Author + information
- Received September 20, 1999
- Revision received April 20, 2000
- Accepted June 16, 2000
- Published online November 1, 2000.
- Charles M Baker, MDa,
- Francis X McGowan Jr., MD†,
- John F Keane, MD, FACCa and
- James E Lock, MD, FACCa,*
- ↵*Reprint requests and correspondence:
Dr. James E. Lock, Department of Cardiology, The Children’s Hospital, 300 Longwood Avenue, Boston, Massachusetts 02115
We reviewed the management and outcome of patients experiencing pulmonary artery (PA) trauma during balloon dilation (BD).
Balloon dilation of the PA is important in the management of peripheral pulmonary stenosis. Successful BD requires a controlled tear of the PA; excessive tearing can produce complications ranging from pseudoaneurysms to rupture and death. The incidence and optimum management for such complications are unreported.
All records of patients who underwent branch PA dilation between June 1984 and October 1997 were reviewed; those with a significant complication were analyzed.
Of 1,286 catheterizations in 782 patients, PA trauma (excluding isolated pulmonary edema and PA aneurysms) was identified in 29 catheterizations in 26 patients. Tears occurred distal to the area of stenosis in most cases (62%). Intensive medical management, with and without catheter directed therapy, was employed. The damaged PA was successfully coil embolized in five patients, four of whom survived; temporary balloon occlusion did not prevent death in two patients. There were six deaths from pulmonary hemorrhage. A case control analysis demonstrated that PA trauma was significantly associated with pulmonary hypertension.
Pulmonary artery trauma associated with BD occurs mostly distal to the site of narrowing, is associated with underlying pulmonary hypertension and is frequently (5/12 or 42%) fatal in those with unconfined tears. Intensive management strategies as well as attention to distal balloon position may reduce incidence and mortality. Coil occlusion of the damaged PA appears to be a valuable strategy to prevent fatal hemorrhage.
Surgical management of distal stenosis and hypoplasia of branch pulmonary arteries (PPS) is frequently ineffective, and balloon dilation (BD) with and without stent placement has become the treatment of choice for many of these patients. Success rates of BD between 38% to 100% have been reported and have improved with the advent of high-pressure balloons and intravascular stents (1–9). Successful dilations require disruption of the intima and, usually, the media to effect a lasting increase in the diameter of the pulmonary artery (PA) (10). Not surprisingly, transcatheter management of PA obstruction is associated with mortality rates of 1% to 9% (2,3,5,7,8,11–13), especially when the distal vasculature requires enlargement. Pulmonary artery rupture has emerged as the most common cause of mortality and serious morbidity. As experience continues to accrue with this modality of therapy, methods of avoiding serious PA disruption are becoming increasingly important. The purpose of this report is to describe the recognition of PA trauma as well as potential strategies for avoidance and management of significant PA disruption during BD of the PA. This retrospective chart review study was completed with the approval of the institutional review board at The Children’s Hospital, Harvard Medical School, Boston Massachusetts.
The indications for BD of stenotic PAs and the techniques used have been previously described (3,5,14) and include the following:
1. Significantly elevated right ventricular pressure,
2. Hypertension in the unaffected segments of pulmonary artery,
3. Marked decrease in flow in the affected segments of lung as assessed by radionuclide scan,
4. Right ventricular dysfunction or cyanosis aggravated by PA stenosis, and
This review spans a 15-year period, and, while the techniques used to dilate stenotic PAs have undergone changes and evolution, in general they have followed our original guidelines (14). After standard premedication, vascular access is obtained percutaneously, using the femoral vessels in most cases. A venous sheath affords easy exchange of multiple catheters for hemodynamic measurements and angiography. A small pigtail catheter is placed in the descending aorta for blood gas sampling and as a continuous blood pressure monitor during the procedure. Hemodynamic evaluations are completed using appropriate end-hole catheters for pressure measurements. Angiographic delineation of the PA anatomy is accomplished using selective angiograms. The dimensions of the stenotic segments are determined by comparing their size to the known dimensions of catheters in optimally angled views. Whenever possible the most severe lesion is dilated first because balloon occlusion of that segment should have the smallest effect on total pulmonary blood flow. An end-hole catheter is advanced distal to the stenosis and a 0.018 to 0.035 inch exchange length wire is advanced as far as possible into the PA. After removal of the catheter and sometimes removal of the sheath and replacement with a larger/longer sheath, the angioplasty catheter is advanced to the area of stenosis. The initial balloon size is chosen to be 3 to 4 times the diameter of the segment to be dilated and is inflated partially so that refinement of position can be accomplished centering the waist on the balloon. If the waist is very tight relative to the balloon size (less than 50% of the inflated diameter of the balloon), full inflation of the balloon may expand the stenotic vessel so much that rupture ensues. A smaller balloon is then substituted and inflated to the maximum pressure to expand the stenotic vessel. After each dilation the angioplasty catheter is removed, leaving the wire in place, and an angiographic catheter (usually a cutoff pigtail) is advanced to the area of interest. Attaching a Y-adapter allows for both pressure measurement as well as angiographic assessment (15). Crossing the dilated segment with wires should be avoided if at all possible but, if necessary, may be done using a soft-tipped torque control wire. Review of the pressure and angiographic data determine whether further dilation or stent placement is necessary. If the initial results are unsatisfactory, as indicated by the inability to eliminate the balloon waist or inability to improve caliber of the stenotic vessel, subsequent dilations are pursued using larger or higher-pressure balloons. The procedure is repeated for other identified stenotic vessels, usually at the same catheterization. Patients considered at increased risk for hemodynamic instability were often electively intubated before intervention.
From June 1983 through October 1997, percutaneous BD of the branch PAs was performed during 1,286 catheterizations of 782 patients at our institution. This population in the computerized database was searched using diagnostic codes for complications that included aneurysms, pulmonary edema, pulmonary hemorrhage, hemoptysis, hemothorax, balloon rupture, blood loss, hypotension, cardiac arrest, resuscitation and death. From this population 29 catheterizations in 26 patients had angiographic or fluoroscopic evidence of extravasation of contrast or blood, indicating PA trauma, and those cases are the subject of this report (Table 1). Of those cases so identified, patients with complications limited to pulmonary edema or pseudoaneurysms without extravasation of dye or blood were excluded. All available data from this group of patients were then reviewed in detail. They were divided into two groups: those with confined tears (Fig. 1) and those with unconfined tears (Fig. 2) , based largely on angiographic features. Confined tears (CT) were characterized by extravasation of blood or contrast that was limited to the local area of dilation and, more importantly, that did not expand and usually resolved after BD. Conversely, unconfined tears (UT) were characterized by free extravasation of contrast or blood communicating with the pleural space, airway or leading to progressive lobar opacification in the distribution supplied by the dilated PA.
To attempt to identify risk factors for PA trauma, a control group was selected from catheterizations that included BD of PAs. Catheterizations of patients who were included in the case group were excluded. Two controls were randomly selected for each case matched for date of procedure (± 2 weeks). In the three patients who had more than one catheterization with evidence of PA trauma, only the first catheterization was compared. The following predilation items were evaluated: age, weight, diagnosis (isolated PPS vs. postoperative PPS associated with other structural heart disease), cardiac index, previous BD of PAs, mean main PA pressure, right ventricle/aorta pressure ratio, previous surgery involving PAs and recent surgery involving PAs (<6 weeks).
The primary outcome variable was the presence of significant PA trauma of either type (CT or UT). Predilation risk factors were compared for cases and controls using the Fisher exact test for categorical variables and Student t test (two sample, two-tailed) for continuous variables. Pearson’s chi-square analysis was used to compare categorical data between controlled tear and uncontrolled tear groups. All data are expressed as mean ± standard deviation.
Details of the 26 patients who underwent 29 BD procedures that resulted in PA trauma are outlined in Table 1. The ages at the time of BD ranged from 3 months to 41 years (median 54 months). Diagnoses consisted of Tetralogy of Fallot with pulmonary atresia (TOF/PA) (10 patients), TOF (5 patients), isolated PPS (4 patients) and PPS/Williams syndrome (4 patients). There were single instances of PPS/Alagille syndrome, tricuspid atresia and heterotaxy syndrome, the last two having previously undergone Fontan procedures.
There were 17 cases with PA tears defined as confined (CT) and 12 cases with PA tears defined as unconfined (UT). The tears occurred distal to the site of obstruction in the majority of the cases in both the UT and CT groups (Table 2). The location of the tear could not be determined in a single case from each group.
General patient management
Intensive medical management of the PA trauma was necessary in most of these patients (Table 3). In addition to the interventions listed, selective mainstem intubation for management of unilateral pulmonary hemorrhage was carried out in two patients. Compared with the CT group, those in the UT category required far more intensive therapy, including resuscitations, large volume transfusions, emergency intubation and catheter delivered therapy. In addition, 42% of the procedures in the UT group lead to a patient death in contrast with 6% in the CT group (p < 0.02).
Therapeutic interventions during catheterization after the PA rupture were carried out in eight patients (Fig. 3). Coils were placed proximal to the PA tear in five patients to control hemorrhage; four survived. Two patients who had only balloon tamponade proximal to the site of rupture (both unconfined tears) died. One of these ruptures occurred early in our experience before the use of coils. The rupture was located in the proximal right PA, which would have precluded the option of coil embolization. The other was a distal unconfined tear, where no attempt to coil occlude the torn branch was made because of immediate and persistent severe hemodynamic collapse that could not be reversed. Another patient had a stent placed over a tear and had extracorporeal membrane oxygenation (ECMO) instituted for low cardiac output due to severe PA vasospasm; this patient survived. Those five in whom coils were placed are now presented in more detail.
Two patients (a 7-year-old with TOF/PA status after repair and a 4-year-old with Williams syndrome) had instantaneous evidence of PA rupture with dramatic extravasation identified immediately after BD (unconfined tear). The tears occurred distal to the stenotic segment in both patients. Both required cardiopulmonary resusitation, inotropic support, drainage of hemothoraces with chest tubes and large volume transfusions. In both patients the PA proximal to the torn segment was occluded initially by balloon tamponade and followed by occlusion with coils. Both were transferred to the CICU and then discharged 10 and 7 days later, respectively.
A 15-year-old with PPS and Alagille syndrome had hemoptysis develop during dilation of a right upper lobe PA. Contrast was then seen in the airway, and a right upper lobe PA vessel tear was identified. The patient was emergently intubated, transfused, and the torn vessel was occluded with coils. The patient was observed overnight in the CICU and discharged four days later.
An 18-year-old male patient with isolated PPS had proximal left PA dilation and stent implantation complicated by balloon rupture with a retained fragment without immediate evidence of PA rupture. He was emergently intubated due to development of pulmonary edema and managed overnight in the cardiac intensive care unit (CICU) because he remained intubated at the completion of the catheterization. The following morning extravasation of contrast into the left pleural space was noted on angiographic review, and he returned to the catheterization laboratory. A left PA tear was identified distal to the stent with extravasation of contrast into a left mediastinal hematoma. The torn vessel was successfully coil occluded. The patient required no cardiopulmonary resusitation or inotropic support but did, however, receive a large volume transfusion. He was extubated after two days and was discharged six days later.
The last patient, age 4 years, with TOF/PA status after placement of a right ventricle to PA conduit developed a tear in the right lower lobe PA after BD. In addition, severe vasospasm of the pulmonary arteries was noted. The tear was occluded with a stent, but the patient required initial hemodynamic support with ECMO. She was discharged 18 days later and returned after six months for VSD closure and conduit replacement, at which time the right PA stent was removed.
There were six deaths, five of which were in the UT group for a mortality rate due to PA trauma of 0.77% (6/782). Three of these six deaths occurred before 1992 (Fig. 4). Since that year, and with the use of coils, the mortality rate has been 0.57% (3/528) despite the fact that the incidence of PA trauma remains 2.5 per 100 catheterizations involving PA dilation. Four of the deaths occurred on the day of catheterization. The remaining two died two and seven days later. The following are brief summaries of those who died.
Dilations were performed in two patients within two weeks of cardiac operations that included PA procedures. The first, a 3-year-old with tricuspid atresia who had had a Fontan procedure, struggled with low cardiac output. Catheterization 14 days after surgery demonstrated bilateral proximal branch PA obstruction. Dilation of the left PA near a surgical repair site resulted in a tear, hypotension and a dramatic bloody effusion. After prolonged resuscitation, an operative exploration revealed rupture of the proximal left PA, which was then surgically repaired. The patient died two days later when support was discontinued after a severe neurologic insult was diagnosed. The second, a 7-year-old with TOF/PA, was catheterized three days after surgical placement of a right ventricle to PA homograft. The surgery included left PA dilation during which a PA tear occurred. During recovery pulmonary hemorrhage recurred, and the patient was taken to catheterization, which included right PA dilation. Proximal right PA rupture occurred after which operative repair of both tears was unsuccessful.
The third patient was a 6-month-old infant with TOF/PA in whom a right ventricle to PA conduit was placed five months before. At catheterization bilateral proximal PA stenoses were identified. After right PA dilation, PA rupture occurred just distal to the stenotic segment with immediate hemodynamic decompensation. Despite balloon tamponade, chest tube placement and maximal medical resuscitation, the patient died. In this baby the wire position was lost after the initial unsuccessful dilation. The wire was readvanced across the dilation site, and the BD was repeated using a larger balloon, after which the rupture occurred.
None of the remaining three patients, ranging from 5 to 33 years of age, had had prior cardiac surgery. All had PPS, associated with Alagille syndrome in one and Williams syndrome in another. Immediately after BD two of the three had progressive opacification of a region of lung associated with the dilation. These patients were emergently intubated, ventilated with positive pressure and had circulation supported with inotropes and red blood cell transfusion as needed. These two cases occurred before the use of coils to embolize torn PAs. Both died within 6 h of the catheterization after massive hemoptysis while being managed in the CICU. The other had hemoptysis and minimal extravasation of contrast with a PA wedge injection, which lead to identification of a confined tear. The torn vessel was occluded with a coil. The patient was then transferred to the CICU and was later extubated. She died two days after discharge, five days after the catheterization, at an outside hospital where she presented with respiratory distress and a hemothorax.
By univariate analysis (Table 4)of predilation characteristics, the following showed significant differences between cases and controls: diagnosis, right ventricle/aorta pressure ratio and mean main PA pressure. Since mean PA pressure correlated strongly with both diagnosis and right ventricular/aorta pressure ratio, neither remained significant after controlling for mean main PA pressure. Recent PA surgery approached statistical significance when mean PA pressure was controlled for (p = 0.08). The odds of PA trauma increase an estimated 97% for each 10 mm Hg increase in mean PA pressure. No significant differences were found between groups for balloon to stenosis ratio (34/78 patients) or the maximum inflation pressure (55/78 patients) in those patients where those data were recorded.
Percutaneous BD of PPS, now in use for some 15 years, remains a more effective treatment modality than surgery, especially in distal vessels. The degree of success has been enhanced in recent years after the introduction of high-pressure balloons and stents. While a safe procedure in the large majority of cases, catastrophic events do occur, with earlier mortality rates of 1% to 9% being reported in several series (2,3,5,7,8,11–13). In the current series the incidence of significant PA trauma (excluding PA aneurysm and pseudoaneurysm) was 2%, and the mortality rate due to PA trauma was 0.77% (6/782) during the time period under review.
We were unable to define any technical risk factors in those analyzed, including balloon lesion ratio. However, two deaths occurred in patients who underwent catheterization and dilation within days after surgery involving the PAs. Since recent surgery involving the PAs approached statistical significance as a risk factor when controlled for PA pressure, dilating lesions in such patients would seem best postponed for at least two months to allow healing. Attention to balloon position and size relative to the area of stenosis is crucial to avoid PA rupture. This is illustrated by the preponderance of tears located distal to the obstruction in the patients in our series. The post-stenotic vessels are thin walled (7,10) and are more susceptible to rupture when large or unnecessarily long balloons are inflated or migrate distally. Such ruptures exposed to high pressure/flow after dilation are more likely to lead to significant hemorrhage. Attention to avoiding distal positioning and migration of the balloon during inflation will reduce the incidence of PA rupture. The case control analysis revealed that patients with significantly elevated mean main PA pressures are at increased risk for developing PA trauma. The case control evaluation was matched only for date of the procedure in order to investigate the potential influence of the factors listed in Table 4 on the occurrence of PA trauma. Since main pulmonary artery pressure was highly correlated with elevated right ventricle pressure, and with diagnosis the independent contribution to PA trauma of diagnosis or elevated right ventricle pressure are unknown. Elevated main PA pressure is the unifying end result of diseased distal pulmonary vasculature and is likely the most important factor, regardless of etiology of the distal disease.
Diagnosis: confined versus unconfined tears
Among the 29 catheterizations in 26 patients, the occurrence of the PA tear was readily manifest by either angiography, hemoptysis, evidence of rapid blood loss or evidence of increasingly dense opacification on fluoroscopy in the lung segment dilated. Angiographically, it was possible to identify two groups. The first group consisted of those patients with a confined tear, where contrast was localized only to the tissues surrounding the dilated vessel. The second group consisted of those patients in whom the tear was unconfined, where contrast was seen to leave the vessel lumen and dissipate into the ipsilateral pleural space or produce an enlarging collection of contrast or blood. Also considered unconfined were those patients with progressive lobar opacification after BD, without frank extravasation of contrast. In our series two patients with progressive opacification died up to 6 h after catheterization of massive sudden hemorrhage while being managed in the CICU.
Intensive medical resuscitative management, including intubation, hemothorax evacuation and transfusions, is urgently required in patients with an unconfined tear in concert with temporary balloon occlusion of the damaged vessel. As demonstrated in Table 3, patients with UT received much more intensive therapy than patients with CT. Initially reported as a therapy for a mycotic pseudoaneurysm in the PA by Remy in 1984 (16), coil occlusion of traumatized PAs associated with Swan-Ganz catheters has been reported (17–20). Our population of patients is quite different and presents the challenge of differentiating signs and symptoms indicative of markedly increased local pulmonary blood flow (segmental pulmonary edema, hemoptysis) that indicate a successful dilation from those of an unconfined PA tear. The angiographic appearance after dilation and the patient’s hemodynamic status are the key pieces of information that direct timely diagnosis and therapy. Obvious extravasation and more than brief hemodynamic instability are certainly indicators of PA trauma rather than increased local pulmonary blood flow. In those patients with a confined tear and without hemodynamic compromise or progressive pulmonary lobar opacification, careful management in the intensive care unit seems to be adequate in most cases.
After recognition of an unconfined tear (free extravasation of contrast or blood that does not decrease or disappear with time or progressive lobar opacification), initial stabilization and permanent coil occlusion of the damaged vessel seems to be indicated. This is because balloon occlusion alone did not avert a fatal outcome in the two patients in whom it was used. Although occlusion of a torn segment of a PA is not desirable because many of these patients have a limited total PA distribution, it indeed is life-saving in this situation.
☆ Dr. Baker was, in part, supported by NIH training grant T-32 HL07572.
Presented as part of the American College of Cardiology 47th Annual Scientific Session, March 1998, Atlanta, Georgia.
- balloon dilation
- cardiac intensive care unit
- confined tear
- extracorporeal membrane oxygenation
- pulmonary artery
- peripheral pulmonary stenosis
- Tetralogy of Fallot with pulmonary atresia
- unconfined tears
- Received September 20, 1999.
- Revision received April 20, 2000.
- Accepted June 16, 2000.
- American College of Cardiology
- Mendelsohn A.M,
- Bove E.L,
- Lupinetti F.M,
- et al.
- Gentles T.L,
- Lock J.E,
- Perry S.B
- Fogelman R,
- Nykanen D,
- Smallhorn J.F,
- McCrindle B.W,
- Freedom R.M,
- Benson L.N
- Kreutzer J,
- Landzberg M.J,
- Preminger T.J,
- et al.
- Edwards B.S,
- Lucas R.V,
- Lock J.E,
- Edwards J.E
- Shaffer K.M,
- Mullins C.E,
- Grifka R.G,
- et al.
- Lock J.E,
- Keane J.F,
- Fellows K.E