A comparison of coarctation resection and subclavian flap angioplasty using ultrasonographically monitored postocclusive reactive hyperemia

J

THoRAc CARDIOVASC SURG

1990;100:817-29

A comparison of coarctation resection and subclavian flap angioplasty using ultrasonographically monitored postocclusive reactive hyperemia The reported relath'ely high incidence of early restenosis at the coarctation repair site with subclavian Rap angioplasty, especially in infants less than 3 months of age,.prompted a physiologically oriented analysis of relief of obstructionfrom coarctation after subclavian Rap angioplasty verses resection and end-to-end anastomosis in infancy. Twenty-one patients who bad undergone repair of coarctation in infancy by either subclavian Rap angioplasty (nine patients) (median age 8 years) or resection and end-to-end anastomosis (12 patients) (median age 8 years)wereevaluated by Doppler spectrum analysis of the blood Row velocities in the femoral artery at rest and during reactive hyperemia. The median resting right upper to lower limb systolic pressure difference (with interquartile range) was similar in the angioplasty, resection and anastomosis, and control groups: -5 mm Hg (18 mm Hg), 0 mm Hg (12 mm Hg), and -2.5 mm Hg (10 mm Hg), respectively. Also, similar resting values for the maximum frequency of the advancing curve and the pulsatitity and resistance indices were measured in the three groups. During reactive hyperemia of the leg, however, a significant hemodynamic obstruction across the repair. ~ite became clinically manifest in the angioplasty group only, as documented by a lower puhltitity index in comparison with the control group (p = 0.01, Mann-Whitney U test). Comparison of the hemodynamic results between the angioplasty and resection and anastomosis groups in subdivisions of infants operated on at an age of less or greater than 3 months, both at rest and during reactive hyperemia, showed, already at rest, a significantly lower value for the pulsatitity index in the former angioplasty subdivision (P = 0.05, Student's t test), indicating a significant resistance at the coarctation repair site in the angioplasty patients operated on before the third month of life. A disadvantage of angioplasty (compared with resection and anastomosis) was noted when angioplasty was performed before the third month of life, and an unequivocal lack of advantage was noted when performed beyond that period regarding reliefof obstructionfrom coarctation. In addition, a definite potentialfor adverse long-term effects on the hemodynamics of the left upper 6mb after subclavian Rap angioplasty in infancy bas been documented. For these reasons we prefer to perform resection and end-to-end anastomosis for repair of coarctation in infancy.

Jacques A. M. van Son, MD, Wim N. J. C. van Asten, PhD, Henk J. J. van Lier, MSc, Otto Daniels, MD, Stefan H. Skotnicki, MD, and Leon K. Lacquet, MD, Nijmegen, The Netherlands

From the Departments of Thoracic and Cardiac Surgery, Pediatric Cardiology, and Statistical Consultation, University Hospital Nijmegen St. Radboud, Nijmegen, The Netherlands. Received for publication Aug. 3, 1989. Accepted for publication March 12, 1990. Addressfor reprints: Jacques A. M. van Son, MD, Department of Thoracic and Cardiac Surgery, University Hospital Nijmegen St. Radboud, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands. 12/1/21377

Amost half a century after the first successful resection and end-to-end anastomosis (RETE) for treatment of aortic coarctation,I strong and divergent preferences still exist as to the optimal management of this common congenital anomaly, particularly in early infancy. High restenosis rates for neonates treated by the conventional technique of RETE2-5 were attributed to inadequate growth of a circumferential suture line, which led to 817

8 18

The Journal of Thoracic and Cardiovascular Surgery

van Son et al.

frequency (kHzl - - - -- - --

75

- - - -- - - - - - - - ,

DOPPLE~ SPECT~UM

50

2.5

0

- 2.5

A 0

1

0.5

frequency (kHz) -

7.5

-

-

-

-

-

-

t ime (sec)

-

-

-

-

-

-

-

-

-

--.,

FRONE!<' S P':'RAMETERS OF OOPPLER 5PECTRUl1

FMAX

5.0

25

0

-2 .5

0

8

FIHN

0.5

1

t ime (s ec)

Fig. 1. Example of a Doppler spectrum with MAX-curves (A and B) and parameters derived from the MAX-curves (B). In the advancing MAX-curve, FMAX indicates the maximum frequency of the curve, FMEAN+ the mean frequency, and FDIA the end-diastolic frequency, In the receding MAX-curve, FMIN is the maximum frequency of the curve,

introduction of the subclavian flap angioplasty (SFA).6 Although excellent long-term results have been reported with this technique in terms of reduced mortality and restenosis rates.?'!? in the recent literature support has been lent to the concept that recrudescence and retraction of residual ductal tissue in the partially resected diaphragmatic ridge after SFA at neonatal age may be more important than growth failure of a circumferential anastomosis after RETE in causing restenosis. 11-14 Considering this potential disadvantage of the S FA technique, we believed that an analysis of the merits of SFA versus RETE regarding relief of obstruction from coarctation was warranted. Because Doppler spectrum analysis has proved to be an accurate and reliable technique for the detection of altered hemodynamics distal to coarctation of the aorta in infants and young children, 15 we applied it in two groups

of children who in infancy had been subjected to either RETE or SFA. In addition, we used the technique of postocclusive reactive hyperemia of the lower limb as a powerful diagnostic criterion to detect clinically significant stenoses at the coarctation repair site that are not apparent at rest. To allow a true comparison of the longterm hemodynamic results of both operative techniques, we subdivided the SFA and RETE groups according to age at the time of repair.

Patients and methods Twenty-one children were selected at random for the study from a total of 56 long-term survivors operated on in infancy, from 1973 to 1989, for coarctation of the aorta, The only criteria for selection were that they had undergone repair of coarctation of the aorta within the first year of life by either RETE or SFA and that at least 5 years had elapsed since the operation, Twelve children had undergone RETE and nine SFA in infancy. In the former group five patients had had repair at an age of less than 3 months and seven at an age of 3 months or greater, The corresponding numbers in the SFA group were three and six. In the RETE group there were seven boys and five girls, whose ages ranged from 6 to 15 years (median age 8 years; interquartilerange [IQR] 5 years); in the SFA group there were four boys and five girls, whose ages ranged from 6 to 10 years (median age 8 years; IQR 2.5 years), The control group consisted of 10 children, five boys and five girls, whose ages ranged from 6 to 14 years (median age 9.5 years; IQR 4.2 years); they were similar in size to the patients in the two operative groups but did not have coarctation of the aorta or cardiac disease. For the purpose of having reference systolic blood pressure values, resting systolic blood pressures were measured in right upper and lower limbs with the children in the supine position; a sphygmomanometer and bidirectional8-MHz Doppler ultrasound velocity detector were used. Right arm and right thigh pressures were representative for the right brachial and right femoral arteries, respectively. Pneumatic cuff sizes were chosen to cover approximately 75% of limb length." In case of local stenosis of the right femoral artery, the resting systolic blood pressure of the left thigh was determined. Recurrent stenosis at the coarctation repair site is defined as the presence of a right brachial-femoral systolic blood pressure difference of 20 mm Hg or greater, after its initial absence. Blood flow velocities in the right brachial and right femoral arteries were measured by Doppler ultrasonography. In case of local stenosis of the right femoral artery, Doppler signals on the left side were recorded. Doppler signals were obtained both at rest and during reactive hyperemia, using an 8-MHz bidirectional continuous-wave probe (Meda Sonies, Mountain View, Calif), with the patient in a supine position, Doppler probes were placed at the right antecubital fossa and directly distal to the inguinal ligament and were kept at an angle relative to the vessel axis that corresponded with optimal Doppler spectra for all measurements. To produce reactive hyperemia, we placed a sufficiently wide pneumatic cuff l6 around the limb of interest, inflated it 50 mm Hg above systolic pressure for 4 minutes, and then deflated it. Doppler spectra were continuously observed immediately after release of the pneumatic cuff until the veloc-

Volume 100 Number 6

Coarctation of aorta

December 1990

8 19

/ j'

Fig. 2. Resection and end-to-end anastomosis. Incision lines in coarctation of the aorta with hypoplasia of the distal aortic arch. ity value returned to preocclusion levels. After elicitation of a peak vasodilatory stimulus, an increase in blood flow velocity is found, followed by a gradual decrease to the pre-stimulus level. In our young patient population we found that peak reactive hyperemia occurs within IS seconds after cuff release. The spectra observed during maximum vasodilation (from 5 to 15 seconds after cuff release) were processed and stored by a realtime spectrum analyzer (model SA 8000, Radionics Medical, Scarborough, Ont., Canada). The spectrum with the highest velocity value was subsequently selected and fed into a Digital MNC 11/23 computer (Digital Equipment Corp., Maynard, Mass.) on the basis of electrocardiographic triggering for off-line analysis. Maximum frequency waveforms (MAXcurves) during the heart cycle were calculated from the spectra by a sophisticated local algorithm.l? Processing the advancing and receding flowseparately yields two MAX-curves from each spectrum. These indicate the maximum advancing and receding velocity as a function of time (Fig. I, A). Four parameters were calculated to describe the shape of the curves (Fig. I, B).18 The pulsatility and resistance indices (PI and RI, respectively) were derived by combining different parameters. In case of forward and receding flowduring the same heart cycle PI is defined as the ratio between the peak to peak difference of the advancingand receding curves and the mean frequency of both MAXcurves (FMEAN)19.22: PI

=

FMAX - FMIN FMEAN

In case of continuous forward flow, as occurs during reactive hyperemia, PI is defined as the difference between the maximum frequency (FMAX) and the end-diastolic frequency (FDIA) divided by the mean forward frequency (FMEAN+):

Fig. 3. The resection is extended just distal to the origin of the left subclavian artery, and the inferior aspect of the aortic arch is incised in a cephalad direction to achieve a maximally wide anastomotic diameter. PI = FMAX - FDIA FMEAN+ Both in case of forward and receding flow during the same heart cycle and in case of continuous forward flow, RI is defined as the difference between the maximum frequency (FMAX) and the end-diastolic frequency (FDIA) divided by the maximum frequency-': I _ FMAX - FDIA R FMAX Both RI and PI are dimensionless figures. Three parameters, namely, FMAX, PI, and RI, were used to describe the Doppler spectra in the present study.

Operative technique

RETE. The chest is opened through a standard left posterolateral thoracotomy in the third (in case of expected hypoplasia of the distal aortic arch) or fourth intercostal space. The mediastinal pleura is opened and the upper half of the descending aorta, left subclavian artery, distal aortic arch, and ductus arteriosus are isolated. Intercostal arteries are sacrificed only as required for adequate mobilization of the descending aorta and control of the anastomotic site. The ductus arteriosus is divided and oversewn, or the ligamentum arteriosum is divided. The proximal vascular occlusive clamp is placed across the aortic arch and the left subclavian artery so that patency of the left common carotid artery is maintained. The distal clamp is placed

The Journal of Thoracic and Cardiovascular

8 2 0 van Son et al.

I

Surgery

I

',' Iii

rc-... ,. )

)0

Fig. 4. A matching anastomosis is created after oblique transection of the distal aorta.

across the upper descending aorta distal to the coarctation. The coarctation and, when present, the stenotic isthmic segment are completely excised, as well as any residual ductal tissue. In patients with tubular hypoplasia of the distal aortic arch, the proximal clamp is positioned just distal to the innominate artery, occluding the left common carotid artery. Subsequently, the resection is extended just distal to the origin of the left subclavian artery, and the inferior aspect of the aortic arch is incised in a cephalad direction to achieve a maximally wide anastomotic diameter (Figs. 2 and 3). After the resection an end-to-end anastomosis has to be established, which should fulfill two criteria. To begin with, it should be at least as large as the aortic arch. This is ensured by making the proximal aperture in the aorta so large that the outer curvature of the aorta straightens as it continues into the descending aorta. Second, efforts must be made to prevent turbulence in the blood flow. This is achieved by ensuring that the outer and inner curvatures of the anastomosis coincide with the contours of the aorta. If necessary, the distal aorta is trimmed obliquely to create a matching anastomosis (Fig. 4). In the past, the anastomosis has been performed either with interrupted sutures placed circumferentially or with a running suture on the posterior aspect and interrupted ones anteriorly (6-0 or 7-0 polypropylene). Currently, we use a circumferentially running suture (6-0 or 7.0 polypropylene or polydioxanone). In a less common situation, when extreme hypoplasia of the aortic arch segment between the left common carotid and left subclavian arteries is encountered or when this segment is very long and hypoplastic, the previously described technique of

Fig. 5. In the subclavian flap angioplasty technique, a longitudinal incision is carried along the lateral wall of the left subclavian artery toward the aorta, through the stenotic isthmus, and well beyond onto the lateral wall of the descending aorta.

extended RETE can be unsatisfactory. In this case we perform, in addition to an extended RETE, an enlargement of the hypoplastic aortic arch segment, as described by. Amato.I" Vincent,25 and their associates. SFA. The chest is opened through a standard left posterolateral thoracotomy in the third or fourth intercostal space. The mediastinal pleura is opened vertically over the aorta, superiorly over the isthmus, the distal aortic arch, and left subclavian artery, medially over the ductus arteriosus, and inferiorly well below the coarctation. The distal transverse aortic arch and isthmus, proximal descending aorta, patent ductus arteriosus, and left subclavian artery are mobilized and clamped. Occasionally it may be necessary to sacrifice one or two pairs of intercostal arteries above the level of the distal clamp. The ductus arteriosus is divided and oversewn, or the ligamentum arteriosum is divided. The vertebral artery is identified and ligated at its origin to prevent subsequent steal from the cerebral circulation to the left upper limb. The left subclavian artery is ligated and transected distally at the thoracic outlet. A longitudinal incision is carried along the lateral wall of the left subclavian artery toward the aorta, through the stenotic isthmus, and well beyond onto the lateral wall of the descending aorta to avoid residual narrowing at the site of (or just distal to) the coarctation (Fig. 5). In all instances, the posterior diaphragm is excised as completely as possible, without weakening the vessel wall in this area unduly. Even if the flap is wide, the distal corners are not trimmed so that all flap tissue is utilized in the reconstruction. Subsequently, a 6-0 or 7-0 polypropylene or polydioxanone suture is passed through the apex of the flap which is approxi-

Volume 100 Number 6 December 1990

Coarctation ofaorta 8 2 I

,) Fig. 6. A suture is passed through the apex of the flap, which is approximated to the lowest point of incision in the descending aorta.

Fig. 7. Two sutures are started at the proximal end of the flap and extended to the apex.

mated to the lowest point of incision in the descending aorta (Fig. 6). Finally, two similar running sutures are started at the proximal end of the flap and extended to the apex (Fig. 7). Statistical analysis. Quantitative data were compared with the Kruskal- Wallis test, Student's t test, and Mann-Whitney U test. Grouped data are given as median with IQR. A P value less than 0.05 indicates a difference unlikely to be due to chance alone.

showed a recurrent diaphragm at the repair site. Subsequent balloon angioplasty was initially successful, although 5 years later we detected a recurrent pressure difference of 60 mm Hg, also indicated by resting FMAX and PI values in the femoral artery of 3066 Hz and 1.68, respectively. In the SFA group operated on at an age of 3 months or older, no recurrent stenosis was found. In the RETE group five patients had had coarctation repair at an age of less than 3 months, and a recurrent right brachial-femoral pressure difference did not develop in any of them. In the RETE group operated on at an age of 3 months or older, one patient, who had undergone repair at an age of 6 months (Table lA, case 4), had a right brachial-femoral systolic pressure difference of 20 mm Hg 10.5 years after the operation. The resting FMAX and PI values in this patient were 3352 Hz and 1.83, respectively. In both patients with a recurrent pressure difference, satisfactory augmentation of flow in the femoral artery was documented during reactive hyperemia, simultaneously with a more than average decrease of the peripheral vascular resistance, as indicated by FMAX and RI values of 8430 Hz and 0.43 and 6853 Hz and 0.47, respectively. Doppler spectrum analysis. At rest, biphasic and triphasic spectra were observed in the femoral artery in 11

Results Systo6c blood pressure difference between right upper and lower Hmbs. The data at rest for the patients and control subjects are shown in Tables IA and lB. There was no significant difference between the patients and control subjects with regard to age, systolic blood pressure in the right brachial and femoral arteries, and right brachialfemoral systolic blood pressure difference (Table II). Also, there were no significant differences in right brachial-femoral systolic blood pressure difference between patients operated on at an age less than or greater than 3 months. Recurrent stenosis at the coarctation repair site. In the SFA group, three children had been subjected to coarctation repair at an age ofless than 3 months. Fifteen months postoperatively, in one of these (Table lA, case 7), a recurrent right brachial-femoral systolic pressure difference of 50 mm Hg had been detected. Angiography

The Journal of

van Son et al.

822

Thoracic and Cardiovascular Surgery

Table IA. Individual baseline values in coarctation patients

Case

Age at operation (mo)

Age at examination (yr)

RB-F SBP difference (mmHg)

RBA

FA

RBA

FA

RBA

FA

0.25 3 5 6 6 3 2 I 5 0.5 5 1

6 11 8 11 11 7 6 6 6 15 8 8

0 -5 -15 20 -10 15 0 0 -10 0 -5 0

5465 6546 1685 6185 7024 3078 4073 5301 3868 5130 5240 6630

4968 4435 5100 3352 4342 4838 6486 6072 7897 8938 6266 6642

5.05 10.39 4.63 6.64 2.68 6.28 6.70 5.25 4.24 7.53 2.99 6.24

3.33 3.36 2.45 1.83 4.40 1.24 4.31 3.68 3.66 4.67 5.21 2.31

0.95 1.00 0.99 1.00 0.98 1.00 1.00 0.99 0.99 1.00 0.98 0.99

0.85 1.00 0.83 0.93 1.00 0.69 1.00 1.00 1.00 1.00 1.00 0.85

0.25 4 6 3.5 3 7.5 0.25 3.5 2

6 8 10 7 8 8 6 10 8

10 -5 0 -10 -5 10 60 15 -5

6113 7579 4237 3924 5453 5334 7667 7410 4509

6509 11184 6292 10698 7880 8700 3066 4438 5412

5.07 4.16 6.32 5.15 4.97 7.05 5.58 3.43 2.16

2.50 4.38 3.60 4.51 4.45 3.83 1.68 4.50 2.44

0.98 1.00 0.97 0.97 1.00 0.99 0.98 0.97 0.99

0.89 1.00 0.99 0.99 1.00 1.00 0.83 1.00 0.86

FMAX(Hz)

PI

RI

RETE I

2 3 4 5 6 7 8 9 10 11 12

SFA I

2 3 4 5 6 7 8 9

RETE, Resection and end-to-end anastomosis; SFA. subclavian flap angioplasty; RB-F SBP, right brachial-femoral systolic blood pressure; FMAX, maximum frequency of advancing curve; PI, pulsatility index; RI, resistance index; RBA, right brachial artery; FA, femoral artery.

Table lB. Individual baseline values in control subjects

Case

Age at examination (yr)

RB-F SBP difference (mmHg)

RBA

FA

RBA

FA

RBA

FA

I 2 3 4 5 6 7 8 9 10

9 7 14 9 8 10 12 6 10 12

-10 0 0 0 -10 - 5 -15 0 0 -10

5908 4284 3571 4731 3736 4136 5144 5569 10132 6818

8730 8153 4674 4661 4261 6476 4967 5823 7169 6093

1.86 3.27 5,21 5.47 5.51 4.14 7.00 6.26 7.16 5.12

6,04 1.91 2.55 4.06 4.98 2.80 4.63 4.95 4.26 2.15

0.75 0.89 0.87 1.00 0.94 0.93 0.89 1.00 1.00 1.00

1.00 0.78 0.89 1.00 1.00 0.89 1.00 1.00 1.00 0.98

FMAX(Hz)

PI

RI

RB-F SBP, Right brachial-femoral systolic blood pressure; FMAX, maximum frequency of advancing curve; PI, pulsatility index; RI, resistance index; RBA, right brachial artery; FA, femoral artery.

patients of the RETE group and in eight patients of the SFA group. During reactive hyperemia, we invariably documented loss of the reverse flow component and occurrence of a monophasic spectrum. The only patients with a monophasic spectrum at rest were both aforementioned children with a recurrent stenosis (Fig. 8). The median values of the FMAX, PI, and RI param-

eters, as obtained from the right brachial and femoral arteries in the RETE, SFA, and control groups, both at rest and during reactive hyperemia, are presented in Tables III and IV. A significant difference in the median values of the FMAX, PI, and RI parameters in the brachial artery and in the median values of the FMAX and RI parameters in

Volume 100 Number 6

Coarctation of aorta

December 1990

823

Table II. Resting values in coarctation patients and control subjects RETE

In =

12)

SFA

In =

9)

Controls

In =

10)

p

8 (2.5)

Upper limb SBP (mm Hg)

120 (15)

120 (32)

Lower limb SBP (mm Hg) RB-F SBP difference (mm Hg)

129 (19)

125 (30)

120 (16)

NS

0(12)

-5 (18)

-2.5 (10)

NS

9.5 (4.2)

NS

(yr) 120 (21)

-, Ii- ....\\

Value"

8 (5.0)

Age at examination

Frequency 1kHz) 7.5 r-~-~---"--""--..,.----r----r----...----...---,

so

(

NS

,1, ,I

25

Values are expressed as median with interquartile range. RETE. Resection and end-to-end anastomosis; SFA. subclavian nap angioplasty; RB-F SBP. right brachial-femoral systolic blood pressure; NS. not significant. *Kruskal-Wallis test.

the femoral artery was not detected among the three groups, neither at rest nor during reactive hyperemia.PI, as obtained from the femoral artery in the SFA group during reactive hyperemia, was the only parameter that reached statistical significance in comparison with the control group (p = om, Mann-Whitney U test) (Table IV) (Fig. 9). Comparison of the various parameters between the corresponding subdivisions of the SFA and RETE groups revealed lowerPI valuesat rest and during reactivehyperemia in the childrenof the SFA group whohad had repair before the third month of life. This difference became significant at rest (p = 0.05, Student's t test). Discussion Critique of methods. Neglecting gravitational potential energy, with the flow horizontal, any pressure difference across a vascular stenosiscan be attributed to frictionallosses and increased kinetic energy in the random motions of turbulence.26 Both frictionalenergy losses and turbulence increase with velocity. The highest pressure difference will, therefore,appear at the highestbloodflow velocity. Thus an abnormal pressure difference, which is not apparent at rest, may develop during and after exercise when blood flow velocity has increased. During the past decade, exercisestress testing in infants and young children has been advocated as an accurate technique to increase blood flow acrossthe coarctation repair site.27-31 We have some concern, however, about reproducibility and quantification of exercisestress testing at this young age. The degree of hyperemia caused by exercisestress testingdepends on the general condition of the infant or childand, therefore,may differconsiderablyamong individuals. Therefore we have selected elicitationof postoc-

o

A

. 2.5 7.5

so

o

B

• 2.5 '---'--'--'--'-------'-----"----'-----"----'-----'--~

o

0.5

Frequency 1kHz) 7.5 r--~-~-~-~~-~-~-----,.-___,

so 2.5

o . 2.5

75

C

so

o

os

1

time Is)

Fig. 8. Representative Doppler spectra as recorded from the femoral artery after successful RETE at rest (A) and during reactive hyperemia (B~ after SFA with a recurrent right brachial-femoral systolic pressure difference of 60 mm Hg, at rest (C) and during reactive hyperemia (D~

824

The Journal of Thoracic and Cardiovascular Surgery

van Son et al.

Table III. Doppler spectrum parameters from right brachial artery at rest and during reactive hyperemia in the RETE, SFA, and control groups During reactive hyperemia

At rest Groups

FMAX (Hz)

5270 (2536) 5453 (3121) Control 4938 (2100) RETE

SFA

p Value*

NS

PI 5.75 (2.35) 5.07 (2.16) 5.34 (2.52)

P Value*

NS

RI

P Value*

FMAX (Hz)

P Value*

NS

9241 (6692) 10141 (2318) 10546 (2641)

NS

0.99 (0.04) 0.98 (0.03) 0.94 (0.1 I)

PI 1.17 (0.37) 1.03 (0.66) 1.01 (0.33)

P Value*

NS

RI 0.68 (0.08) 0.63 (0.18) 0.62 (0.1 I)

P Value*

NS

Values are expressed as median with interquartile range. For abbreviations see Tables I and II. *Kruskal-Wallis test.

elusive reactive hyperemia.P'" which is independent of this variable, to maximally increase the blood flow across the coarctation repair site. Further, we have not measured systolic blood pressure differences but blood flow velocities during reactive hyperemia, because the former technique admits ambiguous results, which are caused by the difference in recovery phase between the normal and stenotic inflow tracts of the right brachial and femoral arteries, respectively. Because the recovery time after peak reactive hyperemia is short and the blood pressure in infants and young children is extremely labile, timing of the pressure measurement during maximum blood flow velocity is crucial. Therefore, by nonsimultaneous blood pressure measurement in these patients, as performed in some of the previously mentioned studies, errors already introduced by using exercise stress testing may be aggravated. In the underlying study we have stored the Doppler spectra of all cardiac cycles from 5 to 15 seconds after cuff release because the influence of reactive hyperemia on the hemodynamics of the underlying vascular system reaches its peak during this period. The cycle with the highest velocity value was subsequently selected for off-line analysis. We strove to maintain the same Doppler angle relative to the axis of the vessel for both the measurements at rest and during reactive hyperemia. We are aware of the fact that this angle may be variable and influences FMAX. PI and RI, however, are independent of the insonation angle. All aforementioned factors considered, we believe that ultrasonographic analysis of the blood flow velocity in the descending aorta at rest and during reactive hyperemia is a sophisticated discriminative technique in the detection of hemodynamic obstructions at the coarctation repair site. In this study we have performed Doppler spectrum analysis to assess the hemodynamics in the right brachial and femoral arteries of children who had been subjected to either RETE or SFA for repair of coarctation in infancy. Doppler spectrum analysis provides information about

the proximal and distal resistance (measured from the recording site) and the total impedance of the vascular bed. The resistance index (RI) is frequently used as an indicator of the degree of distal vascular resistance. Our study clearly shows that at rest the peripheral vascular resistance in the upper and lower extremities is high in both surgical groups and in the control subjects (RI is 1.0 or approximates 1.0). After release of the cuff the resistance decreases drastically, because the peripheral vascular bed dilates after elicitation of peak reactive hyperemia (RI is approximately 0.60). Because RI was not significantly different in both operative groups and the control group, we assume that there is no significant difference in peripheral vascular resistance among the three groups and that therefore none of the patients in this study had distal vascular pathology. The PI has proved to assess accurately arterial inflow.21• 22 In the absence of multisegment vascular disease, PI is a reliable indicator not only of proximal arterial stenosis, but also of the ratio between proximal and distal resistances. Reactive hyperemia hardly influences the proximal vascular resistance, but elicitates a considerable decrease in the peripheral vascular resistance as a result of tissue hypoxia and the accumulation of metabolic products.F During the period of increased limb blood flow, clinically significant stenoses that are not critical at rest may become apparent. This is manifested by an increase of the influence of the proximal resistance on the hemodynamics of the underlying vascular bed. Hence proximal obstructions can be detected more accurately. Review of previous studies. The initial approach to surgical repair of aortic coarctation was resection of the coarctation with end-to-end anastomosis,' but this technique in particular has been associated with high early mortality and restenosis rates. 2- S SFA6 was developed to overcome these problems. SFA is an expeditious procedure because it avoids extensive dissection and mobilization of the aorta, frequently does not require sacrifice of intercostal arteries, and produces a noncircurnferential

Volume 100 Number 6

Coarctation of aorta

December 1990

825

Pulsat il ity index - - - - - - - - - - - - - - ,

6

5

3

2

Fig. 9. Median pulsatility index, at rest and during reactive hyperemia, as obtained from the femoral artery in patients treated with SFA,in those treated with RETE, and in control subjects. I = At rest: 0 , pulsatility index in RETE; D, pulsatility index in controls.A. pulsatility index inSFA. 2 = During reactive hyperemia:., pulsatility index in RETE;. pulsatility index in controls; ., pulsatility index in SFA.

Fig. 10. Transverse section of the periductal aorta (Ao) of a young infant with preductal coarctation. Underneath theaortic internal elastic lamina (i.e.l.) the lightly stained ductal tissue is clearly seen to completely encircle the lumen of the aorta. An intimal cushion (i.e.) is present in thisspecimen. At the site of theentrance of theductus arteriosus (D.A .). all along onethird ofthetotal circumference of theaorta,ductal tissue constitutes the full thickness of the aortic media (elastictissue stain; original magnification X 10). (Reprinted with permission of the British Heart Journal. D. KrikJer, MD, ed. 1983;49:317-23, courtesy of N. J. Elzenga, MD.)

suture line. A particular advantage of SFA is its use of native tissue in a vascularized pedicle, thus offering the potential for growth.10 A loweroperative mortality and a lower restenosis rate have been reported for SFA than for RETE.7. 10 However, many initial reports included few infants younger than 3 months of age.30•31. 38. 39 As soon as larger seriesof infants who had undergone SFA under 3 months of age were reported, a relativelyhigh incidence of early restenosis became apparent;'! '!" despite the advantage of coincidental improvement in preoperative management, primarily the introduction of prostaglandin E1.40-43 Although at the time of the original operation the subclavian flap had been carried well beyond the area of coarctation, reoperation invariably revealed that the residual posterior diaphragm had continued to involute and obliterate the aortic lumen, the process stopping at the subclavian flap suture line.I I. 14

These clinical data are corroborated by morphologic studies,which showthat the abundant presenceof ductal tissue in the periductal aorta 44 • 45 (Fig. 10) leads to an excessive age-dependentmoldingprocessof this area .45•46 The implicationof all these data is that the coarctogenetic potential of residual ductal tissue may circumvent the advantage of SFA because the mechanism of recurrence may outstrip the growth potential of the tissue flap. 12 This hypothesisis consistent with the fact that the majority of restenoses after SFA performed during the first 3 months of life occur within the first year after operation.P?" A second factor that must be considered in evaluating the use of SFA in coarctation of the aorta is the effect of ligationof the left subclavianartery . We noted detrimental effects after SFA in infancy, consisting of left upper arm shortening and complaints of claudication in the left arm during strenuous exercise.!" Subsequent hemody-

2

The Journal of Thoracic and Cardiovascular

8 2 6 van Son et al.

Surgery

Table IV. Doppler spectrum parameters from femoral artery at rest and during reactive hyperemia in the RETE, SFA, and control groups At rest Groups

RETE

FMAX (Hz)

5586 (2052) 6509 (4784) Control 5958 (2744)

SFA

p Value"

NS

PI 3.51 (2.02) 3.83 (2.01) 4.16 (2.51)

During reactive hyperemia P Value"

NS

RI 1.00 (0.15) 1.00 (0.13) 1.00 (0.14)

Values are expressed as median with interquartile range. For abbreviations see Tables 1 and II. *Kruskal-Wallis test. tMann-Whitney U test: RETE-SFA. p = NS; RETE-control, p

= NS;

P Value"

FMAX (Hz)

P Value·

NS

8291 (1750) 9066 (1790) 10114 (1783)

NS

SFA-control, p

namic analysis of the blood flow velocity in the left brachial artery during reactive hyperemia in this subset of children showed that the capacity of physiologicaugmenta tion of the blood flow velocity in the left brachial artery was marginal to absent."? This finding emphasizes that the balance between a reduced left brachial artery perfusion pressure and compensatory mechanisms is a delicate one and may easily be disturbed. Other studies found that interruption of the subclavian artery causes a substantial diminution in the longitudinal growth of the long bones or diminution in the muscle thickness of the corresponding arm, or both. 48-5o A third drawback of SFA is inherent in the need to excise as completely as possible the diaphragm of the coarctation to remove the ductal tissue. The remaining rough surface may induce thrombosis and granulation, with a resultant increased risk of retraction and restenosis. Conversely, after too radical resection, a weakened area can be created with potential for aneurysm development.51 The implication of these data is that all ductal tissue should be removed to prevent recurrent stenosis. This pertains not only to the preductal subtype of coarctation, in which ductal tissue is almost invariably present,44.45 but also to the paraductal subtype, in which ductal tissue is also usually present.v' 52, 53 RETE, although not being a panacea for the surgical treatment of coarctation of the aorta, may avoid the problems inherent to SFA. High restenosis rates with RETE in the past have been the result not only of less-refined suture techniques, but also of the fear of removing too long a segment of the aortic isthmus. Because of the elasticity of the aorta at neonatal and infantile ages, an end-to-end anastomosis can always be performed, even after resection of a long segment. 54 Since the introduction of refined microvascular techniques and materials, the disadvantage related to the circumferential suture line may be less problematic, as evidenced by the neonatal arterial switch experience.55 In

PI

P Value"

RI

0.97 (0.39) 0.61 (0.16) 0.90 (0.35) 0.034t 0.58 (0.15) 1.15 (0.28) 0.63 (0.12)

P Value"

NS

= 0.01.

this perspective, several series have reported excellent results with RETE in terms of long-term freedom from obstruction from restenosis.11-14,56,57 An additional advantage of the RETE technique is that, in case of hypoplasia of the distal aortic arch, an extended resection can be performed. 14,58,59 The resultant extended anastomosis is limited only by the diameter of the aortic arch itself, thus allowing for maximum growth potential. In this context, extended RETE relieves hypoplasia of the distal arch, whereas SFA may lead to further narrowing of this area. The hemodynamic importance of this afterload-reducing effect of extended RETE in the outcome of additional complex cardiac disease has recently been emphasized by Vouhe and associates.-? In the latter situation an approach through a single sternotomy incision to simultaneously' repair the coarctation and the intracardiac anomaly has become fashionable.v-"' This approach allows optimal mobilization of the ascending aorta and proximal aortic arch, in case creation of an end-to-end anastomosis at the levelof the proximal aortic arch is necessary. Selection of the surgical method for treating coarctation of the aorta in infancy requires documentation of the results from the techniques considered by means of the most sensitive clinical methods available. Because significant right brachial-femoral systolic pressure differences in patients without such differences at rest may be revealed during exercise, exercise testing has increasingly been used to compare surgical techniques for repair of coarctation of the aorta. Several studies have used treadmill or bicycle exercise stress testing for this purpose.F'!' Fripp and associates'? exercised children who had undergone SFA in infancy and compared them with a control group; a significantly increased pressure difference between upper and lower limbs was not detected. They concluded that SFA can be expected to result in a low incidence of subclinical residual obstruction or restenosis. Waldman and co-workers" evaluated a group of patients

Volume 100 Number 6 December 1990

who had had RETE, SFA, or patch angioplasty in infancy. Inasmuch as they noted no significant arm-to-leg systolic blood pressure differences among the three groups, they believed that optimal surgical treatment for coarctation in infancy seems to be an eclectic approach, based on intraoperative anatomic findings. As outlined before, we have some concern regarding exercise stress testing in infants and young children, as performed in these studies. Because with this technique true peak hyperemic responses may not always be elicited, the right brachial-femoral systolic blood pressure differences, as obtained during stress testing in these studies, may not in all cases have been an accurate reflection of the degree of hemodynamic obstruction at the coarctation repair site. Furthermore, as indicated previously, in stress testing absolute simultaneity of upper and lower limb pressure measurements is crucial, which criterion was not fulfilled by all of these studies. Interpretations and inferences from the present study. Comparison of the values of the FMAX, PI, and RI parameters, as obtained from the right brachial and femoral arteries in both operative groups and the control group, revealed only a significant difference in PI, as obtained fr:om the femoral artery during reactive hyperemia, between the SFA and control groups. Because PI appears to have a good correlation with the hemodynamic severity of a proximal arterial obstruction, as outlined before, our observation may reflect a subclinical hemodynamic obstruction at the coarctation repair site after SFA in infancy, becoming ultrasonographically manifest during increased blood flow. The similarity of the blood flow velocitiesin the aortoiliac system in the three groups during reactive hyperemia in the presence of a significant stenosis in the SFA group demonstrates a remarkable capacity for augmentation of the blood flow despite an obstruction. In the absence of a significant difference in the peripheral vascular resistance among the three groups, this capacity is most likely allowed by the recruitment of existing collateral connections'S bypassing the stenosis. In two patients with an angiographically proved restenosis, we documented a satisfactory augmentation of the blood flowvelocity distal to the coarctation repair site during reactive hyperemia, as evidenced by an adequate increase of FMAX. Simultaneously, lower than average RI values were documented. These findings are consistent with our hypothesis concerning the mechanisms of maintenance of peak aortic flow after long-term reductions in perfusion pressure, namely the recruitment of collateral vascular pathways and microcirculatory vasodilation.63 Our study suggests that after RETE in infancy excel-

Coarctation of aorta 8 2 7

lent hemodynamics distal to the coarctation repair site can be anticipated, whereas this effect may be less optimal after SFA during the first 3 months of life. Although a restenosis was clinically detected in only one patient of the SFA group, ultrasonographically, in comparison with a control group, a hemodynamically significant aortic obstruction was documented during reactive hyperemia in this group as a whole. Further analysis revealed that an obstruction at the coarctation repair site was only apparent in the SFA subdivision that had been operated on before the third month of life. Although we realize that the number of patients in both operative groups is small, we hypothesize that the lessoptimal hemodynamic results in theSFA group after operation during the first 3 months of life may have been caused by a subclinical stenosis of the periductal aorta as a consequence of retraction and fibrosis of residual ductal tissue. Such a subclinical stenosis becomes apparent at rest by a certain degree of reactive hyperemia distal to the coarctation repair site, as evidenced by a lower than average peripheral vascular resistance in this subset of patients. In conclusion, this study offers unequivocal evidence for lack of advantage of SFA versus RETE concerning the relief of obstruction from coarctation in infancy. In addition, in comparison with RETE, our study reveals less optimal hemodynamics beyond the coarctation repair site after SFA performed before the third month of life. Adding this information to data obtained from our hemodynamic study of the circulation in the left brachial artery after SFA in infancy, which showed a definite potential for adverse hemodynamic effects on that limb.'? and also considering morphologic research data from the literature that suggest an age-dependent potential for coarctogenesis by residual ductal tissue,44.45 we conclude that RETE seems to be the most appropriate procedure for repair of coarctation of the aorta in infancy, especially during the first 3 months of life. Besides being a hemodynamically adequate repair with excellent long-term results regarding the relief of obstruction from coarctation, it also avoids potential adverse effects on the hemodynamics of the left upper limb. We thank Gijs Stenger, MD, for preparing the illustrations. REFERENCES 1. Crafoord C, Nylin G. Congenital coarctation of the aorta and its surgical treatment. J THORAC SURG 1945;14:34761. 2. Tawes RL Jr, Aberdeen E, Waterston DJ, Bonham Carter RE. Coarctation of the aorta in infants and children: a review of 333 operative cases, including 179 infants. Circulation 1969;39(Pt 2):1173-84. 3. Williams WG, Shindo G, Trusler GA, Dische MR, Olley

8 2 8 van Son et al.

PM. Results of repair of coarctation of the aorta during infancy. J THORAC CARDIOVASC SURG 1980;79:603-8. 4. Connors JP, Hartmann AF Jr, Weldon CS. Considerations in the surgical management of infantile coarctation of aorta. Am J Cardiol 1975;36:489-92. 5. Shinebourne EA, Tam ASY, ElseedAM, et al. Coarctation of the aorta in infancy and childhood. Br Heart J 1976; 38:375-80. 6. Waldhausen JA, Nahrwold DL. Repair of coarctation of the aorta with a subclavian flap. J THORAC CARDIOVASC SURG 1966;51 :532-3. 7. Bergdahl LAL, Blackstone EH, Kirklin JW, PacificoAD, Bargeron LM Jr. Determinants of early successin repair of aortic coarctation in infants. J THORAC CARDIOVASC SURG 1982;83:736-42. 8. Campbell DB, Waldhausen JA, Pierce WS, Fripp R, Whitman V. Should elective repair of coarctation of the aorta be done in infancy? J THORAC CARDIOVASC SURG 1984;88:929-38. 9. Penkoske PA, Williams WG, Olley PM, et al. Subclavian arterioplasty: repair of coarctation of the aorta in the first year of life. J THORAC CARDIOVASC SURG 1984;87:894900. 10. Moulton AL, Brenner J I, Roberts G, et al. Subclavian flap repair of coarctation of the aorta in neonates: Realization of growth potential? J THORAC CARDIOVASC SURG 1984; 87:220-35. II. Cobanoglu A, Teply JF, Grunkemeier GL, Sunderland CO, Starr A. Coarctation of the aorta in patients younger than three months: a critique of the subclavian flap operation. J THORAC CARDIOVASC SURG 1985;89: 128-35. 12. Metzdorff MT, Cobanoglu A, Grunkemeier GL, Sunderland CO, Starr A. Influence of age at operation on late results with subclavian flap aortoplasty. J THORAC CARDIOVASC SURG 1985;89:235-41. 13. ZiemerG,JonasRA,PerrySB,FreedMD,CastanedaAR. Surgery for coarctation of the aorta in the neonate. Circulation 1986;74(Pt 2):125-31. 14. Van Son JAM, Daniels 0, Vincent JG, van Lier HJJ, Lacquet LK. Appraisal of resection and end-to-end anastomosis for repair of coarctation of the aorta in infancy: preference for resection. Ann Thorac Surg 1989;48:496-502. 15. Van Son JAM, Skotnicki SH, van Asten WNJC, Daniels 0, van Lier HJJ, Lacquet LK. Quantitative assessment of coarctation in infancy by Doppler spectral analysis. Am J Cardiol 1989;63: 1282-5. 16. Second task force on blood pressure control in children. Report. Bethesda,Maryland: National Institutesof Health, 1987. 17. Wijn PFF, van der Sar P, Gootzen THJM, Tilmans MHJ, Skotnicki SH. Value of the spectral broadening index in continuous wave Doppler measurements. Med Bioi Eng Comput 1987;25:377-85. 18. Fronek A, Coel M, Bernstein EF. Quantitative ultrasonographic studies of lower extremity flow velocities in health and disease. Circulation 1976;53:957-60. 19. Gosling RG, Dunbar G, King DH, et al. The quantitative

The Journal of Thoracic and Cardiovascular Surgery

analysisof occlusive peripheral arterial diseaseby a non-intrusive ultrasonic technique. Angiology 1971;22:52-5. 20. Baird RN, Bird DR, Clifford PC, Lusby RJ, Skidmore R, WoodcockJP. Upstream stenosis: its diagnosisby Doppler signals from the femoral artery. Arch Surg 1980;115: 131622. 21. Ward AS, Martin TP. Some aspects of ultrasound in the diagnosis and assessment of aortoiliac disease. Am J Surg 1980;140:260-5. 22. DeMorais 0, Johnston KW. Assessmentof aorto-iliac disease by non-invasive quantitative Doppler waveformanalysis. Br J Surg 1981;68:789-92. 23. Planiol T, Pourcelot L. Doppler effect study of the carotid circulation. In: de Vlieger M, White ON, McCready VR, eds. Proceedingsof the second world congress on ultrasonics in medicine (Rotterdam). Amsterdam: Excerpta Medica, 1974:104-11. 24. Amato JJ, Rheinlander HF, Cleveland RJ. A method of enlarging the distal transversearch in infants with hypoplasia and coarctation of the aorta. Ann Thorac Surg 1977; 23:261-3. 25. Vincent JG, Daniels 0, van Oort A, Lacquet LK. Hypoplastic aortic arch with aortic coarctation: surgical correction. J THORAC CARDIOVASC SURG 1985;89:465-8. 26. Schultz RD, Hokanson DE, Strandness DE Jr. Pressureflow and stress-strain measurementsof normaland diseased aortoiliac segments. Surg Gynecol Obstet 1967;124:126776. 27. Freed MD, RocchiniA, Rosenthal A, Nadas AS, Castaneda AR. Exercise-induced hypertension after surgical repair of coarctation of the aorta. Am J Cardiol 1979;43: 253-8. 28. Connor TM, Baker WP. A comparison of Coarctation resection and patch angioplasty using postexercise blood pressure measurements. Circulation 1981 ;64:567-72. 29. Smith RT Jr, Sade RM, Riopel DA, Taylor AB, Crawford FA Jr, Hohn AR. Stress testing for comparisonof synthetic patch aortoplasty with resection and end to end anastomosis for repair of coarctation in childhood. J Am Coll Cardiol 1984;4:765-70. 30. Fripp RR, Whitman V, Werner JC, Nicholas GG, Waldhausen JA. Bloodpressure responseto exercise in children following the subclavian flap procedure for coarctation of the aorta. J THORAC CARDIOVASC SURG 1983;85:682-5. 31. Waldman JD, Lamberti JJ, Goodman AH, et al. Coarctation in the first year of life: patterns of postoperative effect. J THORAC CARDIOVASC SURG 1983;86:9-17. 32. Patterson GC, Whelan RF. Reactive hyperaemia in the human forearm. Clin Sci 1955;14:197-211. 33. Folkow B, Grimby G, Thulesius O. Adaptive structural changes of the vascular walls in hypertension and their relation to the control of the peripheral resistance. Acta Physiol Scand 1958;44:255-72. 34. Fronek A, Johansen K, Dilley RB, Bernstein EF. Ultrasonographically monitored postocclusive reactive hyperemia in the diagnosisof peripheral arterial occlusive disease. Circulation 1973;48: 149-52.

Volume 100 Number 6 December 1990

35. Brener BJ, Raines JK, Darling RC, Austen WG. Measurement of systolic femoral arterial pressure during reactive hyperemia: an estimate of aortoiliac disease. Circulation 1974;50(3 Pt 2):1259-67. 36. Sinoway LI, Musch TI, Minotti JR, Zelis R. Enhanced maximal metabolic vasodilatation in the dominant forearms of tennis players. J Appl Physiol 1986;61:673-8. 37. Barcroft H. The mechanism of vasodilatation in the limbs during and after arrest of the circulation. Angiology 1972;23:595-9. 38. Pierce WS, Waldhausen JA, Berman W Jr, Whitman V. Late results of the subclavian flap procedure in infants with coarctation of the thoracic aorta. Circulation 1978;58(Pt 2):178-82. 39. Kamau P, Miles V, Toews W, et al. Surgical repair of coarctation of the aorta in infants less than six months of age: including the question of pulmonary artery banding. J THORAC CARDIOVASC SURG 1981;81:171-9. 40. Elliott RB, Starling MB, Neutze JM. Medical manipulation of the ductus arteriosus. Lancet 1975;1:140-2. 41. Olley PM, Coceani F, Bodach E. E-type postaglandins: a new emergency therapy for certain cyanotic congenital heart malformations. Circulation 1976;53:728-31. 42. Neutze JM, Starling MB, Elliott RB, Barratt-Boyes BG. Palliation of cyanotic congenital heart disease in infancy with Estype prostaglandins. Circulation 1977;55:238-41. 43. Leoni F, Huhta JC, Douglas J, et al. Effect of prostaglandin on early surgical mortality in obstructive lesions of the systemic circulation. Br Heart J 1984;52:654-9. 44. Ho SY, Anderson RH. Coarctation, tubular hypoplasia, and the ductus arteriosus: histological study of 35 specimens. Br Heart J 1979;41:268-74. 45. Elzenga NJ, Gittenberger-de Groot AC. Localised coarctation of the aorta: an age dependent spectrum. Br Heart J 1983;49:317-23. 46. Gillman RG, Burton AC. Constriction of the neonatal aorta by raised oxygen tension. Circ Res 1966;19:755-65. 47. Van Son JAM, van Asten WNJC, van Lier HJJ, et al. Detrimental sequelae on the hemodynamics of the upper left limb after subclavian flap angioplasty in infancy. Circulation 1990;81:996-1004. 48. Currarino G, Engle MA. The effects of ligation of the subclavian artery on the bones and soft tissues of the arms. J Pediatr 1965;67:808-11. 49. Skovranek J, Goetzova J, Samanek M. Changes in muscle blood flow and development of the arm following the Blalock-Taussig anastomosis. Cardiology 1975;61:131-7. 50. Lodge FA, Lamberti JJ, Goodman AH, et al. Vascular

Coarctation of aorta

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62. 63.

829

consequences of subclavian artery transection for the treatment of congenital heart disease. J THORAC CARDIOVASC SURG 1983;86:18-23. Hehrlein FW, Mulch J, Rautenburg HW, Schlepper M, Scheid HH. Incidence and pathogenesis of late aneurysms after patch graft aortoplasty for coarctation. J THORAC CARDIOVASC SURG 1986;92:226-30. Elzenga NJ, Gittenberger-de Groot AC. Coarctation and related aortic arch anomalies in hypoplastic left heart syndrome. Int J Cardiol 1985;8:379-89. Elzenga NJ, Gittenberger-de Groot AC, OppenheimerDekker A. Coarctation and other obstructive aortic arch anomalies: their relationship to the ductus arteriosus. Int J Cardiol 1986;13:289-308. Brom AG. Narrowing of the aortic isthmus and enlargement of the mind. J THORAC CARDIOVASC SURG 1965; 50:166-80. Arensman FW, Sievers HH, Lange P, et al. Assessment of coronary and aortic anastomoses after anatomic correction of transposition of the great arteries. J THORAC CARDIavASC SURG 1985;90:597-604. Harlan JL, Doty DB, Brandt B III, Ehrenhaft JL. Coarctation of the aorta in infants. J THORAC CARDIOVASC SURG 1984;88:1012-9. KorferR, Meyer H, KleikampG, Bircks W. Early and late results after resection and end-to-end anastomosis of coarctation of the thoracic aorta in early infancy. J THORAC CARDIOVASC SURG 1985;89:616-22. Lansman S, Shapiro AJ, Schiller MS, et al. Extended aortic arch anastomosis for repair of coarctation in infancy. Circulation 1986;74(Pt 2):137-41. Vouhe PR, Trinquet F, Lecompte Y, et al. Aortic coarctation with hypoplastic aortic arch: results of extended endto-end aortic arch anastomosis. J THORAC CARDIOVASC SURG 1988;96:557-63. Deleon SY, Idriss FS, Ilbawi MN, Tin N, Berry T. Transmediastinal repair of complex coarctation and interrupted aortic arch. J THORAC CARDIOVASC SURG 1981; 82:98-102. Ungerieider RM, Ebert PA. Indications and techniques for midline approach to aortic coarctation in infants and children. Ann Thorac Surg 1987;44:517-22. Levin PM, Rich NM, Hutton JE Jr. Collateral circulation in arterial injuries. Arch Surg 1971;102:392-9. Hogan RD, Hirschmann L. Arteriolar proliferation in the rat cremaster muscle as a long-term autoregulatory response to reduced perfusion. Microvasc Res 1984;27: 290-6.