Predictors of Serious Adverse Events and High-Level Cardiorespiratory Support in Patients Undergoing Transcatheter Pulmonary Vein Interventions

Patients with pulmonary vein stenosis (PVS) often require frequent transcatheter pulmonary vein (PV) interventions for management of restenosis. Predictors of serious adverse events (AEs) and need for high-level cardiorespiratory support (mechanical ventilation, vasoactive support, and/or extracorporeal membrane oxygenation) 48 h after transcatheter PV interventions have not been reported. This is a single-center retrospective cohort analysis of patients with PVS who underwent transcatheter PV interventions from 3/1/2014 to 12/31/2021. Univariate and multivariable analyses were performed using generalized estimating equations to account for within-patient correlation. 240 patients underwent 841 catheterizations involving PV interventions (median 2 catheterizations per patient [1,3]). At least one serious AE was reported in 100 (12%) cases, the most common of which were pulmonary hemorrhage (n = 20) and arrhythmia (n = 17). There were 14 severe/catastrophic AEs (1.7% of cases) including three strokes and one patient death. On multivariable analysis, age less than 6 months, low systemic arterial saturation (< 95% in patients with biventricular [BiV] physiology, < 78% in single ventricle [SV] physiology), and severely elevated mean PA pressure (≥ 45 mmHg in BiV, ≥ 17 mmHg in SV) were associated with SAEs. Age less than 1 year, hospitalization prior to catheterization, and moderate–severe RV dysfunction were associated with high-level support after catheterization. Serious AEs during transcatheter PV interventions in patients with PVS are common, although major events such as stroke or death are uncommon. Younger patients and those with abnormal hemodynamics are more likely to experience serious AEs and require high-level cardiorespiratory support after catheterization.


Introduction
Pulmonary vein stenosis (PVS) continues to have high morbidity and mortality despite advances in medical therapy and interventional techniques, particularly in patients with bilateral disease, younger age, lower weight, and higher severity stenosis [1][2][3][4][5][6][7][8][9]. Restenosis is a hallmark of aggressive PVS, prompting recurrent transcatheter pulmonary vein (PV) interventions [7,10]. Adverse events (AEs) related to transcatheter PV interventions are common with an estimated frequency of 25%, but precipitating risk factors remain largely unknown [11]. A thorough understanding of predictors of AEs is crucial for risk stratification and optimization of care before, during, and after catheterization.
There are limited data addressing the need for high-level cardiorespiratory support (e.g., mechanical ventilation, vasoactive medications) after transcatheter PV interventions. Studies focusing on predictors of admission to the intensive care unit (ICU) after PV catheterization estimate a 50% admission rate, predominantly in patients with low weight, pre-catheterization oxygen supplementation, elevated right ventricular pressure, and need for intraoperative inotropic support [12,13]. Risk assessment of patients undergoing transcatheter PV interventions should additionally include predictors of high-level support to further inform patient counseling, resource allocation, and post-catheterization care.
The primary aim of this study was to determine the frequency and predictors of serious AEs in patients undergoing transcatheter PV interventions. The secondary aim was to determine the risk factors associated with need for high-level support including mechanical ventilation, vasoactive support, and/or extracorporeal membrane oxygenation (ECMO) at 48 h after transcatheter PV interventions. We hypothesized that serious AEs were associated with younger age and a higher number of PVs undergoing intervention.

Study Design and Participants
This is a retrospective cohort analysis of patients of all ages at a tertiary freestanding children's hospital who underwent transcatheter PV interventions from March 1, 2014, to December 31, 2021. PV interventions included balloon venoplasty, stent implantation, and/or stent re-dilation. Data were obtained from catheterization reports and an institutional database for catheterization cases. Variables entered into the database are curated for submission into the Congenital Cardiac Catheterization Project on Outcomes (C3PO) registry, and data were, thus, limited to that found in the registry. This study was approved by the Institutional Review Board of Boston Children's Hospital (IRB-P00041119).

Variables
The primary outcome was serious AEs defined as level 3 (moderate), 4 (severe), and 5 (catastrophic) events occurring within 14 days of the catheterization and prior to the next intracardiac procedure [14]. The secondary outcome of high-level cardiorespiratory support at 48 h was defined as a composite outcome of need for mechanical ventilation, vasoactive support (dopamine, epinephrine, norepinephrine, dobutamine, milrinone, vasopressin, phenylephrine, ephedrine, isoproterenol), and/or ECMO at 48 h post-catheterization. This sub-analysis used a cohort of patients from January 1, 2019, to December 31, 2021, given data for these variables were only entered into the institutional database during this time period. Patients who had an existing endotracheal tube or tracheostomy were excluded from the secondary outcome analysis.
The predictors were age at intervention, classified as younger than 6 months, 6 to 11.9 months, and older than 1 year of age; number of PVs undergoing intervention including interventions on all lobar and segmental veins and classified as 1 to 2, 3 to 4, and greater than 4 veins; gender, weight, PV diagnosis, cardiac diagnosis, single or biventricular physiology, presence of chronic lung disease, pre-catheterization right ventricle (RV) function as determined by pre-catheterization echocardiogram, case characteristics (number of PV interventions, case duration), pre-and post-catheterization location (home, cardiology ward, or intensive care unit (ICU)), and hospital length of stay (LOS) prior to catheterization. The following hemodynamic data obtained during the catheterization were also included: severely elevated mean pulmonary artery (PA) pressure ≥ 45 mmHg in biventricular physiology (BiV) and ≥ 17 mmHg in single ventricle physiology (SV); low mixed venous saturation < 60% in BiV and < 50% in SV; low systemic arterial saturation < 95% in BiV and < 78% in SV; elevated pulmonary to systemic blood flow ratio (Qp/ Qs) > 1.5; and elevated systemic ventricle end-diastolic pressure ≥ 18 mm Hg. These hemodynamic variables have been established and tested as indicators of patient vulnerability for occurrence of adverse events as part of the Congenital Heart Adjustment for Risk Method (CHARM) model [15].

Statistical Analysis
Patient characteristics were summarized overall and stratified by age at intervention. Categorical variables were described using frequencies and percentages and continuous variables using medians with interquartile ranges. The incidence of serious AEs were estimated with an exact binomial 95% confidence interval and univariable and multivariable analysis were performed using generalized estimating equations (GEE) to identify predictors of serious AEs. GEE was chosen to account for within-patient correlation given that the unit of analysis is the PV intervention, and a patient can have multiple lifetime interventions. Odds ratios are reported with 95% confidence intervals. A p value less than 0.05 based on the likelihood ratio test was required for retention in the final multivariable model. The association between predictors and the secondary outcome were analyzed in a similar manner. All statistical analysis were performed using Stata Statistical Software version 16.

Population Characteristics
There were 240 patients who met eligibility criteria with a total of 841 catheterizations involving PV interventions (median 2 interventions per patient [IQR 1,3], range 1-20) ( Table 1). Most cases were performed in patients 1 year of age or older with a median age at intervention of 12 months (IQR 9,36; range 7 days-63 years). Ten (1%) catheterizations included patients who were 18 years or older at the time of intervention. Most patients had PVS related to congenital heart disease (84%), and anomalous pulmonary venous return was the most common diagnosis (41%). Prior to catheterization, 55% were at home and 20% were in the ward. After catheterization, 23% of patients coming from home and 58% of those from the ward were admitted to the ICU. Approximately one-third of interventions involved five or more pulmonary veins and 57% involved five or more balloon dilations or stents. Regarding severity of disease, 14% of cases were performed on patients with moderate-to-severe RV dysfunction and 47% on patients with a severely elevated mean PA pressure.

Primary Outcome Analysis
Of 841 catheterizations involving PV interventions, at least one AE was reported in 208 (25%) cases and at least one serious AE was reported in 100 cases (12%, 95% CI 10-14%). Eighteen cases had greater than one serious AE. Major AEs (level 4) occurred in 14 cases (1.7%) and catastrophic AEs (level 5) in 1 case (0.1%) (Fig. 1). When evaluating the data at the patient level, 76 of the 240 patients experienced a serious AE (32%) during the study period, with major or catastrophic (level 4/5) AEs occurring in 15 patients (6%). Age less than 6 months, weight below five kilograms, low systemic arterial saturation, severely elevated mean pulmonary artery pressure, and pre-catheterization hospital LOS ≥ 1 day were associated with higher odds of serious AEs in univariate analysis (Table 2). There was no association between higher number of PVs undergoing intervention and serious AEs. On multivariable analysis, patients with age less than 6 months (OR 2.05, 95% CI 1.11-3.81, p = 0.02), low systemic arterial saturation (OR 1.52, 95% CI 1.02-2.27, p = 0.04), and severely elevated mean PA pressure (OR 1.74, 95% CI 1.16-2.63, p = 0.008) were more likely to have serious AEs (Table 3).

Secondary Outcome Analysis
In 333 transcatheter PV interventions performed in 121 patients between 2019 and 2021, high-level cardiorespiratory support at 48 h after the intervention was required in 30 cases (9%). Mechanical ventilation was needed after 20 cases (6%), vasoactive support after 21 cases (6.3%), and ECMO after 1 case (0.3%). Factors associated with this composite outcome included age younger than one year, weight below five kilograms, moderate-severe pre-catheterization RV dysfunction, low mixed venous saturation, pre-catheterization hospital LOS equal to or greater than one day, and ICU stay prior to catheterization. The number of PVs undergoing intervention or case duration were not associated with the composite outcome (Table 4). On multivariable analysis, hospital LOS prior to catheterization was a statistically significant predictor of the composite outcome (OR 33.1, 95% CI 7.1-155, p < 0.001). Patients younger than 1 year at time of intervention (OR 7.12, 95% CI 2.05-24.8, p = 0.002 for patients < 6 months; OR 5.41, 95% CI 1.98-14.8, p = 0.001 for patients 6-11.9 months) or with moderate-severe RV dysfunction (OR 3.97, 95% CI 1.52-10.4, p = 0.005) were more likely to achieve the composite outcome ( Table 5).

Discussion
In cardiac catheterizations involving PV interventions at a high-volume referral center, we found serious AEs were common (12%) and major/catastrophic AEs, including one patient death, were uncommon (1.8%). On a patient level, as most patients underwent multiple transcatheter interventions, 32% of patients experienced a serious AE and 6% experienced a major/catastrophic event. Most of the serious AEs were short lived and non-life threatening and their incidence should be balanced against the high mortality of aggressive disease. Young patients, patients with low systemic arterial saturations and patients with severely elevated mean pulmonary artery pressure at the time of catheterization were particularly vulnerable. Additionally, analysis of a smaller, more recent sample of this cohort found that hospitalized patients, younger patients, and patients with significant RV dysfunction were more likely to require highlevel cardiorespiratory support 48 h after the catheterization. Both the large sample size predictive analysis for the primary outcome and the 48-h high-level support outcome are, to our knowledge, the first to be performed in patients with PVS. These data may assist clinicians in anticipating peri-catheterization care needs and counseling families on the potential risks and expectations of the treatment plan for patients with PVS. The incidence of all level AEs of 25% is similar to a previous study from our institution analyzing a cohort of patients receiving PV interventions between 2005 and 2014 (28%; 114/406 cases, excluding access-related issues not accounted for by the authors) [11]. Notably, the incidence of stroke in our cohort (0.36% [3/841] of cases, 1.25% [3/240] of patients) was lower than that reported in this earlier, non-overlapping cohort from our center (2.5% [10/406] of cases, 6.25% [9/144] of patients) [11]. While not assessed prospectively, the findings of the prior study drove a change in institutional practice to continue anticoagulation 18 to 24 h post-catheterization using heparin or enoxaparin in the absence of bleeding; in theory to prevent post-intervention thrombus formation within the pulmonary veins. Further, all patients with PVS were managed with aspirin therapy with dosing continued through catheterization. There was only one stroke in the last 5 years (0.15% [1/648] of cases) following these institutional changes.
Younger age at the time of intervention was associated with both study outcomes. Prior studies have identified younger age at the time of diagnosis as a risk factor for mortality in patients with PVS [4,16,17], suggesting this may be a higher risk cohort. In addition, lower weight and smaller PVs in younger patients likely contribute to a more technically demanding procedure, potentially increasing the chances of serious AEs. It is reported that higher complexity surgical or transcatheter interventions are associated with a higher likelihood of AEs [18,19]. Lastly, more patients in this age group were admitted to the ICU or had a longer LOS prior to catheterization suggesting higher severity of illness that was not captured in this study.
Low systemic arterial saturation, a marker of severity of lung disease in PVS, was associated with serious AEs. Pulmonary venous obstruction precipitates capillary remodeling, flow redistribution, and pulmonary edema, leading to impaired oxygenation and V/Q mismatch [20,21]. All of these factors can worsen at the time of catheterization. Though we did not find an association between low systemic arterial saturation and need for high-level support after PV interventions, PVS patients with hypoxemia may require closer monitoring after interventions as suggested a Adverse events listed under both level 3 and level 4 severity differ in that level 4 events required invasive monitoring, unanticipated intubation, and/or major invasive procedures or transcatheter interventions to correct the condition by a recent study reporting higher ICU admission in this population [12]. Severely elevated PA pressures were also associated with a higher frequency of serious AEs. Pulmonary hypertension is an independent risk factor for mortality in PVS [22] and increases the risk of AEs during anesthesia, especially in younger patients undergoing interventional catheterizations [23]. Elevated PA pressures can additionally place patients at risk of pulmonary hemorrhage after balloon angioplasty due to increased pressure in the lung segment with balloon inflation. Notably, pulmonary hemorrhage was the most common serious AE in this cohort, occurring in 2.3% of cases and accounting for 17% (20/120) of the serious AEs.

Fig. 1 Frequency of Serious Adverse Events by Severity Level
Contrary to our hypothesis, the number of PVs undergoing intervention and the number of balloon and stent angioplasties were not predictors of serious AEs. Our findings may suggest that multiple interventions on a higher number of PVs are safe to pursue, further supported by the lack of association between longer case duration and serious AEs. It should be considered that longer-term risks associated with longer PVS cases, including exposure to higher doses of radiation and contrast, remain unexplored.
High-level cardiorespiratory support at 48 h after PV intervention was required in 9% of cases. A recent study from our institution using an overlapping cohort of patients found that ICU admission and mechanical ventilation immediately post-catheterization was as high as 50% in a cohort of patients from 2011 to 2021 [12]. However, the authors noted a significant decrease in recent years, coinciding with the period captured in our analysis (2019-2021). We found that younger age and worse RV dysfunction were predictors of highlevel cardiorespiratory support 48 h after catheterization. In patients with severe PVS and biventricular circulation, longstanding pulmonary hypertension can lead to RV dysfunction, significantly increasing the risk of cardiorespiratory failure and death in patients with PVS [22]. This is especially true in patients that are ill enough to require hospitalization, as suggested by the strong association of pre-catheterization hospital LOS of one day or greater with need for highlevel support. Patients that require inpatient therapy and monitoring may have advanced-stage PVS and/or additional comorbidities contributing to their illness severity that were not captured in this study.
The strengths of our study include the use of a comprehensive dataset from a referral center specialized in PVS management with a large sample size for a relatively rare disease. The limitations of this study include the use of an institutional catheterization database with retrospectively collected data, which limited available variables to those entered into the database. This also limited the sample size for the secondary outcome analysis of higher-level cardiorespiratory support after catheterization. The single-center design may limit the generalizability of results. Patient outcomes may be influenced by operator-related factors and advancement of catheterization techniques over time. Need for mechanical ventilation, vasoactive support, and ECMO were relatively rare outcomes in our cohort, and this may have limited identification of risk factors. Future studies should aim to include multicenter data to consider variation in center practices.

Conclusion
Serious adverse events during transcatheter PV interventions in patients with PVS are common, although major events such as stroke or death are uncommon. Younger patients and those with severe secondary hemodynamic sequelae of PVS are more likely to experience serious adverse events or require high-level cardiorespiratory support 48 h post-catheterization. These findings can be used to complement risk stratification tools to appropriately identify patients who may require particular attention during and after transcatheter pulmonary vein interventions.