Author + information
- Received January 7, 2018
- Revision received May 3, 2018
- Accepted May 8, 2018
- Published online August 20, 2018.
- Erin A. Fender, MD, MSa,
- R. Jay Widmer, MD, PhDa,
- David O. Hodge, MSb,
- Douglas L. Packer, MDa and
- David R. Holmes Jr., MDa,∗ ()
- aDepartment of Cardiovascular Diseases, Mayo Clinic, Rochester, Minnesota
- bDepartment of Health Sciences Research, Mayo Clinic, Jacksonville, Florida
- ↵∗Address for correspondence:
Dr. David R. Holmes, Jr., Department of Cardiovascular Diseases, Mayo Clinic, 200 First Street Southwest, Rochester, Minnesota 55905.
Objectives This study sought to evaluate the sensitivity of noninvasive imaging in the assessment of severely stenosed and occluded pulmonary veins, and examine clinical outcomes following percutaneous intervention.
Background PV stenosis (PVS) is a rare complication of atrial fibrillation ablation, but is associated with significant morbidity. Patients present with nonspecific pulmonary symptoms that can result in delayed diagnosis and progression to PV occlusion. The assessment and management of PV occlusion has rarely been described.
Methods This was a prospective observational study performed from 2000 to 2014.
Results Computed tomography identified 124 patients with severe PVS, including 46 patients with at least 1 occluded vein. Patients with PV occlusion more frequently presented with cough (64.1% vs. 32.8%; p = 0.002) and hemoptysis (39.1% vs. 14.1%; p = 0.0015) and were more likely to have pulmonary parenchymal consolidation (77.3% vs. 41.7%; p = 0.0002). Intervention was attempted in 65 occluded veins and a residual microchannel was identified in 22 (34.0%). Balloon angioplasty was performed in 11, and 11 were treated with stenting. Over 3 years the rates of restenosis were similar for patients with PVS and PV occlusion (47.0% vs. 35.0%; p = 0.24). Among patients with PV occlusion, stenting significantly reduced the rate of restenosis (hazard ratio: 3.97; 95% confidence interval: 1.14 to 13.85; p = 0.03).
Conclusions Veins deemed occluded on noninvasive imaging require invasive characterization, as residual microchannels may be present in one-third of patients. In patients with a microchannel, intervention can be performed with either balloon angioplasty or stenting. Recurrence remains a common problem; however, stenting significantly reduces the rate of subsequent restenosis.
Pulmonary vein stenosis (PVS) is a rare complication of ablation for atrial fibrillation (AF), but is associated with significant morbidity (1). The incidence of PVS has decreased substantially with improvements in ablation techniques, and is now estimated to complicate between 0.3% to 3.4% of AF ablation procedures (2–4). As the incidence of PVS has fallen, use of routine post-ablation computed tomography (CT) screening has also decreased. The Heart Rhythm Society consensus statements no longer recommend universal screening for all post-ablation patients, but rather endorse a symptom guided approach to imaging (5,6). However, a significant portion of patients with severe PVS may have few symptoms, or present with insidious onset of nonspecific pulmonary symptoms resulting in misdiagnosis (1,7–12). Delayed diagnosis can result in critical luminal loss and progression to total pulmonary vein (PV) occlusion. The optimal method for diagnosing and managing PV occlusion has not previously been well described.
PVS is frequently diagnosed with a contrast enhanced CT scan timed for PV opacification, as it provides excellent spatial resolution and allows for 3-dimensional reconstructions and generation of en face views of the venous lumen for more accurate determination of luminal diameter (13,14). However, CT may be insensitive for near-total occlusions due to sluggish contrast flow through the residual microlumen. Therefore, we examined our own center’s experience with the diagnosis of PV occlusion using CT versus invasive angiography. Additionally, we evaluated acute procedural success for interventions to veins deemed occluded by CT, and defined long-term patency following balloon angioplasty (BA) and stenting.
Our study and procedural methods have been previously described in detail (1,15–18). In brief, this was a prospective, single-center study of 124 consecutive patients with PVS who were evaluated at the Mayo Clinic (Rochester, Minnesota) between February 2000 and November 2014. Only patients with severe PVS after AF ablation were included; patients with congenital PVS or mediastinal fibrosis were excluded. Severe stenosis was defined as >75% luminal narrowing on a dedicated contrast enhanced CT scan with image acquisition timed for PV enhancement. Fifty-three of the 124 subjects had undergone AF ablation at the Mayo Clinic, whereas the other 71 patients were referred from outside institutions for treatment of PVS. Local patients who had AF ablation at Mayo Clinic underwent routine post-ablation CT screening for PVS. Patients who were referred to Mayo Clinic for ablation were encouraged to undergo CT screening with their referring providers. Patients referred from outside institutions after the diagnosis of PVS typically underwent CT assessment only after symptoms had developed. Two patients had been treated with cryoballoon ablation; all others had received radiofrequency ablation. Following intervention all patients underwent scheduled follow up clinical assessment and imaging at 3 to 4 months and 9 to 12 months. Patients provided written consent at the time of enrollment in this Institutional Review Board–approved study. Patients were followed from the time of initial PVS diagnosis until either the date of last clinical follow-up, or through April 2015 if still receiving ongoing care.
Statistical analysis was carried out on both a per-patient and a per-vein level. Patient characteristics were compared between subjects with severe nonocclusive PVS and those who had at least 1 occluded PV on the diagnostic CT. If a patient with PV occlusion also had 1 or more severely stenosed veins, the patient was analyzed as part of the PV occlusion cohort. Vein-specific characteristics were compared between severely stenosed veins and veins characterized as occluded on the index CT, such that a single patient might have had veins analyzed in 2 different cohorts.
All variables were analyzed using software from SAS version 9.4 (SAS Institute, Cary, North Carolina). Normally distributed continuous variables are reported as a mean ± SD, whereas nonuniformly distributed continuous variables are reported as median and interquartile range (IQR). Nonparametric estimates were compared using the Wilcoxon test. Discrete variables are reported as a frequency. Fisher exact testing was used to assess differences in binary variables. Survival free from PV restenosis was assessed using Kaplan-Meier curves and a Cox proportional hazards model. The risk for restenosis was evaluated on a per-vein level. To account for multiple vessels in the same individual the robust sandwich estimate of SE was used to test individual parameters. Significance of difference was defined as a p value of <0.05.
Our procedural approach has been reported in detail (17,18). Briefly, patients were loaded with 300 to 600 mg clopidogrel before the procedure. Vascular access was obtained using 2 femoral venous sheaths (8-F and 10-F) to facilitate passage of a 10-F intracardiac echocardiography catheter, and an 8-F transseptal sheath (SL1, St. Jude Medical, St. Paul, Minnesota). After crossing the septum, heparin was administered to a target activated clotting time of 250 to 350 s. A 6-F multipurpose catheter was advanced via the transseptal sheath to the orifice of the affected PV. A combination of contrast injections and intracardiac echocardiography were used to identify the venous ostium. The severity of stenosis was visually assessed by contrast injection, and by Doppler assessment of PV flow (Figure 1A, Online Video 1). A 0.035-inch guidewire was then inserted into the PV. If the 0.035-inch guidewire could not be passed, cannulation with a Glidewire (Terumo Cardiovascular Systems Corporation, Ann Arbor, Michigan) and 0.014-inch coronary wire was attempted (Figure 1B, Online Video 2). If a patent lumen was not identified with contrast injections, Doppler imaging failed to show venous return, and no wires could be passed after multiple attempts, the vein was determined to be fully occluded and no further intervention was attempted. If a wire was successfully passed into the vein and a catheter could be advanced beyond the stenosis, a pressure gradient was measured. If the lumen was too small to permit passage of a catheter, direct angioplasty was carried out. Serial balloon dilations were performed to achieve a goal diameter of 10 mm (Figure 1C, Online Video 3). The response to intervention was assessed by repeat angiography and reassessment of the pressure gradient (Figure 1D, Online Video 4). If there was persistent stenosis >20%, or significant venous elastic recoil, a stent was deployed. Stents were sized by the largest balloon diameter achieved, typically this was a 10-mm diameter bare-metal stent intended for use in peripheral interventions. In some cases, smaller bare-metal stents for biliary interventions or coronary drug-eluting stents were used when a densely fibrotic vein could not be dilated to a larger size (1).
One hundred and twenty-four patients with severe PVS were identified and followed prospectively for a median duration of 4.6 (IQR: 1.8 to 9.5) years (Figure 2). Seventy-eight patients had severe nonocclusive PVS, and 46 patients had at least 1 occluded PV by CT assessment. Baseline and demographic characteristics for all patients are available in Table 1. There was no difference in the time from ablation to diagnosis for patients with PV occlusion versus PVS (3.5 ± 5.3 months vs. 5.5 ± 5.3 months; p = 0.06). Symptoms were present in 89.0% of patients with PV occlusion and in 78.0% of those with severe PVS (p = 0.12) (Figure 3). Compared with patients with severe PVS, patients with PV occlusion more frequently presented with cough (64.1% vs. 32.8%; p = 0.002) and hemoptysis (39.1% vs. 14.1%; p = 0.0015). Additionally, patients with PV occlusion more frequently had radiographic parenchymal lung consolidation on chest CT (77.3% vs. 41.7%; p = 0.0002).
Of the 46 patients with PV occlusion by CT, intervention was not attempted in 4 patients (n = 5 PVs) due to perceived futility. The remaining PV occlusion patients went onto invasive catheterization with an intervention attempted in 65 occluded veins. Forty-three veins were confirmed to be occluded by a combination of angiography, intracardiac Doppler, and direct wire probing. However, in 22 veins (33.8%) a microchannel was identified. The pre-intervention pressure gradients were significantly higher in veins deemed occluded by CT (16.7 mm Hg vs. 11.6 mm Hg in severely stenosed vein; p = 0.03). In all 22 of these veins (100%) a successful intervention was performed.
Eleven occluded veins were treated with BA (median diameter 8.00 [IQR: 6.00 to 10.00] mm), and 11 were treated with stenting (median diameter 7.00 mm [IQR: 3.50 to 10.00] mm). Compared with veins with severe PVS, the maximum achieved balloon diameter was smaller (median diameter 8.00 [IQR: 6.00 to 10.00] mm vs. 10.00 [IQR: 8.00 to 10.00] mm; p = 0.03), and the stent size was smaller (7.00 [IQR: 3.50 to 10.00] mm vs. 10.00 [IQR: 9.75 to 10.00] mm; p = 0.01).
At 3 years there was no significant difference in the rate of restenosis for occluded and severely stenosed veins (47.0% vs. 35.0%; p = 0.24) (Figure 4). However, for occluded veins treated with BA, restenosis occurred in 64.0% as compared with 30.0% for veins treated with stent (hazard ratio: 3.97; 95% confidence interval: 1.14 to 13.85; p = 0.03) (Figure 5).
There were 18 procedural complications in the overall cohort, including 14 in patients treated for PVS and 4 in patients treated for PV occlusion (p = 0.19). Complications in the PV occlusion group included PV perforation requiring emergent surgical intervention (n = 1), pericardial tamponade (n = 1), and stent dislodgement (n = 2).
PVS has a nonspecific presentation and symptoms can be mistaken for other more common pulmonary illnesses (1). Delayed diagnosis allows for progression to PV occlusion which is associated with development of lung infarction manifesting as severe chest pain, cough, and hemoptysis. Although total occlusion may preclude percutaneous intervention, one must be careful to avoid preliminary misdiagnoses of PV occlusion based on a CT scan alone. This study demonstrates that invasive angiography is an essential tool for further assessment of veins deemed occluded by CT, as residual microchannels may be present in one third of patients, allowing for successful PV intervention to be performed. Additionally, we found that the long-term risk of restenosis after a successful intervention is similar in occluded and severely stenotic veins, underscoring the importance of prompt diagnosis and treatment before total occlusion. Finally, we demonstrate that stenting significantly reduces the risk of recurrence for both severe PVS and PV occlusion.
Previous small studies have suggested that CT may be insensitive to subtotal PV occlusion; however, no previous study has systematically evaluated the role of invasive PV angiography, procedural outcomes, or the long-term efficacy of stenting or BA in veins deemed occluded by CT.
In a study using pulmonary artery wedge angiography in 80 patients with PV occlusion by CT, 27 patients (34.0%) were found to have a microchannel (9). Additionally, 2 previous small studies have documented successful interventions in veins deemed occluded by CT. In a study of 18 patients with PV occlusion by CT, an intervention was attempted in 15, including 8 patients who were treated with stenting and 7 who were treated with BA (19). At 14-month follow-up, restenosis occurred in 46% of treated veins; however, the recurrence rate was not analyzed according to the type of intervention. A second series of 3 patients with 5 occluded PVs reported a 100% acute procedural success rate (4 veins treated with BA, 1 vein treated with a 7-mm stent) (7). However, all of these veins developed multiple recurrences and required repeat interventions with stent implantation or BA to areas of in-stent stenosis, and 1 stented vein progressed to a chronic total occlusion. These studies confirm the importance of invasive characterization of veins thought to be occluded by CT, and when feasible the importance of percutaneous interventions to maintain venous patency. Recurrence remains a common problem in patients with both PVS and PV occlusion; however, stenting can significantly reduce the likelihood of restenosis.
This study has several important limitations. This was an observational cohort study with no control arm. At our institution all post-ablation patients are screened for PVS with a CT scan at 3 months. However, a significant portion of the study population was referred from outside institutions where CT scans were obtained only in symptomatic patients. Therefore, we do not know the true incidence of PVS or PV occlusion in post-ablation patients. Additionally, within our cohort not all patients with PV occlusion were taken for invasive characterization. In 4 patients (veins), intervention was never attempted and therefore we do not know if a microchannel was present. Next, we defined PV occlusion as being identified at the time of the post-ablation CT scan. Some patients analyzed in the severe PV stenosis group may have experienced disease progression resulting in new venous occlusion in the interval between the diagnostic CT scan and the interventional procedure. Fortunately, the delay between CT and intervention is typically short (2 to 5 days); however, we cannot exclude that some cases of PV occlusion were analyzed in the severe PVS group. A final limitation was the sample size. Specifically, we report an absolute difference of 12% in the 3-year rate of restenosis between occluded and severely stenosed veins (p = 0.24); however, to have 80% power to detect that magnitude of a difference would require 250 veins in each of the 2 study groups. It is possible that we did not observe a significant difference in this outcome due to lack of power.
Delayed recognition of PV stenosis after AF ablation can result in severe stenosis and progression to symptomatic venous occlusion. Noninvasive imaging relies on delayed image acquisition timed for contrast enhancement of the PVs. However, CT is insensitive for near total occlusion, and invasive assessment of PV occlusion is required, as small microchannels may be present allowing for successful percutaneous intervention in one third of patients. Recurrent PVS is common and difficult to treat; however, index stenting can significantly reduce the rates of restenosis.
WHAT IS KNOWN? PVS is a rare but morbid complication of AF ablation. Previous studies have demonstrated that percutaneous interventions can successfully treat PVS; however, restenosis is common. PV occlusion is less frequently observed and invasive characterization may not be performed due to a perceived low likelihood for successful intervention. The optimal method of assessing and managing patients with PV occlusion has not previously been described.
WHAT IS NEW? Patients were characterized as having PV occlusion when contrast-enhanced CT failed to demonstrate opacification of the venous lumen. Patients were then taken for invasive characterization, which included angiography, Doppler imaging, and wire probing. Using this technique, we identified a residual microchannel permitting successful intervention in 34% of patients who had been initially characterized as having PV occlusion. Over 3 years, the rates of restenosis were similar for patients with PVS and PV occlusion (47.0% vs. 35.0%; p = 0.24). Among patients with PV occlusion stenting resulted in a lower rate of restenosis compared with balloon angioplasty (hazard ratio: 3.97; 95% confidence interval: 1.14 to 3.85; p = 0.03).
WHAT IS NEXT? This study highlights the need for invasive characterization in patients diagnosed with PV occlusion after AF ablation. Percutaneous interventions provide excellent acute success in subjects with PV occlusion; however, restenosis remains common. Further research into methods of reducing restenosis such as the use of large-diameter drug-eluting stents and drug-coated balloon is needed.
Dr. Packer has served as a consultant for Abiomed, Biosense Webster, Boston Scientific, CardioDX, CardioFocus, CardioInsight Technologies, InfoBionic, Johnson & Johnson Healthcare Systems, Johnson & Johnson, MediaSphere Medical, Medtronic CryoCath, Sanofi, Siemens, St. Jude Medical, and Topera Medical (he received no personal compensation for these consulting activities); has received research funding from an AHA Foundation Award, Biosense Webster, Boston Scientific/EPT, CardioInsight, CardioFocus, Endosense, EpiEP, EP Rewards, Hansen Medical, Medtronic CryoCath LP, National Institutes of Health, St. Jude Medical, and Siemens. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Drs. Fender and Widmer contributed equally to this paper.
- Abbreviations and Acronyms
- atrial fibrillation
- balloon angioplasty
- computed tomography
- interquartile range
- pulmonary vein
- pulmonary vein stenosis
- Received January 7, 2018.
- Revision received May 3, 2018.
- Accepted May 8, 2018.
- 2018 American College of Cardiology Foundation
- Fender E.A.,
- Widmer R.J.,
- Hodge D.O.,
- et al.
- Saad E.B.,
- Rossillo A.,
- Saad C.P.,
- et al.
- Dong J.,
- Vasamreddy C.R.,
- Jayam V.,
- et al.
- Cappato R.,
- Calkins H.,
- Chen S.A.,
- et al.
- Calkins H.,
- Kuck K.H.,
- Cappato R.,
- et al.
- Calkins H.,
- Hindricks G.,
- Cappato R.,
- et al.
- Fender E.A.,
- Widmer R.J.,
- Packer D.L.,
- Holmes D.R. Jr..
- Gebhard C.,
- Krasniqi N.,
- Stahli B.E.,
- et al.
- ↵Fender E, Widmer RJ, Monahan KH, et al. Pulmonary vein stenosis: symptoms, management, and outcomes. 2015.
- Holmes D.R. Jr..,
- Monahan K.H.,
- Packer D.
- Packer D.L.,
- Keelan P.,
- Munger T.M.,
- et al.
- Di Biase L.,
- Fahmy T.S.,
- Wazni O.M.,
- et al.