Author + information
- Received February 21, 2017
- Revision received May 24, 2017
- Accepted May 24, 2017
- Published online August 10, 2017.
- Melinda J. Cory, MD,
- Yinn Khurn Ooi, MD,
- Michael S. Kelleman, MS, MSPH,
- Robert N. Vincent, MD,
- Dennis W. Kim, MD, PhD and
- Christopher J. Petit, MD∗ ()
- Division of Pediatric Cardiology, Department of Pediatrics, Children’s Healthcare of Atlanta, Sibley Heart Center, Emory University School of Medicine, Atlanta, Georgia
- ↵∗Address for correspondence:
Dr. Christopher J. Petit, Emory University School of Medicine, 1405 Clifton Road NE, Atlanta, Georgia 30322.
Objectives The aim of this study was to evaluate survival following catheter intervention in pediatric patients with pulmonary vein stenosis (PVS).
Background Despite aggressive surgical and catheter intervention on PVS in children, recurrence and progression of stenosis can lead to right heart failure and death. Clinicians continue to seek effective treatment options for PVS.
Methods A single-center, retrospective study was performed including all patients <18 years of age who underwent catheter intervention (balloon angioplasty and bare-metal stent and drug-eluting stent insertion) on PVS. Endpoints included death, vein loss, and rate of reintervention.
Results Thirty patients underwent intervention (balloon angioplasty, n = 9; bare-metal stent, n = 5; drug-eluting stent, n = 16) at a median age of 6.4 months (4.3 to 9.9 months). Median follow-up duration was 30.6 months (77 days to 10.5 years). Fourteen patients (47%) died at a median of 2.0 months (0.4 to 3.2 months) following intervention. There was no association between DES placement and survival (p = 0.067). Reintervention (catheter or surgical) was associated with improved survival (p = 0.001), with a 1-year survival rate of 84% compared with 25% for no reintervention. Vein loss occurred in 34 of 58 (59%) veins at a median of 3.3 months (1.0 to 5.0 months). One-year vein survival was higher with DES implantation (p = 0.031) and with reintervention (p < 0.001).
Conclusions DES implantation at first catheter intervention appears to be associated with improved vein survival but may not result in improved patient survival. However, reintervention appears to be associated with improved patient survival and vein patency, suggesting that despite mode of treatment, frequent surveillance is important in the care of these patients.
In children, pulmonary vein stenosis (PVS) is uncommon, presenting either as primary (idiopathic) PVS or secondary (post-surgical) PVS, usually following repair of total or partial anomalous pulmonary venous return. PVS is progressive in nature and can result in secondary pulmonary hypertension, right heart failure, and death (1–4). Surgical intervention at some centers is considered the primary treatment for both types of PVS (4). Despite aggressive surgical intervention, PVS is often recurrent (5).
Transcatheter interventions have been used to treat PVS, although outcomes have been variable. Outcomes following balloon angioplasty (BA) for PVS were reported as early as 1982, with poor long-term pulmonary vein (PV) patency (6–8). Stent implantation has also been reported, although many investigators report a high rate of neointimal proliferation within the implanted PV stents (9). The largest study evaluating stent efficacy for PVS in children found stent diameter to be a predictor of freedom from occlusion and reintervention (10). Studies looking at adult patients with acquired PVS found stent placement to be superior to BA, especially when larger stents can be deployed (11,12).
Histologic evaluation of PVS pathology suggests that myofibroblastic proliferation contributes importantly to the disease and may play a role in aggressive recurrence (13). Recent reports have suggested that inhibiting proliferation could temper the progression of disease. As a result, many centers have used antiproliferative and anti-inflammatory strategies when treating PVS. Drug-eluting stents (DES), initially introduced to alleviate coronary artery neointimal hyperplasia (14–17), have been implanted in infants and small children with PVS, who have had particularly poor outcomes (1,3,4,10,18). In a porcine model, DES were found to reduce in-stent stenosis compared with bare-metal stents (BMS) when placed in PVs (19). Outcomes following DES implantation in children with PVS are limited to case reports and small series that constitute a subset of larger reviews. The objective of our study was to compare survival following transcatheter intervention, including BA and BMS and DES implantation, in pediatric patients with PVS.
We retrospectively reviewed data from all patients <18 years of age who underwent catheter intervention for PVS from January 2005 to April 2016 at Children’s Healthcare of Atlanta. The patients were identified through review of our database of catheterization procedures. Patient characteristics and demographics were noted, including age and weight at first PV intervention, the presence of associated congenital heart disease, and history of prior PV intervention. PV disease was characterized as: 1) unilateral or bilateral; 2) primary or secondary (defined as post-surgical, either for total or partial anomalous pulmonary venous return or primary PVS); 3) and by the number of PVs intervened upon both at index catheter-based intervention and later. Approval was obtained from the Institutional Review Board at Children’s Healthcare of Atlanta.
Techniques and indications for PV dilation and stent implantation have been described previously (20–24). PVS was confirmed during catheterization with pulmonary artery wedge angiography or with directed injections into each PV. Access to the left atrium was through either a patent foramen ovale, atrial septal defect, or via transseptal puncture. Choice of intervention was determined by the operator. Balloon or stent diameter was also determined by the operator, on the basis of the angiographic size of the vessel. Type of DES (i.e., everolimus eluting vs. paclitaxel eluting vs. zotarolimus eluting) was not chosen by the operator for specific immunomodulatory effect, but rather DES type was selected by availability. All digital angiograms from the index case were reviewed. Areas of obstruction were classified as discrete or long segment. Measurement of the area of stenosis was made before and after angioplasty or stent placement. For long-segment stenosis, the location of the balloon or stent waist was used as the site of vessel measurement.
Patient demographics, diagnosis, mortality, and total reintervention were reported with each patient (i.e., not each treated PV) as the index unit. Rate of atresia and reintervention on individual veins were reported with the stented or angioplastied vein as the index unit.
The primary outcome assessed was mortality following transcatheter PV intervention. Secondary outcomes included PV reintervention (both transcatheter and surgical) and vein loss (a composite of vein atresia or patient death). For the purpose of testing the hypothesis that immunomodulatory stents (i.e., DES) have superior efficacy, the cohorts undergoing BMS implantation and undergoing BA were combined as the “nonimmunomodulatory” cohort. Data are expressed as median (interquartile range) or as number (percentage). When comparing 2 variables, categorical variables were analyzed using the Fisher exact test, and nonparametric continuous variables were assessed using the Mann-Whitney U test. Time-dependent survival analyses were performed using Kaplan-Meier log rank analysis. A p value of <0.05 was considered to indicate statistical significance. Data analysis for 2 variables was performed using SPSS version 23.0 (IBM, Armonk, New York). When comparing 3 variables, standardized mean differences (SMD) were calculated in R version 3.3.2 using the package “tableone”. An SMD of <0.49 was considered a slight or small difference, an SMD of 0.50 to 0.79 was considered a moderate difference, and an SMD >0.80 was considered a large difference.
Between 2005 and 2016, 30 patients underwent catheter-based intervention for PVS at a median age of 6.4 months (4.3 to 9.9 months). Fifteen of the patients (50%) had undergone previous surgical PV repair, including total or partial anomalous pulmonary venous return repair (n = 12 [40%]) and surgical repair of primary PVS (n = 3 [10%]). The remaining 15 patients (50%) underwent primary catheter intervention, and these patients were all diagnosed with idiopathic PVS.
Among the 30 patients, 9 (30%) underwent BA, 5 (17%) underwent BMS implantation, and 16 (53%) underwent DES implantation. Four patients (13%) had multiple therapies for different PVs at initial catheter-based intervention. Demographic data are summarized in Table 1.
Among the 30 patients, 58 individual veins were intervened upon, with a median of 2 (range 1 to 5) PVs per patient. The veins were characterized as either discrete (n = 42 [72%]) or long segment (n = 16 [28%]). Thirty veins (52%) underwent DES implantation. The DES agents included everolimus (n = 21 [70%]), zotarolimus (n = 7 [23%]), and paclitaxel (n = 2 [7%]). Ten veins (17%) underwent BMS implantation, and 18 (31%) underwent BA alone. Only 1 patient underwent cutting BA of a single PV; the remaining BA interventions used conventional balloons. The DES were smaller in diameter than the BMS and balloons for BA (SMD = 0.890) (Table 2). Although the DES had smaller diameters than the balloons used for BA, veins that underwent BA had smaller immediate post-intervention diameters than veins that underwent DES or BMS implantation (SMD = 0.201) (Table 2). There was no difference in the rate of reintervention by manner of catheter intervention. Vein intervention characteristics are summarized in Table 2. Detailed information on intervention type provided in Online Table 1.
Follow-up duration for the cohort was a median of 30.6 months (range 77 days to 10.5 years). Fourteen patients (47%) died at a median of 2.0 months (0.4 to 3.2 months) following intervention. The causes of death for all of these patients were directly related to PVS, including right heart failure, pulmonary hypertension, and respiratory failure. Overall survival rates at 1 month and 1 year were 83% and 51%, respectively (Figure 1). Mortality was not associated with age or weight at first catheter-based intervention, single-ventricle physiology, bilateral vein involvement, or prior surgical intervention. There was also no significant association between intervention type and mortality. Survival was better in patients who underwent repeat intervention (surgical or catheter based) following primary catheter intervention (p = 0.001). Mortality data are summarized in Table 3.
The DES cohort was compared with patients undergoing BA and BMS placement, who combined were considered the nonimmunomodulatory cohort in a Kaplan-Meier survival analysis. The survival rates at 1 year were 68% for the DES group and 33% for the nonimmunomodulatory group (log-rank p = 0.067) (Figure 2).
The survival rate at 1 year for patients who underwent reintervention (surgical or catheter) was 84% compared with 25% for patients who did not undergo reintervention (log-rank p = 0.007). The relative risk for 1-year survival in patients who underwent reintervention compared with those who did not was 4.09 (95% confidence interval: 1.43 to 11.69). This survival advantage was found to be greater with increasing number of interventions, with survival rate at 1 year of 100% in patients with ≥3 interventions, 53% in patients with 2 interventions, and 25% in patients with 1 intervention (Figures 3A and 3B).
In an attempt to understand if reintervention was truly protective against mortality or if survivors merely were afforded an opportunity for reintervention, we excluded patients who died within 1 month of primary catheter intervention and performed Kaplan-Meir analysis on the remaining cohort. This was based on our clinical practice of encouraging repeat catheterization 6 weeks after initial intervention. This analysis likewise demonstrated a survival advantage in patients who underwent reintervention, with a 1-year survival rate of 84% compared with 36% in patients who did not undergo reintervention (log-rank p = 0.013).
Among the 31 veins in 16 survivors, 7 veins (22%) were found to be atretic in 7 patients (44%). Atresia was noted by angiography and occurred at a median time of 12.1 months (7.3 to 31.2 months). One of the patients who developed left upper and lower PV atresia ultimately underwent pneumonectomy because of recurrent, severe hemoptysis. In 2 of the 14 patients who died, 4 of 27 veins (15%) had previously been noted to be atretic by angiography before death. Hence, a total of 11 veins in 9 patients became atretic following intervention; the median time to discovery of atresia in these 11 veins was 7.3 months (2.1 to 14.8 months). Ten patients who died did not undergo repeat catheterization following the intervention and thus had no documentation of the status of the PV at the time of death.
Of the 9 patients with documented atresia, 7 (78%) had evidence of decompressing collateral formation by angiography. Collateral formation included interlobar pulmonary venovenous collateral vessels (n = 4) (Figures 4A and 4B) and PV–to–systemic vein collateral vessels (n = 3) (Figure 4C). Of the 7 patients with collateral formation, 5 (71%) were alive at most recent follow-up.
Vein loss, defined as actual PV atresia or PVs of uncertain status in deceased patients, occurred in 34 veins (59%) at a median time of 3.3 months (1.0 to 5.0 months). Overall vein patency rates at 1 month and 1 year were 86% and 44%, respectively. Vein loss was not associated with type of stenosis (discrete or long segment), intervention type, stent or balloon diameter, or pre-intervention measurement of stenosis. Post-surgical PVS and veins that had larger post-intervention measurements were associated with vein survival (p = 0.03 and p = 0.013, respectively). Consistent with overall patient mortality, veins that underwent reintervention (catheter based or surgical) had lower rates of vein loss (p < 0.001). Vein survival is summarized in Table 4.
The vein survival rate at 1 year for individual veins that underwent more than 1 procedure was 88% compared with a 1-year survival rate of 18% for veins that did not undergo reintervention (log-rank p < 0.001). The relative risk for 1-year vein survival when a vein had more than 1 intervention was 5.00 (95% confidence interval: 2.37 to 10.55) compared with the veins that had only 1 catheter intervention. This survival advantage was again noted to be associated with increasing number of interventions, with 1-year survival for veins undergoing ≥3 procedures of 100% compared with 75% for 2 procedures and 18% for 1 procedure (Figures 5A and 5B).
When comparing vein loss by intervention type, 1-year vein survival in PVs treated with primary DES implantation was higher than PVs treated with BA or BMS (nonimmunomodulatory group) (59% vs. 26%, log-rank p = 0.031) (Figure 6).
Fourteen patients (47%) underwent a total of 25 reinterventions, including additional DES placement (n = 5 [20%]), BMS placement (n = 2 [8%]), stent dilation or BA (n = 13 [52%]), and PV surgery (n = 5 [20%]) for recurrent PVS. Among the patients who had surgical reintervention, 3 patients initially had catheter reintervention ultimately followed by surgical PV repair, and 2 patients had surgical PV repair after first catheter-based intervention. Surgery included pericardial well procedure in all 5 patients with stent removal in all 3 patients with existing PV stents. In the 12 patients who underwent catheter reintervention, 19 veins were reintervened upon because of recurrent stenosis, and 5 of these patients had new vein involvement that required intervention. Median time to first reintervention was 3.1 months (2.2 to 7.0 months).
Within the DES cohort, 8 of 16 patients (50%) underwent catheter reintervention, and 3 of 16 patients (18%) had PV surgery, with a median time between interventions of 3.5 months (1.3 to 7.7 months) after DES placement. Within the BMS cohort, 3 of 5 patients (60%) underwent catheter reintervention, with a median of 2.4 months (2.4 to 17.5 months) between interventions. Within the BA cohort, 1 of 9 patient (11%) underwent catheter reintervention, and 2 of 9 patients (22%) underwent PV surgery, with a median time between interventions of 4.8 months (3.9 to 5.7 months). Rates of reintervention are summarized in Table 5. Reintervention rates within individual veins are summarized in Table 6.
When looking at rates of surveillance following the initial catheter intervention, 19 patients (63%) had at least 1 follow-up catheterization at a median of 2.9 months (2.1 to 3.3 months). At first follow-up catheterization, only 8 of 19 patients (42%) underwent a reintervention, and 2 of these 8 patients required intervention on newly affected veins. The remaining 11 patients (58%) did not undergo repeat intervention at first follow-up catheterization. The time to first catheterization was no different in patients who underwent reintervention (2.4 months [2.1 to 3.1 months]) compared with those who only had diagnostic catheterization (2.9 months [2.1 to 5.9 months]) (p = 0.351). There was no difference in mortality among the patients who had an intervention (n = 2 of 8 [25%]) at first follow-up catheterization compared with those who did not have an intervention (n = 2 of 11 [18%]) (p = 0.999); however, vein survival was greater in veins that underwent intervention at first follow-up catheterization (p = 0.04). Rates of reintervention are summarized in Table 5.
This retrospective study was undertaken to evaluate the outcomes of children who underwent catheter-based intervention for PVS. We had a relatively large cohort of pediatric patients at a single institution with both primary and post-surgical PVS. Although many different types of intervention have been attempted for PVS, all of our patients underwent BA, BMS placement, or DES placement.
Our patient population was diagnosed with PVS at a young age, which has been shown in previous studies to be an independent predictor of increased mortality compared with older age at diagnosis (8,18,25). Similar to prior studies, single-ventricle physiology was not associated with increased mortality in our patient population. Interestingly, our data indicated that the presence of bilateral PVS was not associated with mortality. This finding is different from previous publications, which indicate that diffuse, bilateral PVS is a risk factor for mortality (5,18,25).
Although transcatheter procedures for PVS have historically been thought of as primarily palliative (1,4), we have shown reasonable outcomes in our patients who underwent DES placement. Consistent with previous reports, BA alone was shown to be acutely successful, but midterm success is limited (6,8,11,20). Our survival analysis shows early mortality with all types of transcatheter interventions and stable survival beyond 6 months from intervention. Once patients reach 1 year post-intervention, survival appears likewise to stabilize. We found that patients who return for surveillance catheterization and undergo reintervention, when indicated, have improved survival than those who do not. Because of this, rates of reintervention in our patient population for all intervention types are high in survivors, which is consistent with previous reports (6,9,10). The early mortality we have seen in all intervention types in combination with the survival advantage we found in patients who receive reintervention, would suggest that early routine surveillance catheterization should be undertaken so that restenosis can be detected and intervened on early. However, our study is not powered to determine the precise relationship between reintervention and survival. That is, reintervention may reasonably be considered an important element in the transcatheter approach to PVS to maintain or optimize PV patency. In contrast, one could argue that the survival bias colors the current analysis and that nonsurvivors never had an opportunity for reintervention. Our focused analysis excluding early post-intervention mortality was aimed at addressing this significant limitation and also showed improved survival in patients who underwent reintervention.
Although mortality rates decreased significantly by 1 year post-intervention, the rates of vein loss persisted beyond 1 year. That vein loss rates do not mirror survival is interesting and, perhaps, indicates intrinsic compensation to PVS and atresia. The late and possibly gradual development of PV atresia may provide sufficient time for the patient to develop venovenous collateral vessels to compensate for the change in their physiology, as we noted in 5 of our survivors with documented atresia. This collateralization may in fact be protective; however, our study is insufficiently powered to address this question.
Balasubramanian et al. (10) showed that a stent diameter >7 mm was a predictor of patency following stent placement for PVS (10). Accordingly, adult studies have also shown a size-dependent success rate for larger stents (11,12). Although it is likely beneficial, with regard to rates of restenosis and the need for reintervention, to have a larger stent diameter, there are situations in which patients have small vessels that cannot accommodate a larger stent. That was, of course, the case in the majority of our patients. The majority of stents placed in our patient population were coronary stents, with only 1 stent >7 mm in diameter. We were interested in success rates in different types of stents and/or BA. Although we did not evaluate stent diameter as a marker of success, the median stent lumen and balloon diameter among the groups was similar, so the survival rate is primarily showing the difference between the interventions as opposed to size.
There is an inherent limitation to DES due to the small maximum luminal diameter in currently available DES, which have a maximal rated diameter of 4.0 mm. In our study, DES implantation appears to be superior to BMS implantation or BA when assessing the rate of vein loss. This suggests that the immunomodulatory effects of DES may be beneficial for PVS when comparing with BMS of similar diameter. This is consistent with previous reports in adult coronary studies and a porcine model with stented PVs, which demonstrated decreased in-stent stenosis in DES compared with BMS (16,19). Additionally, the pathological evaluation by Sadr et al. (13) showing myofibroblastic proliferation as the etiology of PVS suggests that antiproliferative therapy would provide benefit to patients with PVS. The local release of antiproliferative drugs with DES may attenuate the progression of disease, resulting in improvement in survival. Although the mechanism of PVS in adults is different from that in our pediatric population, De Potter et al. (26) also found DES to be an effective treatment in adults for PVS following radiofrequency ablation despite the limitation in the stent luminal diameter. More studies will need to be undertaken to further evaluate the efficacy of DES for inhibiting the progression of PVS. This observation also suggests that it may be warranted to evaluate drug-eluting balloons as a therapy for PVS, as this may provide the benefit of delivering antiproliferative drugs to the site of injury without the inherent limitation of a small luminal diameter with DES.
As this was a retrospective study, the patients were not randomized to each intervention type. There were no standardized criteria for intervention on the stenotic PVs, and we were unable to control for interoperator variability and decision making, including which veins to intervene on and which method to use; however, at our institution, surgical repair of PVS is usually considered first-line. All post-surgical patients included in our study were discussed with the surgical team regarding candidacy for surgical repair of PVS before undergoing catheter-based intervention. Additionally, we were unable to control for both practitioner- and patient-specific factors that resulted in variations in follow-up time. Our overall sample size was small, but for this rare disease entity, this is a relatively large cohort that has undergone catheter-based intervention for PVS.
WHAT IS KNOWN? PVS in pediatric patients is a progressive disease that is difficult to treat. Evaluation of catheter-based intervention has been limited and has shown suboptimal survival in pediatric patients.
WHAT IS NEW? Despite size limitations of DES, this intervention is a reasonable treatment option in pediatric patients with PVS. Frequent surveillance catheterization is important to identify the need for reintervention early when restenosis occurs, as this may help improve survival in these patients.
WHAT IS NEXT? Because of the limited studies available, larger multicenter trials would be beneficial in the further evaluation of catheter-based intervention on PVS in pediatric patients. Additionally, evaluation of newer modalities, including drug-eluting BA, should be further evaluated.
For a supplemental table, please see the online version of this article.
All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- balloon angioplasty
- bare-metal stent(s)
- drug-eluting stent(s)
- pulmonary vein
- pulmonary vein stenosis
- standardized mean difference
- Received February 21, 2017.
- Revision received May 24, 2017.
- Accepted May 24, 2017.
- 2017 American College of Cardiology Foundation
- Latson L.A.,
- Prieto L.R.
- Devaney E.J.,
- Ohye R.G.,
- Bove E.L.
- Drossner D.M.,
- Kim D.W.,
- Maher K.O.,
- Mahle W.T.
- Seale A.N.,
- Webber S.A.,
- Uemura H.,
- et al.
- Balasubramanian S.,
- Marshall A.C.,
- Gauvreau K.,
- et al.
- Sousa J.E.,
- Costa M.A.,
- Abizaid A.C.,
- et al.
- Sonoda S.,
- Morino Y.,
- Ako J.,
- et al.
- Suzuki T.,
- Kopia G.,
- Hayashi S.,
- et al.
- Lock J.E.,
- Bass J.L.,
- Castaneda-Zuniga W.,
- Fuhrman B.P.,
- Rashkind W.J.,
- Lucas R.V. Jr..
- O’Laughlin M.P.,
- Perry S.B.,
- Lock J.E.,
- Mullins C.E.
- O’Laughlin M.P.,
- Slack M.C.,
- Grifka R.G.,
- Perry S.B.,
- Lock J.E.,
- Mullins C.E.
- Feltes T.F.,
- Bacha E.,
- Beekman R.H. III.,
- et al.