- Email: [email protected]
The results of a surgical program for interrupted aortic arch
J
THoRAc CARDIOV ASC SURG
1988;96:864-77
The results of a surgical program for interrupted aortic arch Seventy-one patients with interrupted arch entered the Boston Children's Hospital between Jan. 1, 1974, and Jan. 1, 1987, of whom 63 underwent an operation. Type B was the most prevalentform of interrupted arch, and ventricular septal defect alone was the commonest coexisting cardiac anomaly. Among the 63, the 3O-day and the 1-, 5-, and IO-year survival rates were 61 %,52%,48%, and 47%, respectively. The mortality rate declined strikingly during the experience, and by multivariate analysis in patients with coexisting ventricular septal defect operated on in 1986, the probability of death within 2 weeks of repair was only 7 %. Also, preoperative therapy became progressively more Iateese and more prolonged. The complication of left ventricular outflow tract obstruction developed in eight of the 33 patients undergoing repair of interrupted arch and of isolated ventricular septal defect The time-related freedom from this complication was 97%, 78%, and 58% at 1 month, 1 year, and 3 years, respectively. Seven of the eight patients underwent a surgical procedure directed against the left ventricular outflowtract obstruction, and all have survived. Recurrent or persistent aortic arch obstruction became evidentafter repair in 15 patients and appeared more frequently and earlier after direct anastomosis than after tube graft repair. AU patients had either reoperation or balloon dilation, but all were alive at foUow-up. Most surviving patients are active and without symptoms. Inferences: An aggressivesurgical program can result in survival and a good clinical state for at least 10 years after birth of over 40% of patients born with interrupted arch. Multiple anatomic bases account for the development of left ventricular outflowtract obstruction in about 50 % of the patients undergoing repair of interrupted arch with coexisting ventricular septal defect. Repair by direct anastomosis combined with repair of the coexisting defect whenever possible is optimal therapy.
Jeffrey E. Sell, MD" (by invitation), Richard A. Jonas, MD" (by invitation), John E. Mayer, MD" (by invitation), Eugene H. Blackstone, MD,b John W. Kirklin, MD,b and AIdo R. Castaneda, MD,a Boston, Mass., and Birmingham, Ala.
Numerous reports of the surgical treatment of interrupted aortic arch have appeared in the literature in the past 20 years, but few have contained enough cases followed a reasonable period of time to allow inferences as to the spectrum of the disease, optimal therapy, and early and late results in terms of survival and complications. Specifically, little reference has been made to the appearance after repair of left ventricular outflow tract From the Boston Children's Hospital and the Harvard Medical School, Boston, Mass.," and the Medical Center of the University of Alabama at Birmingham, Birmingham, Ala." Read at the Sixty-eighth Annual Meeting of The American Association for Thoracic Surgery, Los Angeles, Calif., April 18-20, 1988.
Address for reprints: Richard A. Jonas, MD, The Boston Children's Hospital, 300 Longwood Ave., Boston, MA 02115.
864
obstruction (LYOTO) or of persistent or recurrent aortic arch obstruction. A careful study of a 13-year experience, including detailed follow-up of all patients not known to be dead, has provided new but as yet incomplete information about these matters.
Patients and methods Patients. All 71 patients with interrupted aortic arch entering the Boston Children's Hospital between Jan. 1, 1974, and Jan. 1, 1987, are included in the study. Sixty-three patients underwent an initial operation, and eight others either died before an operation could be done (three patients) or were considered inoperable because of extremely complex major associated cardiac or noncardiac anomalies, or both (five patients). One patient in the latter group is alive at follow-up 23 months later. The median age at entry was 4 days (Fig. I).
Volume 96 Number 6
Interrupted aortic arch
December 1988
865
100
.90 m
4(eo
•
'0
ai 70
US
% Entry
Age (days)
>060
lD
2
~50
4
-.:
2S% SO% 7SO/. 90 %
20
S40 c
at~
Stated Age
54
W
iu
30
12O 10 10
20
30
40
50
60
Age (days) at Entry
eo
70
90
Fig. 1. Cumulative frequency distribution plot depicting age at entry into the hospital (and thus into the study). Four patients were older than 90 days at entry, being 123 days, 299 days, 489 days, and 1371 days of age.
Table I. Surgical treatment for interrupted aortic arch (Boston Children's Hospital; 1974-1987; n = 71; deaths = 39) Time period Events Death at any time Preoperative PGE , Preoperative low-dose dopamine Preoperative intubation Mean immediately preoperative pH Median interval between admission and operation (days) One-stage repair Direct anastomosis
/974-/979
12/15 3/13 1/13 6/13 7.26 0.6
(80%.) (23%) (8%) (46%)
4/13 (31'it) 3/13 (23%)
/98/-/983
/979-/98/
10/14 7/12 0/12 2/12 7.34 0.7
/983-/985
(71%) (58%) (0%) (17%)
4/9 (44%) 5/9 (56'1C) 0/9 (0%) 5/9 (56%) 7.41 1.8
12/21 17/18 6/18 13/18 7.41 1.4
7/12 (58%) 1/12 (8%)
4/9 (44%) 0/9 (0%)
12/18 (67%) 4/18 (22o/r)
(57%) (94%) (33%) (72%)
/985-/987
1/12 9/11 7/11 9/11 7.43 2.0
(8%) (82%) (64%) (82%)
9/11 (82%) 9/11 (82%)
P for difference 0.0002 <0.0001 <0.0001 0.003 0.0006 0.002 0.004 0.004
Sole: The event death at all.. lime pertained to the entire group of 71 patients; all other events pertained to the 63 who underwent an initial repair.
Type B was the most prevalent form of interrupted arch (Appendix A, Table AI). Twelve of the patients with this type clearly had the DiGeorge syndrome, and others may have had it as well; the DiGeorge syndrome did not coexist with types A or C. The commonest coexisting cardiac anomaly was ventricular septal defect (VSD) (Appendix A, Table A2). Multiple VSDs were present in five patients, including two with VSD as the only coexisting cardiac anomaly, two with double-outlet right ventricle, and one with truncus arteriosus. All five patients have died. Preoperative management. Most of the patients were seriously ill on admission, 30'k having an arterial pH equal to or less than 7.3 at the time of admission (range 6.9 to 7.6). Patient management protocols in the current era, which have evolved gradually through the l4-year period (Table I),
included the use of prostaglandin E" intubation and ventilation, and low-dose dopamine. The prostaglandin E, was infused through a secure intravenous line and usually at a rate of 0.05 J.lg . kg-I. min-I (earlier, 0.1 J.lg. kg-I. min-I), to maintain ductal patency and thereby good perfusion of the organs and tissues of the lower part of the body. This, plus the aggressive use of sodium bicarbonate and tromethamine, usually effectively combated metabolic acidosis. After intubation, ventilation was conducted so as to maximize the pulmonary /systemic resistance ratio and to encourage flow through the ductus arteriosus and into the descending thoracic aorta. Thus the ventilatory gas was room air, arterial carbon dioxide tension was maintained at about 40 mm Hg, end-expiratory pressure of 5 to 10 em H 20 was used. Arterial blood pressure was usually monitored by a catheter in the umbilical artery or
The Journal of
866
Thoracic and Cardiovascular Surgery
Sell et al.
B Fig. 2. Depiction of technique of repair by direct anastomosis of interrupted arch type B. A, Purse strings on the aorta and pulmonary arteries indicate the siting of the two arterial cannulas. Before cannulation, all of the vessels are thoroughly dissected to maximize their mobility and minimize tension on the anastomosis. B, The completed anastomosis is shown. The position of the arterial cannula in the ascending aorta is illustrated, as the precisely correct positioning of this cannula is critical for even perfusion of the body and the coronary arteries.
Table II. Surgical treatment for interrupted aortic arch (Boston Children's Hospital; 1974-1987) Deaths Type of repair
n
No.
%
CL
One stage Multiple stage
36 27
13 19
36 70
27-46 59-80
Total
63
32
51
44-58
P (X')
0.007
Note: The deaths include all deaths. early and late, after repair.
an artery in the foot. Cardiac output and renal blood flow were usually maintained with dopamine infused at a rate of about 5 J.Lg • kg-I. min-I. Surgical methods. Like the preoperative management, the methods used in the current era have gradually evolved over the period of the study. In the current era, a one-stage repair was performed as the initial surgical procedure unless the intracardiac morphology demanded a Fontan type of repair (Table II). The one-stage repair was performed during profoundly hypothermic total circulatory arrest, the technique of which has been described previously.I In most patients, the technique was modified for patients with interrupted arch by
placing an arterial perfusion cannula not only in the ascending aorta, but also in the pulmonary artery. The right and left pulmonary arteries were temporarily occluded while this latter cannula perfused the lower part of the body through the ductus arteriosus. Often, the aortic interruption was repaired by direct anastomosis (Fig. 2), although a tube graft reconstruction was more frequently used in the early part of the experience.' The technique for direct anastomosis has been described in detail elsewhere.' Miscellaneous other types of reconstruction were also employed (Table III). The VSDs were repaired with a right atrial approach, and in the last 10 years of the experience cold cardioplegia was used. Data collection and follow-up. The details of the preoperative condition and management of the patient, of the operations, of postoperative care, and of subsequent studies and procedures, both at the initial and subsequent hospitalizations and at outpatient visits, were obtained from the records of the Boston Children's Hospital. In October 1987, a formal follow-up was performed of all patients not known to be dead; Both letters and telephone calls to families and to the patients' physicians at home were used. Two patients could not be reached, one of whom had not been able to be contacted since hospital dismissal 19 days after repair and the other of whom was known to be alive 26 months after repair but has been untraced since then. The median follow-up time was 43 months and the range, 10 months to II years. Six years or
Volume 96 Number 6
Interrupted aortic arch 8 6 7
December 1988
Table
m. Surgical treatment for interrupted aortic arch (Boston Children's Hospital;
/974-/987) Deaths
Type of initial repair Direct anastomosis + Correction of cardiac anomaly + Palliation of cardiac anomaly Isolated TGR + Correction of cardiac anomaly + Palliation of cardiac anomaly Anomaly LSCA to descending aorta + Palliation of cardiac anomaly Isolated LSCA flap aortoplasty (with palliation of cardiac anomaly) Preservation of ductal continuity (with correction of cardiac anomaly) LV-aortic conduit as repair of IAA (with correction of cardiac anomaly) TGR + LV-aortic conduit as repair of IAA (plus palliation of cardiac anomaly) TGR + LSCA to proximal aorta as repair of IAA (plus correction) No operation Total
% of63
n
17
No. 4 2
I
0
35
29
100 0 19
56
4
86
33
6
3
I
I
0
54
33
6 12 I
14 3
35
6
27
14 2 18
%
33 0
2
25 100 50 100
2
2
100 100
8
7
88
71
39
55
Key: IAA. Interrupted aortic arch: LSCA. left subclavian artery; LV. left ventricle: TGR. tube graft reconstruction .
.VOle: The table includes
late.
all patient' entering the hospital. including the eight (seven deaths) who did not reach operation. The deaths include all deaths. early and
more of follow-up information is available for 22% of the 71 patients. Data analysis. The data were sorted and examined from numerous points of view. Multiple contingency tables were constructed. Incremental risk factors for the presence or absence of an event were sought through logistic regression analysis: Time-related events were examined by both actuarial' and parametric methods, and risk factors were sought in the parametric (hazard function) domain." The usual methods were used to determine the likelihood that differences were due to chance alone.
Results Survival. Among the 71 unselected patients, 39 died during the initial hospitalization or later (Fig. 3). The prevalence of death has declined strikingly across the time period of the study (Table I). Among the 63 patients who underwent an initial operation, the 3Q-day survival rate was 61%. The time-related survival rate (Fig. 4) had a single phased, rapidly declining hazard function (Fig. 5), no deaths having occurred more than 19 months after repair (24 patients traced beyond 19 months). Solution of the equation describing the risk factors for death at any time after repair (Table IV)
emphasized the improvement in survival across the time period of the study, emphasized the more favorable outcome in the patients in whom (VSD) was the sole coexisting cardiac anomaly (Fig. 6), and also demonstrated the importance of the immediately preoperative arterial pH (Fig. 7). The earlier preoperative pHs, including that on admission, did not influence outcome. Although a good result was obtained infrequently in patients with the other associated cardiac anomalies (see again Appendix Table A2), the possibility existed. Thus one of the two patients with coexisting transposition of the great arteries and VSD underwent initial interrupted arch repair by direct anastomosis, then 4 days later a pulmonary artery banding, and at 6 months of age an arterial switch repair; one year later the child was in good health. The other child with transposition and VSD has a hypoplastic right ventricle and is scheduled for a Fontan repair. Changes during the 13-year period. Many preoperative and intraoperative methods changed during the time period of the study, and particularly the frequency of intense preoperative therapy (see again Table I). The
The Journal of Thoracic and Cardiovascular
8 6 8 Sell et al.
Surgery
100
tr---------. Years
90 80
70
..
'ii
~60
-e ::I
(/)50
c: 1)40
I)
Q.
% Survival
1/52
72%
1/12
60%
6/12
50%
1
48%
2 5 10
46% 44% 42%
---------------------------------------------------------------------
30
20 10 Ol:.-_--'--_---J._ _..L..-_--'--_---J._ _..L..-_--'--_---J._ _..L..-_--'--_--' 11 10 456 7 9 8 2 o
Years After Entry
Fig. 3. Time-related survival of patients (n = 71) entering with interrupted arch, including all deaths. Time zero is the time of entry. The open circles represent individual deaths, actuarially positioned. The solid line indicates the parametrically determined survival. The dashed lines enclose the 70% confidence intervals. The dash-dot-das]i line represents the survival of an age-matched general population, U.S. Life Tables, 1976.
100
90 80
'ii >
.
70
>60
-
Years
% Survival
1/52
70%
1/12
61%
6/12
54%
1
52%
2
50%
5
48%
10
47%
::::I
(/)50'
c:
~40
~
-----------------------------------
30 20 10 oL-_...L..._.J..._---I._ _L-_..l....-_--'--_--l.._---.JL-_..l....-_--'--_--' 11 9 10 4 5 6 7 8 2 o
Years After Repair
Fig. 4. Actuarial and parametric time-related survival after initial operation (n = 63) for interrupted aortic arch. Time zero is the time of the initial operation. The depiction is as in Fig. 3.
prevalence of the type of aortic arch and of the various types of associated cardiac anomaly did not change across time, nor did the prevalence of patients not undergoing operation (P::::: 0.4 in each instance).
LVOTO. Systemic ventricular outflow tract obstruction (LVOTO) became evident soon after repair in two patients with associated "single ventricle," both of whom died. (LVOTO was considered to be present when the
Volume 96 Number 6 December 1988
Interrupted aortic arch
869
0.10 0.09 0.08
...
-
I.e
c
0
~
III
.e
0.07
·•
0.06
•I
Years
" "
.,
1/52 1/12
"
6/12
0.05
cu 0.04 cu C
0.03
•I
" " " • I 'I 'I I
0.02 0.01
I \
Hazard IX 1000) 410 83 10 4.5 1.9 0.61 0.26
1 2 5 10
,
, ,
, ,, ,, , ,
- ---:-:: -18
24
:30
36
Months After Repair
42
48
54
60
Fig. 5. Hazard function for death in the 63 depicted in Fig. 4. The dashed lines enclose the 70% confidence intervals of the hazard function. The barely visible dot-dash-dot line is the hazard function of an age-matched general population.
gradient between left ventricle and ascending aorta was 40 mm Hg or more by echocardiography or cardiac catheterization.) The mechanism of the obstruction has been described earlier.' LVOTO also became evident in eight patients with coexisting isolated VSD (Fig. 8), all ofwhom were alive at the time of follow-up. The hazard function for the appearance of evident LVOTO in this group had a single declining hazard phase (Fig. 9), which had reached a low level by 36 months after repair. The complication of LVOTO obstruction developed only in the subset of 33 patients with the VSD in the outlet portion of the septum (n = 33), cradled in the limbs of the septal band (or trabecula septomarginalis), and in many of these patients the surgeon had noted posterior malalignment of the outlet (conal, infundibuIar) septum. Seven of the eight patients underwent a surgical procedure directed against the LVOTO (Table V). Among the three in whom subvalvular obstructing muscle or fibrous tissue was resected, one had a 10 rom Hg gradient 23 months later, one a 60 rom Hg gradient 28 months later, and one a 100 mm Hg gradient and a small aortic anulus 21 months later. Among the three treated by placement of a left ventricular-aortic conduit, one had an 80 rom Hg gradient between the left ventricle and aorta 1 year later; this child underwent revision of the proximal conduit anastomosis, 11 months after which a 50 rom Hg gradient existed. A second
Table IV. Surgical treatment for interrupted aortic arch (Boston Children's Hospital; 1974-1987) Incremental risk factors for death at any time after repair In = 63) Demographic Date of operation (earlier) Morphologic "Single ventricle" Truncus arteriosus Multiple VSDs Preoperative condition Immediately preoperative pH (lower) Surgical LV-aortic conduit See Appendix B for variables entered into the analysis. coefficients. and P values.
patient among the three had an 80 rom Hg gradient 23 months after placement of the left ventricular-aortic conduit, 28 months later had transaortic resection of subaortic stenosis, and 22 months after that had a 65 rom Hg gradient between the left ventricle and ascending aorta. The third patient is without symptoms 5 years later but is known to have had a gradient of 60 mm Hg at catheterization 11 days after placement of the conduit. The patient in whom aortic valvotomy was performed had a dysmorphic, bicuspid valve, and a second surgical valvotomy was performed 14 months
The Journal of
870
Thoracic and Cardiovascular Surgery
Sell et al.
100 III
~
at 90
~
N 80
-
c :E 70
~60
-
.c
: 50
C
040
--1976
A
1978
1980
1982
1984
1986
Date of Operation
100
; 90
~
M80 c
£ ~
-
70
.c as
--
at C 50 0
40
~
!...
>-
30
:!::
:cas
.0 0
~
20 10 0
B
1974
--1976
1978
1980
1982
1984
1986
Date of Operation
Fig. 6. A, Nomogram from the multivariate equation (Table IV) illustrating the effect of date of operation, in patients with interrupted aortic arch and VSD, on the probability of death within 2 weeks. (See Appendix B for details.) B, Nomogram similar to that in A, except illustration is of the probability of death within 3 years according to the date of operation.
later. The patient receiving no treatment had a 45 mm Hg gradient 11 months after the initial one-stage repair and is well but without further studies 2 years later. Persistence or redevelopment of evident aortic arch
obstruction. Aortic arch obstruction became evident sometime (0.2 to 62 months) after repair in 15 patients (Table VI). (This complication was considered to be present when the gradient between the ascending and
Volume 96 Number 6
Interrupted aortic arch
December 1988
100 U) ~
90
~
SO
QI
---- .. --.
...
-- .. --
'- '-
........
........
N
C
70
~
60
tV QI
50
:E or:
-C
........
........
40
0
-
30
0~
~ 20
:s
~ 10
...o no
------ ------~~=-=-:-::::-::-::-::-:-:-:~;.;.~...::..::...:..::...:
------------0 1:..L............u.....& ....................................................u.....& ....................................................U-J................................L..L..J.......... U-J................................L..L..J..........U-J....J...J 7.1 7.2 7.3 7.4 7.0 7.5 6.9 Immediately Prerepair pH
Fig. 7. Nomogram from the multivariate equation (Table IV). The small arrow indicates the "point of evident difference" in the case of VSD, that is, the point (7.32) at which, compared with a pH of 6.9, the probability of death is reasonably certainly (nonoverlapping 70% confidence intervals less. (See Appendix B for details.)
100
o
90
b SO
--- --- --- ----
~
_70 c QI :260 > w
'0
(13r - _ cia)---_
50
:40
~
Outlet VSD Repair Subgroup (n=33; events=8l
'ECIl 30 o
~ l1.
20 10
6
12
--- ---
---------------
6)-
-----Years
% Freedom
1/52 1/12 6/12
99% 97% 87%
1
78%
2
64%
3
58%
18 24 30 36 42 Months After VSD Closure
60
Fig. 8. Actuarial time-related freedom from evident LVOTO, among the 33 patients undergoing repair of a VSD in the outlet portion of the right ventricular aspect of the septum. (In four other patients, the VSD was in the trabecular portion of the septum or in an unknown position.) The circles represent individual events (the appearance of evident LVOTO) and are actuarially positioned. The vertical bars enclose the 70% confidence intervals and the lines merely connect the circles. The last event is at 36 months.
87 I
The Journal of Thoracic and Cardiovascular
8 7 2 Sell et al.
Surgery
0.050 0.045
...
\ \
\
0.040
\
\ \ \
I.e
'E 0.035 o
.
~0.030
\
\ \
" "
~0.025
o
-
Outlet VSD Repair SUbgroup , (n=33; event8=8)
,,
,,
~O.020
,,
Years
Hazard (x 1000)
1/52 1/12
27.0
6/12
20.6
,
c
-80.015
';:
UJ 0.010
~
0.005
-
~
23.9
1
18.9
2
12.1
3
4.6
-
~-
~
~
~
~
~-
~
---------~-~-~--------------
12
6
--
18
24
:30
36
Months After VSD Closure
42
48
54
60
Fig. 9. Hazard function for the appearance of evident LVOTO at each moment in time after the repair in the 33 patients. The hazard function has a single declining phase. The depiction is as in fig. 5.
Table V. Surgical treatment for interrupted aortic arch (Boston Children's Hospital; 1974-1987) Type of treatment for post repair LVOTO
Deaths n
No.
%
Resection (15, 18, and 39 months after VSD repair) LV-aortic conduit (0.4, 10, and 14 months after VSD repair) Aortic valvotomy only (4.4 months after VSD repair) No treatment
3
0
0
3
0
Total
8
0
CL %
0
0 0
0
0-21
descending aorta was 30 mm Hg or more by echocardiography or cardiac catheterization.) The use of a direct anastomosis was a risk factor for postrepair aortic arch obstruction, as was earlier date of repair when a direct anastomosis was used (Table VII). All patients with this complication had either a reoperation or balloon dilation performed for the aortic arch obstruction. Nonetheless, all but two of the patients were alive at the time of follow-up. This event occurred sooner and more frequently after direct anastomosis than after tube graft reconstruction (Fig. 10). Status of surviving patients. Despite treated and
untreated persistent or recurrent obstructive lesions in many patients, most of those surviving at the time of follow-up were active and without symptoms (Table VIII). Discussion Patients. The patients were relatively unselected, the only criteria for inclusion being that the diagnosis be made. Most of the patients were entered into the study when they were only a few days old. Most previous publications reported only a small number of patients undergoing an initial operation, and many have included only one morphologic subset rather than all patients with interrupted aortic arch.v" Type B was the commonest type in this experience, as it was in the experience of Schumacher's group" and others. Although VSD was most commonly the associated cardiac anomaly, a wide variety of other cardiac anomalies coexisted with interrupted arch. Because of the relatively unselected nature and relatively large size of the patient group, the prevalence of the various types of interrupted arch and of the associated cardiac anomalies, and the relation between these two, can be considered representative for patients in general with interrupted arch. Survival. The 5- and I O-yearsurvivorships reported in this paper (44% and 42%, respectively, of all patients with this diagnosis entering the hospital) are distinctly
Volume 96 Number 6
Interrupted aortic arch
December 1988
Table VI. Surgical treatment for interrupted aortic arch (Boston Children's Hospital; 1974-1987) Redevelopment (or persistence) of evident aortic arch obstructioni Type of arch repair (n = 63) Direct anastomosis Subclavian flap aortoplasty
Blalock-Park repair" Formalization of PDA Tube graft repair Tube graft repair plus LV-aortic conduit Tube graft repair plus LSCA to proximal aorta anastomosis l.V-aortic conduit Total
CL
n
No.
%
(%)
17 I 4 2 35 2
5:\: I§ 211
29 100 50 50 14 50
17-45 15-100 18-82 7-93 8-23 7-93
63
I~
1··
5#
15
0
0-85
0
0-85
24
18-31
Key:CL.. 707, confidence limits; LSCA. left subclavian artery; LV.Ieft ventricle; PDA. patent ductus arteriosus. 'End-to-side anastomosis of the left subclavian artery to the descending thoracic aorta. tfhe percents are percents of n. and have no time-related or actuarial implications. +The correction was a patch graft enlargement in two (one died after the procedure; the other lived but is without further information), and by balloon dilation in three (in two, there was no gradient early after dilation; in the third, a second dilation was performed 3 months after the first one). §Balloon dilation was performed 11.5 months after repair, which reduced the gradient to 18 mm Hg (also had anastomosis of proximal pulmonary artery to ascending aorta. for subaortic stenosis). I'nterval was 0.2 to 2.4 months; one patient had aortoplasty of the anastomotic site with a polytetrafluoroethylene graft, with all pulses the same thereafter; the other had an ascending-descending aorta bypass graft. resulting in a 10 mm Hg residual gradient. ~Interval was 4.8 months, at which time a left subclavian artery-descending aorta anastomosis was made followed by a J0 mm Hg gradient, according to extremity of blood pressure.
s Three were treated. one by ascending aorta-descending aorta tube graft, one
bypatch graft enlargement of distal anastomosis, and one by conversion to direct anastomosis. "'nterval was 2.9 months; treatment was by a polytetrafluoroethylene interposition tube graft between the left subclavian artery and ascending aorta, followed in I day by patch graft enlargement of the anastomosis, and then in four days by replacement of the LV-aortic conduit; the patient died 2 months after the last procedure.
encouraging, particularly since no deaths have occurred more than 19 months after admission and 24 patients have been traced beyond this point. The outlook for patients with coexisting VSD is particularly favorable, as can be seen by the projected 3-year survival rate of 88% for such patients operated on in the current era (see again Fig. 6, B). The era effect. Many things changed across the 13-yeartime period of this study, including the mortality rate (see again Table I). Clearly, the decreasing mortal-
873
Table VII, Surgical treatment for interrupted aortic arch (Boston Children's Hospital; 1974-1987) Incremental risk factors for evident recurrent or persistent aortic arch obstruction (n = 63) Direct anastomosis Date of direct anastomosis (earlier) See Appendix C for variables entered into the analysis, coefficients, and P values.
ity rate did not result from the physicians being more selective as to who was treated surgically or from patients with favorable lesions being more frequent in the recent time periods. Probably the decreasing mortality rate is related to more intensive preoperative preparation of the patient with prostaglandin E 1 to improve lower body perfusion and thereby improve renal function and diminish acidosis": 15; with low dose dopamine to improve renal and cardiac function; and with intubation and ventilation to maximize the pulmonary/ systemic resistance ratio and thus the flow across the ductus and into the descending aorta. These same measures have proved to be helpful in patients with the hypoplastic left heart syndrome." Probably an important reason for the continuing high mortality among patients whose coexisting cardiac anomaly is "single ventricle," truncus arteriosus, and other complex conditions is that in the past a strenuous consistent effort to repair all coexisting cardiac anomalies was not made. Methods for doing this safely should improve results in the future. In the current era, a one-stage repair with direct anastomosis of the proximal to the distal aortic segment was the most prevalent technique. Clearly, this has had no adverse effect on survival, but it has not yet been shown that it has had a favorable effect. LVOTO. Becu,' Van Praagh," and their colleagues very early described the tendency toward LVOTO in patients with interrupted aortic arch and coexisting VSD; and more recent studies have refined knowledge in this area.":" According to the present study, in about 90% of patients with this combination of anomalies the VSD is in the outlet portion of the septum, and presumably only in them is there a predisposition to LVOTO. The experience here reported supports the idea that the basis for the obstruction is present at the time of VSD repair and therefore becomes evident early after repair, with the probability of its becoming evident (hazard function) decreasing as time passes. It would appear that about half the patients with interrupted aortic arch and VSD have only mild initial narrowing of
874
The Journal of Thoracic and Cardiovascular Surgery
Sell et al.
100
-... c
.2 90 U
::::I
In
121
80
.Q
0
.r:. u
...
0(
-...
.2
-
70 Tube Graft ("=35; event.=51
60
(51
50
Direct Ana.tomo." ("=17; eventa=SI
0
c(
40
0
Q) Q)
-...
It c
(21
30 20
Q)
u
I-
10 0
2
0
:3
4
Years After Repair
5
6
7
8
Fig. 10. Actuarial time-related freedom from recurrence (or persistence) of evident aortic arch obstruction in patients having direct anastomosis as compared with patients having tube graft reconstruction.
Table VIII. Surgical treatment for interrupted aortic arch (Boston Children's Hospital; /974-1987) NYHA class at follow-up I II III IV V
No.
i
% of 27
81 15
22]
4
21
4
Unknown Dead
32
Total
63
100
Note: The table pertains only to the 63 patients who underwent an initial operation.
the left ventricular outflow tract, and that growth of the left ventricular outflow tract after repair is sufficient to accommodate normal flow without an appreciable pressure gradient. The other half of these patients appear to have a congenitally narrow pathway. The freedom from evident LVOTO of about 50% (more precisely, 58%, CL * 49% to 67%) at 3 years after repair (see again Fig. 8) corresponds rather well to the prevalence of freedom from echocardiographic evidence of narrowing of the left ventricular outflow tract and ascending aorta found *CL
= Confidence limits.
by Ilbawi and colleagues" (67%, CL 56% to 76%, of 30 consecutive neonates with interrupted aortic arch and VSD). These prevalences are similar to those reported by Freedom," Ho,20 and their colleagues. Theoretically, one-stage repair might be expected to reduce the prevalence of LVOTO after repair by forcing all the left ventricular output through the outflow tract, aortic valve, and ascending aorta very early in life. However, this has not been the case, since in the present experience all instances of LVOTO occurred in patients who had undergone one-stage repair. The patients in this study have done sufficiently well with complete repair to cast doubt on the advisability of the program of Ilbawi and colleagues" embracing more prolonged palliation and multiple operations. Persistent or redeveloped aortic arch obstruction. The complication of persistent or recurrent aortic arch obstruction has been more prevalent in patients in whom direct anastomosis was used. However, treatment by percutaneous balloon aortoplasty has usually been successful when direct anastomosis has been used." When the reconstruction has involved a tube graft, often an ascending-descending aorta tube graft must be placed through a right thoracotomy when the first tube becomes obstructed. Also, persistent or redeveloped aortic arch obstruction after direct anastomosis may be developing less frequently after operations performed in the current era.
Volume 96 Number 6
Interrupted aortic arch
December 1988
Inferences. An aggressive surgical program can result in survival and a good clinical state for at least 10 years after birth of over 40% of patients born with interrupted arch. Without operation, nearly all die in the first year of life." In the current era, 82% of patients with interrupted aortic arch and single VSD are predicted to survive at least 3 years after repair (see again Fig. 6, B), and since the hazard function for death is nearly that of the general population by that point (see again Fig. 5) most of these will probably be alive at least 10 years after repair. The failure to accomplish these excellent results when more complex cardiac anomalies coexist with interrupted aortic arch may be due to lack of success to date in evolving for all of them a completely corrective initial operation. The anatomic bases of the L VOTO are often multiple, including posterior malalignment of the outlet septum,I7·I~ hypertrophy of the anterolateral muscle bundle of the left ventricle (muscle of Moulaertj.P" narrowness of the aortic valve anulus and ascending aorta, and bicuspid, dysplastic aortic valve. This is supported by unpublished observations (R. Jonas, S. Van Praagh, R. Van Praagh, 1988; B. Soto, R. Ceballos, 1. W. Kirklin, 1988). Thus surgical measures to overcome the obstruction probably need to vary from case to case. However, when preoperative echocardiographic or cineangiographic evidence of important left ventricular outflow narrowing is obtained, something more than simple transatrial repair of the VSD should be considered. Since repair of the aortic arch by direct anastomosis, rather than by tube graft reconstruction, is not disadvantageous to survival, and since persistent or recurrent obstruction can be well managed by percutaneous balloon dilation," direct anastomosis should be used whenever possible in treating the interruption of the aortic arch. Whenever possible, this should be combined with repair of the coexisting cardiac defect in a onestage repair.
875
Reparative operations for interrupted aortic arch with ventricular septal defect. J THORAC CARDIOVASC SURG 1983;86:832-37.
3. Jonas RA. Interrupted aortic arch. In: Grillo HC. Austen W A, Wilkins EW, Mathisen DJ. Vlahakes AJ. eds. Current therapy in cardiothoracic surgery. Toronto: BC Decker. 1988. 4. Walker SH. Duncan DB. Estimation of the probability of an event as a function of several independent variables. Biometrika 1967;54:167-79. 5. Kaplan EL. Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:45781. 6. Blackstone EH. Naftel DC, Turner ME. The decomposition of time-varying hazard into phases. each incorporating a separate stream of concomitant information. J Am Stat Assoc 1986;81 :615-24. 7. Rothman A, Lang p. Lock JE. Jonas RA. Mayer JE. Castaneda AR. Surgical management of subaortic obstruction in single left ventricle and tricuspid atresia. J Am Coll Cardiol 1987;10:421-6. 8. Fishman NH. Bronstein MH. Berman W. et al. Surgical management of severe aortic coarctation and interrupted aortic arch in neonates. J THORAC CARDIOVASC SL:RG 1976;71 :35-48.
9. Allard JR. Williams RL. Dobell ARC. Interrupted aortic arch: factors influencing prognosis. Ann Thorac Surg 1976;21 :243-6. 10. Higgins CB, French JW. Silverman JF. Wexler L.
Interruption of the aortic arch: preoperative and postoperative clinical. hemodynamic. and angiographic features. Am J Cardiol 1977;39:563-71. II. Moulton AL. Bowman FO. Primary definitive repair of type B interrupted aortic arch. ventricular septal defect. and patent ductus arteriosus. J THORAC CARDIOVASC SL:RG 1981;82:501-10. 12. Turley K. Yee ES. Ebert PA. The total repair of interrupted arch complex in infants: the anterior approach. Circulation 1984;70(Pt 2):116-20. 13. Schumacher G. Schreiber R. Meisner H. Lorenz HP. Sebening F. Biihlmeyer K. Interrupted aortic arch: natural history and operative results. Pediatr Cardiol 1986; 7:89-93. 14. Zahka KG. Roland MA. Cutilletta AF. Gardner TJ.
Donahoo JS. Kidd L. Management of aortic arch interruption with prostaglandin E] infusion and microporous expanded polytetrafluoroethylene grafts. Am J Cardiol
We express our appreciation for the considerable participation of Mr. Rob Brown in all aspects of the study and to Ms. Debbie Nuby for her contribution to the graphics and the manuscript.
1980;46:1001-5. 15. Leoni F. Huhta JC. Douglas J. et al. Effect of prostaglan-
REFERENCES I. Rein JG, Freed MD. Norwood WI. Castaneda AR. Early and late results of closure of ventricular septal defect in infancy. Ann Thorac Surg 1977;24:19-27. 2. Norwood WI. Lang P, Castaneda AR. Hougen TJ.
din on early surgical mortality in obstructive lesionsof the systemic circulation. Br Heart J 1984;52:654-9. 16. Jonas RA. Lang P, Hansen DD. Hickey PRo Castaneda AR. First-stage palliation of hypoplastic left heart syndrome: the importance of coarctation and shunt size. J THORAC CARDIOVASC SL:RG 1986;92:6-13. 17. Becu LM, Tauxe WN. DuShane JW. Edwards JE. A
The Journal of
18.
19.
20.
21.
22.
Thoracic and Cardiovascular Surgery
Sell et al.
876
complex of congenital cardiac anomalies: ventricular septal defects, biventricular origin of the pulmonary trunk, and subaortic stenosis. Am Heart J 1955;50:901-11. Van Praagh R, Bernhard WF, Rosenthal A, Parisi LF, Fyler DC. Interrupted aortic arch: surgical treatment. Am J Cardiol 1971;27:200-11. Freedom RM, Bain HH, Esplugas E, Dische R, Rowe RD. Ventricular septal defect in interruption of aortic arch. Am J Cardiol 1977;39:572-82. Ho SY, Wilcox BR, Anderson RH, Lincoln JCR. Interrupted aortic arch: anatomical features of surgical significance. Thorac Cardiovasc Surg 1983;31:199-205. Sennari E. Morphological study of ventricular septal defect associated with obstruction of aortic arch among Japanese. Jpn Circ J 1985;49:61-7. Ilbawi MN, Idriss FS, DeLeon SY, Muster AJ, Benson OW, Paul MH. Surgical management of patients with
Appendix A: Morphologic details and relation univariateIy to total mortality
Table At Deaths (early and late)
n
% of7l
No.
%
C
20 49 2
28 69 3
12 25 2
60 51 100
Total
71
100
39
55
A B
24. 25.
26.
Table A2 Type of associated cardiac anomalies
Appendixes
Type of interrupted arch
23.
interrupted aortic arch and severe subaortic stenosis. Ann Thorac Surg 1988;45:174-80. Saul JP, Keane JF, Fellows KE, Lock JE. Balloon dilation angioplasty of postoperative aortic obstructions. Am J Cardiol 1987;59:943-8. Kirklin JW, Barratt-Boyes BG: Cardiac surgery. New York: John Wiley, 1986:1046. Moulaert AJ, Oppenheimer-Dekker AJ. Anterolateral muscle bundle of the left ventricle, bulboventricular flange and subaortic stenosis. Am J Cardiol 1976;37:78-81. Moulaert AJ, Bruins CC, Oppenheimer-Dekker A. Anomalies of the aortic arch and ventricular septal defects. Circulation 1976;53:1011-5.
P(X')
0.3
0.3
Note: Tables AI and A2 include all patients entering the hospital. including eight (seven deaths) who did not reach operation.
Deaths (early and late)
n
% of7l
No.
%
VSD (isolated) "Single ventricle" Truncus arteriosus DORV TGA+ VSD Complete AV canal DOlV Isolated ventricular inversion + VSD None (PDA present)
44 8 7 5 2 2 1
62
19 8 6 3 0 1
I
I
43 100 86 60 0 50 100 100
Total
71
II
10 7 3 3 1
1 I
0
0
39
55 .01
P (x') for VSD vs. all others
Key: AV. Atrioventricular; DOL V. double-outlet left ventricle; DORV. doubleoutlet right ventricle; PDA. patent ductus arteriosus; TGA. transposition of the great arteries; VSD. ventricular septal defect.
Note: "Single ventricle" included three patients with complex tricuspid atresia. one patient with complex mitral atresia. and four patients with double-inlet left ventricle.
Table A3
Type of associated cardiac anomalies VSD (isolated) "Single ventricle" Truncus arteriosus DORV TGA+ VSD Complete AV canal DOlV Isolated vs. inversion + VSD None (PDA present) Total
Type C
n
No.
44
7
8
2
5 2 2 2
2
I
7
5
No.
% of 2
2
100
2
100
I 1 I
71
20
AV. Atrioventricular; DOlV. double-outlet left ventricle; DORV. double-outlet right ventricle; fAA. interrupted aortic arch; PDA. patent ductus arteriosus: TGA. transposition of the great arteries; YSD. ventricular septal defect.
Volume 96 Number 6 December 1988
Appendix B: Parameter estimates, P-~alues, and multivariate analysis (variables entered, variables retained, coefficients, and P-values) for death at any time after the initial operation, in a time-related manner (n = 63). Variables entered into the analysis included demographic ones (birth weight, age at admission, date of admission, date of operation); morphologic variables (type of interrupted aortic arch, VSD alone, "single ventricle," as a group, truncus arteriosus, double-inlet left ventricle, ventriculoarterial discordant connection, double-outlet right ventricle, multiple VSDs, hypoplastic ventricle, noncardiac congenital anomaly); preoperative variables (interval from admission to operation, use of prostaglandin E), use of dopamine, use of preoperative intubation and ventilation, pH at admission, pH immediately preoperatively); and surgical variables (operation performed during the night, one-stage repair, type of arch repair). A single declining hazard phase was found. Shaping parameter estimates, the variables retained (P < 0.1) their coefficients plus or minus one standard deviation, and, in
Interrupted aortic arch
877
parentheses, the P values, are as follows: 0 = 0, p = 08868, v = 3912, II. = 0, intercept 2.277, date -0.2051 ± 0.066 (P = 0.(02), "single ventricle" 2.174 ± 0.61 (P = 0.0004), truncus arteriosus 1.661 ± 0.55 (P = 0.003), multiple VSDs 1.984 ± 0.57 (P < 0.0005), immediately preoperative pH - 3.507 ± 1.29 (P = 0.006), and left ventricular-aortic conduit 1.515 ± 0.65 (P = 0.02). Appendix C: Parameter estimates, P-~alues, and multivariate analysis variables entered, variables retained, coefficients, and P-~alues of the event recurrent or persistent aortic arch obstruction in a time-related manner. Variables entered into the analysis included the demographic, morphologic, and surgical variables listed in Appendix B. A single constant hazard phase was found. The variables retained (P < 0.1), their coefficients plus or minus one standard deviation, and, in parentheses, the P values, are as follows: intercept -5.055, direct anastomosis 4.969 ± 0.92 (P < 0.0001), and interaction of direct anastomosis with date of repair -0.3228 ± 0.091 (P = 0.(004).
THoRAc CARDIOV ASC SURG
1988;96:864-77
The results of a surgical program for interrupted aortic arch Seventy-one patients with interrupted arch entered the Boston Children's Hospital between Jan. 1, 1974, and Jan. 1, 1987, of whom 63 underwent an operation. Type B was the most prevalentform of interrupted arch, and ventricular septal defect alone was the commonest coexisting cardiac anomaly. Among the 63, the 3O-day and the 1-, 5-, and IO-year survival rates were 61 %,52%,48%, and 47%, respectively. The mortality rate declined strikingly during the experience, and by multivariate analysis in patients with coexisting ventricular septal defect operated on in 1986, the probability of death within 2 weeks of repair was only 7 %. Also, preoperative therapy became progressively more Iateese and more prolonged. The complication of left ventricular outflow tract obstruction developed in eight of the 33 patients undergoing repair of interrupted arch and of isolated ventricular septal defect The time-related freedom from this complication was 97%, 78%, and 58% at 1 month, 1 year, and 3 years, respectively. Seven of the eight patients underwent a surgical procedure directed against the left ventricular outflowtract obstruction, and all have survived. Recurrent or persistent aortic arch obstruction became evidentafter repair in 15 patients and appeared more frequently and earlier after direct anastomosis than after tube graft repair. AU patients had either reoperation or balloon dilation, but all were alive at foUow-up. Most surviving patients are active and without symptoms. Inferences: An aggressivesurgical program can result in survival and a good clinical state for at least 10 years after birth of over 40% of patients born with interrupted arch. Multiple anatomic bases account for the development of left ventricular outflowtract obstruction in about 50 % of the patients undergoing repair of interrupted arch with coexisting ventricular septal defect. Repair by direct anastomosis combined with repair of the coexisting defect whenever possible is optimal therapy.
Jeffrey E. Sell, MD" (by invitation), Richard A. Jonas, MD" (by invitation), John E. Mayer, MD" (by invitation), Eugene H. Blackstone, MD,b John W. Kirklin, MD,b and AIdo R. Castaneda, MD,a Boston, Mass., and Birmingham, Ala.
Numerous reports of the surgical treatment of interrupted aortic arch have appeared in the literature in the past 20 years, but few have contained enough cases followed a reasonable period of time to allow inferences as to the spectrum of the disease, optimal therapy, and early and late results in terms of survival and complications. Specifically, little reference has been made to the appearance after repair of left ventricular outflow tract From the Boston Children's Hospital and the Harvard Medical School, Boston, Mass.," and the Medical Center of the University of Alabama at Birmingham, Birmingham, Ala." Read at the Sixty-eighth Annual Meeting of The American Association for Thoracic Surgery, Los Angeles, Calif., April 18-20, 1988.
Address for reprints: Richard A. Jonas, MD, The Boston Children's Hospital, 300 Longwood Ave., Boston, MA 02115.
864
obstruction (LYOTO) or of persistent or recurrent aortic arch obstruction. A careful study of a 13-year experience, including detailed follow-up of all patients not known to be dead, has provided new but as yet incomplete information about these matters.
Patients and methods Patients. All 71 patients with interrupted aortic arch entering the Boston Children's Hospital between Jan. 1, 1974, and Jan. 1, 1987, are included in the study. Sixty-three patients underwent an initial operation, and eight others either died before an operation could be done (three patients) or were considered inoperable because of extremely complex major associated cardiac or noncardiac anomalies, or both (five patients). One patient in the latter group is alive at follow-up 23 months later. The median age at entry was 4 days (Fig. I).
Volume 96 Number 6
Interrupted aortic arch
December 1988
865
100
.90 m
4(eo
•
'0
ai 70
US
% Entry
Age (days)
>060
lD
2
~50
4
-.:
2S% SO% 7SO/. 90 %
20
S40 c
at~
Stated Age
54
W
iu
30
12O 10 10
20
30
40
50
60
Age (days) at Entry
eo
70
90
Fig. 1. Cumulative frequency distribution plot depicting age at entry into the hospital (and thus into the study). Four patients were older than 90 days at entry, being 123 days, 299 days, 489 days, and 1371 days of age.
Table I. Surgical treatment for interrupted aortic arch (Boston Children's Hospital; 1974-1987; n = 71; deaths = 39) Time period Events Death at any time Preoperative PGE , Preoperative low-dose dopamine Preoperative intubation Mean immediately preoperative pH Median interval between admission and operation (days) One-stage repair Direct anastomosis
/974-/979
12/15 3/13 1/13 6/13 7.26 0.6
(80%.) (23%) (8%) (46%)
4/13 (31'it) 3/13 (23%)
/98/-/983
/979-/98/
10/14 7/12 0/12 2/12 7.34 0.7
/983-/985
(71%) (58%) (0%) (17%)
4/9 (44%) 5/9 (56'1C) 0/9 (0%) 5/9 (56%) 7.41 1.8
12/21 17/18 6/18 13/18 7.41 1.4
7/12 (58%) 1/12 (8%)
4/9 (44%) 0/9 (0%)
12/18 (67%) 4/18 (22o/r)
(57%) (94%) (33%) (72%)
/985-/987
1/12 9/11 7/11 9/11 7.43 2.0
(8%) (82%) (64%) (82%)
9/11 (82%) 9/11 (82%)
P for difference 0.0002 <0.0001 <0.0001 0.003 0.0006 0.002 0.004 0.004
Sole: The event death at all.. lime pertained to the entire group of 71 patients; all other events pertained to the 63 who underwent an initial repair.
Type B was the most prevalent form of interrupted arch (Appendix A, Table AI). Twelve of the patients with this type clearly had the DiGeorge syndrome, and others may have had it as well; the DiGeorge syndrome did not coexist with types A or C. The commonest coexisting cardiac anomaly was ventricular septal defect (VSD) (Appendix A, Table A2). Multiple VSDs were present in five patients, including two with VSD as the only coexisting cardiac anomaly, two with double-outlet right ventricle, and one with truncus arteriosus. All five patients have died. Preoperative management. Most of the patients were seriously ill on admission, 30'k having an arterial pH equal to or less than 7.3 at the time of admission (range 6.9 to 7.6). Patient management protocols in the current era, which have evolved gradually through the l4-year period (Table I),
included the use of prostaglandin E" intubation and ventilation, and low-dose dopamine. The prostaglandin E, was infused through a secure intravenous line and usually at a rate of 0.05 J.lg . kg-I. min-I (earlier, 0.1 J.lg. kg-I. min-I), to maintain ductal patency and thereby good perfusion of the organs and tissues of the lower part of the body. This, plus the aggressive use of sodium bicarbonate and tromethamine, usually effectively combated metabolic acidosis. After intubation, ventilation was conducted so as to maximize the pulmonary /systemic resistance ratio and to encourage flow through the ductus arteriosus and into the descending thoracic aorta. Thus the ventilatory gas was room air, arterial carbon dioxide tension was maintained at about 40 mm Hg, end-expiratory pressure of 5 to 10 em H 20 was used. Arterial blood pressure was usually monitored by a catheter in the umbilical artery or
The Journal of
866
Thoracic and Cardiovascular Surgery
Sell et al.
B Fig. 2. Depiction of technique of repair by direct anastomosis of interrupted arch type B. A, Purse strings on the aorta and pulmonary arteries indicate the siting of the two arterial cannulas. Before cannulation, all of the vessels are thoroughly dissected to maximize their mobility and minimize tension on the anastomosis. B, The completed anastomosis is shown. The position of the arterial cannula in the ascending aorta is illustrated, as the precisely correct positioning of this cannula is critical for even perfusion of the body and the coronary arteries.
Table II. Surgical treatment for interrupted aortic arch (Boston Children's Hospital; 1974-1987) Deaths Type of repair
n
No.
%
CL
One stage Multiple stage
36 27
13 19
36 70
27-46 59-80
Total
63
32
51
44-58
P (X')
0.007
Note: The deaths include all deaths. early and late, after repair.
an artery in the foot. Cardiac output and renal blood flow were usually maintained with dopamine infused at a rate of about 5 J.Lg • kg-I. min-I. Surgical methods. Like the preoperative management, the methods used in the current era have gradually evolved over the period of the study. In the current era, a one-stage repair was performed as the initial surgical procedure unless the intracardiac morphology demanded a Fontan type of repair (Table II). The one-stage repair was performed during profoundly hypothermic total circulatory arrest, the technique of which has been described previously.I In most patients, the technique was modified for patients with interrupted arch by
placing an arterial perfusion cannula not only in the ascending aorta, but also in the pulmonary artery. The right and left pulmonary arteries were temporarily occluded while this latter cannula perfused the lower part of the body through the ductus arteriosus. Often, the aortic interruption was repaired by direct anastomosis (Fig. 2), although a tube graft reconstruction was more frequently used in the early part of the experience.' The technique for direct anastomosis has been described in detail elsewhere.' Miscellaneous other types of reconstruction were also employed (Table III). The VSDs were repaired with a right atrial approach, and in the last 10 years of the experience cold cardioplegia was used. Data collection and follow-up. The details of the preoperative condition and management of the patient, of the operations, of postoperative care, and of subsequent studies and procedures, both at the initial and subsequent hospitalizations and at outpatient visits, were obtained from the records of the Boston Children's Hospital. In October 1987, a formal follow-up was performed of all patients not known to be dead; Both letters and telephone calls to families and to the patients' physicians at home were used. Two patients could not be reached, one of whom had not been able to be contacted since hospital dismissal 19 days after repair and the other of whom was known to be alive 26 months after repair but has been untraced since then. The median follow-up time was 43 months and the range, 10 months to II years. Six years or
Volume 96 Number 6
Interrupted aortic arch 8 6 7
December 1988
Table
m. Surgical treatment for interrupted aortic arch (Boston Children's Hospital;
/974-/987) Deaths
Type of initial repair Direct anastomosis + Correction of cardiac anomaly + Palliation of cardiac anomaly Isolated TGR + Correction of cardiac anomaly + Palliation of cardiac anomaly Anomaly LSCA to descending aorta + Palliation of cardiac anomaly Isolated LSCA flap aortoplasty (with palliation of cardiac anomaly) Preservation of ductal continuity (with correction of cardiac anomaly) LV-aortic conduit as repair of IAA (with correction of cardiac anomaly) TGR + LV-aortic conduit as repair of IAA (plus palliation of cardiac anomaly) TGR + LSCA to proximal aorta as repair of IAA (plus correction) No operation Total
% of63
n
17
No. 4 2
I
0
35
29
100 0 19
56
4
86
33
6
3
I
I
0
54
33
6 12 I
14 3
35
6
27
14 2 18
%
33 0
2
25 100 50 100
2
2
100 100
8
7
88
71
39
55
Key: IAA. Interrupted aortic arch: LSCA. left subclavian artery; LV. left ventricle: TGR. tube graft reconstruction .
.VOle: The table includes
late.
all patient' entering the hospital. including the eight (seven deaths) who did not reach operation. The deaths include all deaths. early and
more of follow-up information is available for 22% of the 71 patients. Data analysis. The data were sorted and examined from numerous points of view. Multiple contingency tables were constructed. Incremental risk factors for the presence or absence of an event were sought through logistic regression analysis: Time-related events were examined by both actuarial' and parametric methods, and risk factors were sought in the parametric (hazard function) domain." The usual methods were used to determine the likelihood that differences were due to chance alone.
Results Survival. Among the 71 unselected patients, 39 died during the initial hospitalization or later (Fig. 3). The prevalence of death has declined strikingly across the time period of the study (Table I). Among the 63 patients who underwent an initial operation, the 3Q-day survival rate was 61%. The time-related survival rate (Fig. 4) had a single phased, rapidly declining hazard function (Fig. 5), no deaths having occurred more than 19 months after repair (24 patients traced beyond 19 months). Solution of the equation describing the risk factors for death at any time after repair (Table IV)
emphasized the improvement in survival across the time period of the study, emphasized the more favorable outcome in the patients in whom (VSD) was the sole coexisting cardiac anomaly (Fig. 6), and also demonstrated the importance of the immediately preoperative arterial pH (Fig. 7). The earlier preoperative pHs, including that on admission, did not influence outcome. Although a good result was obtained infrequently in patients with the other associated cardiac anomalies (see again Appendix Table A2), the possibility existed. Thus one of the two patients with coexisting transposition of the great arteries and VSD underwent initial interrupted arch repair by direct anastomosis, then 4 days later a pulmonary artery banding, and at 6 months of age an arterial switch repair; one year later the child was in good health. The other child with transposition and VSD has a hypoplastic right ventricle and is scheduled for a Fontan repair. Changes during the 13-year period. Many preoperative and intraoperative methods changed during the time period of the study, and particularly the frequency of intense preoperative therapy (see again Table I). The
The Journal of Thoracic and Cardiovascular
8 6 8 Sell et al.
Surgery
100
tr---------. Years
90 80
70
..
'ii
~60
-e ::I
(/)50
c: 1)40
I)
Q.
% Survival
1/52
72%
1/12
60%
6/12
50%
1
48%
2 5 10
46% 44% 42%
---------------------------------------------------------------------
30
20 10 Ol:.-_--'--_---J._ _..L..-_--'--_---J._ _..L..-_--'--_---J._ _..L..-_--'--_--' 11 10 456 7 9 8 2 o
Years After Entry
Fig. 3. Time-related survival of patients (n = 71) entering with interrupted arch, including all deaths. Time zero is the time of entry. The open circles represent individual deaths, actuarially positioned. The solid line indicates the parametrically determined survival. The dashed lines enclose the 70% confidence intervals. The dash-dot-das]i line represents the survival of an age-matched general population, U.S. Life Tables, 1976.
100
90 80
'ii >
.
70
>60
-
Years
% Survival
1/52
70%
1/12
61%
6/12
54%
1
52%
2
50%
5
48%
10
47%
::::I
(/)50'
c:
~40
~
-----------------------------------
30 20 10 oL-_...L..._.J..._---I._ _L-_..l....-_--'--_--l.._---.JL-_..l....-_--'--_--' 11 9 10 4 5 6 7 8 2 o
Years After Repair
Fig. 4. Actuarial and parametric time-related survival after initial operation (n = 63) for interrupted aortic arch. Time zero is the time of the initial operation. The depiction is as in Fig. 3.
prevalence of the type of aortic arch and of the various types of associated cardiac anomaly did not change across time, nor did the prevalence of patients not undergoing operation (P::::: 0.4 in each instance).
LVOTO. Systemic ventricular outflow tract obstruction (LVOTO) became evident soon after repair in two patients with associated "single ventricle," both of whom died. (LVOTO was considered to be present when the
Volume 96 Number 6 December 1988
Interrupted aortic arch
869
0.10 0.09 0.08
...
-
I.e
c
0
~
III
.e
0.07
·•
0.06
•I
Years
" "
.,
1/52 1/12
"
6/12
0.05
cu 0.04 cu C
0.03
•I
" " " • I 'I 'I I
0.02 0.01
I \
Hazard IX 1000) 410 83 10 4.5 1.9 0.61 0.26
1 2 5 10
,
, ,
, ,, ,, , ,
- ---:-:: -18
24
:30
36
Months After Repair
42
48
54
60
Fig. 5. Hazard function for death in the 63 depicted in Fig. 4. The dashed lines enclose the 70% confidence intervals of the hazard function. The barely visible dot-dash-dot line is the hazard function of an age-matched general population.
gradient between left ventricle and ascending aorta was 40 mm Hg or more by echocardiography or cardiac catheterization.) The mechanism of the obstruction has been described earlier.' LVOTO also became evident in eight patients with coexisting isolated VSD (Fig. 8), all ofwhom were alive at the time of follow-up. The hazard function for the appearance of evident LVOTO in this group had a single declining hazard phase (Fig. 9), which had reached a low level by 36 months after repair. The complication of LVOTO obstruction developed only in the subset of 33 patients with the VSD in the outlet portion of the septum (n = 33), cradled in the limbs of the septal band (or trabecula septomarginalis), and in many of these patients the surgeon had noted posterior malalignment of the outlet (conal, infundibuIar) septum. Seven of the eight patients underwent a surgical procedure directed against the LVOTO (Table V). Among the three in whom subvalvular obstructing muscle or fibrous tissue was resected, one had a 10 rom Hg gradient 23 months later, one a 60 rom Hg gradient 28 months later, and one a 100 mm Hg gradient and a small aortic anulus 21 months later. Among the three treated by placement of a left ventricular-aortic conduit, one had an 80 rom Hg gradient between the left ventricle and aorta 1 year later; this child underwent revision of the proximal conduit anastomosis, 11 months after which a 50 rom Hg gradient existed. A second
Table IV. Surgical treatment for interrupted aortic arch (Boston Children's Hospital; 1974-1987) Incremental risk factors for death at any time after repair In = 63) Demographic Date of operation (earlier) Morphologic "Single ventricle" Truncus arteriosus Multiple VSDs Preoperative condition Immediately preoperative pH (lower) Surgical LV-aortic conduit See Appendix B for variables entered into the analysis. coefficients. and P values.
patient among the three had an 80 rom Hg gradient 23 months after placement of the left ventricular-aortic conduit, 28 months later had transaortic resection of subaortic stenosis, and 22 months after that had a 65 rom Hg gradient between the left ventricle and ascending aorta. The third patient is without symptoms 5 years later but is known to have had a gradient of 60 mm Hg at catheterization 11 days after placement of the conduit. The patient in whom aortic valvotomy was performed had a dysmorphic, bicuspid valve, and a second surgical valvotomy was performed 14 months
The Journal of
870
Thoracic and Cardiovascular Surgery
Sell et al.
100 III
~
at 90
~
N 80
-
c :E 70
~60
-
.c
: 50
C
040
--1976
A
1978
1980
1982
1984
1986
Date of Operation
100
; 90
~
M80 c
£ ~
-
70
.c as
--
at C 50 0
40
~
!...
>-
30
:!::
:cas
.0 0
~
20 10 0
B
1974
--1976
1978
1980
1982
1984
1986
Date of Operation
Fig. 6. A, Nomogram from the multivariate equation (Table IV) illustrating the effect of date of operation, in patients with interrupted aortic arch and VSD, on the probability of death within 2 weeks. (See Appendix B for details.) B, Nomogram similar to that in A, except illustration is of the probability of death within 3 years according to the date of operation.
later. The patient receiving no treatment had a 45 mm Hg gradient 11 months after the initial one-stage repair and is well but without further studies 2 years later. Persistence or redevelopment of evident aortic arch
obstruction. Aortic arch obstruction became evident sometime (0.2 to 62 months) after repair in 15 patients (Table VI). (This complication was considered to be present when the gradient between the ascending and
Volume 96 Number 6
Interrupted aortic arch
December 1988
100 U) ~
90
~
SO
QI
---- .. --.
...
-- .. --
'- '-
........
........
N
C
70
~
60
tV QI
50
:E or:
-C
........
........
40
0
-
30
0~
~ 20
:s
~ 10
...o no
------ ------~~=-=-:-::::-::-::-::-:-:-:~;.;.~...::..::...:..::...:
------------0 1:..L............u.....& ....................................................u.....& ....................................................U-J................................L..L..J.......... U-J................................L..L..J..........U-J....J...J 7.1 7.2 7.3 7.4 7.0 7.5 6.9 Immediately Prerepair pH
Fig. 7. Nomogram from the multivariate equation (Table IV). The small arrow indicates the "point of evident difference" in the case of VSD, that is, the point (7.32) at which, compared with a pH of 6.9, the probability of death is reasonably certainly (nonoverlapping 70% confidence intervals less. (See Appendix B for details.)
100
o
90
b SO
--- --- --- ----
~
_70 c QI :260 > w
'0
(13r - _ cia)---_
50
:40
~
Outlet VSD Repair Subgroup (n=33; events=8l
'ECIl 30 o
~ l1.
20 10
6
12
--- ---
---------------
6)-
-----Years
% Freedom
1/52 1/12 6/12
99% 97% 87%
1
78%
2
64%
3
58%
18 24 30 36 42 Months After VSD Closure
60
Fig. 8. Actuarial time-related freedom from evident LVOTO, among the 33 patients undergoing repair of a VSD in the outlet portion of the right ventricular aspect of the septum. (In four other patients, the VSD was in the trabecular portion of the septum or in an unknown position.) The circles represent individual events (the appearance of evident LVOTO) and are actuarially positioned. The vertical bars enclose the 70% confidence intervals and the lines merely connect the circles. The last event is at 36 months.
87 I
The Journal of Thoracic and Cardiovascular
8 7 2 Sell et al.
Surgery
0.050 0.045
...
\ \
\
0.040
\
\ \ \
I.e
'E 0.035 o
.
~0.030
\
\ \
" "
~0.025
o
-
Outlet VSD Repair SUbgroup , (n=33; event8=8)
,,
,,
~O.020
,,
Years
Hazard (x 1000)
1/52 1/12
27.0
6/12
20.6
,
c
-80.015
';:
UJ 0.010
~
0.005
-
~
23.9
1
18.9
2
12.1
3
4.6
-
~-
~
~
~
~
~-
~
---------~-~-~--------------
12
6
--
18
24
:30
36
Months After VSD Closure
42
48
54
60
Fig. 9. Hazard function for the appearance of evident LVOTO at each moment in time after the repair in the 33 patients. The hazard function has a single declining phase. The depiction is as in fig. 5.
Table V. Surgical treatment for interrupted aortic arch (Boston Children's Hospital; 1974-1987) Type of treatment for post repair LVOTO
Deaths n
No.
%
Resection (15, 18, and 39 months after VSD repair) LV-aortic conduit (0.4, 10, and 14 months after VSD repair) Aortic valvotomy only (4.4 months after VSD repair) No treatment
3
0
0
3
0
Total
8
0
CL %
0
0 0
0
0-21
descending aorta was 30 mm Hg or more by echocardiography or cardiac catheterization.) The use of a direct anastomosis was a risk factor for postrepair aortic arch obstruction, as was earlier date of repair when a direct anastomosis was used (Table VII). All patients with this complication had either a reoperation or balloon dilation performed for the aortic arch obstruction. Nonetheless, all but two of the patients were alive at the time of follow-up. This event occurred sooner and more frequently after direct anastomosis than after tube graft reconstruction (Fig. 10). Status of surviving patients. Despite treated and
untreated persistent or recurrent obstructive lesions in many patients, most of those surviving at the time of follow-up were active and without symptoms (Table VIII). Discussion Patients. The patients were relatively unselected, the only criteria for inclusion being that the diagnosis be made. Most of the patients were entered into the study when they were only a few days old. Most previous publications reported only a small number of patients undergoing an initial operation, and many have included only one morphologic subset rather than all patients with interrupted aortic arch.v" Type B was the commonest type in this experience, as it was in the experience of Schumacher's group" and others. Although VSD was most commonly the associated cardiac anomaly, a wide variety of other cardiac anomalies coexisted with interrupted arch. Because of the relatively unselected nature and relatively large size of the patient group, the prevalence of the various types of interrupted arch and of the associated cardiac anomalies, and the relation between these two, can be considered representative for patients in general with interrupted arch. Survival. The 5- and I O-yearsurvivorships reported in this paper (44% and 42%, respectively, of all patients with this diagnosis entering the hospital) are distinctly
Volume 96 Number 6
Interrupted aortic arch
December 1988
Table VI. Surgical treatment for interrupted aortic arch (Boston Children's Hospital; 1974-1987) Redevelopment (or persistence) of evident aortic arch obstructioni Type of arch repair (n = 63) Direct anastomosis Subclavian flap aortoplasty
Blalock-Park repair" Formalization of PDA Tube graft repair Tube graft repair plus LV-aortic conduit Tube graft repair plus LSCA to proximal aorta anastomosis l.V-aortic conduit Total
CL
n
No.
%
(%)
17 I 4 2 35 2
5:\: I§ 211
29 100 50 50 14 50
17-45 15-100 18-82 7-93 8-23 7-93
63
I~
1··
5#
15
0
0-85
0
0-85
24
18-31
Key:CL.. 707, confidence limits; LSCA. left subclavian artery; LV.Ieft ventricle; PDA. patent ductus arteriosus. 'End-to-side anastomosis of the left subclavian artery to the descending thoracic aorta. tfhe percents are percents of n. and have no time-related or actuarial implications. +The correction was a patch graft enlargement in two (one died after the procedure; the other lived but is without further information), and by balloon dilation in three (in two, there was no gradient early after dilation; in the third, a second dilation was performed 3 months after the first one). §Balloon dilation was performed 11.5 months after repair, which reduced the gradient to 18 mm Hg (also had anastomosis of proximal pulmonary artery to ascending aorta. for subaortic stenosis). I'nterval was 0.2 to 2.4 months; one patient had aortoplasty of the anastomotic site with a polytetrafluoroethylene graft, with all pulses the same thereafter; the other had an ascending-descending aorta bypass graft. resulting in a 10 mm Hg residual gradient. ~Interval was 4.8 months, at which time a left subclavian artery-descending aorta anastomosis was made followed by a J0 mm Hg gradient, according to extremity of blood pressure.
s Three were treated. one by ascending aorta-descending aorta tube graft, one
bypatch graft enlargement of distal anastomosis, and one by conversion to direct anastomosis. "'nterval was 2.9 months; treatment was by a polytetrafluoroethylene interposition tube graft between the left subclavian artery and ascending aorta, followed in I day by patch graft enlargement of the anastomosis, and then in four days by replacement of the LV-aortic conduit; the patient died 2 months after the last procedure.
encouraging, particularly since no deaths have occurred more than 19 months after admission and 24 patients have been traced beyond this point. The outlook for patients with coexisting VSD is particularly favorable, as can be seen by the projected 3-year survival rate of 88% for such patients operated on in the current era (see again Fig. 6, B). The era effect. Many things changed across the 13-yeartime period of this study, including the mortality rate (see again Table I). Clearly, the decreasing mortal-
873
Table VII, Surgical treatment for interrupted aortic arch (Boston Children's Hospital; 1974-1987) Incremental risk factors for evident recurrent or persistent aortic arch obstruction (n = 63) Direct anastomosis Date of direct anastomosis (earlier) See Appendix C for variables entered into the analysis, coefficients, and P values.
ity rate did not result from the physicians being more selective as to who was treated surgically or from patients with favorable lesions being more frequent in the recent time periods. Probably the decreasing mortality rate is related to more intensive preoperative preparation of the patient with prostaglandin E 1 to improve lower body perfusion and thereby improve renal function and diminish acidosis": 15; with low dose dopamine to improve renal and cardiac function; and with intubation and ventilation to maximize the pulmonary/ systemic resistance ratio and thus the flow across the ductus and into the descending aorta. These same measures have proved to be helpful in patients with the hypoplastic left heart syndrome." Probably an important reason for the continuing high mortality among patients whose coexisting cardiac anomaly is "single ventricle," truncus arteriosus, and other complex conditions is that in the past a strenuous consistent effort to repair all coexisting cardiac anomalies was not made. Methods for doing this safely should improve results in the future. In the current era, a one-stage repair with direct anastomosis of the proximal to the distal aortic segment was the most prevalent technique. Clearly, this has had no adverse effect on survival, but it has not yet been shown that it has had a favorable effect. LVOTO. Becu,' Van Praagh," and their colleagues very early described the tendency toward LVOTO in patients with interrupted aortic arch and coexisting VSD; and more recent studies have refined knowledge in this area.":" According to the present study, in about 90% of patients with this combination of anomalies the VSD is in the outlet portion of the septum, and presumably only in them is there a predisposition to LVOTO. The experience here reported supports the idea that the basis for the obstruction is present at the time of VSD repair and therefore becomes evident early after repair, with the probability of its becoming evident (hazard function) decreasing as time passes. It would appear that about half the patients with interrupted aortic arch and VSD have only mild initial narrowing of
874
The Journal of Thoracic and Cardiovascular Surgery
Sell et al.
100
-... c
.2 90 U
::::I
In
121
80
.Q
0
.r:. u
...
0(
-...
.2
-
70 Tube Graft ("=35; event.=51
60
(51
50
Direct Ana.tomo." ("=17; eventa=SI
0
c(
40
0
Q) Q)
-...
It c
(21
30 20
Q)
u
I-
10 0
2
0
:3
4
Years After Repair
5
6
7
8
Fig. 10. Actuarial time-related freedom from recurrence (or persistence) of evident aortic arch obstruction in patients having direct anastomosis as compared with patients having tube graft reconstruction.
Table VIII. Surgical treatment for interrupted aortic arch (Boston Children's Hospital; /974-1987) NYHA class at follow-up I II III IV V
No.
i
% of 27
81 15
22]
4
21
4
Unknown Dead
32
Total
63
100
Note: The table pertains only to the 63 patients who underwent an initial operation.
the left ventricular outflow tract, and that growth of the left ventricular outflow tract after repair is sufficient to accommodate normal flow without an appreciable pressure gradient. The other half of these patients appear to have a congenitally narrow pathway. The freedom from evident LVOTO of about 50% (more precisely, 58%, CL * 49% to 67%) at 3 years after repair (see again Fig. 8) corresponds rather well to the prevalence of freedom from echocardiographic evidence of narrowing of the left ventricular outflow tract and ascending aorta found *CL
= Confidence limits.
by Ilbawi and colleagues" (67%, CL 56% to 76%, of 30 consecutive neonates with interrupted aortic arch and VSD). These prevalences are similar to those reported by Freedom," Ho,20 and their colleagues. Theoretically, one-stage repair might be expected to reduce the prevalence of LVOTO after repair by forcing all the left ventricular output through the outflow tract, aortic valve, and ascending aorta very early in life. However, this has not been the case, since in the present experience all instances of LVOTO occurred in patients who had undergone one-stage repair. The patients in this study have done sufficiently well with complete repair to cast doubt on the advisability of the program of Ilbawi and colleagues" embracing more prolonged palliation and multiple operations. Persistent or redeveloped aortic arch obstruction. The complication of persistent or recurrent aortic arch obstruction has been more prevalent in patients in whom direct anastomosis was used. However, treatment by percutaneous balloon aortoplasty has usually been successful when direct anastomosis has been used." When the reconstruction has involved a tube graft, often an ascending-descending aorta tube graft must be placed through a right thoracotomy when the first tube becomes obstructed. Also, persistent or redeveloped aortic arch obstruction after direct anastomosis may be developing less frequently after operations performed in the current era.
Volume 96 Number 6
Interrupted aortic arch
December 1988
Inferences. An aggressive surgical program can result in survival and a good clinical state for at least 10 years after birth of over 40% of patients born with interrupted arch. Without operation, nearly all die in the first year of life." In the current era, 82% of patients with interrupted aortic arch and single VSD are predicted to survive at least 3 years after repair (see again Fig. 6, B), and since the hazard function for death is nearly that of the general population by that point (see again Fig. 5) most of these will probably be alive at least 10 years after repair. The failure to accomplish these excellent results when more complex cardiac anomalies coexist with interrupted aortic arch may be due to lack of success to date in evolving for all of them a completely corrective initial operation. The anatomic bases of the L VOTO are often multiple, including posterior malalignment of the outlet septum,I7·I~ hypertrophy of the anterolateral muscle bundle of the left ventricle (muscle of Moulaertj.P" narrowness of the aortic valve anulus and ascending aorta, and bicuspid, dysplastic aortic valve. This is supported by unpublished observations (R. Jonas, S. Van Praagh, R. Van Praagh, 1988; B. Soto, R. Ceballos, 1. W. Kirklin, 1988). Thus surgical measures to overcome the obstruction probably need to vary from case to case. However, when preoperative echocardiographic or cineangiographic evidence of important left ventricular outflow narrowing is obtained, something more than simple transatrial repair of the VSD should be considered. Since repair of the aortic arch by direct anastomosis, rather than by tube graft reconstruction, is not disadvantageous to survival, and since persistent or recurrent obstruction can be well managed by percutaneous balloon dilation," direct anastomosis should be used whenever possible in treating the interruption of the aortic arch. Whenever possible, this should be combined with repair of the coexisting cardiac defect in a onestage repair.
875
Reparative operations for interrupted aortic arch with ventricular septal defect. J THORAC CARDIOVASC SURG 1983;86:832-37.
3. Jonas RA. Interrupted aortic arch. In: Grillo HC. Austen W A, Wilkins EW, Mathisen DJ. Vlahakes AJ. eds. Current therapy in cardiothoracic surgery. Toronto: BC Decker. 1988. 4. Walker SH. Duncan DB. Estimation of the probability of an event as a function of several independent variables. Biometrika 1967;54:167-79. 5. Kaplan EL. Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:45781. 6. Blackstone EH. Naftel DC, Turner ME. The decomposition of time-varying hazard into phases. each incorporating a separate stream of concomitant information. J Am Stat Assoc 1986;81 :615-24. 7. Rothman A, Lang p. Lock JE. Jonas RA. Mayer JE. Castaneda AR. Surgical management of subaortic obstruction in single left ventricle and tricuspid atresia. J Am Coll Cardiol 1987;10:421-6. 8. Fishman NH. Bronstein MH. Berman W. et al. Surgical management of severe aortic coarctation and interrupted aortic arch in neonates. J THORAC CARDIOVASC SL:RG 1976;71 :35-48.
9. Allard JR. Williams RL. Dobell ARC. Interrupted aortic arch: factors influencing prognosis. Ann Thorac Surg 1976;21 :243-6. 10. Higgins CB, French JW. Silverman JF. Wexler L.
Interruption of the aortic arch: preoperative and postoperative clinical. hemodynamic. and angiographic features. Am J Cardiol 1977;39:563-71. II. Moulton AL. Bowman FO. Primary definitive repair of type B interrupted aortic arch. ventricular septal defect. and patent ductus arteriosus. J THORAC CARDIOVASC SL:RG 1981;82:501-10. 12. Turley K. Yee ES. Ebert PA. The total repair of interrupted arch complex in infants: the anterior approach. Circulation 1984;70(Pt 2):116-20. 13. Schumacher G. Schreiber R. Meisner H. Lorenz HP. Sebening F. Biihlmeyer K. Interrupted aortic arch: natural history and operative results. Pediatr Cardiol 1986; 7:89-93. 14. Zahka KG. Roland MA. Cutilletta AF. Gardner TJ.
Donahoo JS. Kidd L. Management of aortic arch interruption with prostaglandin E] infusion and microporous expanded polytetrafluoroethylene grafts. Am J Cardiol
We express our appreciation for the considerable participation of Mr. Rob Brown in all aspects of the study and to Ms. Debbie Nuby for her contribution to the graphics and the manuscript.
1980;46:1001-5. 15. Leoni F. Huhta JC. Douglas J. et al. Effect of prostaglan-
REFERENCES I. Rein JG, Freed MD. Norwood WI. Castaneda AR. Early and late results of closure of ventricular septal defect in infancy. Ann Thorac Surg 1977;24:19-27. 2. Norwood WI. Lang P, Castaneda AR. Hougen TJ.
din on early surgical mortality in obstructive lesionsof the systemic circulation. Br Heart J 1984;52:654-9. 16. Jonas RA. Lang P, Hansen DD. Hickey PRo Castaneda AR. First-stage palliation of hypoplastic left heart syndrome: the importance of coarctation and shunt size. J THORAC CARDIOVASC SL:RG 1986;92:6-13. 17. Becu LM, Tauxe WN. DuShane JW. Edwards JE. A
The Journal of
18.
19.
20.
21.
22.
Thoracic and Cardiovascular Surgery
Sell et al.
876
complex of congenital cardiac anomalies: ventricular septal defects, biventricular origin of the pulmonary trunk, and subaortic stenosis. Am Heart J 1955;50:901-11. Van Praagh R, Bernhard WF, Rosenthal A, Parisi LF, Fyler DC. Interrupted aortic arch: surgical treatment. Am J Cardiol 1971;27:200-11. Freedom RM, Bain HH, Esplugas E, Dische R, Rowe RD. Ventricular septal defect in interruption of aortic arch. Am J Cardiol 1977;39:572-82. Ho SY, Wilcox BR, Anderson RH, Lincoln JCR. Interrupted aortic arch: anatomical features of surgical significance. Thorac Cardiovasc Surg 1983;31:199-205. Sennari E. Morphological study of ventricular septal defect associated with obstruction of aortic arch among Japanese. Jpn Circ J 1985;49:61-7. Ilbawi MN, Idriss FS, DeLeon SY, Muster AJ, Benson OW, Paul MH. Surgical management of patients with
Appendix A: Morphologic details and relation univariateIy to total mortality
Table At Deaths (early and late)
n
% of7l
No.
%
C
20 49 2
28 69 3
12 25 2
60 51 100
Total
71
100
39
55
A B
24. 25.
26.
Table A2 Type of associated cardiac anomalies
Appendixes
Type of interrupted arch
23.
interrupted aortic arch and severe subaortic stenosis. Ann Thorac Surg 1988;45:174-80. Saul JP, Keane JF, Fellows KE, Lock JE. Balloon dilation angioplasty of postoperative aortic obstructions. Am J Cardiol 1987;59:943-8. Kirklin JW, Barratt-Boyes BG: Cardiac surgery. New York: John Wiley, 1986:1046. Moulaert AJ, Oppenheimer-Dekker AJ. Anterolateral muscle bundle of the left ventricle, bulboventricular flange and subaortic stenosis. Am J Cardiol 1976;37:78-81. Moulaert AJ, Bruins CC, Oppenheimer-Dekker A. Anomalies of the aortic arch and ventricular septal defects. Circulation 1976;53:1011-5.
P(X')
0.3
0.3
Note: Tables AI and A2 include all patients entering the hospital. including eight (seven deaths) who did not reach operation.
Deaths (early and late)
n
% of7l
No.
%
VSD (isolated) "Single ventricle" Truncus arteriosus DORV TGA+ VSD Complete AV canal DOlV Isolated ventricular inversion + VSD None (PDA present)
44 8 7 5 2 2 1
62
19 8 6 3 0 1
I
I
43 100 86 60 0 50 100 100
Total
71
II
10 7 3 3 1
1 I
0
0
39
55 .01
P (x') for VSD vs. all others
Key: AV. Atrioventricular; DOL V. double-outlet left ventricle; DORV. doubleoutlet right ventricle; PDA. patent ductus arteriosus; TGA. transposition of the great arteries; VSD. ventricular septal defect.
Note: "Single ventricle" included three patients with complex tricuspid atresia. one patient with complex mitral atresia. and four patients with double-inlet left ventricle.
Table A3
Type of associated cardiac anomalies VSD (isolated) "Single ventricle" Truncus arteriosus DORV TGA+ VSD Complete AV canal DOlV Isolated vs. inversion + VSD None (PDA present) Total
Type C
n
No.
44
7
8
2
5 2 2 2
2
I
7
5
No.
% of 2
2
100
2
100
I 1 I
71
20
AV. Atrioventricular; DOlV. double-outlet left ventricle; DORV. double-outlet right ventricle; fAA. interrupted aortic arch; PDA. patent ductus arteriosus: TGA. transposition of the great arteries; YSD. ventricular septal defect.
Volume 96 Number 6 December 1988
Appendix B: Parameter estimates, P-~alues, and multivariate analysis (variables entered, variables retained, coefficients, and P-values) for death at any time after the initial operation, in a time-related manner (n = 63). Variables entered into the analysis included demographic ones (birth weight, age at admission, date of admission, date of operation); morphologic variables (type of interrupted aortic arch, VSD alone, "single ventricle," as a group, truncus arteriosus, double-inlet left ventricle, ventriculoarterial discordant connection, double-outlet right ventricle, multiple VSDs, hypoplastic ventricle, noncardiac congenital anomaly); preoperative variables (interval from admission to operation, use of prostaglandin E), use of dopamine, use of preoperative intubation and ventilation, pH at admission, pH immediately preoperatively); and surgical variables (operation performed during the night, one-stage repair, type of arch repair). A single declining hazard phase was found. Shaping parameter estimates, the variables retained (P < 0.1) their coefficients plus or minus one standard deviation, and, in
Interrupted aortic arch
877
parentheses, the P values, are as follows: 0 = 0, p = 08868, v = 3912, II. = 0, intercept 2.277, date -0.2051 ± 0.066 (P = 0.(02), "single ventricle" 2.174 ± 0.61 (P = 0.0004), truncus arteriosus 1.661 ± 0.55 (P = 0.003), multiple VSDs 1.984 ± 0.57 (P < 0.0005), immediately preoperative pH - 3.507 ± 1.29 (P = 0.006), and left ventricular-aortic conduit 1.515 ± 0.65 (P = 0.02). Appendix C: Parameter estimates, P-~alues, and multivariate analysis variables entered, variables retained, coefficients, and P-~alues of the event recurrent or persistent aortic arch obstruction in a time-related manner. Variables entered into the analysis included the demographic, morphologic, and surgical variables listed in Appendix B. A single constant hazard phase was found. The variables retained (P < 0.1), their coefficients plus or minus one standard deviation, and, in parentheses, the P values, are as follows: intercept -5.055, direct anastomosis 4.969 ± 0.92 (P < 0.0001), and interaction of direct anastomosis with date of repair -0.3228 ± 0.091 (P = 0.(004).