Might Free Arterial Grafts Fail Due to Spasm? Giorgio Massa, MD, Sune Johansson, MD, Per-Ola Kimblad, MD, Trygve Sjoberg, PhD, and Stig Steen, MD, PhD Department of Cardiothoracic Surgery, University Hospital, Lund, Sweden
The rat femoral artery was used as a free graft and was studied after 2, 7,14,30, and 60 days. The patency of the grafts was 100% (2days, n = 61, 78% (7 days, n = 9), 63% (14days, n = 8), 33% (30 days, n = 121, and 18% (60days, n = 11). Histology showed an intimal thickening after 14 days and the media, which in the controls consisted of eight to ten layers of myocytes, was reduced to six to eight cell layers. During the first 2 weeks the graft segments had an impaired contraction when exposed to Krebs solution with 124 mmoYL K+, whereas after 1 month and later the graft segments approached the controls or had even higher contractile force. The throm-
boxane mimic U-46619elicited full contractile force at all times whereas the potency was significantly lower during the first 14 days. Noradrenaline was unable to induce contraction in the graft segments during the first 14 days, but at 30 and 60 days it had regained full contractile force and was significantly more potent (approximately 60 times) in the graft segments compared with the controls. This study suggests that intimal thickening and hypercontractility might be a problem in free muscular arterial grafts.
A
Material and Methods Operation
rm veins used as aortocoronary bypass grafts have a high failure rate compared with saphenous vein grafts [l].The internal mammary artery used as an in situ graft has a much higher late patency rate than a comparable saphenous vein graft [2, 31. Inspired by these facts several surgeons have started to use free arterial grafts in the revascularization of the ischemic heart (4-1 11 hoping that they will be as good as the in situ mammary artery graft or at least better than the saphenous vein graft. However, some factors might play a potentially negative role when comparing a free versus an in situ arterial graft. The free graft is totally denervated, and this might influence the number or the function of the important postjunctional contraction-mediating a-adrenergic receptors. Another factor is that the circulation of blood and lymph in the vasa vasorum is totally disrupted in the free graft, which might cause medial necrosis and intimal thickening [12-141. Will such a changed intima retain its full capacity to produce relaxing or antithrombotic factors [15, 16]? In a postoperative angiographic study an “unexpected finding” was that three of ten free right gastroepiploic artery grafts had major vasospasm [17]. In other studies most of the free radial artery aortocoronary bypass grafts failed within 1 year 15, 131. Might spasm be a problem with free arterial grafts? Spasm is known to occur in mammary and gastroepiploic arteries used as in situ grafts [18-231, but this is not at present generally recognized to be of clinical significance.The present study was designed with the aim of elucidating some contractile properties of a free arterial graft. Accepted for publication Sep 10, 1990. Address reprint requests to Dr Steen, Department of Cardiothoracic Surgery, University Hospital of Lund, 5-221 85 Lund, Sweden.
0 1991 by The Society of
Thoracic Surgeons
(Ann Thorac Surg 1991;51:94-101)
Forty-four male Sprague-Dawleyrats weighing 300 g were anesthetized with 75 mg chloral hydrate intraperitoneally. The left femoral artery was exposed and gently dissected free from the iliofemoral vein. The superficial circumflex iliac artery was ligated and cut and a 10- to 12-mm segment of the femoral artery could thus be extirpated between two arterial clamps placed at the inguinal ligament and at the deep femoral artery. The free arterial graft was kept in chilled (4°C) Ringer’s solution for 30 minutes and then reimplanted into the femoral artery keeping the original orientation. A Zeiss OPMI 6-S operating microscope was used. Eight interrupted stitches of 10-0 polyglycolic acid (Vicryl; Ethicon) were used at each anastomosis. The rats were randomized into five groups. After 2, 7, 14, 30, and 60 days the rats were again anesthetized and the femoral artery was dissected free and cut just distal to the distal anastomosis. If blood ejected from the proximal cut end the graft was deemed patent. If no blood was coming out of the graft, the graft was dissected out and cut into five ring segments, which were investigated under the operative microscope to be sure that the lumen really was occluded. The patent graft was cut into five ring segments (each about 2 mm in length) and immediately brought into tissue baths. A 2-mm ring segment was also taken from the right femoral artery and used as a control. All surgical procedures were done by the same surgeon. The study was approved by the Ethics Committee of the University of Lund. The animals were treated in compliance with the ”Guide for the Care and Use of Laboratory Animals” published by the National Institutes of Health (NIH publication No. 85-23, revised 1985). 0003-4975/91/$3.50
MASSAETAL CONTRACTILE CAPACITY OF FREE ARTERIAL GRAFTS
Ann Thorac Surg 1991;51:94101
95
Experimental Procedure Repeated contractions were induced by a preheated (37°C) and bubbled (95% O2 + 5% CO,) K+-rich Krebs solution (for composition, see "Buffer Solutions" section) until the amplitude of two subsequent contractions varied less than 10%. The maximum response was used as an internal standard to which subsequent agonist-induced contractions were related. First noradrenaline and then, after repeated washes resulting in a return to the basal tension, the thromboxane analogue U-46619 were added cumulatively to all vessel segments.
Contro Control
B B-
Buffer Solutions The Krebs buffer solution contained (in mmoYL): NaCl, 119; NaHCO,, 15; KCl, 4.6; NaH,PO,, 1.2; MgCl,, 1.2; CaCl,, 1.5;and glucose, 11. The K+-rich solution was prepared by substituting all NaCl in the Krebs solution with equimolar amounts of KCl.
Drugs
> \ Y
Dilutions of (-)-noradrenaline hydrochloride (Sigma) were made in NaCl containing 1mmoYL ascorbic acid. U-46619 (9,1l-dideoxy-lla,9a-epoxymethano-prostaglandin F2a; Upjohn) was prepared to a stock solution in absolute ethanol (5mg/mL) and stored at -20°C. Fresh dilutions of U-46619were made with phosphate buffer at neutral pH just before use. The concentrations are given as final molar concentrations in the tissue baths. I
Fig 1 . The graft implanted in the left femoral artery was excised and divided into five segments. These and a control segment obtained from the contralateral side were mounted between two L-shaped metal holders (a) in organ baths @). One of the metal holders is connected to a force displacement transducer (c).
Recording of Mechanical Activity The ring segments were suspended between two L-shaped metal holders (0.17 mm in diameter) arranged in parallel. One of the metal holders was attached to a Grass FT 03 C force-displacement transducer (Fig 1).The output from the transducer was amplified by and displayed on a Grass polygraph. The other holder was fixed to a movable unit. By a micrometer screw the movable unit could be displaced and a resting tension applied to the vessel. The vessel segments were submerged in temperature controlled (37°C) tissue baths containing 5 mL Krebs solution (for composition, see "Buffer Solutions" section). The baths were bubbled with oxygen containing 5% carbon dioxide, giving a pH of approximately 7.4. The vessels were repeatedly stretched until a basal tension of about 6 mN was reached. In separate experiments in healthy 2-mm ring segments from rat femoral arteries it was ascertained that a maximal contractile response was obtained at this basal tension. They were allowed to equilibrate in the baths for 1 to 2 hours before the experiments started. Six vessel segments could be investigated simultaneously.
Analysis of Data The maximum contraction (Emax)produced by each agonist was determined. The contraction was regarded as maximum when two subsequent contractions gave responses of the same amplitude or when the subsequent concentration produced a decreased contraction amplitude. The pEC,, value, ie, the negative logarithm of the EC,, value, was determined graphically from each curve as the concentration giving half-maximal contraction. The pEC,, value was not calculated if the maximum contraction obtained with an agonist was less than 10% of the K+-induced contraction. Wilcoxon's test for unpaired data was used for statistical comparisons.
His tology After the organ bath experiments the specimens were fixed in 4% phosphate-buffered (pH = 7.0) formalin-saline solution. Then they were sectioned and stained with hematoxylin-eosin and van Gieson's stain for elastin.
Results
Patency of the Grafts All grafts were patent after 2 days; thereafter the patency decreased to 18% after 60 days (Table 1).
Histological Examination After 2 and 7 days the intima was not changed as deemed from light microscopic studies with x 400 magnification (Fig 2A-C). After 14,30, and 60 days an intimal thickening
96
Ann Thorac Surg
MASSAETAL CONTRACTILE CAPACITY OF FREE ARTERIAL GRAFTS
1991;51:94101
Table 1. Patency of the Free Arterial Grafts
and 60-day groups were like fibrous strings and no distinct cell layers could be identified.
~~~~~
Days 2 7 14 30 60
No. of Grafts
No. of Patent Grafts
Patency
6 9 8 12 11
6 7 5 4 2
100 78 63 33 18
("/.I
could be seen (Fig 2D-F). The media, which in the controls consisted of eight to ten cell layers of myocytes, was reduced to six to eight cell layers (Fig 2E-F), whereas the adventitia had increased in thickness (Fig 2B.F). The nonpatent grafts at 7 and 14 days were occluded by thrombotic material. The nonpatent grafts in the 30-day
Responses to Potassium-Rich (124 rnmollL) Krebs S o h tion Two days after the grafting procedure the contractile response to potassium in the graft segments was only about one third of the control. Segment 4, which had the highest mean value (2.53 5 0.85 mN), was significantly lower ( p < 0.05) than the control (6.65 2 0.97 mN). On days 7 and 14 the response to potassium in the graft segments had increased to about 50% compared with their controls. However, no significant difference was found between the segment producing the highest contractile response and the control. After 30 days the graft segments approached the control and after 60 days some graft segments elicited even higher contractile force (Fig 3). During the first 2 weeks the segments adjacent to the
Table 2. Drug Effect on Patent Graft Segments and Controls Noradrenaline Day 2 (n = 6)
7 ( n = 7)
14 (n = 5)
30 (n = 4)
60 (n
=
2)
U-46619
Segment
Emax
PECM
Z
Emax
1 2 3 4 5 Control 1 2 3 4 5 Control 1 2 3 4 5 Control 1 2 3 4 5 Control 1 2 3 4 5 Control
0 0 0 0 0
...
0 0 0 0 0 6 0 1 3 3 1 7 0 0 0 1 2 5 2 3 4 4 2 3 1 2 2 2 2 2
66 f 6 107 + 16 95 f 15 92 + 25 51 f 16 106 f 3 56 f 6 89 f 13 86 f 14 86 2 12 80 8 137 2 21
98
f6
0 6+4 10 f 4 20 f 11 3 f 3 96 f 17 0 0 0 11 + 11 926 102 f 10 24 f 14 89 2 30 103 f 8 103 2 13 47 2 23 124 f 2 0; 52 108; 92 118; 58 147; 102 150; 96 150; 101
... ... ...
... 4.6
f 0.5
... 8.5 7.4 f 0.6 6.5 f 0.1 7.0 4.9 ? 0.4
... ... ...
5.5 6.8; 6.1 6.0 +- 0.6 6.5; 6.4 6.7 f 0.4 5.6 f 1.0 5.8 2 0.8 6.6; 6.7 4.8 f 0.6 7.3 6.6; 6.6 6.1; 7.3 6.7; 5.8 6.5; 6.2 4.8; 5.5
PEC,
z
6.1; 7.0 6.5 f 0.2 6.2 f 0.1 6.4 f 0.3 6.5 f 0.3 8.5 f 0.6 6.6 f 0.1 6.9 f 0.1 7.7 f 0.5 7.4 f 0.3 6.8 f 0.1 8.4 f 0.2
2 6 6 5 4 5 7 7 7 7 7 7
0
...
...
72 f 27 121 f 12 125 f 17 110 f 26 146 f 10 100 f 47 182 f 65 146 f 35 142 f 28 96 f 20 148 f 20 0; 135 142; 124 218; 138 200; 148 180; 119 217; 108
7.2 f 0.1 7.7 f 0.1 7.5 f 0.2 7.4 2 0.3 9.3 f 0.3 8.2 f 0.7 8.9 0.6 8.2 f 0.3 8.4 2 0.8 7.7 f 0.7 9.1 f 0.7 8.3 7.8; 8.0 7.7; 7.7 7.7; 7.9 7.6; 7.8 7.8; 7.7
3 4 4 4 5 3 3 4 4 4 4 1 2 2 2 2 2
*
~~
Em,,
n = number of rats with patent grafts; contraction in percent of the maximum K+ (124 mmol/L)-induced contraction; z = number on which pEC,, is based (see Material and Methods). logarithm of concentration giving 50% of E,,,; =
pECM = negative
Ann Thorac Surg 1991;51:94-101
97
MASSAETAL CONTRACTILE CAPACITY OF FREE ARTERIAL GRAFTS
-u P
1
P
c
E
9)
5
98
Day
MASSAETAL CONTRACTILE CAPACITY OF FREE ARTERIAL GRAFTS
Potassium
8
Ann Thorac Surg 1991;51:94-101
U-46619
Noradrenaline
T
1
Z
E .-C K .-0 -Id
u
0
c L .
c
0 0
OAOVOO
Segment
A
8
1
Potassium T
-11
-10
-9
-a
-7
-6
-5
Log agonist concentration (M)
Log agonist concentration (M)
Noradrena I ine
U-46619
1
Z
E .-C
.-0 e0 -Id
0
100
a0
2
3
4
n
x
50
2
sx
+
v
0 O A O V O O
B
Segment
Day l 48
0 0
2
6
C
.w
150
c
Potassium
Log agonist concentration (M)
-a
-9
-7 -6
-5
0
-4 -3
Log agonist concentration (M)
U-46619
Noradrenaline
1
Z
..
E
.-C C .-0
100
a0 2 3
-Id
0
e +J
n
K
0 0
50
s
+
i i v
-9
O A O V O O
C
Segment
Log agonist Concentration (M)
-a
-7 -6 -5
-4
-3 Log agonist concentration (M)
0
Fig 3. The contractile responses obtained in the free arterial graft and the control segments to K+ (124 mmollL), expressed in mN (left), and for U46619 (middle) and noradrenaline (right), expressed as the percentage of the K+-induced contraction. Each point in the curves indicates mean standard error of the mean. For number of segments, see Table 2. (0) segment I , (A) segment 2, (0)segment 3, (V) segment 4 , (0)segment 5, and (0)control.
*
Ann Thorac Surg 1991;51:94-101
U-46619
Day 30
Norad ren a li ne
8
150
z € K
.-K
0 .+ 0 e +
100
3
x
50
2
-11
Segment
D
Day 60 8 €
-a
-7 -6 -5 Log agonist concentration (M)
-10
-9
-a
-9
-Id
-6
-5
-4
Log agonist concentration (M)
U-46619
Noradrenaline
Potassium
150
100
a0 g 3
4
n
x
K 0
E
0 0
2
6
50
2 0
+
-3
-Id
0
x
0 -7
K
.-0 0 e
s
W
O A O V O O
.-c
$
n
4
0
z
sz a
6
c
0 0
99
MASSAETAL CONTRACTILE CAPACITY OF FREE ARTERIAL GRAFTS
0,
x
+
v
3 AAOOVVO O .
Segment
-11
-10
-9
-8
-7
-5 Log agonist concentration (M)
0
-6
Log agonist concentration (M)
Fig 3. Continued.
anastomoses (segments 1 and 5) had lower contractile capacity than the other graft segments (see Fig 3).
Responses to the Thromboxane Analogue U-46619 U-46619 elicited contraction in all graft segments. When comparing the graft segment from each rat having the highest Em,, value with the control segment, no significant differences were found at any time. When the corresponding pEC,, values were compared on day 2 they were significantly lower ( p < 0.01) in the graft segments (pEC,, = 6.6 f 0.3) than in the controls (pEC,, = 8.5 0.6). The same relationship (p < 0.05) was found on day 14 (pEC,, = 7.5 f 0.1 and 9.3 f 0.3, respectively). On day 7 and day 2 3 0 (ie, analyzing day 30 and 60 as one group to get a large enough sample for statistical evaluation) no significant differences were found.
*
Responses to Noradrenaline Noradrenaline induced no or only weak contractions in the graft segments during the first 14 days, but at 30 and
60 days it elicited full contractile force in segments 2 through 4, ie, no significant difference existed between the Em,, values in these graft segments as compared with the controls (see Fig 3). After 30 days noradrenaline was at most 80 times more potent in the graft segments as compared with the controls, and after 60 days it was at most 126 times more potent in the graft segments. Putting together the segment from each rat with the highest contractile capacity on day 30 and 60, their pEC,, values (6.7 ? 0.3) were significantlygreater ( p < 0.01) than in the controls (4.9 2 0.4). The relation between the EC, concentrations was thus: 10-4.9:10-6.7 = 10-4.9 - 10+1.8= 63; ie, noradrenaline was 63 times more potent in the graft segments than in the control segments.
Comment The potassium-rich (124 mmol/L) Krebs solution used in the present study creates a depolarization of the smooth muscle cell membrane; this elicits a constant activation of
100
Ann Thorac Surg
MASSA ET AL CONTRACTILE CAPACITY OF FREE ARTERIAL GRAFTS
potential-operated calcium channels with subsequent entry of calcium into the myocytes, thus eliciting a contraction. When a steady state is reached, a stable, constant vasocontraction will be present. We used the amplitude of this contraction as an internal standard for each vessel segment studied, and expressed the responses obtained with the other drugs as a percentage of this maximal potassium-induced contraction. This is a generally accepted way of presenting vasoconstrictive effects, having the advantage that the responses of different drugs to different vessel segments can be compared independently of the sizes or muscle masses. Radial artery grafts for aortocoronary bypass were used with early favorable results [4]. However, most of these grafts occluded within several months [5, 131. Chiu [14] studied the histology of free femoral arterial and external jugular vein grafts in dogs after 60 days. Half of the grafts were wrapped with plastic sheets to interfere with the regeneration of vasa vasorum. As in the present study (see Fig 2F) he found mild to moderate subintimal hyperplasia in a number of the free femoral arterial grafts, particularly when the grafts were wrapped; the vein grafts had better patency and did not develop intimal hyperplasia except in one graft that was wrapped with the plastic sheet. Extrapolating these results to a clinical situation he suggested that free thick-walled, muscular arterial grafts with the vasa vasorum disrupted at both ends cannot regenerate readily from the adjacent tissue, as is the case in the aortocoronary bypass position, and therefore they are more vulnerable to subintimal hyperplasia and occlusion than thin-walled vein grafts. If his hypothesis is correct, the internal mammary artery, which is a thinwalled artery, should be better as a free aortocoronary bypass graft than the radial artery, which has a thicker wall. The observations of Loop and associates [7] suggest that this might be true, although the number of grafts studied was very small. Our morphological results are in accordance with those of Chiu [14]. Furthermore, our findings indicate that hypercontractility, ie, increased sensitivity to cathecholamines (spasm), might be a problem in free arterial grafts because noradrenaline had full contractile force and was much more potent in the graft segments than in the controls after 30 and 60 days. One explanation of this increased potency of noradrenaline could be an upregulation of adrenergic a-adrenoceptors as a consequence of the total denervation of the free arterial graft. Another explanation might be that the uptake of noradrenaline into sympathetic nerve endings, a physiological event, was abolished due to the lack of these nerves, and all the noradrenaline given was thus available to stimulate the postjunctional a-adrenergic receptors. During the first 14 days the a-adrenergic receptors seemed to be absent or out of function as noradrenaline had no contractile effect, whereas U-46619 had full contractile capacity in the same graft segments. After 1 month there was a hyperfunction of the a-adrenergic receptors because noradrenaline was significantly more potent in the grafts than in the controls. This might have contributed to the disappointing patency after 30 and 60 days, ie, circulating catecholamines, which
1991:51:94101
under normal conditions should be unable to cause pathological vascular constriction, might now have induced spasm in the grafts resulting in occlusion. The intimal thickening (see Fig 2D-F) that is generally seen in free muscular arterial grafts [14] might be another explanation of the poor patency. To our knowledge no study has been done to investigate if such a thickened intima is able to or has retained its normal capacity to produce relaxing or antithrombogenic factors. It has been shown that a regenerated endothelium has a reduced function of endothelium-derived relaxing factors [24]. To test these hypotheses a randomized study with and without a potent a-adrenergic antagonist should be done to asertain if an effective blockade of the a-adrenergic receptors would increase the patency of different free arterial grafts. Furthermore, corresponding studies should be done with, for example, antithrombotic or cytostatic drugs with the aim of preventing occlusion. In conclusion, we think that a skeptical attitude to a liberal use of free arterial grafts in clinical surgery is warranted until more studies on this topic have been published. We thank Prof Jan Kugelberg, our chief, who by his enthusiastic and economic support made this study possible.
References 1. Stoney WS, Alford WC Jr, Burrus GR, Glassford DM Jr, Petracek MR, Thomas CS Jr. The fate of arm veins used for aorto-coronary bypass grafts. J Thorac Cardiovasc Surg 1984; 88:522-6. 2. Loop FD, Lytle BW, Cosgrove DM, et al. Influence of the internal-mammary-artery graft on 10-year survival and other cardiac events. N Engl J Med 1986;314:1-6. 3. Johnson DW, Brenowitz JB, Kayser KL. Factors influencing long-term (10-year to 15-year) survival after a successful coronary artery bypass operation. Ann Thorac Surg 1989;48: 19-25. 4. Carpentier A, Guermonprez JL, Deloche A, Frechette C, DuBost C. The aorta-to-coronary radial artery bypass graft. Ann Thorac Surg 1973;16:111-21. 5. Fisk RL, Brooks CH, Callaghan JC, Dvorkin J. Experience with the radial artery graft for coronary artery bypass. Ann Thorac Surg 1976;21:513-8. 6. Barner HB. The internal mammary artery as a free graft. J Thorac Cardiovasc Surg 1973;66:219-21. 7. Loop FD, Lytle BW, Cosgrove DM, et al. Free (aortacoronary) internal mammary artery graft. Late results. J Thorac Cardiovasc Surg 1986;92:827-31. 8. Lytle BW, Cosgrove DM, Ratliff NB, Loop FD. Coronary artery bypass grafting with the right gastroepiploic artery. J Thorac Cardiovasc Surg 1989;97:826-31. 9. Tanimoto Y, Matsuda Y, Masuda T, et al. Multiple free (aorta-coronary) gastroepiploic artery grafting. Ann Thorac Surg 1990;49:479-80. 10. Puig LB, Ciongolli W, Cividanes GVL, et al. Inferior epigastric artery as a free graft for myocardial revascularization. J Thorac Cardiovasc Surg 1990;99:251-5. 11. Vincent JG, van Son JAM, Skotnicki SH. Inferior epigastric artery as a conduit in myocardial revascularization: the alternative free arterial graft. Ann Thorac Surg 1990;49:323-5.
Ann Thorac Surg 1991;51:94101
12. Fonkalsrud EW, Sanchez M, Zerubavel R, Lassaletta L, Smeesters C , Mahoney A. Arterial endothelial changes after ischemia and perfusion. Surg Gynecol Obstet 1976;142: 715-21. 13. Curtis JJ, Stoney WS, Alford WC, Burrus G, Thomas C. Intimal hyperplasia. A cause of radial artery aorto-coronary bypass graft failure. Ann Thorac Surg 1975;20:628-35. 14. Chiu C-J. Why do radial artery grafts for aortocoronary bypass fail? A reappraisal. Ann Thorac Surg 1976;22:52&3. 15. Angelini GD, Christie MI, Bryan AJ, Lewis MJ. Surgical preparation impairs release of endothelium-derived relaxing factor from human saphenous vein. Ann Thorac Surg 1989; 48:417-20. 16. Johns RA, Peach MJ, Flanaagan T, Kron IL. Probing of the canine mammary artery damages endothelium and impairs vasodilation resulting from prostacyclin and endotheliumderived relaxing factor. J Thorac Cardiovasc Surg 1989;97 252-8. 17. Mills NL, Everson CT. Right gastroepiploic artery: a third arterial conduit for coronary artery bypass. Ann Thorac Surg 1989;47706-11.
MASSA ET AL CONTRACTILE CAPACITY OF FREE ARTERIAL GRAFE
101
18. Sarabu MR, McClung JA, Fass A, Reed GE. Early postoperative spasm in left internal mammary artery bypass grafts. Ann Thorac Surg 1987;44:199-200. 19. Blanche C, Chaux A. Spasm in mammary artery grafts. Ann Thorac Surg 1988;45:586. 20. Mills NL, Bringaze WL 111. Preparation of the internal mammary artery graft. Which is the best method? J Thorac Cardiovasc Surg 1989;98:73-9. 21. Von Segesser L, Simonet F, Meier B, Finci L, Faidutti B. Inadequate flow after internal mammary-coronary artery anastomoses. J Thorac Cardiovasc Surg 1987;35:3524. 22. Von Segesser L, Lehmann K, Turina M. Deleterious effect of shock in internal mammary artery anastomoses. Ann Thorac Surg 1989;47575-9. 23. Suma H. Spasm of the gastroepiploic artery graft. Ann Thorac Surg 1990;49:168-9. 24. Shimokawa H, Aarhus LL, Vanhoutte PM. Porcine coronary arteries with regenerated endothelium have a reduced endothelium-dependent responsiveness to aggregating platelets and serotonin. Circ Res 1987;61:25670.
The rat femoral artery was used as a free graft and was studied after 2, 7,14,30, and 60 days. The patency of the grafts was 100% (2days, n = 61, 78% (7 days, n = 9), 63% (14days, n = 8), 33% (30 days, n = 121, and 18% (60days, n = 11). Histology showed an intimal thickening after 14 days and the media, which in the controls consisted of eight to ten layers of myocytes, was reduced to six to eight cell layers. During the first 2 weeks the graft segments had an impaired contraction when exposed to Krebs solution with 124 mmoYL K+, whereas after 1 month and later the graft segments approached the controls or had even higher contractile force. The throm-
boxane mimic U-46619elicited full contractile force at all times whereas the potency was significantly lower during the first 14 days. Noradrenaline was unable to induce contraction in the graft segments during the first 14 days, but at 30 and 60 days it had regained full contractile force and was significantly more potent (approximately 60 times) in the graft segments compared with the controls. This study suggests that intimal thickening and hypercontractility might be a problem in free muscular arterial grafts.
A
Material and Methods Operation
rm veins used as aortocoronary bypass grafts have a high failure rate compared with saphenous vein grafts [l].The internal mammary artery used as an in situ graft has a much higher late patency rate than a comparable saphenous vein graft [2, 31. Inspired by these facts several surgeons have started to use free arterial grafts in the revascularization of the ischemic heart (4-1 11 hoping that they will be as good as the in situ mammary artery graft or at least better than the saphenous vein graft. However, some factors might play a potentially negative role when comparing a free versus an in situ arterial graft. The free graft is totally denervated, and this might influence the number or the function of the important postjunctional contraction-mediating a-adrenergic receptors. Another factor is that the circulation of blood and lymph in the vasa vasorum is totally disrupted in the free graft, which might cause medial necrosis and intimal thickening [12-141. Will such a changed intima retain its full capacity to produce relaxing or antithrombotic factors [15, 16]? In a postoperative angiographic study an “unexpected finding” was that three of ten free right gastroepiploic artery grafts had major vasospasm [17]. In other studies most of the free radial artery aortocoronary bypass grafts failed within 1 year 15, 131. Might spasm be a problem with free arterial grafts? Spasm is known to occur in mammary and gastroepiploic arteries used as in situ grafts [18-231, but this is not at present generally recognized to be of clinical significance.The present study was designed with the aim of elucidating some contractile properties of a free arterial graft. Accepted for publication Sep 10, 1990. Address reprint requests to Dr Steen, Department of Cardiothoracic Surgery, University Hospital of Lund, 5-221 85 Lund, Sweden.
0 1991 by The Society of
Thoracic Surgeons
(Ann Thorac Surg 1991;51:94-101)
Forty-four male Sprague-Dawleyrats weighing 300 g were anesthetized with 75 mg chloral hydrate intraperitoneally. The left femoral artery was exposed and gently dissected free from the iliofemoral vein. The superficial circumflex iliac artery was ligated and cut and a 10- to 12-mm segment of the femoral artery could thus be extirpated between two arterial clamps placed at the inguinal ligament and at the deep femoral artery. The free arterial graft was kept in chilled (4°C) Ringer’s solution for 30 minutes and then reimplanted into the femoral artery keeping the original orientation. A Zeiss OPMI 6-S operating microscope was used. Eight interrupted stitches of 10-0 polyglycolic acid (Vicryl; Ethicon) were used at each anastomosis. The rats were randomized into five groups. After 2, 7, 14, 30, and 60 days the rats were again anesthetized and the femoral artery was dissected free and cut just distal to the distal anastomosis. If blood ejected from the proximal cut end the graft was deemed patent. If no blood was coming out of the graft, the graft was dissected out and cut into five ring segments, which were investigated under the operative microscope to be sure that the lumen really was occluded. The patent graft was cut into five ring segments (each about 2 mm in length) and immediately brought into tissue baths. A 2-mm ring segment was also taken from the right femoral artery and used as a control. All surgical procedures were done by the same surgeon. The study was approved by the Ethics Committee of the University of Lund. The animals were treated in compliance with the ”Guide for the Care and Use of Laboratory Animals” published by the National Institutes of Health (NIH publication No. 85-23, revised 1985). 0003-4975/91/$3.50
MASSAETAL CONTRACTILE CAPACITY OF FREE ARTERIAL GRAFTS
Ann Thorac Surg 1991;51:94101
95
Experimental Procedure Repeated contractions were induced by a preheated (37°C) and bubbled (95% O2 + 5% CO,) K+-rich Krebs solution (for composition, see "Buffer Solutions" section) until the amplitude of two subsequent contractions varied less than 10%. The maximum response was used as an internal standard to which subsequent agonist-induced contractions were related. First noradrenaline and then, after repeated washes resulting in a return to the basal tension, the thromboxane analogue U-46619 were added cumulatively to all vessel segments.
Contro Control
B B-
Buffer Solutions The Krebs buffer solution contained (in mmoYL): NaCl, 119; NaHCO,, 15; KCl, 4.6; NaH,PO,, 1.2; MgCl,, 1.2; CaCl,, 1.5;and glucose, 11. The K+-rich solution was prepared by substituting all NaCl in the Krebs solution with equimolar amounts of KCl.
Drugs
> \ Y
Dilutions of (-)-noradrenaline hydrochloride (Sigma) were made in NaCl containing 1mmoYL ascorbic acid. U-46619 (9,1l-dideoxy-lla,9a-epoxymethano-prostaglandin F2a; Upjohn) was prepared to a stock solution in absolute ethanol (5mg/mL) and stored at -20°C. Fresh dilutions of U-46619were made with phosphate buffer at neutral pH just before use. The concentrations are given as final molar concentrations in the tissue baths. I
Fig 1 . The graft implanted in the left femoral artery was excised and divided into five segments. These and a control segment obtained from the contralateral side were mounted between two L-shaped metal holders (a) in organ baths @). One of the metal holders is connected to a force displacement transducer (c).
Recording of Mechanical Activity The ring segments were suspended between two L-shaped metal holders (0.17 mm in diameter) arranged in parallel. One of the metal holders was attached to a Grass FT 03 C force-displacement transducer (Fig 1).The output from the transducer was amplified by and displayed on a Grass polygraph. The other holder was fixed to a movable unit. By a micrometer screw the movable unit could be displaced and a resting tension applied to the vessel. The vessel segments were submerged in temperature controlled (37°C) tissue baths containing 5 mL Krebs solution (for composition, see "Buffer Solutions" section). The baths were bubbled with oxygen containing 5% carbon dioxide, giving a pH of approximately 7.4. The vessels were repeatedly stretched until a basal tension of about 6 mN was reached. In separate experiments in healthy 2-mm ring segments from rat femoral arteries it was ascertained that a maximal contractile response was obtained at this basal tension. They were allowed to equilibrate in the baths for 1 to 2 hours before the experiments started. Six vessel segments could be investigated simultaneously.
Analysis of Data The maximum contraction (Emax)produced by each agonist was determined. The contraction was regarded as maximum when two subsequent contractions gave responses of the same amplitude or when the subsequent concentration produced a decreased contraction amplitude. The pEC,, value, ie, the negative logarithm of the EC,, value, was determined graphically from each curve as the concentration giving half-maximal contraction. The pEC,, value was not calculated if the maximum contraction obtained with an agonist was less than 10% of the K+-induced contraction. Wilcoxon's test for unpaired data was used for statistical comparisons.
His tology After the organ bath experiments the specimens were fixed in 4% phosphate-buffered (pH = 7.0) formalin-saline solution. Then they were sectioned and stained with hematoxylin-eosin and van Gieson's stain for elastin.
Results
Patency of the Grafts All grafts were patent after 2 days; thereafter the patency decreased to 18% after 60 days (Table 1).
Histological Examination After 2 and 7 days the intima was not changed as deemed from light microscopic studies with x 400 magnification (Fig 2A-C). After 14,30, and 60 days an intimal thickening
96
Ann Thorac Surg
MASSAETAL CONTRACTILE CAPACITY OF FREE ARTERIAL GRAFTS
1991;51:94101
Table 1. Patency of the Free Arterial Grafts
and 60-day groups were like fibrous strings and no distinct cell layers could be identified.
~~~~~
Days 2 7 14 30 60
No. of Grafts
No. of Patent Grafts
Patency
6 9 8 12 11
6 7 5 4 2
100 78 63 33 18
("/.I
could be seen (Fig 2D-F). The media, which in the controls consisted of eight to ten cell layers of myocytes, was reduced to six to eight cell layers (Fig 2E-F), whereas the adventitia had increased in thickness (Fig 2B.F). The nonpatent grafts at 7 and 14 days were occluded by thrombotic material. The nonpatent grafts in the 30-day
Responses to Potassium-Rich (124 rnmollL) Krebs S o h tion Two days after the grafting procedure the contractile response to potassium in the graft segments was only about one third of the control. Segment 4, which had the highest mean value (2.53 5 0.85 mN), was significantly lower ( p < 0.05) than the control (6.65 2 0.97 mN). On days 7 and 14 the response to potassium in the graft segments had increased to about 50% compared with their controls. However, no significant difference was found between the segment producing the highest contractile response and the control. After 30 days the graft segments approached the control and after 60 days some graft segments elicited even higher contractile force (Fig 3). During the first 2 weeks the segments adjacent to the
Table 2. Drug Effect on Patent Graft Segments and Controls Noradrenaline Day 2 (n = 6)
7 ( n = 7)
14 (n = 5)
30 (n = 4)
60 (n
=
2)
U-46619
Segment
Emax
PECM
Z
Emax
1 2 3 4 5 Control 1 2 3 4 5 Control 1 2 3 4 5 Control 1 2 3 4 5 Control 1 2 3 4 5 Control
0 0 0 0 0
...
0 0 0 0 0 6 0 1 3 3 1 7 0 0 0 1 2 5 2 3 4 4 2 3 1 2 2 2 2 2
66 f 6 107 + 16 95 f 15 92 + 25 51 f 16 106 f 3 56 f 6 89 f 13 86 f 14 86 2 12 80 8 137 2 21
98
f6
0 6+4 10 f 4 20 f 11 3 f 3 96 f 17 0 0 0 11 + 11 926 102 f 10 24 f 14 89 2 30 103 f 8 103 2 13 47 2 23 124 f 2 0; 52 108; 92 118; 58 147; 102 150; 96 150; 101
... ... ...
... 4.6
f 0.5
... 8.5 7.4 f 0.6 6.5 f 0.1 7.0 4.9 ? 0.4
... ... ...
5.5 6.8; 6.1 6.0 +- 0.6 6.5; 6.4 6.7 f 0.4 5.6 f 1.0 5.8 2 0.8 6.6; 6.7 4.8 f 0.6 7.3 6.6; 6.6 6.1; 7.3 6.7; 5.8 6.5; 6.2 4.8; 5.5
PEC,
z
6.1; 7.0 6.5 f 0.2 6.2 f 0.1 6.4 f 0.3 6.5 f 0.3 8.5 f 0.6 6.6 f 0.1 6.9 f 0.1 7.7 f 0.5 7.4 f 0.3 6.8 f 0.1 8.4 f 0.2
2 6 6 5 4 5 7 7 7 7 7 7
0
...
...
72 f 27 121 f 12 125 f 17 110 f 26 146 f 10 100 f 47 182 f 65 146 f 35 142 f 28 96 f 20 148 f 20 0; 135 142; 124 218; 138 200; 148 180; 119 217; 108
7.2 f 0.1 7.7 f 0.1 7.5 f 0.2 7.4 2 0.3 9.3 f 0.3 8.2 f 0.7 8.9 0.6 8.2 f 0.3 8.4 2 0.8 7.7 f 0.7 9.1 f 0.7 8.3 7.8; 8.0 7.7; 7.7 7.7; 7.9 7.6; 7.8 7.8; 7.7
3 4 4 4 5 3 3 4 4 4 4 1 2 2 2 2 2
*
~~
Em,,
n = number of rats with patent grafts; contraction in percent of the maximum K+ (124 mmol/L)-induced contraction; z = number on which pEC,, is based (see Material and Methods). logarithm of concentration giving 50% of E,,,; =
pECM = negative
Ann Thorac Surg 1991;51:94-101
97
MASSAETAL CONTRACTILE CAPACITY OF FREE ARTERIAL GRAFTS
-u P
1
P
c
E
9)
5
98
Day
MASSAETAL CONTRACTILE CAPACITY OF FREE ARTERIAL GRAFTS
Potassium
8
Ann Thorac Surg 1991;51:94-101
U-46619
Noradrenaline
T
1
Z
E .-C K .-0 -Id
u
0
c L .
c
0 0
OAOVOO
Segment
A
8
1
Potassium T
-11
-10
-9
-a
-7
-6
-5
Log agonist concentration (M)
Log agonist concentration (M)
Noradrena I ine
U-46619
1
Z
E .-C
.-0 e0 -Id
0
100
a0
2
3
4
n
x
50
2
sx
+
v
0 O A O V O O
B
Segment
Day l 48
0 0
2
6
C
.w
150
c
Potassium
Log agonist concentration (M)
-a
-9
-7 -6
-5
0
-4 -3
Log agonist concentration (M)
U-46619
Noradrenaline
1
Z
..
E
.-C C .-0
100
a0 2 3
-Id
0
e +J
n
K
0 0
50
s
+
i i v
-9
O A O V O O
C
Segment
Log agonist Concentration (M)
-a
-7 -6 -5
-4
-3 Log agonist concentration (M)
0
Fig 3. The contractile responses obtained in the free arterial graft and the control segments to K+ (124 mmollL), expressed in mN (left), and for U46619 (middle) and noradrenaline (right), expressed as the percentage of the K+-induced contraction. Each point in the curves indicates mean standard error of the mean. For number of segments, see Table 2. (0) segment I , (A) segment 2, (0)segment 3, (V) segment 4 , (0)segment 5, and (0)control.
*
Ann Thorac Surg 1991;51:94-101
U-46619
Day 30
Norad ren a li ne
8
150
z € K
.-K
0 .+ 0 e +
100
3
x
50
2
-11
Segment
D
Day 60 8 €
-a
-7 -6 -5 Log agonist concentration (M)
-10
-9
-a
-9
-Id
-6
-5
-4
Log agonist concentration (M)
U-46619
Noradrenaline
Potassium
150
100
a0 g 3
4
n
x
K 0
E
0 0
2
6
50
2 0
+
-3
-Id
0
x
0 -7
K
.-0 0 e
s
W
O A O V O O
.-c
$
n
4
0
z
sz a
6
c
0 0
99
MASSAETAL CONTRACTILE CAPACITY OF FREE ARTERIAL GRAFTS
0,
x
+
v
3 AAOOVVO O .
Segment
-11
-10
-9
-8
-7
-5 Log agonist concentration (M)
0
-6
Log agonist concentration (M)
Fig 3. Continued.
anastomoses (segments 1 and 5) had lower contractile capacity than the other graft segments (see Fig 3).
Responses to the Thromboxane Analogue U-46619 U-46619 elicited contraction in all graft segments. When comparing the graft segment from each rat having the highest Em,, value with the control segment, no significant differences were found at any time. When the corresponding pEC,, values were compared on day 2 they were significantly lower ( p < 0.01) in the graft segments (pEC,, = 6.6 f 0.3) than in the controls (pEC,, = 8.5 0.6). The same relationship (p < 0.05) was found on day 14 (pEC,, = 7.5 f 0.1 and 9.3 f 0.3, respectively). On day 7 and day 2 3 0 (ie, analyzing day 30 and 60 as one group to get a large enough sample for statistical evaluation) no significant differences were found.
*
Responses to Noradrenaline Noradrenaline induced no or only weak contractions in the graft segments during the first 14 days, but at 30 and
60 days it elicited full contractile force in segments 2 through 4, ie, no significant difference existed between the Em,, values in these graft segments as compared with the controls (see Fig 3). After 30 days noradrenaline was at most 80 times more potent in the graft segments as compared with the controls, and after 60 days it was at most 126 times more potent in the graft segments. Putting together the segment from each rat with the highest contractile capacity on day 30 and 60, their pEC,, values (6.7 ? 0.3) were significantlygreater ( p < 0.01) than in the controls (4.9 2 0.4). The relation between the EC, concentrations was thus: 10-4.9:10-6.7 = 10-4.9 - 10+1.8= 63; ie, noradrenaline was 63 times more potent in the graft segments than in the control segments.
Comment The potassium-rich (124 mmol/L) Krebs solution used in the present study creates a depolarization of the smooth muscle cell membrane; this elicits a constant activation of
100
Ann Thorac Surg
MASSA ET AL CONTRACTILE CAPACITY OF FREE ARTERIAL GRAFTS
potential-operated calcium channels with subsequent entry of calcium into the myocytes, thus eliciting a contraction. When a steady state is reached, a stable, constant vasocontraction will be present. We used the amplitude of this contraction as an internal standard for each vessel segment studied, and expressed the responses obtained with the other drugs as a percentage of this maximal potassium-induced contraction. This is a generally accepted way of presenting vasoconstrictive effects, having the advantage that the responses of different drugs to different vessel segments can be compared independently of the sizes or muscle masses. Radial artery grafts for aortocoronary bypass were used with early favorable results [4]. However, most of these grafts occluded within several months [5, 131. Chiu [14] studied the histology of free femoral arterial and external jugular vein grafts in dogs after 60 days. Half of the grafts were wrapped with plastic sheets to interfere with the regeneration of vasa vasorum. As in the present study (see Fig 2F) he found mild to moderate subintimal hyperplasia in a number of the free femoral arterial grafts, particularly when the grafts were wrapped; the vein grafts had better patency and did not develop intimal hyperplasia except in one graft that was wrapped with the plastic sheet. Extrapolating these results to a clinical situation he suggested that free thick-walled, muscular arterial grafts with the vasa vasorum disrupted at both ends cannot regenerate readily from the adjacent tissue, as is the case in the aortocoronary bypass position, and therefore they are more vulnerable to subintimal hyperplasia and occlusion than thin-walled vein grafts. If his hypothesis is correct, the internal mammary artery, which is a thinwalled artery, should be better as a free aortocoronary bypass graft than the radial artery, which has a thicker wall. The observations of Loop and associates [7] suggest that this might be true, although the number of grafts studied was very small. Our morphological results are in accordance with those of Chiu [14]. Furthermore, our findings indicate that hypercontractility, ie, increased sensitivity to cathecholamines (spasm), might be a problem in free arterial grafts because noradrenaline had full contractile force and was much more potent in the graft segments than in the controls after 30 and 60 days. One explanation of this increased potency of noradrenaline could be an upregulation of adrenergic a-adrenoceptors as a consequence of the total denervation of the free arterial graft. Another explanation might be that the uptake of noradrenaline into sympathetic nerve endings, a physiological event, was abolished due to the lack of these nerves, and all the noradrenaline given was thus available to stimulate the postjunctional a-adrenergic receptors. During the first 14 days the a-adrenergic receptors seemed to be absent or out of function as noradrenaline had no contractile effect, whereas U-46619 had full contractile capacity in the same graft segments. After 1 month there was a hyperfunction of the a-adrenergic receptors because noradrenaline was significantly more potent in the grafts than in the controls. This might have contributed to the disappointing patency after 30 and 60 days, ie, circulating catecholamines, which
1991:51:94101
under normal conditions should be unable to cause pathological vascular constriction, might now have induced spasm in the grafts resulting in occlusion. The intimal thickening (see Fig 2D-F) that is generally seen in free muscular arterial grafts [14] might be another explanation of the poor patency. To our knowledge no study has been done to investigate if such a thickened intima is able to or has retained its normal capacity to produce relaxing or antithrombogenic factors. It has been shown that a regenerated endothelium has a reduced function of endothelium-derived relaxing factors [24]. To test these hypotheses a randomized study with and without a potent a-adrenergic antagonist should be done to asertain if an effective blockade of the a-adrenergic receptors would increase the patency of different free arterial grafts. Furthermore, corresponding studies should be done with, for example, antithrombotic or cytostatic drugs with the aim of preventing occlusion. In conclusion, we think that a skeptical attitude to a liberal use of free arterial grafts in clinical surgery is warranted until more studies on this topic have been published. We thank Prof Jan Kugelberg, our chief, who by his enthusiastic and economic support made this study possible.
References 1. Stoney WS, Alford WC Jr, Burrus GR, Glassford DM Jr, Petracek MR, Thomas CS Jr. The fate of arm veins used for aorto-coronary bypass grafts. J Thorac Cardiovasc Surg 1984; 88:522-6. 2. Loop FD, Lytle BW, Cosgrove DM, et al. Influence of the internal-mammary-artery graft on 10-year survival and other cardiac events. N Engl J Med 1986;314:1-6. 3. Johnson DW, Brenowitz JB, Kayser KL. Factors influencing long-term (10-year to 15-year) survival after a successful coronary artery bypass operation. Ann Thorac Surg 1989;48: 19-25. 4. Carpentier A, Guermonprez JL, Deloche A, Frechette C, DuBost C. The aorta-to-coronary radial artery bypass graft. Ann Thorac Surg 1973;16:111-21. 5. Fisk RL, Brooks CH, Callaghan JC, Dvorkin J. Experience with the radial artery graft for coronary artery bypass. Ann Thorac Surg 1976;21:513-8. 6. Barner HB. The internal mammary artery as a free graft. J Thorac Cardiovasc Surg 1973;66:219-21. 7. Loop FD, Lytle BW, Cosgrove DM, et al. Free (aortacoronary) internal mammary artery graft. Late results. J Thorac Cardiovasc Surg 1986;92:827-31. 8. Lytle BW, Cosgrove DM, Ratliff NB, Loop FD. Coronary artery bypass grafting with the right gastroepiploic artery. J Thorac Cardiovasc Surg 1989;97:826-31. 9. Tanimoto Y, Matsuda Y, Masuda T, et al. Multiple free (aorta-coronary) gastroepiploic artery grafting. Ann Thorac Surg 1990;49:479-80. 10. Puig LB, Ciongolli W, Cividanes GVL, et al. Inferior epigastric artery as a free graft for myocardial revascularization. J Thorac Cardiovasc Surg 1990;99:251-5. 11. Vincent JG, van Son JAM, Skotnicki SH. Inferior epigastric artery as a conduit in myocardial revascularization: the alternative free arterial graft. Ann Thorac Surg 1990;49:323-5.
Ann Thorac Surg 1991;51:94101
12. Fonkalsrud EW, Sanchez M, Zerubavel R, Lassaletta L, Smeesters C , Mahoney A. Arterial endothelial changes after ischemia and perfusion. Surg Gynecol Obstet 1976;142: 715-21. 13. Curtis JJ, Stoney WS, Alford WC, Burrus G, Thomas C. Intimal hyperplasia. A cause of radial artery aorto-coronary bypass graft failure. Ann Thorac Surg 1975;20:628-35. 14. Chiu C-J. Why do radial artery grafts for aortocoronary bypass fail? A reappraisal. Ann Thorac Surg 1976;22:52&3. 15. Angelini GD, Christie MI, Bryan AJ, Lewis MJ. Surgical preparation impairs release of endothelium-derived relaxing factor from human saphenous vein. Ann Thorac Surg 1989; 48:417-20. 16. Johns RA, Peach MJ, Flanaagan T, Kron IL. Probing of the canine mammary artery damages endothelium and impairs vasodilation resulting from prostacyclin and endotheliumderived relaxing factor. J Thorac Cardiovasc Surg 1989;97 252-8. 17. Mills NL, Everson CT. Right gastroepiploic artery: a third arterial conduit for coronary artery bypass. Ann Thorac Surg 1989;47706-11.
MASSA ET AL CONTRACTILE CAPACITY OF FREE ARTERIAL GRAFE
101
18. Sarabu MR, McClung JA, Fass A, Reed GE. Early postoperative spasm in left internal mammary artery bypass grafts. Ann Thorac Surg 1987;44:199-200. 19. Blanche C, Chaux A. Spasm in mammary artery grafts. Ann Thorac Surg 1988;45:586. 20. Mills NL, Bringaze WL 111. Preparation of the internal mammary artery graft. Which is the best method? J Thorac Cardiovasc Surg 1989;98:73-9. 21. Von Segesser L, Simonet F, Meier B, Finci L, Faidutti B. Inadequate flow after internal mammary-coronary artery anastomoses. J Thorac Cardiovasc Surg 1987;35:3524. 22. Von Segesser L, Lehmann K, Turina M. Deleterious effect of shock in internal mammary artery anastomoses. Ann Thorac Surg 1989;47575-9. 23. Suma H. Spasm of the gastroepiploic artery graft. Ann Thorac Surg 1990;49:168-9. 24. Shimokawa H, Aarhus LL, Vanhoutte PM. Porcine coronary arteries with regenerated endothelium have a reduced endothelium-dependent responsiveness to aggregating platelets and serotonin. Circ Res 1987;61:25670.
