing biases that could be introduced by events occurring before the landmark time. In doing so, this analysis allows for a more accurate comparison of the 2 strategies' outcomes after patients have survived to this time point.^15,16 Our findings show that the risk of RI continues to decrease at each landmark. Although the risk of RI is generally higher for those undergoing an SR strategy at any given time, the difference is small and not statistically significant. Overall, these estimates present a consistent message about RI risk in this population: RI rates are high for both groups and vary over time but not significantly by initial management strategy.

Not all RIs are the same, and understanding the specific types of RIs encountered by this population is crucial. Different types of RIs can have varying implications for patient outcomes and may reflect distinct underlying issues related to the initial surgical strategy or subsequent hemodynamic changes. Transcatheter interventions were the most common form of RI across all groups and time frames, indicating an obvious preference for less invasive methods, whether palliative or corrective, when possible. Surgical RIs, while less common, showed a relative increase in frequency over time, which may suggest that more complex issues requiring surgical intervention either emerge later or are only addressable at later stages, such as refractory PA stenoses or pulmonary valve replacement in smaller children. Support for this concept, while circumstantial, is suggested by a transient inverse relationship with PA interventions during the same period.

During a follow-up period of ∼5.2 years, mortality was rare and similar across both treatment strategies. However, it is important to view these results in the context of the study's design, which only included patients who survived beyond 1 year of age. That is, sTOF survivors at 1 year of age were very likely to survive in the midterm. Our findings suggest that in sTOF patients who live past their first year, while RI is common, it is not significantly different between treatment strategies, and mortality is very low. While a more detailed analysis of the causes and rates of death in children with sTOF is necessary, the low incidence of late mortality makes it challenging to identify predictive factors without a significantly larger patient population.

The late echocardiographic data primarily highlight similarities rather than differences between treatment strategies. The notable exception is the association between PR and frequency of ≥moderate PI, a finding supported by the trend of greater right ventricular enlargement in patients in this group. While this later finding failed to meet statistical significance, it is a reasonable association to consider. The hypothesis is that early CR in the PR group leads to a longer period of progressive PI and consequent right ventricular dilation. It is also possible that later CR in the SR group was more likely to involve valve-sparing technique in TOF surgical repair. Evidence for this is circumstantial. While we did observe a difference between CR and last echo, the difference was not statistically significant, and it is important to note that this study was neither designed nor powered to trace the physiological steps leading to a statistically significant association between these mechanisms.

Notably, while most patients exhibited no or mild TR at their last assessment, about 10% of patients had moderate to severe TR. Determinants of TR in this age group have received little recent attention. Excluding those patients with concurrent AV septal defect and other (primary) structural abnormalities of the tricuspid valve, the most common acquired (secondary) etiology is damage to the septal leaflet of the tricuspid valve during the placement of the VSD patch.¹⁷ While the actual approach to VSD repair has not been shown to predict postoperative TR,¹⁸ presumably the challenges of VSD visualization and immature AV valve tissue manipulation that come with complete neonatal repair would place patients at higher risk for early TR, but this has not been examined adequately, is not apparent from our data, and therefore remains speculative. In contrast, residual hemodynamic burdens of elevated right ventricular pressure, residual VSDs, and pulmonary regurgitation are correlated with TR.¹⁸ Our finding of the association between elevated right ventricular pressures and TR highlights the impact of residual lesions and may provide some guidance in intraoperative decisions and postoperative management.

Finally, despite similar initial PA Z-scores, we observed a positive relationship between PA dimensions and employment of the PR strategy. This finding is consistent with previous studies in similar populations, although there have been inconsistent or weak associations with LPA growth.^5,19, 20, 21 To place this in the correct context, it is worth noting that differences in early RI primarily involved surgical and transcatheter interventions on surgical shunts, presumably aimed at increasing pulmonary blood flow and improving PA size, while interventions directly on PAs themselves were less common. While the factors influencing PA growth and symmetry are numerous and interact in complex ways, our findings suggest that, despite a higher rate of early RI in patients undergoing SR, PA dimensions are better with PR.^22,23 On this point, historical wisdom remains prescient in holding that for infants with sTOF there remains “no advantage (and a possible disadvantage) of a protocol of preliminary shunting and later repair.”^21,24,25 However, there are alternative explanations. Our method of measuring PA size followed the convention of measurement at the hilum in systole. If early PR results in longer periods and greater severity of PI, the resulting exaggerated pulse pressure may yield greater measured values, measurements that, while valid methodologically, may not provide as much characterization of PA health as desired.²⁶ Finally, it is possible that unmeasured patient-level factors influenced the decision to offer neonatal sTOF PR and that those factors are responsible for enhanced PA growth.

Study Limitations

This study has limitations beyond its retrospective design. Although a relatively large patient population and multicenter design lend statistical power and generalizability, they do not preclude unmeasured confounding. Duration of follow-up was >5 years, which may be inadequate to capture late outcomes of interest, such as timing and manner of pulmonary valve replacement, as well as the time course and clinical significance of early TR. Conversely, the relatively long interval may limit accurate representation of current practice, which leans toward ductal stent/right ventricular outflow tract stent in patients undergoing an SR strategy. Finally, as noted above, while the use of landmark analysis offers several advantages, including ease of graphical interpretation and mitigation of certain statistical biases, it does have some disadvantages, including sensitivity to choice of landmark and loss of statistical power at later timepoints. It is worth mentioning that this analysis likely captures outcomes among patients that are more closely followed by the health system (ie, those not lost to follow-up) and introduces follow-up bias. As such, rates of RI in this subset of patients are likely higher than the overall TOF cohort. However, this bias should be nondifferential, being equally present in both SR and PR groups, thereby preserving the validity of these comparisons via the landmark approach.

Conclusions

Taken in the context of prior comparisons of these 2 treatment strategies, these findings suggest that while the vast majority of early morbidity and risk exposure favors PR, this pathway is not without its perils. Mortality is uncommon but is also an insufficient metric for evaluation of outcomes in contemporary management of this population. Subsequent RI is required for most, irrespective of management strategy. Residual hemodynamic burdens bear heavily, but largely without discrimination between treatment strategies (Central Illustration).

Perspectives.

COMPETENCY IN MEDICAL KNOWLEDGE: In symptomatic neonates with tetralogy of Fallot, both primary and staged repair strategies are associated with substantial long-term reintervention burden. While early anatomic correction with primary repair may offer a slight reduction in late reintervention, this advantage does not reach statistical significance. Individualized surgical planning--accounting for patient comorbidities and center expertise--remains critical, as long-term clinical and echocardiographic outcomes are similar between strategies.

TRANSLATIONAL OUTLOOK: While clinical outcomes continue to improve, future studies should evaluate strategies to mitigate reintervention risk regardless of initial surgical approach, including novel valve-sparing techniques, improvement of existing transcatheter valve therapy together with a reduction in infections complications of valve implants and imaging-based surveillance protocols. Multi-institutional, prospective registries could clarify patient selection criteria and optimize timing for definitive repair to improve long-term durability of repair.

Central Illustration.

Clinical Status and Reintervention in Neonates With Symptomatic Tetralogy of Fallot: A Landmark Analysis

This retrospective comparison of primary vs staged repair of infants with sTOF finds that after the high-risk perioperative period, significant hemodynamic burdens remain, largely independent of initial management strategy. PR = primary repair; RI = reintervention; SR = staged repair; sTOF = symptomatic tetralogy of Fallot.

Funding support and author disclosures

This research is supported, in part, by donations to the CCRC from the Kennedy Hammill Pediatric Cardiac Research Fund, the Liam Sexton Foundation, and A Heart Like Ava. Dr Shahanavaz has reported consulting relationships with Medtronic, W.L. Gore & Associates, and Edwards Lifesciences. Dr Goldstein has consulting relationships with Medtronic, W.L. Gore & Associates, PECA Labs, and Mezzion Pharma. Dr Qureshi has reported consulting relationships with W.L. Gore & Associates, Medtronic Inc., and B. Braun. Dr Zampi has consulting relationships with Medtronic and W.L. Gore & Associates. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.