CARDIOVASCULAR
J. MAXWELL CHAMBERLAIN MEMORIAL PAPER
Pulmonary Endarterectomy: Experience and Lessons
Learned in 1,500 Cases
Stuart W. Jamieson, MB, FRCS, David P. Kapelanski, MD, Naohide Sakakibara, MD,
Gerard R. Manecke, MD, Patricia A. Thistlethwaite, MD, PhD, Kim M. Kerr, MD,
Richard N. Channick, MD, Peter F. Fedullo, MD, and William R. Auger, MD
Division of Cardiothoracic Surgery, Department of Anesthesia, and Division of Pulmonary and Critical Care Medicine, UCSD
Medical Center, San Diego, California
Background. The incidence of pulmonary hypertension
resulting from chronic thrombotic occlusion of the pul-
monary arteries is significantly underestimated. Al-
though medical therapy for the condition is supportive
only, surgical therapy is curative. Our pulmonary endar-
terectomy program was begun in 1970, and 188 patients
were operated on in the subsequent 20 years. With the
increased recognition of the disease and the success of
operative therapy, however, more than 1,400 operations
have been done since 1990 at our center.
Methods. The safety and efficacy of the operation was
assessed with changes made through increased experi-
ence. We examined in detail the results of our last 500
consecutive patients.
Results. Median sternotomy, cardiopulmonary bypass,
profound hypothermia, and circulatory arrest were found
to be essential to the success of the operation. All
occluding material could be removed at operation. We
M edical therapy for pulmonary hypertension caused
by pulmonary vascular occlusion is generally un-
satisfactory and palliative. Therapy by pulmonary end-
arterectomy offers a surgical cure, yet this procedure is
not commonly applied.
This operation is a technically demanding yet highly
successful treatment for pulmonary hypertension that is
caused by chronic pulmonary thromboembolic disease.
Proper patient selection, meticulous surgical technique,
and careful postoperative management have now clearly
shown that it is an effective therapy. A true endarterec-
tomy (not an embolectomy) of all affected parts of the
lung is essential. Circulatory arrest is fundamental for the
visibility necessary to clear all affected areas of the
pulmonary vasculature.
Although pulmonary transplantation is still used in
some centers for patients with thromboembolic disease,
we consider that treatment to be outdated. In good
hands, the endarterectomy operation has a lower opera-
tive mortality rate than lung transplantation, it can be
Presented at the Thirty-ninth Annual Meeting of The Society of Thoracic
Surgeons, San Diego, CA, Jan 31–Feb 2, 2003.
Address reprint requests to Mr Jamieson, Cardiothoracic Surgery, UCSD
Medical Center, 200 West Arbor Drive, San Diego, CA 92103; e-mail:
sjamieson@ucsd.edu.
© 2003 by The Society of Thoracic Surgeons
Published by Elsevier Inc
currently believe that there is no degree of embolic
occlusion within the pulmonary vascular tree that is
inaccessible and no degree of right ventricular impair-
ment or any level of pulmonary vascular resistance that is
inoperable. With shorter cardiac arrest periods and the
use of a cooling jacket to the head, cerebral impairment
has been eliminated. The pulmonary artery pressures
and pulmonary vascular resistance in a recent cohort of
500 patients is examined. The mortality rate for the
operation has been reduced steadily, and was 22 of the
last 500 patients operated on (4.4%).
Conclusions. The operation is considered curative and
therefore greatly superior to transplantation for this
condition. Current techniques of operation make the
procedure relatively safe.
(Ann Thorac Surg 2003;76:1457– 64)
© 2003 by The Society of Thoracic Surgeons
done electively without the wait for a donor, and the
long-term problems associated with rejection and immu-
nosuppressive drugs are eliminated. In addition, the
mortality rate for transplantation (and especially double
lung transplantation) as a therapeutic strategy should
include patients on the waiting list and is therefore much
higher than is generally reported.
Pulmonary endarterectomy appears to be permanently
curative. However, probably fewer than 2,500 of these
operations have been performed, and most of them (1,500
cases) were done at the University of California San
Diego (UCSD). The recent experience using this proce-
dure at UCSD provides the basis of this report.
Patients and Methods
Patient Selection
The first pulmonary thromboendarterectomy operation
at UCSD was done in 1970 [1], and 188 operations were
carried out in the ensuing 20 years. The program has
grown steadily since 1990, and 1,580 patients had been
operated on by December 2002. Patients are referred
nationally and internationally to this center, and the
number of referrals increases yearly, indicating an in-
creased awareness of the disease and the effectiveness of
0003-4975/03/$30.00
doi:10.1016/S0003-4975(03)00828-2
 1458 CHAMBERLAIN MEMORIAL PAPER JAMIESON ET AL
CARDIOVASCULAR
PULMONARY ENDARTERECTOMY
therapy rather than an increased incidence. The clinical
presentation of the patients varied, from patients with
dyspnea on exertion but otherwise well to patients trans-
ferred by air ambulance in respiratory failure with florid
right heart failure and ascites. At least half the patients
did not have a history of deep venous thrombosis, and
many gave no history of pulmonary embolism, thus
contributing to the difficulty in diagnosis.
Patients accepted for operation included those with
pulmonary hypertension and evidence of thromboem-
bolic disease on pulmonary angiography. In general,
pulmonary vascular resistance was higher than 300
dynes · second⫺1 · cm⫺5 and in some cases was more
than 2,000 dynes · second⫺1 · cm⫺5. This group of pa-
tients, spanning ages 8 to 86 years, reflects a broad
clinical range. On one side of the spectrum, young
patients with dyspnea on exertion have been accepted for
operation to remove unilateral pulmonary artery throm-
boembolic disease. On the other side of the spectrum, we
now accept high-risk patients who have vascular disease
in subsegmental pulmonary vessels only, as well as
patients presenting with advanced multisystem organ
failure from cardiovascular collapse.
Contraindications were limited only to conditions
judged to be unrelated to pulmonary hypertension or
deep venous thrombosis that were likely to indepen-
dently limit patient survival in the near future. The
expansion of patient population considered acceptable
for this operation is based largely on accumulated surgi-
cal experience, improved operative technique, and a
multidisciplinary approach to postoperative care.
Operative Technique
Coronary angiography was performed in all patients
more than 45 years old. A Greenfield inferior vena cava
filter is always placed preoperatively.
The operation has been described in detail [2]. The
essential features are that a median sternotomy incision
is made, with cardiopulmonary bypass enabling hypo-
thermia to 20°C. During the precardiopulmonary bypass
period, the head is wrapped in a circulating cold-water
blanket (Polar Care; Breg, Inc., Vista, CA). Water, main-
tained at approximately 4°C, is circulated through this
blanket by an electrical pump. This system, originally
designed as a knee-wrap for orthopedic and physical
medicine purposes, is easily applied to the head. It
contains a thermometer within the fluid circulation sys-
tem for confirmation of adequate blanket cooling, as well
as a flow control dial.
After the institution of cardiopulmonary bypass, the
patient is cooled, and vents are placed in both the
pulmonary artery and the right upper pulmonary vein.
The right pulmonary artery is exposed between the aorta
and superior vena cava and mobilized within the peri-
cardial reflection. The pleura are not entered on either
side. The aorta is cross-clamped, and antegrade cold
cardioplegia is administered. Additional myocardial pro-
tection is then provided by the use of a cooling jacket,
and no further cardioplegia is administered during the
arrest period.
Ann Thorac Surg
2003;76:1457– 64
Fig 1. An endarterectomy specimen is shown. Some of the thrombus
shown proximally may be secondary to stasis rather than being em-
bolic. The essential part of the operation consists of removal of the
distal fibrotic component in all segmental and subsegmental vessels
by endarterectomy.
An incision is made in the right pulmonary artery and
continued into the right lower lobe. Any thrombus en-
countered, be it loose or organized and attached to the
vessel wall, is now removed. The endarterectomy cannot
be performed in the presence of thrombus because it
obscures the plane and prevents collapse of the endar-
terectomized specimen, hindering distal exposure. Cir-
culatory arrest is begun when bronchial back bleeding
obscures good visualization.
Excellent visibility of the entire vascular tree can now
be obtained with appropriate positioning of the incision
and the patient. Once the endarterectomy plane is devel-
oped, gentle traction with forceps while sweeping away
the outer vessel wall layer will result in the progressive
withdrawal of the endarterectomy specimen. The endar-
terectomy is performed primarily with a long miniature
sucker with a rounded tip [2]. As each lobar branch
appears it is grasped individually, and the specimen is
withdrawn until each segmental vessel branches again.
Each of these subsegmental specimens is then removed.
Removal of each lobar and then segmental branch
makes subsequent distal dissection easier. If a large mass
of endarterectomized tissue begins to obscure visibility, it
is excised. The entire specimen can thus be removed
progressively for a length of approximately 20 cm (Fig 1).
All mechanically occluding material can be removed with
these techniques, and visibility can be obtained in sub-
segmental vessels. The term “surgically accessible
thrombus” is no longer appropriate, because the plane
can be raised, with experience, even at segmental and
subsegmental levels. Although pulmonary vascular resis-
tance does not always decrease to normal levels postop-
eratively, provided a full endarterectomy has been per-
formed, residual pulmonary hypertension is caused by
secondary damage to the vessel at the arteriolar and
capillary levels, not by residual embolic material.
Circulatory arrest is limited to 20-minute intervals, and
Ann Thorac Surg CHAMBERLAIN MEMORIAL PAPER JAMIESON ET AL 1459

CARDIOVASCULAR
2003;76:1457– 64 PULMONARY ENDARTERECTOMY

Table 1. Mean Hemodynamic Values in 500 Patients

All patients Type 1 Type 2 Type 3 Type 4
(n ⫽ 500, 100%) (n ⫽ 187, 37.4%) (n ⫽ 245, 49%) (n ⫽ 60, 12%) (n ⫽ 8, 1.6%)

PVR (dynes/sec/cm⫺5) 893 ⫾ 443.5 973 ⫾ 444.3 809 ⫾ 419.2 973 ⫾ 487.6 1163 ⫾ 490.5
285 ⫾ 214.7 275 ⫾ 201.2 273 ⫾ 219.2 344 ⫾ 193.9 868 ⫾ 421.6
CO (L/min) 3.8 ⫾ 1.3 3.6 ⫾ 1.4 4.1 ⫾ 1.2 3.8 ⫾ 1.5 3.4 ⫾ 1.2
5.5 ⫾ 1.5 5.6 ⫾ 1.5 5.6 ⫾ 1.5 5.3 ⫾ 1.6 4.2 ⫾ 1.6
Systolic PA pressure (mm Hg) 77 ⫾ 18.1 78 ⫾ 16.2 76 ⫾ 19.6 77 ⫾ 18.1 86.5 ⫾ 11.4
47 ⫾ 17.6 46 ⫾ 15.1 45 ⫾ 16.7 53 ⫾ 17.2 101 ⫾ 38.9
Diastolic PA pressure (mm Hg) 30 ⫾ 9.8 30.3 ⫾ 8.9 28.8 ⫾ 10.2 30.1 ⫾ 10.3 36.9 ⫾ 11.1
19 ⫾ 7.6 18.1 ⫾ 6.9 18.1 ⫾ 7.4 20.5 ⫾ 9.1 37.8 ⫾ 11.7
Mean PA pressure (mm Hg) 46 ⫾ 11.0 47 ⫾ 9.7 46 ⫾ 11.7 47 ⫾ 11.7 55 ⫾ 9.8
28 ⫾ 10.1 28 ⫾ 8.9 28 ⫾ 10.3 32 ⫾ 10.9 55 ⫾ 13.2
Circulatory arrest time (min) 35.7 ⫾ 11.9 35.8 ⫾ 12.1 33.8 ⫾ 10.4 44.5 ⫾ 12.8 27.9 ⫾ 12.8
Mortality 22 (4.4%) 4 (2.1%) 13 (5.3%) 3 (5.0%) 2 (25%)

Data are given as mean ⫾ standard deviation. Top numbers are preoperative values and bottom numbers are postoperative values obtained just prior
to removal of the Swan-Ganz cathether. The change in all values was significant at p ⱕ 0.0001, except in type 4 where they were not significant.
CO ⫽ cardiac output; PA ⫽ pulmonary artery; PVR ⫽ pulmonary vascular resistance.

with experience the entire unilateral endarterectomy can
generally be accomplished within that time. Reperfusion
is then initiated while the arteriotomy is closed. Attention
is then turned to the left side. This side is generally not as
affected as the right and is usually easier to endarterec-
tomize, with the exception of the left lower lobe. Circu-
latory arrest times are generally shorter on the left side.
At the completion of the left endarterectomy, circula-
tion is recommenced and the patient rewarmed. The left
pulmonary arteriotomy is closed. The atrial septum is
always inspected; an atrial septal defect or persistent
foramen ovale was seen in approximately 25% of cases. If
additional procedures, such as valve replacement or
coronary bypass operation, are required they are conve-
niently performed during the rewarming period. Al-
though tricuspid regurgitation was invariable in these
patients, tricuspid repair was not done unless the valve
itself was damaged. Right ventricular remodeling after
reduction in right ventricular pressure results in tricus-
pid competence.
Postoperative care is similar to that of routine open
heart operations, except that aggressive diuresis is insti-
tuted to remove fluid in the third space that accumulated
as a result of prolonged bypass and hypothermia and to
reduce the incidence of pulmonary edema [2]. Nitric
oxide and other pulmonary vasodilating agents have not
been used routinely and should not be necessary if the
mechanical obstruction has been removed.
A cohort of 500 consecutive patients (numbers 1,000 to
1,500 in our series) operated on between July 22, 1998 and
July 11, 2002 were analyzed in detail for hemodynamic
data and survival.
Statistical Analysis
Results are summarized as mean ⫾ standard deviation or
n (%). Tympanic, bladder, and rectal temperatures were
compared using one-way analysis of variance. The asso-
ciation between preoperative and postoperative pulmo-
nary vascular resistance (PVR) and mortality rate was
assessed using an exact likelihood ratio test. Paired
Student’s t tests were performed to identify significant
changes in hemodynamic and other variables. Statistical
tests were two-sided and assumed a 5% level of signifi-
cance. Software used for the data and statistical analyses
included Axis Clinical Software, Inc, JMP 5.0 (SAS Insti-
tute Inc., Cary, NC) and StatXact-5 Release 5.0.3 (Cytel
Software Corporation, ).
Results
There is undoubtedly a learning curve for the operation.
The operative mortality rate was 17% for the first 200
patients in the beginning of the UCSD series (1970 to
1990). Changes were made in the operative technique in
1990 [3], and the perioperative mortality rate has since
steadily declined, to 8.8% for the 500 individuals who had
the operation between 1994 and 1998, and to 4.4% for the
500 patients operated on between 1998 and 2002.
The cohort of 500 consecutive patients was analyzed in
detail for hemodynamic data and survival. Patients in
this group ranged in age from 8 to 84 years (median, 52
years), with a slight preponderance of males (256 of 500
patients, 51.2%). Duration of preoperative symptoms
ranged from 6 months to 25 years. Only 229 patients
(45%) gave a history of prior deep venous thrombosis,
and 150 (30%) had no history suggestive of pulmonary
embolism.
Mean cardiopulmonary bypass time was 218.7 ⫾ 40.5
minutes. The mean cardiac arrest time was 88 ⫾ 24.6
minutes, and the mean circulatory arrest time was 35.7 ⫾
11.9 minutes (19.8 ⫾ 8.0 minutes for the right side and
15.8 ⫾ 6.3 minutes for the left side).
The effect of the cooling jacket on tympanic membrane
temperature was measured in 55 patients. Temperatures
were measured at the commencement of circulatory
arrest; the mean tympanic membrane temperature was
15.1 ⫾ 1.1°C, rectal 20.8 ⫾ 1.5°C, and bladder 19.8 ⫾ 1.1°C.
The difference between tympanic membrane tempera-
 1460 CHAMBERLAIN MEMORIAL PAPER JAMIESON ET AL
CARDIOVASCULAR
PULMONARY ENDARTERECTOMY
ture and that measured in the bladder or rectum was
apparent in all cases and was statistically significant (p ⬍
0.0001).
Tricuspid annuloplasty was not performed, even when
severe tricuspid regurgitation was documented preoper-
atively. Two patients had a tricuspid valve replacement
for leaflet structural damage caused by past endocarditis.
One patient had an aortic valve replacement; one patient
with an absent pulmonary valve had a pulmonary valve
replacement. Thirty-six of 500 (7%) patients had concom-
itant coronary artery operation.
The median stay in the intensive care unit was 96
hours, and the median number of days the patients were
intubated was 1 ⫾ 9.6 (longest, 86 days). The median
length of hospital stay postoperatively was 10 days.
Twenty-two of 500 patients died (either in the hospital
or within 30 days of discharge) for an overall mortality
rate of 4.4%. Death was related to residual high pulmo-
nary pressures in 17 of 22 patients (77%). In 18 patients
who died, the preoperative PVR was higher than 1,000
dynes · second⫺1 · cm⫺5, making the mortality rate for
patients with this degree of preoperative impairment
10.1% (18 of 179 patients). In individuals in whom the
preoperative PVR was less than 1,000 dynes · second⫺1 ·
cm⫺5, the mortality rate was 1.3% (4 of 309 patients, p ⬍
0.0001). The degree of postoperative residual PVR most
strongly correlated with mortality rate. Patients with a
postoperative PVR higher than 500 dynes · second⫺1 ·
cm⫺5 had a mortality rate of 30.6% (15 of 49 patients),
whereas individuals with a postoperative PVR less than
500 had a mortality rate of 0.9% (4 of 434 patients, p ⫽ ⬍
0.0001).
During the time of study, 17 additional patients who
had pulmonary endarterectomy were excluded from this
series. These patients did not have thromboembolic
pulmonary disease; rather 9 were found to have pulmo-
nary artery neoplasia (angiosarcoma), 6 had pulmonary
arteritis (diagnosed by microscopy) with stricture forma-
tion, and 2 had fibrosing mediastinitis. In all these
patients, because of atypical appearances on angiogra-
phy, the correct diagnosis was suspected but not proven
before operation. One patient of these 17 (6%) died.
We previously described a method of patient classifi-
cation based on the location and morphology of throm-
boembolic and vascular wall disease found at the time of
operation [2, 4]. The results of operation correlated with
the classification type, as confirmed in this series. In this
cohort of 500 patients, 187 (37.4%) were type 1—fresh or
organized clot in the main or lobar pulmonary arteries;
245 (49%) were type 2—intimal thickening and fibrosis
without visible thrombus; 60 (12%) were type 3—fibrosis,
intimal webbing, and thickening within distal segmental
arteries only (no visible thrombus); and 8 (1.6%) were
type 4 —microscopic distal arteriolar vasculopathy. Type
4 patients represent a group with pulmonary hyperten-
sion not directly related to pulmonary embolism and are
not benefited by pulmonary endarterectomy [2]. Table 1
shows the hemodynamics and results for all patients
overall and as classified by type.
Ann Thorac Surg
2003;76:1457– 64
Comment
Unappreciated Incidence
Pulmonary hypertension as a result of pulmonary embo-
lism is an unappreciated but major cause of morbidity
and mortality. Its incidence is difficult to calculate be-
cause of the uncertainty regarding the frequency of acute
pulmonary embolism and the proportion of patients with
acute pulmonary emboli in whom embolic residua fail to
resolve. In 1975, Dalen and Alpert [5] calculated that
pulmonary embolism resulted in 630,000 symptomatic
episodes in the United States, making it about half as
common as acute myocardial infarction and three times
more common than cerebrovascular accidents. They also
estimated that pulmonary embolism caused approxi-
mately 200,000 deaths yearly in the United States (the
sole contributing cause of death in 100,000 and the major
contributing cause in another 100,000) and as such was
the third most frequent cause of death. However, the true
incidence of acute pulmonary embolism is unknown
because many (if not most) of the episodes are
asymptomatic.
Because emboli resolve completely in most patients, it
is unclear why some individuals progress to chronic
pulmonary hypertension. Approximately 10% of patients
with chronic thromboembolic pulmonary hypertension
have a coagulation abnormality, such as protein S or C
deficiency, whereas a smaller cohort have defects in the
fibrinolytic system [2]. In addition, some cases may be
caused by repetitive pulmonary emboli, in situ propaga-
tion of a clot into branch pulmonary vessels, or emboli-
zation of partially organized thrombus. In many patients,
chronic pulmonary hypertension may be a combination
of these factors. The pathologic progression from acute
embolism to end-stage pulmonary hypertension is there-
fore thought to be a multifactorial process that occurs
over months to years. Failure of the initial embolus to
resolve can also lead to secondary vasculopathy within
the arterioles of the lung, resulting in microvascular
muscular thickening and occlusion, similar to that seen in
primary pulmonary hypertension and Eisenmenger syn-
drome [2, 4, 6]
Whatever the incidence of acute pulmonary emboli,
chronic thromboembolic pulmonary hypertension is es-
timated to result after only 1% to 5% of all cases of acute
pulmonary embolism [2, 7, 8]. However, the chronic
disease process, even when established and symptom-
atic, is commonly unrecognized because of the nonspe-
cific nature of the two major symptoms, effort dyspnea
and fatigue, and the fact that physical findings may be
elusive until right heart failure supervenes. Autopsy
series concentrating specifically on pulmonary artery
pathology found that 30% of cases had gross evidence of
recurrent emboli and an additional 34% had stigmata of
partial resolution [2]. Calculations extrapolated from
mortality rates and the random incidence of major
thrombotic occlusion found at autopsy could support a
estimate that more than 100,000 people in the United
States currently suffer from pulmonary hypertension that
could be relieved by operation.
Ann Thorac Surg
2003;76:1457– 64
Chronic pulmonary hypertension from thrombotic dis-
ease carries a poor prognosis, which is proportional to
the severity of pulmonary hypertension. Reidel and col-
leagues [9] followed up 147 patients with pulmonary
hypertension with serial right heart studies and pulmo-
nary arteriograms, and they found that patients with
mean pulmonary artery pressure more than 30 mm Hg
had a 30% 5-year survival rate. Those with pressures
more than 50 mm Hg had a 10% survival rate at 5 years.
These findings were confirmed by a more recent study
[10].
Treatment
Medical therapy with anticoagulant drugs, thrombolytic
agents, or vasodilator drugs has not been shown to affect
the prognosis [2]. However, the lungs are unique in that
embolization uncommonly results in tissue necrosis (be-
cause of the bronchial circulation); therefore, subsequent
endarterectomy of occluded pulmonary vessels allows
the distal pulmonary tissue to be recruited once more to
assist in gas exchange. Surgical therapy is curative, and
with few exceptions [11] is regarded as permanent. Long-
term outcome after pulmonary thromboendarterectomy
has also been studied by our group. In 308 patients (mean
age, 56 years [range, 19 to 89 years]) with a mean of 3.3 ⫾
2.7 years since the operation (range, 1 to 16 years),
survival, functional status, quality of life, and the subse-
quent use of medical help were assessed. Survival after
pulmonary thromboendarterectomy was 75% at 6 years
or more. Ninety-three percent of the patients were found
to be in New York Heart Association class I or II [12].
Historical Perspective of Surgical Treatment
Trendelenburg described an operative approach for
acute pulmonary embolism in 1908 [13] using inflow
occlusion. However, success with this operation was rare,
and a significant milestone in the application of acute
pulmonary embolectomy was the use of cardiopulmo-
nary bypass to allow a careful and more thorough ap-
proach, as first described by Cooley and associates in
1961 [14].
The entity of chronic occlusion of the pulmonary arter-
ies was not recognized until its description at autopsy in
1928 by Lungdahl [15]. A review in 1956 stated, “it is
probable that no more than 200 cases of the syndrome
have been reported in the medical literature to date” [7].
In any event, the impression at that time was that chronic
thrombotic occlusion was not amenable to surgical
correction.
Carroll [16] described the first operation on a patient
with chronic thrombotic occlusion of the pulmonary
arteries. The patient underwent a left thoracotomy at
Johns Hopkins Hospital in January 1948 by Dr Alfred
Blalock. The left pulmonary artery was found to be small,
with proximal occlusion, although aspiration with a nee-
dle produced red blood. The artery was divided and
found to contain organized thrombus. No attempt was
made to relieve the obstruction, and the situation was
thought to be inoperable. The patient was discharged
from the hospital unimproved.
CHAMBERLAIN MEMORIAL PAPER JAMIESON ET AL 1461
CARDIOVASCULAR
PULMONARY ENDARTERECTOMY
In May 1962, Dr Charles Hufnagel operated on proba-
bly the first patient in whom the diagnosis was made
preoperatively, and successful correction was achieved
[17]. The operation showed that extensive, well-
organized thrombi, which obstructed major pulmonary
vessels for months to years, could be removed success-
fully. Furthermore, the reclaimed lung areas were shown
to be able to accept safely the sudden return of blood flow
and to resume adequate respiratory function. In 1984,
Chitwood and colleagues [18] reviewed the world’s liter-
ature and found 85 cases managed surgically, with a
mortality rate of 22%.
In 1970, Nina Braunwald performed the first operation
at UCSD [1]. This patient was a 67-year-old human
whose operation was done through a right lateral thora-
cotomy; cardiopulmonary bypass was used. The patient
was discharged from the hospital and returned to full
activity.
Moser and Braunwald observed that the patient had a
“two compartment pulmonary vascular bed.” The open
pulmonary arteries probably had advanced changes of
pulmonary hypertension, but the closed vascular bed,
which had never been exposed to high pressures, had
retained normal structure. In this patient the thrombo-
embolism was known to have been present for 10 years
or more, and this was the first documentation of the
ability to remove such material and to achieve long-term
patency at operation. The pulmonary vascular resistance
decreased from 1208 to 640 dynes · second⫺1 · cm⫺5, and
the patient was discharged well.
Drs Daily, Utley, and Dembitsky, who together per-
formed the next 187 cases in the UCSD series in the 20
years between 1970 and 1989, made progressive modifi-
cations of the surgical technique, including the use of a
median sternotomy and hypothermic circulatory arrest.
Two of the authors of this paper did the subsequent 1,400
cases.
Improvements in Technique
In 1993, the first 323 patients in the UCSD series were
described [3]. Changes had been made in the surgical
management of the cohort comprising the last 150 pa-
tients in that group, which improved results, with more
expeditious removal of occluding material and shorter
circulatory arrest times. These changes included more
proximal incisions, an approach to the right side beneath
the superior vena cava rather than above it, and avoid-
ance of more than one arteriotomy on each side. The
method of raising the endarterectomy plane posteriorly
and leaving normal pulmonary artery in the region of the
incision resulted in less leakage from suture lines.
When we compared circulatory arrest times in the 150
patients operated on between November 1989 and March
1992 with the 100 immediately preceding patients oper-
ated on between October 1987 and October 1989, we
found a reduction in circulatory arrest times from a mean
of 59.16 ⫾ 23.03 minutes to 36.46 ⫾ 16.60 minutes (p ⬍
0.0001). The mortality rate decreased from 17.0% to 8.7%.
In the present series, the mean circulatory arrest time
was 36 ⫾ 11.89 minutes. The mortality rate was 4.4%.
 1462 CHAMBERLAIN MEMORIAL PAPER JAMIESON ET AL
CARDIOVASCULAR
PULMONARY ENDARTERECTOMY
Although the circulatory arrest times have remained the
same, the mortality rate has gradually decreased. The
mortality rate for the entire 1,500 cases in our experience
was 7.5%, compared with 6.7% for the last 1,000. We have
progressively accepted more complicated patients for
operation, including patients in whom it is clear that
thromboembolic disease is only part of the pulmonary
hypertension they present. Operation on more distal
disease (type 3) [2, 4] is more time-consuming and
requires longer periods on bypass but is still able to
restore the pulmonary circulation.
Neurologic injury preventing extubation and normal
patient recovery is no longer seen with this operation.
Although reasons for this are multifactorial, we believe
that a substantive advance in cerebral protection during
pulmonary endarterectomy is the use of the cooling
jacket placed around the head. This head-wrap had been
used in more than 1,000 pulmonary endarterectomies at
our institution without complication, and it provides
even cooling to the surface of the cranium, particularly in
the posterior location. It is easier to apply circumferen-
tially than ice bags and has not been associated with skin
necrosis, as has been seen with other topical cooling
modalities.
The largest risk factor for operation remains the degree
of operability as assessed by PVR. A high PVR without
gross changes on angiogram signifies secondary vascu-
lopathy –an inoperable change and a degree of postop-
erative pulmonary hypertension that will likely hinder
recovery. The mortality rate was 30.6% when the residual
PVR was higher than 500 dynes · second⫺1 · cm⫺5 but
only 0.9% when it was below this level. This disease
represents not direct thromboembolic occlusion, but
rather vascular obstruction at the capillary level, be this a
reflection of primary pulmonary hypertension or changes
in the pulmonary vascular bed as a result of redirection
of pressure or flow as a result of mechanical obstruction
in the closed bed. Interestingly, this observation was
made by Moser and Braunwald after their first San Diego
case in 1971 [1]. Current studies in the laboratory may
provide insight into the molecular pathogenesis of these
changes [6, 19].
Cleonice Gordon and David Garcia were invaluable in data
analysis. Reena Deutsch assisted in the statistical analysis.
Reena Deutsch is funded by General Clinical Research Center
grant National Institutes of Health M01 RR00827.
DISCUSSION
DR CHRISTOPHER G. MCGREGOR (Rochester, MN): First of
all, I would like to thank the authors for sending me the
manuscript in a timely fashion. I would like to congratulate Dr
Jamieson and his team for the development of the world’s
leading pulmonary endarterectomy practice, which is testimony
not only to your surgical skill but to the organization of the team
approach to the management of this demanding disease. In this
paper the authors have demonstrated progressive excellence in
outcomes of pulmonary endarterectomy, establishing the gold
Ann Thorac Surg
2003;76:1457– 64
References
1. Moser KM, Braunwald NS. Successful surgical intervention
in severe chronic thromboembolic pulmonary hypertension.
Chest 1973;64:29 –35.
2. Jamieson SW, Kapelanski DP. Pulmonary endarterectomy.
Curr Prob Surg 2000;37:165–252.
3. Jamieson SW, Auger WR, Fedullo PF, et al. Experience and
results of 150 pulmonary thromboendarterectomy opera-
tions over a 29 month period. J Thorac Cardiovasc Surg
1993;106:116 –27.
4. Thistlethwaite PA, Mo M, Madani MM, et al. Operative
classification of thromboembolic disease determines out-
come after pulmonary endarterectomy. J Thorac Cardiovasc
Surg 2002;124:1203–11.
5. Dalen JE, Alpert JS. Natural history of pulmonary embolism.
Prog Cardiovasc Dis 1975;17:257–70.
6. Thistlethwaite PA, Lee SH, Du L, et al. Human angiopoietin
gene expression is a marker for severity of pulmonary
hypertension in patients undergoing pulmonary thrombo-
endarterectomy. J Thorac Cardiovasc Surg 2001;122:65–73.
7. Hollister LE, Cull VL. The syndrome of chronic thromboem-
bolism of the major pulmonary arteries. Am J Med 1956;21:
312–20.
8. Freiman DG, Suyemoto J, Wessler S. Frequency of pulmo-
nary thromboembolism in man. N Engl J Med 1965;272:
1278 –80.
9. Riedel M, Stanek V, Widimsky J, Prerovsky I. Long term
follow-up of patients with pulmonary embolism. Late prog-
nosis and evolution of hemodynamic and respiratory data.
Chest 1982;81:151–8.
10. Lewczuk J, Piszko P, Jagas J, et al. Prognostic factors in
medically treated patients with chronic pulmonary embo-
lism. Chest 2001;119:818 –23.
11. Mo M, Kapelanski DP, Mitruka SN, et al. Reoperative
pulmonary thromboendarterectomy. Ann Thorac Surg 1999;
68:1770 –7.
12. Archibald CJ, Auger WR, Fedullo PF, et al. Long-term
outcome after pulmonary thromboendarterectomy. Am J
Respir Crit Care Med 1999;160:523–8.
13. Trendelenburg F. Uber die operative behandlung der em-
bolie derlungarterie. Arch Klin Chir 1908;86:686 –700.
14. Cooley DA, Beall AC, Alexander JK. Acute massive pulmo-
nary embolism: Successful surgical treatment using tempo-
rary cardiopulmonary bypass. JAMA 1961;177:283.
15. Lungdahl M. Gibt es eine chroniche embolisierg der lunge-
narterie? Deutsches Archiv Klin Med 1928;160:1.
16. Carroll D. Chronic obstruction of major pulmonary arteries.
Am J Med 1950;9:175–85.
17. Moser KM, Rhodes PG, Hufnagel CC. Chronic unilateral
pulmonary artery thrombosis. N Engl J Med 1965;272:1195–9.
18. Chitwood WR, Sabiston DC, Wechsler AS. Surgical treat-
ment of chronic unresolved pulmonary embolism. Clin
Chest Med 1984;5:507–36.
19. Du L, Sullivan CC, Chu D, et al. Signaling molecules in
non-familial pulmonary hypertension. N Engl J Med 2003;
348:500 –9.
standard for both surgical mortality rate and hemodynamic
outcomes.
I and others in the field would like to acknowledge Dr
Jamieson’s group for education of colleagues in the United
States and overseas about the disease itself, patient selection,
and surgical techniques. They have been gracious hosts to many
of us in San Diego. In addition, you are to be congratulated not
just for the development of the clinical practice but also for
developments in the basic laboratory research— by Dr Thistle-
Ann Thorac Surg
2003;76:1457– 64
thwaite and other members of Dr Jamieson’s team—in the
pathogenesis of chronic thromboembolic pulmonary hyperten-
sion. I would like to focus on the specific lessons from this paper
regarding the surgical risks for individual patients who undergo
pulmonary endarterectomy.
You outlined that a learning curve exists for this operation,
with a 17% mortality in the San Diego group in the first 200
patients in the series, decreasing to 4.4% for the last 500 patients.
This experience reflects our own smaller experience where the
operative mortality rate in our own program decreased from
19% in the first 21 patients operated on before 1997 to 4.6% in the
most recent 43 consecutive patients. Hemodynamic outcomes,
interestingly, have been almost identical between our own
program and the UCSD program, with a dramatic decrease in
mean pulmonary artery pressure to a range where we know that
prognosis is good.
In your presentation, Dr Jamieson, you stated that major
preoperative risk factors included a pulmonary vascular resis-
tance of 1,000, especially in the context of limited angiographic
disease, but that preoperative right ventricular function was not
a risk factor. Because of the availability of ultrafast computed
tomography at our institution, we have been able to develop
some unique data confirming that improvement in right ven-
tricular ejection fraction is consistent no matter what the original
preoperative calculation is.
As you can see in the slide, right ventricular ejection fraction
in a recent cohort of 37 patients improved significantly, from
36% to 51%. Similar improvement occurred across the spectrum
of preoperative right ventricular ejection fraction levels. So no
matter how low the initial right ventricular ejection fraction was,
there is a real possibility for improvement after pulmonary
endarterectomy.
Associated with this increase in right ventricular ejection
fraction is a decrease in right ventricular end-diastolic volume,
from a mean of 254 mL before pulmonary thoracoendarterec-
tomy PTE to 157 mL after the procedure. You also describe the
high operative risk when residual pulmonary vascular resistance
is greater than 500, especially if only a limited endarterectomy is
possible.
I have a number of questions for the speaker. First, can you
comment on the wider application of this surgical procedure in
terms of appropriate medical and surgical resources, commit-
ment, and number of cases?
Second, do you believe that pulmonary angiography remains
absolutely necessary as part of the preoperative evaluation of
the patient, or can computed tomography be substituted?
Third, what happens to patients who do have a residual
pulmonary vascular resistance greater than 500? How do we
follow up these patients?
With the great improvement in outcomes that you have
described today, should the threshold in terms of symptoms for
recommending pulmonary endarterectomy be changed, ie, low-
ered, in the setting of significant elevation of a pulmonary
vascular resistance as we know that the disease will likely
progress and be fatal? In other words, should patients be
operated on earlier, if possible, even with limited symptoms?
And finally, is there still a place in rare circumstances for lung
transplantation in patients with a pulmonary vascular resistance
greater than 1,000 and the presence of limited angiographic
disease?
DR ALVAN W. ATKINSON (Raleigh, NC): I have several
questions. One, would magnetic resonance imaging be more
favorable than computed tomography for staging these patients?
Could you give us a little insight on some stratification of risk in
CHAMBERLAIN MEMORIAL PAPER JAMIESON ET AL 1463
CARDIOVASCULAR
PULMONARY ENDARTERECTOMY
terms of magnetic resonance imaging, computed tomography,
or angiographic findings?
Would all patients need a filter postoperatively, or is there
some selection to that, whether anticoagulation is tolerated or
not?
And finally, I know there is a new drug out for treatment of
pulmonary hypertension, and I wondered if you have any
experience with it preoperatively perioperatively, or
postoperatively?
DR JOSEPH E. BAVARIA (Philadelphia, PA): I congratulate you
on a fantastic series. We have performed about 200 of those
operations since Dr L. Henry Edmunds visited Dr Daily in San
Diego during the mid-1980s. One of the things we have noticed
in last 5 years or so is that our results are better if there is no
significant concomitant pulmonary disease. In other words, we
do better if we make sure there is no IPF, chronic obstructive
pulmonary disease, or other nonvascular pulmonary diagnosis
which will compromise postoperative results. Would you like to
comment on your preoperative pulmonary risk evaluation?
DR ANDREA M. D’ARMINI (Pavia, Italy): Congratulations on
your excellent presentation. I have a question. Did you see any
difference in outcome based on duration of the disease and how
long the patient was classified New York Heart Association class
III or IV? Considering the natural progression of chronic throm-
boembolic pulmonary hypertension to irreversible right cardiac
failure, did you see better results in patients with a shorter
duration of the disease? If so, could this be a reason to perform
this operation in patients still classified as New York Heart
Association class II?
DR JAMIESON: I would like to thank you very much for your
kind comments. I think we all realize that it is our responsibility
to teach what we learn, and we have tried to do this. If any credit
is to be given to our effort with this disease, it of course goes to
the team, without whom success would not be possible.
With regard to the first question about the wider application
of pulmonary endarterectomy in terms of resources and com-
mitment, this is a difficult question because there is definitely a
learning curve to the procedure and, unquestionably, a team
approach is necessary. Perhaps at least a commitment to the
procedure with a minimum volume of 10 or 20 cases a year
would be a reasonable objective.
Regarding pulmonary angiography, there is some risk with
the procedure but it is not high. We have done probably 5,000
pulmonary angiograms with no mortality and very limited
morbidity. There are other techniques, as the subsequent dis-
cussant mentioned. I think whether you absolutely need to use
angiography really depends on the pulmonary vascular resis-
tance, because the most fundamental risk is to operate on a
patient in whom the pulmonary vascular resistance is out of
context with the degree of angiographic obstruction. I would
suggest, perhaps, as a compromise that if the pulmonary vascu-
lar resistance is below 500, it is not absolutely necessary to do an
angiogram. Otherwise it probably is, in order to get an exact
road map for the technical portions of the procedure and to
match the degree of angiographic obstruction with the pulmo-
nary vascular resistance.
With regard to the threshold of when we should be doing
pulmonary endarterectomy, I believe absolutely that we should
be doing it earlier. We have become increasingly aware of the
secondary changes that occur in the pulmonary vascular bed
because of increased pressure and flow in the open vessels, and
that, of course, is eventually inoperable.
 1464 CHAMBERLAIN MEMORIAL PAPER JAMIESON ET AL
CARDIOVASCULAR
PULMONARY ENDARTERECTOMY
With regard to the place of lung transplantation, we have
applied lung transplantation in perhaps half a dozen patients in
whom we were not able to get a good result because of type 3 or
type 4 disease. I suppose that if patients were clearly at very high
risk and the angiographic appearances suggested that they
would get limited benefit, then serious consideration of trans-
plantation should be given at that time. Other factors to be
weighed in the decision here would be the projected length of
time on the waiting list and the likelihood of a suitable donor.
With regard to follow-up, I think follow-up with pulmonary
artery pressures could largely be done with echocardiography.
To answer the question about the filter, we believe it is
important that all patients have an inferior vena cava filter
placed preoperatively.
With regard to the newer drugs coming out for treatment of
pulmonary hypertension, these are immensely promising. But of
course, this is a mechanical condition. It cannot be treated by
angioplasty; it can only be removed by open operation. And you
Ann Thorac Surg
2003;76:1457– 64
might tide somebody through temporarily and improve the
condition somewhat before the operation by using pulmonary
vasodilating agents, but in the end the only thing that will help
is physical removal of the obstruction.
Dr Bavaria brought up an important question about the
comorbidities of the patient, particularly with regard to lung
function, and I think the answer depends on the assessment of
how much residual lung function will be left once you have a
normal vascular supply to the lung.
I am sorry, Dr D’Armini, that we did not have time to discuss
in more detail the preoperative status of the patients. However,
there is no question that if there is mechanically occluding
disease causing pulmonary hypertension, no matter where it is
in the pulmonary vascular bed, this can be removed at opera-
tion, and delay is not in the patient’s interest. I think occluding
material can be removed at any time period, even many years
after obstruction, and there is ample justification for doing this
before the patient becomes severely ill.