- Email: [email protected]
Weak β-adrenoceptor-mediated relaxation in the human internal mammary artery
J THoRAc
CARDIOVASC SURG
1989;97:259-66
Weak fj-adrenoceptor-mediated relaxation in the human internal mammary artery The fimction of p-adrenoceptors in the human internal mammary artery was studiedin vitro to predict the way in which the internal mammary artery graft would respond to p-adrenergic agonists and antagonists given in the perioperative period. Ring segments of the distal internal mammary artery obtained from patients not receiving p-blocker therapy were mounted in organ baths and isometric wall force was measured. For comparison,similar experimentswere conductedon segmentsof canine coronary artery, a vessel known to havepowerful p-adrenoceptor function.AU arteries wereprecontracted with potassium or the thromboxane mimetic agent, U46619, before isoproterenol cumulative concentration-relaxation curves were constructed, In the human internal mammary artery, the maximum relaxation induced by isoproterenol was only 14% of the potassium-induced contraction and 24% of the U46619-induced contraction. These responses were weak compared with 54 % and 86 % for p-adrenoceptor relaxation measured in corresponding experiments in the canine coronary artery. In all experiments, propranolol antagonized the relaxation induced by isoproterenol These studies suggested that the human internal mammary artery has only a small number of p-adrenoceptors. We concludethat p-adrenoceptors would contribute little to the reactivity of the human internal mammary artery graft to sympathomimetic drugs.
Gu
Although the long-term patency of the internal mammary artery (IMA) graft is superior to that of the saphenous vein graft, the IMA has a smaller diameter and a greater tendency to go into spasm during the perioperative period.' Adrenoceptor agonists such as epinephrine, norepinephrine, dopamine, and phenylephrine are commonly used as vasoactive agents in the perioperative period. There is controversy as to whether the IMA can respond to administered vasoactive agents' or simply acts as a passive conduit leading to the more responsive coronary bed.3 In large-diameter arteries, «-adrenoceptors mediate contraction and ,8-adrenoceptors mediate relaxation. Therefore it is important to
know whether or not administered agents can stimulate ,8-adrenoceptors in the IMA. Furthermore, ,8-adren
Methods From the Baker Medical Research Institute and Epworth Hospital, Melbourne, Australia. This work was supported by the National Health and Medical Research Council of Australia and the Australia-China Medical and Scientific Research Foundation. Received for publication Dec. 30, 1987. Accepted for publication Aug. 5, 1988. Address for reprints: Dr. James A. Angus, Baker Medical Research Institute, Commercial Road, Prahran, Victoria 3181, Australia.
Preparation of vessels Human [MA. Human IMA segements were taken from 12 patients undergoing IMA--coronary artery graft operations. These patients were not receiving {1-adrenoceptor antagonist therapy. The clinical characteristics of these patients and the drugs they received are listed in Table I. Seven patients had single IMA grafts and five had bilateral IMA grafts. Any redundant length at the distal end of either the left or the right IMA pedicle was taken for study. Approval to use discarded IMA tissue was given by the Human Ethics Committee of the Epworth Hospital. The IMA specimens were placed imrnedi-
259
260
The Journal of Thoracic and Cardiovascular
He et al.
Surgery
Table I. Clinical characteristics and prior therapy of 12 patients undergoing IMA grafting Clinical characteristics
No. of patients
Age 59.5 ± 3.8 years Sex Male Female Procedure Single IMA graft Double IMA graft Drug therapy Verapamil Nifedipine Diltiazem GTN {3-adrenoceptor antagonists
Force F
+
n ~~
t
f
.~ L...---.-.I
2cm
I
I I
I
I
l' - /j
Fig. 1. Technical drawings of the organ bath, transducer, and micrometer-controlled support leg for the vessel segments. The stainless-steel (SS) wire hooks were arranged so as to measure the passive force-internal circumference relationship. g, Length of vessel segment cut by a double-bladed scalpel; f, separation of wires measured by the micrometer; F, isometric force measured by the transducer. (From He G-W, Angus JA, Rosenfeldt FL. Reactivity of the canine isolated internal mammary artery, saphenous vein and coronary artery to constrictor and dilator substances: relevance to coronary bypass graft surgery. J Cardiovasc Pharmacol 1988;12:12-22.)
ately in a container with oxygenated Krebs solution (4 0 C) and transferred to the laboratory. The time delay between excision in the operating room and the laboratory experiment was Ih to 2 hours. The IMA was placed in a dish coated with silicone rubber and the connective tissue was carefully removed. The artery was cut precisely with a double-bladed scalpel into ring segments, 3 mm in length, which were suspended on wires in organ baths (see below). The number of rings taken from each IMA varied from 1 to 4.
11 I
7 5 3 4 3
8
o
Canine coronary artery. Twenty greyhound dogs of either sex, weighing 20 to 30 kg, were anesthetized with sodium pentobarbital (40 mg/kg), The hearts were removed and placed in cold oxygenated Krebs solution. The animal holding and experimental procedures were approved by the Baker Medical Research Institute Ethics Committee under the guidelines of the Code of Practice for the Care and Use of Animals for Experimental Purposes of the National Health and Medical Research Council of Australia. A segment of the proximal circumflex artery,S to 8 ern in length, was excised and cleared of connective tissue, and the ring segments were prepared. From each dog, 6 to 12 coronary artery rings were obtained. Organ bath technique. The arterial rings were suspended on wire hooks in a 25 ml jacketed glass organ bath maintained at 37° C ± 0.1 0 C (standard error of the mean). The Krebs solution had the followingcomposition (in millimoles per liter): Na+ 144, K+ 5.9, Ca2+ 2.5, Mg2+ 1.2, cr- 128.7, H,PO:; 1.2, HC0 3 25, SO':; 1.2 and glucose 11. It was aerated with a gas mixture of 95% oxygen and 5% carbon dioxide. Hooks were fashioned from a length of surgical steel suture wire (B&S 20, Ethicon, Inc., Somerville, N.J.) 500 ~m in diameter. The upper hook was suspended from a Grass FT03C force transducer (Grass Instrument Company, Quincy, Mass.), and the lower hook was fixed to a methacrylate leg attached to a micrometer (Mitutoyo, Tokyo, Japan) (Fig. I). Force was recorded on single channel flat bed recorders (Rikadenki, Japan). Six organ bath arrangements were run concurrently. Passive force-diameter relationships. We have developed a procedure to stretch the vessel rings to an optimal point on the length-passive wall tension relationship. This technique allows a reproducible wall tension to be applied that corresponds with arterial pressure in the body. Full details of this technique have been published.':" In brief, the vesselrings were stretched in steps and the passive wall tension (T) recorded (Fig. 2). From the Laplace relationship,
P = 211" T/L the corresponding transmural pressure (P, in millimeters of mercury) was determined from the circumference of the ring
Volume 97 Number 2 February 1989
26 1
Weak {3-adrenoceptor-mediated relaxation in IMA
(L) and tension (T). Applying a number of stretch steps (see Fig. 2, top) and using a computer-based iterative fitting technique allows the intersection of the theoretical isobar (100 mm Hg) and the exponential line (fitting the data points for the stretches) to be determined to indicate the internal circumference (L lOO) at an equivalent transmural pressure of 100 mm Hg (see Fig. 2, bottom). The artery was then relaxed a little to a circumference equal to 0.9 X LuX) and held at this degree of stretch for the remainder of the experiment. The aim of this method was to determine the optimum passive resting force for each vessel from the knowledge of its own lengthtension curve. This method normalizes all the vessels of different sizes and varying smooth muscle content to a comparable degree of passive stretch."! Protocol. After the normalization procedure, the ring segments were left resting at 0.9L uX) for 30 minutes. A steady level of active contraction was then established by adding potassium (as potassium chloride) or U46619 (a thromboxane A2 mimetic agent). The concentrations of potassium (25 mrnol/L) and U46619 (20 nmol/L), subsequently used to precontract the vessels, were determined from pilot experiments and from previous work." These concentrations gave 50% to 80% of the maximum contraction for that agent in a particular artery. This was designed as the submaximal concentration. In some rings, propranolol (0.1 J.Lmol/L) was added to the organ bath 10 minutes before the addition of the constrictor agent. Cumulative concentration-response curves for isoproterenol were then obtained by adding this agent to the organ bath (Figs. 3 to 5). Only one concentration-response curve was obtained in each ring. After the isoproterenolresponsecurve was completed, prazosin (1 J.Lmol/L) was added to antagonize any o-adrenoceptor effect produced by the highest dose of isoproterenol given (30 J.Lmol/L). Finally, glyceryl trinitrate (GTN), 10 J.Lmol/L was added to determine whether the rings were still capable of further relaxation (Fig. 3). To establish the time course of the contraction of the canine coronary artery to potassium or U46619, a separate set of 6 rings was contracted but then left alone without any addition of isoproterenol. This provided an average time trend to properly estimate the degree of relaxation by isoproterenol from the contracted force. The force was measured in the "time-trend" ring each time the concentration of isoproterenol was raised and has therefore been plotted on the same graph (Fig. 5). These time trends could not be determined in human IMA because of the paucity of tissue. Data analysis. The l3-adrenoceptor function in these vessels was estimated from the percentage relaxation induced by isoproterenol of the precontraction force at the start of the relaxation curve (100%). Student's unpaired t test was used to test for statistical significance between treatments. The average standard error of the mean (SEM) for all points making up a concentration-response curve was calculated from the analysis of variance as (error mean square/number of rings)" after subtracting the sum of squares "between rings" and "between concentrations" from the "total" sum of squares for each concentration. This error bar (± 1 average SEM) is located on the average concentration-response curve (Figs. 4 and 5). The concentration of isoproterenol that caused 50% of the relaxation (EC5Q) was determined from each concentration-response curve by a logistic curve fitting equation
E=MN/N+K P
20
Human
Canine
IMA
CA
-~
Reltlng force
....
2mln
30
•
....... ~ 20
I
•
!
I ~
.. ~ ..
o
2
i
/J
•
/
,
345 Diameter (mm)
,
6
Fig. 2. Top, Chart records of isometric force developed by a ring segment of human IMA and canine coronary artery (CA) during the stretch procedure to normalize the vessels. Resting force was set to the level indicated, with the internal circumference stretched to 90% of the circumference that would occur at a transmural pressure of 100 mm Hg for the IMA and coronary artery. Bottom, Graphic display of the wall tension (force per unit length) and internal circumference and diameter from the chart records above. The straight line is the isobar relating wall tension to radius from the Laplace relationship for a transmural pressure of 100 mm Hg. The intersection of this isobar with the exponential curves indicates the internal circumference (x axis) for these segments at 100 mmHg.
262
The Journal of Thoracic and Cardiovascular Surgery
He et al.
Human IMA
Isoproterenol 4
-8,5 -7,5 -9 -8
-7
2
-15.15~~~ -
Prazosln -4.l5
-8",,",
QTN -15
l-.-..J 10min
Fig. 3. Representative traces of isometric force developed in a human IMA ring segment (top) and in a canine coronary artery ring (bottom). Numbers are the cumulative log molar concentration. The arteries were precontracted with the thromboxane mimetic U46619. Note the very weak relaxation to isoproterenol in the human IMA in comparison with the powerful effect in the canine coronary artery.
where E is response, M is maximum relaxation, A is concentration, K is EC so concentration, and P is the slope parameter." From these fitted equations, the average EC so value and I SEM were calculated.
Results Resting vessel parameters. Representative diameter-wall tension relationships of the human IMA and the canine coronary artery are shown in Fig. 2. The average vessel diameter at a transmural pressure of 100 mm Hg, the transmural pressure at 0.9L lOo (see Methods), and the resting force at 0.9L lOo are shown in Table II. The diameter of the rings of human IMA were significantly less than the diameter of those from the canine coronary artery (p < 0.001). Contraction. The increases in contractile force induced by the submaximal concentration of potassium or U46619 in the human IMA and canine coronary artery are listed in Table III. In the human IMA, the force of contraction induced by U46619 (20 nmol/L) was significantly greater than that induced by potassium (25 mmol/L) (p < 0.05). In the canine coronary artery,
the contraction induced by potassium was not significantly different from that induced by U46619 (p > 0.05). Relaxation Human [MA. Fig. 3 shows typical examples of relaxation in vessel rings precontracted by U46619. The human IMA showed a weak relaxation to isoproterenol whereas the canine coronary artery showed marked relaxation. The group mean results for the human IMA precontracted with potassium or U46619 and relaxed with isoproterenol are shown in Fig. 4. In the potassiumcontracted human IMA, the maximum relaxation induced by isoproterenol was 13.9 ± 6.5% (n = 7) (Fig. 4, left). This relaxation was almost completely abolished by propranolol (0.1 ~mol/L). At high concentrations, isoproterenol caused slight contraction. This was presumably due to o-adrenoceptor stimulation, since it was antagonized by the o-adrenoceptor antagonist, prazosin. Near-maximum relaxation could be induced by GTN 10 ~mol/L. In the human IMA precontracted by U46619 (Fig. 4, right), isoproterenol produced more
Volume 97 Number 2 February 1989
Weak {3-adrenoceptor-mediated relaxation in IMA
263
U46619
o
propranolol
.-O.":'~~~---
25
0·,0
--·-· "~~l.;~;r~~~;~-
c.ntr.~l.:"~
\\
\1
\
\
50
I
75
100
'0.0\\
\1 !
I
I
I
I
9
7
9
7
Isoproterenol
-logM
Isoproterenol
-logM
Fig. 4. Average concentration-relaxation curves for isoproterenol in human isolated IMA ring segments precontracted with potassium (K+) in seven rings (left) or by thromboxane mimetic U46619 in 9 rings (right). In separate experiments (7 rings) propranolol (0.1 J.Lmol/L) was added before the contraction. Relaxation was measured as a percentage of the precontraction. Error bars on the lines are average SEM from analysis of variance (see Methods) and ± I SEM for the maximum relaxation to GTN (10 J.Lmol/L) treatment.
relaxation (24.1% ± 5.3%, n = 9) than in the IMA precontracted by potassium (p> 0.05). This relaxation was partially antagonized by propranolol in a further 7 rings. Canine coronary artery. The canine large coronary artery showed relaxation in response to isoproterenol. The maximum relaxation for isoproterenol against potassium was 37.6% ± 3.2%, n = 13 (Fig. 5, left). Pretreatment with propranolol (0.1 J.Lmol/L) displaced the isoproterenol relaxation curve to the right so that the concentration required to produce 50% of the maximum relaxation response (EC so) was 21-fold higher than in the absence of propranolol (p < 0.001). In all cases, GTN (10 umol/L) could still induce further relaxation. The relaxation induced by isoproterenol in the canine coronary artery was most pronounced when the rings were contracted by U46619 (Fig. 5, right). The maximum relaxation was 88.6% ± 1.1 %, n = 9, significantly more than that for potassium-contracted rings (37.6% ± 3.2%, p < 0.001). In separate U46619-contracted coronary artery rings (n = 9), propranolol (0.1 J.LmoljL) again displaced the isoproterenol concentra-
tion-response curve to the right (as for potassiumprecontracted arteries). The increase in EC so was from 7.69 (-logM) to 6.53, a shift of 14.5-fold (p < 0.001). Therefore propranolol significantly antagonized the {3adrenoceptor-mediated relaxation in the canine coronary artery contracted by either potassium or U46619. GTN further relaxed the canine coronary artery to near maximum. Discussion
The human IMA relaxed very weakly in response to {3-adrenoceptor agonists, which suggests that functional {3-adrenoceptors are sparse in this artery. In comparison with dog coronary arteries, the human IMA showed a much weaker {3-adrenoceptor dilatation. Others have demonstrated powerful relaxation of segments of human and monkey large coronary arteries in vitro in response to {3-adrenoceptor stimulation.v" The use of short ring segments of arteries suspended in the controlled environment of an organ bath under isometric tension is an ideal method to analyze adrenoceptor function. The specific agonists and antagonists can be applied under controlled conditions since the
264
The Journal of Thoracic and Cardiovascular Surgery
He et al.
U46619
o
~~ . tr.~ ~·I"'· '· '
-. ~~-o~~~.
,
time trend
".. . ,.. ,.. ,.. ''l'.......,.. ,',.,.,
,.,."
•.'o.'o~~-~------------'',
----------
.'.
" 'e \
0\
\, prcfpranolol
28
control
~
\
\
io...-."\
propranolol
\J
~
contro,,\
\
\\i
\\ \\
\\ 75
• ....~·o
..........
100 I
I
9 7 Isoproterenol
_--L.
I~
5 J GTN -logM
9
.L.....
I~/L...l..-
5
7
Isoproterenol
-logM
,.'.
'/GTN
Fig. 5. Average concentration-response curves for isoproterenol in canine isolated coronary artery ring segments precontracted with potassium (left. n = 13) or U46619 (right. n = 9). Propranolol (0.1 ,umol/L) caused iJ-adrenoceptor antagonism displayed as rightward shifts in the isoproterenol curve in potassium-contracted rings (n = 7) and in U46619-contracted rings (n = 9). Errors bars: Horizontal bars are placed on EC so values (± 1 SEM); vertical bars are average SEM (see legend of Fig. 4) or SEM of the relaxation to GTN (10 ,umol/L). Time trend lines are from 6 separate rings (see text).
Table Il, Summary of vessel parameters Human IMA
Canine coronary artery
n
D/OO (mm)
P (mm Hg)
Resting force (g)
30 50
2.02 ± 0.11 3.85 ± 0.07
75,88 ± 1.17 63.09 ± 0.57
4.15 ± 0.30 6.87 ± 0.20
n, Number of rings; D100. internal diameter (mm) at a theoretical transmural pressure of 100 mm Hg, After the vessels were normalized and the circumference was set to O,9LuX) (see Methods), the equivalent transmural pressure (P [mm Hg]) and resting force (g) are indicated. Values are mean ± I SEM,
bathing solution and the artery wall are allowed to come into concentration equilibrium as judged by a steady plateau response. The normalization technique for each artery segment ensures that physiologic conditions of passive stretch are applied to all arteries regardless of diameter. This is important since variations in initial length of the smooth muscle cells can alter the range and sensitivity (EC so) of agonist concentration-response curves." In the present work, the human IMA segments had small internal diameters as measured at an equivalent transmural pressure of 100 mm Hg, namely 2.02 mm compared with nearly double the diameter in the canine coronary artery (3.85 mm). Correspondingly, the
resting force that was applied at the optimal length was 4.15 g for the human IMA but 6.87 g for the dog vessel (Table II). The isolated ring segments studied in this way allow precise assay of the isometric force developed by the smooth muscle in response to chosen agents under physiologic conditions without the effect of circulating hormones, neural stimulation, or blood flow. By this method, the endothelium is preserved in isolated ring segments. This was tested previously in our laboratory by the endothelium-dependent relaxation to acetylcho-
line." We used isoproterenol to assay /1-adrenoceptor activity since this agent is considered to be a full agonist at
Volume 97
Weak fJ-adrenoceptor-mediated relaxation in IMA
Number 2 February 1989
26 5
Table ill. Increase in force (g) caused by submaximal concentration of potassium (25 mmoljL) or U46619 (20 nmoljL) before relaxation by isoproterenol Human IMA Canine coronary artery
Potassium
U466I 9
p Value
3.66 ± 0.65 (14) 10.90 ± 0.95 (26)
6.34 ± 1.03 (16) 13.37 ± 1.30 (24)
<0.05
NS
Number of rings indicated in parentheses. NS. Not significantly different.
both fJl-adrenoceptors and fJ2-adrenoceptors. The stimulus that causes IMA spasm during coronary bypass grafting is unknown. In the present work two potential stimuli were used: (1) potassium depolarization, which causes calcium entry into smooth muscle via a voltagedependent calcium channel and also via the release of norepinephrine from sympathetic nerve terminals, and (2) U46619, which is a chemically stable analog of thromboxane A 2. Thromboxane A 2 is a potent vasoconstrictor that is released from aggregating platelets during surgery; elevated plasma levels have been found during cardiopulmonary bypass. 14 Compared with potassium-contracted vessels, the contraction to U46619 in dog coronary artery was more readily relaxed by iJ-adrenoceptor stimulation, even though the starting contractions were not significantly different (Table III). The weak response to isoproterenol in the human IMA was not due to persisting effects of fJ-blockade, since we used only tissues that was free from prior exposure to fJ-adrenoceptor antagonist therapy. In human IMA, propranolol (0.1 umol/L), at a concentration within the range of levels occurring clinically, I 5 prevented the small relaxation (14% to 24%) to isoproterenol. Evidence that these arteries could respond further to a vasodilator agent acting via a non-dadrenoceptor mechanism was the 50% to 60% relaxation to GTN in potassium-contracted rings and greater than 80% relaxation in the U46619-contracted preparations. Therefore, under these conditions, the small relaxation in response to isoproterenol is compelling evidence that the functional fJ-adrenoceptors were sparse in the human IMA. Autoradiography has demonstrated that l1-adrenoceptors are localized, in relatively low concentrations, in the media of the human IMA.6 These functional and receptor density observations suggest that the poor response to isoproterenol is directly caused by a low number of receptors rather than by alterations in the intracellular coupling with adenylate cyclase and the generation of cyclic adenosine monophosphate. By contrast, in the dog coronary artery, which was included in these studies to compare with the reactivity of the human IMA, we found evidence for marked iJ-adrenoceptor activity.
Clinical implications. The human isolated IMA responds poorly to fJ-adrenoceptor stimulation. This implies that perioperatively administered catecholamines or sympathomimetic amines are most unlikely to dilate the IMA. As a corollary, fJ-adrenoceptor antagonists will not uncover constrictor activity on the IMA and are therefore unlikely to induce spasm, as has been reported for coronary arteries." We thank Mrs. Clara Chan for secretarial assistance. REFERENCES 1. 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. 2. Jett GK, Arcidi JM, Dorsey LMA, Hatcher CR, Guyton RA. Vasoactive drug effects on blood flow in internal mammary artery and saphenous vein grafts. J THORAC CARDIOVASC SURG 1987;94:2-11. 3. McCormick JR, Kaneko M, Baue AE, Geha AS. Blood flow and vasoactive drug effects in internal mammary artery and venous bypass grafts. Circulation 1975; 51,52(Pt 2):172-9. 4. Thadani U. Assessment of "optimal" beta blockade in treating patients with angina pectoris. Acta Med Scand I984;suppI694: 178-87. 5. Hjalmarson A. Early intervention with a beta-blocking drug after acute myocardial infarction. Am J Cardiol 1984;54: 11 E-3E. 6. Molenaar P, Malta E, Jones CR, Buxton BF, Summers RJ. Autoradiographic localization of beta-adrenoceptors in human internal mammary artery and saphenous vein. Br J Pharmacol 1988;95:225-33. 7. Angus JA, Cocks TM, Satoh K. a2-Adrenoceptors and endothelium-dependent relaxation in canine large arteries. Br J Pharmacol 1986;88:767-77. 8. He G-W, Angus JA, Rosenfeldt FL. Reactivity of the canine isolated internal mammary artery, saphenous vein and coronary artery to constrictor and dilator substances: relevance to coronary bypass graft surgery. J Cardiovasc PharmacoI1988;12:12-22. 9. Nakashima A, Angus JA, Johnston Cl. Comparison of angiotensin converting enzyme inhibitors captopril and MK421-diacid in guinea pig atria. Eur J Pharmacol 1982;81 :487-92. 10. Anderson R, Holmberg S, Svedmyr N, Aberg G. Adren-
266 He et al.
ergic a- and J3-receptors in coronary artery vesselsin man. Acta Med Scand 1972;191:241-4. 11. Toda N. Isolated human coronary arteries in response to vasoconstrictor substances. Am J Physiol1983;245:H93741. 12. Toda N. Response of isolated monkey coronary arteries to catecholarnines and to transmural electrical stimulation. Circ Res 1981;49:1228-36. 13. Tallarida RJ, Sevy RW, Harakal C, Hendrick J, Faust R.
The Journal of Thoracic and Cardiovascular Surgery
The effect of preload on the dissociation constant of norepinephrine in isolated strips of rabbit thoracic aorta. Arch Int Pharmacodyn Ther 1974;210:67-74. 14. Davies GC, Sobel M, Salzman EW. Elevated plasma fibrinopeptide A and thromboxane 8 2 levels during cardiopulmonary bypass. Circulation 1980;61:808-13. 15. Robertson RM, Wood AJJ, Vaughn WK, Robertson D. Exacerbation of vasotonic angina pectoris by propranolol. Circulation 1982;65:281-5.
Bound volumes available to subscribers Bound volumes of THE JOURNAL OF THORACIC AND CARDIOVASCULAR SURGERY are available to subscribers (only) for the 1989 issues from the Publisher, at a cost of $55.00 ($72.00 international) for Yol. 97 (January-June) and Yol. 98 (July-December). Shipping charges are included. Each bound volume contains a subject and author index and aU advertising is removed. Copies are shipped within 60 days after publication of the last issue of the volume. The binding is durable buckram with the JOURNAL name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact The C.Y. Mosby Company, Circulation Department, 11830 Westline Industrial Drive, St. Louis, Missouri 63146-3318, USA; phone (800) 325-4177, ext. 351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular JOURNAL subscription.
CARDIOVASC SURG
1989;97:259-66
Weak fj-adrenoceptor-mediated relaxation in the human internal mammary artery The fimction of p-adrenoceptors in the human internal mammary artery was studiedin vitro to predict the way in which the internal mammary artery graft would respond to p-adrenergic agonists and antagonists given in the perioperative period. Ring segments of the distal internal mammary artery obtained from patients not receiving p-blocker therapy were mounted in organ baths and isometric wall force was measured. For comparison,similar experimentswere conductedon segmentsof canine coronary artery, a vessel known to havepowerful p-adrenoceptor function.AU arteries wereprecontracted with potassium or the thromboxane mimetic agent, U46619, before isoproterenol cumulative concentration-relaxation curves were constructed, In the human internal mammary artery, the maximum relaxation induced by isoproterenol was only 14% of the potassium-induced contraction and 24% of the U46619-induced contraction. These responses were weak compared with 54 % and 86 % for p-adrenoceptor relaxation measured in corresponding experiments in the canine coronary artery. In all experiments, propranolol antagonized the relaxation induced by isoproterenol These studies suggested that the human internal mammary artery has only a small number of p-adrenoceptors. We concludethat p-adrenoceptors would contribute little to the reactivity of the human internal mammary artery graft to sympathomimetic drugs.
Gu
Although the long-term patency of the internal mammary artery (IMA) graft is superior to that of the saphenous vein graft, the IMA has a smaller diameter and a greater tendency to go into spasm during the perioperative period.' Adrenoceptor agonists such as epinephrine, norepinephrine, dopamine, and phenylephrine are commonly used as vasoactive agents in the perioperative period. There is controversy as to whether the IMA can respond to administered vasoactive agents' or simply acts as a passive conduit leading to the more responsive coronary bed.3 In large-diameter arteries, «-adrenoceptors mediate contraction and ,8-adrenoceptors mediate relaxation. Therefore it is important to
know whether or not administered agents can stimulate ,8-adrenoceptors in the IMA. Furthermore, ,8-adren
Methods From the Baker Medical Research Institute and Epworth Hospital, Melbourne, Australia. This work was supported by the National Health and Medical Research Council of Australia and the Australia-China Medical and Scientific Research Foundation. Received for publication Dec. 30, 1987. Accepted for publication Aug. 5, 1988. Address for reprints: Dr. James A. Angus, Baker Medical Research Institute, Commercial Road, Prahran, Victoria 3181, Australia.
Preparation of vessels Human [MA. Human IMA segements were taken from 12 patients undergoing IMA--coronary artery graft operations. These patients were not receiving {1-adrenoceptor antagonist therapy. The clinical characteristics of these patients and the drugs they received are listed in Table I. Seven patients had single IMA grafts and five had bilateral IMA grafts. Any redundant length at the distal end of either the left or the right IMA pedicle was taken for study. Approval to use discarded IMA tissue was given by the Human Ethics Committee of the Epworth Hospital. The IMA specimens were placed imrnedi-
259
260
The Journal of Thoracic and Cardiovascular
He et al.
Surgery
Table I. Clinical characteristics and prior therapy of 12 patients undergoing IMA grafting Clinical characteristics
No. of patients
Age 59.5 ± 3.8 years Sex Male Female Procedure Single IMA graft Double IMA graft Drug therapy Verapamil Nifedipine Diltiazem GTN {3-adrenoceptor antagonists
Force F
+
n ~~
t
f
.~ L...---.-.I
2cm
I
I I
I
I
l' - /j
Fig. 1. Technical drawings of the organ bath, transducer, and micrometer-controlled support leg for the vessel segments. The stainless-steel (SS) wire hooks were arranged so as to measure the passive force-internal circumference relationship. g, Length of vessel segment cut by a double-bladed scalpel; f, separation of wires measured by the micrometer; F, isometric force measured by the transducer. (From He G-W, Angus JA, Rosenfeldt FL. Reactivity of the canine isolated internal mammary artery, saphenous vein and coronary artery to constrictor and dilator substances: relevance to coronary bypass graft surgery. J Cardiovasc Pharmacol 1988;12:12-22.)
ately in a container with oxygenated Krebs solution (4 0 C) and transferred to the laboratory. The time delay between excision in the operating room and the laboratory experiment was Ih to 2 hours. The IMA was placed in a dish coated with silicone rubber and the connective tissue was carefully removed. The artery was cut precisely with a double-bladed scalpel into ring segments, 3 mm in length, which were suspended on wires in organ baths (see below). The number of rings taken from each IMA varied from 1 to 4.
11 I
7 5 3 4 3
8
o
Canine coronary artery. Twenty greyhound dogs of either sex, weighing 20 to 30 kg, were anesthetized with sodium pentobarbital (40 mg/kg), The hearts were removed and placed in cold oxygenated Krebs solution. The animal holding and experimental procedures were approved by the Baker Medical Research Institute Ethics Committee under the guidelines of the Code of Practice for the Care and Use of Animals for Experimental Purposes of the National Health and Medical Research Council of Australia. A segment of the proximal circumflex artery,S to 8 ern in length, was excised and cleared of connective tissue, and the ring segments were prepared. From each dog, 6 to 12 coronary artery rings were obtained. Organ bath technique. The arterial rings were suspended on wire hooks in a 25 ml jacketed glass organ bath maintained at 37° C ± 0.1 0 C (standard error of the mean). The Krebs solution had the followingcomposition (in millimoles per liter): Na+ 144, K+ 5.9, Ca2+ 2.5, Mg2+ 1.2, cr- 128.7, H,PO:; 1.2, HC0 3 25, SO':; 1.2 and glucose 11. It was aerated with a gas mixture of 95% oxygen and 5% carbon dioxide. Hooks were fashioned from a length of surgical steel suture wire (B&S 20, Ethicon, Inc., Somerville, N.J.) 500 ~m in diameter. The upper hook was suspended from a Grass FT03C force transducer (Grass Instrument Company, Quincy, Mass.), and the lower hook was fixed to a methacrylate leg attached to a micrometer (Mitutoyo, Tokyo, Japan) (Fig. I). Force was recorded on single channel flat bed recorders (Rikadenki, Japan). Six organ bath arrangements were run concurrently. Passive force-diameter relationships. We have developed a procedure to stretch the vessel rings to an optimal point on the length-passive wall tension relationship. This technique allows a reproducible wall tension to be applied that corresponds with arterial pressure in the body. Full details of this technique have been published.':" In brief, the vesselrings were stretched in steps and the passive wall tension (T) recorded (Fig. 2). From the Laplace relationship,
P = 211" T/L the corresponding transmural pressure (P, in millimeters of mercury) was determined from the circumference of the ring
Volume 97 Number 2 February 1989
26 1
Weak {3-adrenoceptor-mediated relaxation in IMA
(L) and tension (T). Applying a number of stretch steps (see Fig. 2, top) and using a computer-based iterative fitting technique allows the intersection of the theoretical isobar (100 mm Hg) and the exponential line (fitting the data points for the stretches) to be determined to indicate the internal circumference (L lOO) at an equivalent transmural pressure of 100 mm Hg (see Fig. 2, bottom). The artery was then relaxed a little to a circumference equal to 0.9 X LuX) and held at this degree of stretch for the remainder of the experiment. The aim of this method was to determine the optimum passive resting force for each vessel from the knowledge of its own lengthtension curve. This method normalizes all the vessels of different sizes and varying smooth muscle content to a comparable degree of passive stretch."! Protocol. After the normalization procedure, the ring segments were left resting at 0.9L uX) for 30 minutes. A steady level of active contraction was then established by adding potassium (as potassium chloride) or U46619 (a thromboxane A2 mimetic agent). The concentrations of potassium (25 mrnol/L) and U46619 (20 nmol/L), subsequently used to precontract the vessels, were determined from pilot experiments and from previous work." These concentrations gave 50% to 80% of the maximum contraction for that agent in a particular artery. This was designed as the submaximal concentration. In some rings, propranolol (0.1 J.Lmol/L) was added to the organ bath 10 minutes before the addition of the constrictor agent. Cumulative concentration-response curves for isoproterenol were then obtained by adding this agent to the organ bath (Figs. 3 to 5). Only one concentration-response curve was obtained in each ring. After the isoproterenolresponsecurve was completed, prazosin (1 J.Lmol/L) was added to antagonize any o-adrenoceptor effect produced by the highest dose of isoproterenol given (30 J.Lmol/L). Finally, glyceryl trinitrate (GTN), 10 J.Lmol/L was added to determine whether the rings were still capable of further relaxation (Fig. 3). To establish the time course of the contraction of the canine coronary artery to potassium or U46619, a separate set of 6 rings was contracted but then left alone without any addition of isoproterenol. This provided an average time trend to properly estimate the degree of relaxation by isoproterenol from the contracted force. The force was measured in the "time-trend" ring each time the concentration of isoproterenol was raised and has therefore been plotted on the same graph (Fig. 5). These time trends could not be determined in human IMA because of the paucity of tissue. Data analysis. The l3-adrenoceptor function in these vessels was estimated from the percentage relaxation induced by isoproterenol of the precontraction force at the start of the relaxation curve (100%). Student's unpaired t test was used to test for statistical significance between treatments. The average standard error of the mean (SEM) for all points making up a concentration-response curve was calculated from the analysis of variance as (error mean square/number of rings)" after subtracting the sum of squares "between rings" and "between concentrations" from the "total" sum of squares for each concentration. This error bar (± 1 average SEM) is located on the average concentration-response curve (Figs. 4 and 5). The concentration of isoproterenol that caused 50% of the relaxation (EC5Q) was determined from each concentration-response curve by a logistic curve fitting equation
E=MN/N+K P
20
Human
Canine
IMA
CA
-~
Reltlng force
....
2mln
30
•
....... ~ 20
I
•
!
I ~
.. ~ ..
o
2
i
/J
•
/
,
345 Diameter (mm)
,
6
Fig. 2. Top, Chart records of isometric force developed by a ring segment of human IMA and canine coronary artery (CA) during the stretch procedure to normalize the vessels. Resting force was set to the level indicated, with the internal circumference stretched to 90% of the circumference that would occur at a transmural pressure of 100 mm Hg for the IMA and coronary artery. Bottom, Graphic display of the wall tension (force per unit length) and internal circumference and diameter from the chart records above. The straight line is the isobar relating wall tension to radius from the Laplace relationship for a transmural pressure of 100 mm Hg. The intersection of this isobar with the exponential curves indicates the internal circumference (x axis) for these segments at 100 mmHg.
262
The Journal of Thoracic and Cardiovascular Surgery
He et al.
Human IMA
Isoproterenol 4
-8,5 -7,5 -9 -8
-7
2
-15.15~~~ -
Prazosln -4.l5
-8",,",
QTN -15
l-.-..J 10min
Fig. 3. Representative traces of isometric force developed in a human IMA ring segment (top) and in a canine coronary artery ring (bottom). Numbers are the cumulative log molar concentration. The arteries were precontracted with the thromboxane mimetic U46619. Note the very weak relaxation to isoproterenol in the human IMA in comparison with the powerful effect in the canine coronary artery.
where E is response, M is maximum relaxation, A is concentration, K is EC so concentration, and P is the slope parameter." From these fitted equations, the average EC so value and I SEM were calculated.
Results Resting vessel parameters. Representative diameter-wall tension relationships of the human IMA and the canine coronary artery are shown in Fig. 2. The average vessel diameter at a transmural pressure of 100 mm Hg, the transmural pressure at 0.9L lOo (see Methods), and the resting force at 0.9L lOo are shown in Table II. The diameter of the rings of human IMA were significantly less than the diameter of those from the canine coronary artery (p < 0.001). Contraction. The increases in contractile force induced by the submaximal concentration of potassium or U46619 in the human IMA and canine coronary artery are listed in Table III. In the human IMA, the force of contraction induced by U46619 (20 nmol/L) was significantly greater than that induced by potassium (25 mmol/L) (p < 0.05). In the canine coronary artery,
the contraction induced by potassium was not significantly different from that induced by U46619 (p > 0.05). Relaxation Human [MA. Fig. 3 shows typical examples of relaxation in vessel rings precontracted by U46619. The human IMA showed a weak relaxation to isoproterenol whereas the canine coronary artery showed marked relaxation. The group mean results for the human IMA precontracted with potassium or U46619 and relaxed with isoproterenol are shown in Fig. 4. In the potassiumcontracted human IMA, the maximum relaxation induced by isoproterenol was 13.9 ± 6.5% (n = 7) (Fig. 4, left). This relaxation was almost completely abolished by propranolol (0.1 ~mol/L). At high concentrations, isoproterenol caused slight contraction. This was presumably due to o-adrenoceptor stimulation, since it was antagonized by the o-adrenoceptor antagonist, prazosin. Near-maximum relaxation could be induced by GTN 10 ~mol/L. In the human IMA precontracted by U46619 (Fig. 4, right), isoproterenol produced more
Volume 97 Number 2 February 1989
Weak {3-adrenoceptor-mediated relaxation in IMA
263
U46619
o
propranolol
.-O.":'~~~---
25
0·,0
--·-· "~~l.;~;r~~~;~-
c.ntr.~l.:"~
\\
\1
\
\
50
I
75
100
'0.0\\
\1 !
I
I
I
I
9
7
9
7
Isoproterenol
-logM
Isoproterenol
-logM
Fig. 4. Average concentration-relaxation curves for isoproterenol in human isolated IMA ring segments precontracted with potassium (K+) in seven rings (left) or by thromboxane mimetic U46619 in 9 rings (right). In separate experiments (7 rings) propranolol (0.1 J.Lmol/L) was added before the contraction. Relaxation was measured as a percentage of the precontraction. Error bars on the lines are average SEM from analysis of variance (see Methods) and ± I SEM for the maximum relaxation to GTN (10 J.Lmol/L) treatment.
relaxation (24.1% ± 5.3%, n = 9) than in the IMA precontracted by potassium (p> 0.05). This relaxation was partially antagonized by propranolol in a further 7 rings. Canine coronary artery. The canine large coronary artery showed relaxation in response to isoproterenol. The maximum relaxation for isoproterenol against potassium was 37.6% ± 3.2%, n = 13 (Fig. 5, left). Pretreatment with propranolol (0.1 J.Lmol/L) displaced the isoproterenol relaxation curve to the right so that the concentration required to produce 50% of the maximum relaxation response (EC so) was 21-fold higher than in the absence of propranolol (p < 0.001). In all cases, GTN (10 umol/L) could still induce further relaxation. The relaxation induced by isoproterenol in the canine coronary artery was most pronounced when the rings were contracted by U46619 (Fig. 5, right). The maximum relaxation was 88.6% ± 1.1 %, n = 9, significantly more than that for potassium-contracted rings (37.6% ± 3.2%, p < 0.001). In separate U46619-contracted coronary artery rings (n = 9), propranolol (0.1 J.LmoljL) again displaced the isoproterenol concentra-
tion-response curve to the right (as for potassiumprecontracted arteries). The increase in EC so was from 7.69 (-logM) to 6.53, a shift of 14.5-fold (p < 0.001). Therefore propranolol significantly antagonized the {3adrenoceptor-mediated relaxation in the canine coronary artery contracted by either potassium or U46619. GTN further relaxed the canine coronary artery to near maximum. Discussion
The human IMA relaxed very weakly in response to {3-adrenoceptor agonists, which suggests that functional {3-adrenoceptors are sparse in this artery. In comparison with dog coronary arteries, the human IMA showed a much weaker {3-adrenoceptor dilatation. Others have demonstrated powerful relaxation of segments of human and monkey large coronary arteries in vitro in response to {3-adrenoceptor stimulation.v" The use of short ring segments of arteries suspended in the controlled environment of an organ bath under isometric tension is an ideal method to analyze adrenoceptor function. The specific agonists and antagonists can be applied under controlled conditions since the
264
The Journal of Thoracic and Cardiovascular Surgery
He et al.
U46619
o
~~ . tr.~ ~·I"'· '· '
-. ~~-o~~~.
,
time trend
".. . ,.. ,.. ,.. ''l'.......,.. ,',.,.,
,.,."
•.'o.'o~~-~------------'',
----------
.'.
" 'e \
0\
\, prcfpranolol
28
control
~
\
\
io...-."\
propranolol
\J
~
contro,,\
\
\\i
\\ \\
\\ 75
• ....~·o
..........
100 I
I
9 7 Isoproterenol
_--L.
I~
5 J GTN -logM
9
.L.....
I~/L...l..-
5
7
Isoproterenol
-logM
,.'.
'/GTN
Fig. 5. Average concentration-response curves for isoproterenol in canine isolated coronary artery ring segments precontracted with potassium (left. n = 13) or U46619 (right. n = 9). Propranolol (0.1 ,umol/L) caused iJ-adrenoceptor antagonism displayed as rightward shifts in the isoproterenol curve in potassium-contracted rings (n = 7) and in U46619-contracted rings (n = 9). Errors bars: Horizontal bars are placed on EC so values (± 1 SEM); vertical bars are average SEM (see legend of Fig. 4) or SEM of the relaxation to GTN (10 ,umol/L). Time trend lines are from 6 separate rings (see text).
Table Il, Summary of vessel parameters Human IMA
Canine coronary artery
n
D/OO (mm)
P (mm Hg)
Resting force (g)
30 50
2.02 ± 0.11 3.85 ± 0.07
75,88 ± 1.17 63.09 ± 0.57
4.15 ± 0.30 6.87 ± 0.20
n, Number of rings; D100. internal diameter (mm) at a theoretical transmural pressure of 100 mm Hg, After the vessels were normalized and the circumference was set to O,9LuX) (see Methods), the equivalent transmural pressure (P [mm Hg]) and resting force (g) are indicated. Values are mean ± I SEM,
bathing solution and the artery wall are allowed to come into concentration equilibrium as judged by a steady plateau response. The normalization technique for each artery segment ensures that physiologic conditions of passive stretch are applied to all arteries regardless of diameter. This is important since variations in initial length of the smooth muscle cells can alter the range and sensitivity (EC so) of agonist concentration-response curves." In the present work, the human IMA segments had small internal diameters as measured at an equivalent transmural pressure of 100 mm Hg, namely 2.02 mm compared with nearly double the diameter in the canine coronary artery (3.85 mm). Correspondingly, the
resting force that was applied at the optimal length was 4.15 g for the human IMA but 6.87 g for the dog vessel (Table II). The isolated ring segments studied in this way allow precise assay of the isometric force developed by the smooth muscle in response to chosen agents under physiologic conditions without the effect of circulating hormones, neural stimulation, or blood flow. By this method, the endothelium is preserved in isolated ring segments. This was tested previously in our laboratory by the endothelium-dependent relaxation to acetylcho-
line." We used isoproterenol to assay /1-adrenoceptor activity since this agent is considered to be a full agonist at
Volume 97
Weak fJ-adrenoceptor-mediated relaxation in IMA
Number 2 February 1989
26 5
Table ill. Increase in force (g) caused by submaximal concentration of potassium (25 mmoljL) or U46619 (20 nmoljL) before relaxation by isoproterenol Human IMA Canine coronary artery
Potassium
U466I 9
p Value
3.66 ± 0.65 (14) 10.90 ± 0.95 (26)
6.34 ± 1.03 (16) 13.37 ± 1.30 (24)
<0.05
NS
Number of rings indicated in parentheses. NS. Not significantly different.
both fJl-adrenoceptors and fJ2-adrenoceptors. The stimulus that causes IMA spasm during coronary bypass grafting is unknown. In the present work two potential stimuli were used: (1) potassium depolarization, which causes calcium entry into smooth muscle via a voltagedependent calcium channel and also via the release of norepinephrine from sympathetic nerve terminals, and (2) U46619, which is a chemically stable analog of thromboxane A 2. Thromboxane A 2 is a potent vasoconstrictor that is released from aggregating platelets during surgery; elevated plasma levels have been found during cardiopulmonary bypass. 14 Compared with potassium-contracted vessels, the contraction to U46619 in dog coronary artery was more readily relaxed by iJ-adrenoceptor stimulation, even though the starting contractions were not significantly different (Table III). The weak response to isoproterenol in the human IMA was not due to persisting effects of fJ-blockade, since we used only tissues that was free from prior exposure to fJ-adrenoceptor antagonist therapy. In human IMA, propranolol (0.1 umol/L), at a concentration within the range of levels occurring clinically, I 5 prevented the small relaxation (14% to 24%) to isoproterenol. Evidence that these arteries could respond further to a vasodilator agent acting via a non-dadrenoceptor mechanism was the 50% to 60% relaxation to GTN in potassium-contracted rings and greater than 80% relaxation in the U46619-contracted preparations. Therefore, under these conditions, the small relaxation in response to isoproterenol is compelling evidence that the functional fJ-adrenoceptors were sparse in the human IMA. Autoradiography has demonstrated that l1-adrenoceptors are localized, in relatively low concentrations, in the media of the human IMA.6 These functional and receptor density observations suggest that the poor response to isoproterenol is directly caused by a low number of receptors rather than by alterations in the intracellular coupling with adenylate cyclase and the generation of cyclic adenosine monophosphate. By contrast, in the dog coronary artery, which was included in these studies to compare with the reactivity of the human IMA, we found evidence for marked iJ-adrenoceptor activity.
Clinical implications. The human isolated IMA responds poorly to fJ-adrenoceptor stimulation. This implies that perioperatively administered catecholamines or sympathomimetic amines are most unlikely to dilate the IMA. As a corollary, fJ-adrenoceptor antagonists will not uncover constrictor activity on the IMA and are therefore unlikely to induce spasm, as has been reported for coronary arteries." We thank Mrs. Clara Chan for secretarial assistance. REFERENCES 1. 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. 2. Jett GK, Arcidi JM, Dorsey LMA, Hatcher CR, Guyton RA. Vasoactive drug effects on blood flow in internal mammary artery and saphenous vein grafts. J THORAC CARDIOVASC SURG 1987;94:2-11. 3. McCormick JR, Kaneko M, Baue AE, Geha AS. Blood flow and vasoactive drug effects in internal mammary artery and venous bypass grafts. Circulation 1975; 51,52(Pt 2):172-9. 4. Thadani U. Assessment of "optimal" beta blockade in treating patients with angina pectoris. Acta Med Scand I984;suppI694: 178-87. 5. Hjalmarson A. Early intervention with a beta-blocking drug after acute myocardial infarction. Am J Cardiol 1984;54: 11 E-3E. 6. Molenaar P, Malta E, Jones CR, Buxton BF, Summers RJ. Autoradiographic localization of beta-adrenoceptors in human internal mammary artery and saphenous vein. Br J Pharmacol 1988;95:225-33. 7. Angus JA, Cocks TM, Satoh K. a2-Adrenoceptors and endothelium-dependent relaxation in canine large arteries. Br J Pharmacol 1986;88:767-77. 8. He G-W, Angus JA, Rosenfeldt FL. Reactivity of the canine isolated internal mammary artery, saphenous vein and coronary artery to constrictor and dilator substances: relevance to coronary bypass graft surgery. J Cardiovasc PharmacoI1988;12:12-22. 9. Nakashima A, Angus JA, Johnston Cl. Comparison of angiotensin converting enzyme inhibitors captopril and MK421-diacid in guinea pig atria. Eur J Pharmacol 1982;81 :487-92. 10. Anderson R, Holmberg S, Svedmyr N, Aberg G. Adren-
266 He et al.
ergic a- and J3-receptors in coronary artery vesselsin man. Acta Med Scand 1972;191:241-4. 11. Toda N. Isolated human coronary arteries in response to vasoconstrictor substances. Am J Physiol1983;245:H93741. 12. Toda N. Response of isolated monkey coronary arteries to catecholarnines and to transmural electrical stimulation. Circ Res 1981;49:1228-36. 13. Tallarida RJ, Sevy RW, Harakal C, Hendrick J, Faust R.
The Journal of Thoracic and Cardiovascular Surgery
The effect of preload on the dissociation constant of norepinephrine in isolated strips of rabbit thoracic aorta. Arch Int Pharmacodyn Ther 1974;210:67-74. 14. Davies GC, Sobel M, Salzman EW. Elevated plasma fibrinopeptide A and thromboxane 8 2 levels during cardiopulmonary bypass. Circulation 1980;61:808-13. 15. Robertson RM, Wood AJJ, Vaughn WK, Robertson D. Exacerbation of vasotonic angina pectoris by propranolol. Circulation 1982;65:281-5.
Bound volumes available to subscribers Bound volumes of THE JOURNAL OF THORACIC AND CARDIOVASCULAR SURGERY are available to subscribers (only) for the 1989 issues from the Publisher, at a cost of $55.00 ($72.00 international) for Yol. 97 (January-June) and Yol. 98 (July-December). Shipping charges are included. Each bound volume contains a subject and author index and aU advertising is removed. Copies are shipped within 60 days after publication of the last issue of the volume. The binding is durable buckram with the JOURNAL name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact The C.Y. Mosby Company, Circulation Department, 11830 Westline Industrial Drive, St. Louis, Missouri 63146-3318, USA; phone (800) 325-4177, ext. 351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular JOURNAL subscription.