morphine-induced venodilation in humans*

Morphine-induced venodilation in humans

Morphine has been extensively used in the treatment of pulmonary edema, and its action is believed to be mediated in part by its ability to produce peripheral venodilation. This study investigated whether opiates produce venodilation in human hand veins and explored the underlying mechanism(s). Fifteen healthy volunteers ( 11 men and four women) were studied with use of the dorsal hand vein compliance technique. After preconstriction with the selective c~,-adrenergic receptor agonist phenylephrine, dose-response curves were constructed to (1) opiate receptor agonists morphine (1 to 30 pg/min) or fentanyl(O.07 to 1 pg/min), (2) a combination of morphine and the p-opiate receptor antagonist naloxone, and (3) morphine and a combination of histamine (H, and H,) receptor antagonists. Infusion of morphine caused venodilation in a dose-dependent manner, whereas fentanyl did not produce venodilation. Coin&ion of naloxone and morphine impaired the venodilation only slightly. Coinfusion of the H,- and H,- antagonists completely abolished the venodilatory effect of morphine. These results suggest that the venodilatory effect of morphine is mediated through histamine release and that F-opiate receptors have little or no involvement in this process. (Clin Pharmacol Ther 1996;60:554-60.)

Matthias Grossmann, MD, Ademola Abiose, MD, Oranee Tangphao, MD, Terrence F. Blaschke, MD, and Brian B. Hoffman, MD Stanford and Palo Alto, Calif:

Morphine has been one of the most widely used drugs available for the treatment of acute pulmonary edema.’ The capacity of morphine to induce venodilation is one of the properties thought to be beneficial in the treatment of acute pulmonary edema because an increase in venous pooling leads to a reduction in ventricular filling pressures. The mechanism by which morphine induces venodilation in humans is uncertain.2’3 There is considerable evidence that morphine induces smooth muscle relaxation in blood vessels from laboratory animals4-6 and humans.2,7 It has been suggested that the effects of morphine on blood vessels are indirect, being due to stimulation of histamine release.8-12

From the Division of Clinical Pharmacology, Stanford University Medical Center, Stanford, and Geriatric Research, Education and Clinical Center, Veterans Affairs Medical Center, Palo Alto.

Supported by the National Institutes of Health, National Institute on Aging (AG05627-11). Dr. Grossmann was supported by the Boehringer Ingelheim Foundation, Stuttgart, Germany; Dr. Abiose was supported by a Merck Sharp and Dohme Interna- tional Fellowship, Merck Fund, Whitehouse Station, N.J.; and Dr. Tangphao was supported by the Anandhamahidol Foun- dation, Bangkok, Thailand.

Received for publication April 29, 1996; accepted July 10, 1996. Reprint requests: Terrence F. Blaschke, MD, Stanford University

Medical Center, Room S-169, Stanford, CA 943055113. 13/1/76396

554

Morphine is known to cause the release of histamine in vivo,8-12 including histamine originating from perivascular mast cells.13 Intradermally admin- istered morphine given to humans produces the “tri- ple response,” the same response known to be characteristic for histamine release; namely, redness, flare, and wheal.14 Morphine selectively induces histamine release from isolated human skin and car- diac mast cells but not from human lung mast cells in vitro.13,15 Opiates differ in their capacity to induce the release of histamine.8,15 Stellato et a1.16 showed in mast cell preparations in vitro that morphine is a potent histamine releasing compound, whereas fentanyl, a synthetic opiate, does not induce histamine release.

The purpose of this study was to investigate opiate-induced venodilation with use of the human hand vein as a model and to explore the underlying mechanisms. The significance of studying opiate- induced venodilatory effects in particular is based on the importance of the capacitance blood vessels in controlling the venous return to the heart.

MATERIAL AND METHODS Subjects. Fifteen healthy volunteers (11 men and

four women), 20 to 49 years old (mean 2 SD age, 34 -+ 4 years) took part in this study. There were 11 white subjects, three black subjects, and one His-

CLINICAL PHAKMACOLOGY & THERAPEUTICS VOI.UMF ho, NUiMBER 5 Grosswumn et al. 555

panic subject. Their weight ranged from 48 to 88 kg (mean + SD weight, 70 ‘-t 10 kg). They were all found to be healthy as determined by complete physical examination, electrocardiogram, and rou- tine laboratory tests (SMA-20, complete blood count, and urinalysis). Each subject gave informed written consent before taking part in the study, which was approved by the Stanford Administrative Panel on Human Subjects in Medical Research (Stanford, Calif.). Exclusion criteria included a his- tory of any significant disease state, use of illicit drugs, alcoholism, or long-term use of any medica- tion, including over-the-counter drugs. All subjects were nonsmokers except for two subjects; subjects were asked to refrain from caffeine and smoking for at least 12 hours before the study.

Dorsal hand-vein technique. The dorsal hand vein compliance technique was used as we have previ- ously described in detail.17 This technique has the advantage of allowing infusion of very small amounts of vasoactive substances, thus avoiding po- tential confounding systemic hemodynamic effects.

after inflation of a blood pressure cuff positioned on the upper portion of the same arm to 45 mm Hg. Doses of drugs are expressed as rates of infusion (in nanograms per minute). Each dose was infused for a minimum of 7 minutes; the cuff was inflated for 2 minutes at intervals during each infusion period. The dilation of the vein at baseline after 30 minutes of saline infusion was defined as 100% dilation. Phenylephrine, a selective cw,-adrenergic receptor agonist, was used to preconstrict the hand vein. Phenylephrine was infused in the dose range of 50 to 7917 ng/min and the dose of phenylephrine that produced 80% constriction was determined. This degree of preconstriction is defined as 0% dilation. This dose of phenylephrine was then infused at a constant rate during the subsequent administration of all other ad- ministered drugs throughout the study. Previous studies have shown that phenylephrine-induced venocon- striction is stable during this time period.17

Subjects were studied while they were in a supine position in a quiet room kept at a constant temper- ature of 72” -+ 2” F. An arm was placed on a padded arm rest with an upward angle of 30 degrees from the horizontal to allow complete emptying of the superficial veins. A suitable vein was chosen on the dorsum of the hand, and a 21-gauge needle was inserted. Normal saline solution (150 mmol/L so- dium chloride) was infused for at least 30 minutes to allow for equilibrium of the vein after initial vaso- constriction caused by the needle insertion. The flow rate was kept constant throughout the study at 0.3 ml/min by use of a Harvard infusion pump (model 975, Harvard Apparatus Inc., South Natick, Mass). A small tripod, holding a linear variable differential transducer (LVDT, model 100 MHR Schaevitz Sen- sor, Lucas Control System Products, Hampton, Va.), was mounted on the back of each subject’s hand with the central aperture of the LVDT over the top of the vein under investigation at a distance of 10 mm downstream from the tip of the needle. The central aperture of the LVDT contains a freely movable metallic core that weighs 0.5 gm. When the core is properly centered within the transducer, there is a linear relationship over the range used between the vertical movement of the core and voltage output, which is recorded on a strip chart recorder.

Study design. After an 80% preconstriction of the vein was obtained with phenylephrine, a dose- response curve to morphine was constructed (doses ranging from 968 to 30,000 ng/min). A washout period of 30 minutes followed; during this time the vein was preconstricted again to 80% and a second response to the maximal dose of morphine (30,000 ng/min) was obtained in five subjects to demonstrate the reproducibility of the venodilatory response to morphine.

To investigate the hypothesis that opiate-induced vasodilatation does not parallel analgesic potency, we investigated the venodilatory effect of a synthetic opiate, fentanyl, which belongs to the piperidine group. In those studies, carried out under similar conditions, dose-response curves were generated to fentanyl (66 to 1050 ng/min) and morphine (968 to 30,000 ng/min) either in the same session (four subjects) or on a different occasion (three subjects).

To evaluate the role of the k-opiate receptors in mediating morphine-induced venodilation, we infused the opiate receptor antagonist naloxone in 10 subjects (30,000 ngimin). Naloxone was preinfused for 25 minutes, followed by coinfusion of naloxone (30,000 g/ ’ ) d n mm an morphine (30,000 ng/min) for 10 minutes.

An estimate of the shift in the dose-response curve to morphine in the presence of naloxone was made using the dose ratio (DR) according to the following equation:

Each reading was taken by measuring the differ- ence between the position of the core before and

k d

= I_@-tagonist concentration] [DR - l]

556 Grossmann etal. CLINICAL PHARMACOLOGY &THERAPEUTICS

NOVEMBER 1996

in which k, is the dissociation constant of the nal- curve-fitting program provides an objective measure oxone, and dose ratio (DR) is the ratio of the con- of the maximal response (E,,) and of the dose at centration producing the half-maximal response which the response was 50% of the maximal re- (EC,,) of the agonist (morphine) in the presence of sponse (EDso). A log-transformation was performed antagonist (naloxone) and the EC,, of the agonist at base1ine.i’

on individual ED,, values to obtain the geometric mean.

At an infusion rate for naloxone of 30 kg/min, given an estimated local venous blood flow of approximately 3 ml/min (unpublished observation, Erny, Hoffman, and Blaschke, July 1996), the local venous concentration of naloxone achieved was 30 pmol/L.

Because the kd for naloxone at the p-receptor is in the nanomolar range (0.8 to 6.0 nmol/L),” and given our estimated baseline EC& of 5 kmol/L for morphine (see below), the calculated EC,, in the presence of naloxone is >25 X lo3 kmol/L. Thus there is effective blockade of the l.r=opioid receptor effect of morphine at the infusion rate of naloxone used in these studies.

A Student paired two-tailed t test was used to compare the individual venodilatory responses before and after coinfusion of morphine with naloxone or with diphenhydramine and famotidine, respec- tively. Results are expressed as mean + SD unless otherwise stated. A value ofp < 0.05 was considered to be significant.

RESULTS

To test the hypothesis that morphine-induced venodilation is mediated through the release of histamine, a combination of Hi- and Hz-receptor antagonists were given in the presence of morphine. In nine subjects we preinfused the Hi-blocker diphenhydramine (7.5 &min) and the Hz-receptor blocker famotidine (45 t&min) for 30 minutes. The venodilatory response to morphine (30,000 ng/min) was then determined in the presence of both Hi- and Hz-receptor antagonists. The doses of diphenhydramine and famotidine used in this study were shown by our group in earlier experiments to inhibit the venodilatory response to exogenously infused histamine.20

Infusion of morphine into phenylephrine- preconstricted veins induced venodilation in a dose- dependent manner. The venodilatory response to the highest morphine dose (30,000 ng/min) ranged from 41% to 101% (77% + 19%, 12 = 10). Fig. l,A, illustrates a representative morphine dose-response curve. The individual E,, and ED,, values for all the 10 subjects are provided in Table I. The E,, for the 10 subjects was 77% + 19% (mean + SD). The mean log ED,, was 4.1 f 0.10 (arithmetic value, 11,500 ng/min) and the EC,, based on local blood flow was 5 kmol/L. The morphine infusion not only induced venodilation, but also produced, to varying degrees, itching and redness in and around the vein being infused.

There was no change in the blood pressure and heart rates of the study subjects during the infusion of these doses of phenylephrine, morphine, naloxone, and antihistamines (data not shown).

Material. All drugs were diluted in normal saline solution. The following drugs were used: phenylephrine hydrochloride (1% injection; Winthrop Labo- ratories, New York, N.Y.), morphine sulfate (Elkins-Sinn, Inc., Cherry Hill, N.J.), fentanyl citrate (Janssen Pharmaceutics Inc., Titusville, N.J.), phenol-free famotidine (Smith Kline Beecham Lab- oratories, Philadelphia, Pa.), diphenhydramine (Wyeth Laboratories Inc., Philadelphia, Pa.), and naloxone hydrochloride (DuPont Pharmaceuticals, Manati, Puerto Rico).

The response to the maximal dose of morphine was measured twice in five subjects, separated by an interval of 30 minutes; during the 30-minute interval morphine was washed out to determine the reproducibility of morphine-induced venodilation. The maximum venodilation after the second infusion of 30,000 ng/min morphine in five subjects was not significantly different from the response to the same morphine dose in the initial dose-response curve performed 30 minutes earlier (68% + 21% versus 74% ? 22%,p = NS, n = 5). This result indicates the absence of acute desensitization of morphine- induced venodilation and demonstrates reproducibility of the response to morphine.

Data analysis. Individual dose-response curves to morphine were fitted to a four-parameter logistic equation with the ALLFIT program.21 This iterative

Seven subjects from the first experiment were studied with fentanyl on a separate day. Their maximal vasodilatory response was 7% + 23% compared with 77% ? 19% in the morphine group. Fig. 1, B, illustrates the dose-response curve to the infusion of the piperidine opiate fentanyl in a subject responding to morphine. Furthermore, none of the subjects reported clinical manifestations of hista-

(:l.INICAL PHARMACOLOGY &THERAPEUTICS VOLLIME 60, NUMRER 5 Grossmann et al. 557

50

Fentanyl dose (ng/min)

300

80 -

O-

I I I I

2000 10000 30000

Morphine dose (ng/min)

Fig. 1. Morphine- and fentanyl-induced venodilation in a representative single subject. A dorsal hand vein was preconstricted with a dose of phenylephrine producing 80% constriction of the vein (ED& as described in the Methods section. In this subject, morphine (A) markedly induced venodilation with a maximal response (E,,) of 91% and the dose producing a half-maximal response (ED5,,) of 12,600 nglmin. Fentanyl (B) did not induce venodilation. This experiment was repeated nine times for morphine (Table I). The dose-response curve for fentanyl was repeated six times with similar results.

mine release, such as itching or redness, in the vein under investigation at any fentanyl dose given.

Pretreatment with naloxone, the p-opiate receptor antagonist, and coinfusion of naloxone and morphine (30,000 ng/min) reduced the venodilation only slightly compared to the response to morphine in the absence of naloxone (80% t 21% versus 66% t 22%,p < 0.05, n = 10). However, pretreatment and coinfusion with the Hi-antagonist diphenhydramine (7.5 pg/min) and the Hz-antagonist famotidine (45 pg/min) completely abolished the venodilatory effect of morphine (30,000 ng/min) (73% t 20% versus 2% f- 4%,p < 0.0001, n = 9), as shown in Fig. 2. This suggests that the venodilatory effect of morphine is mediated through the release of histamine. Fentanyl produced no venodilation at the doses infused (66 to 1050 ng/min, y1 = 6).

DISCUSSION Increased histamine levels in plasma or serum

after systemic administration of morphine has pre- viously been shown in humans,1’312 but the role of

Table I. Responsiveness to the hand vein to morphine in 10 normal subjects

E log ED,, Subject No. (tq (nglmin)

1 91 4.1 2 53 4.1 3 106 4.1 4 52 4.1 5 76 4.2 6 78 4.1 7 80 4.0 8 50 3.9 9 74 3.9

10 101 4.1

Geometric mean 77 -+ 19 4.1 2 0.1 -t SD

Antilog value - 11,500 -~

E,,, Maximum response: ED,,, dose producing half-maximal effect.

this mediator in causing hemodynamic effects is con- troversial. In normal volunteers, although there was a fourfold increase in plasma histamine, no significant hemodynamic changes were observed after

558 Grossmann et al. CLINICAL PHARMA COLOGY & THERAPEUTICS

NOVEMBER 1996

100

80

Mor Nal Nal + Mor D+F D+F+Mor

Fig. 2. Effects of opiate antagonist and HI- and HZ-antagonists on morphine-induced venodilation. The effect of morphine (Mor, 30 t.&min); naloxone (Nal, 30 pg/min); naloxone (Nal, 30 kg/min) and morphine (Mor, 30 p,g/min); diphenhydramine (D, 7.5 t.&min) and famotidine (F, 45 Fg/min); and diphenhydramine (D, 7.5 kg/min), famotidine (F, 45 pg/min), and morphine (Mor, 30 pg/min) were measured in dorsal hand veins.

doses of 0.15 mg/kg; it was suggested that histamine that this can be ameliorated, at least in part, by was released mainly from cutaneous mast cells.i2 In naloxone infusion.23,24 Blockade of Hi- and HT a study carried out in patients undergoing surgery in receptors by coinfusion of diphenhydramine and which morphine was used as the analgesic compo- famotidine completely antagonized the venodilation nent of balanced anesthesia, histamine concentra- in our study, indicating that histamine is the final tions in serum increased approximately fourfold af- mediator of the venodilation induced by morphine. ter administration of 0.3 mg/kg morphine, and this Earlier experiments by our group showed that exog- was associated with decreased systemic vascular re- enous infusions of histamine activated both H,- and sistance.ll H,-receptors in promoting venodilation.20

The objective of this study was to measure directly venous responsiveness to opiates in humans and to characterize the underlying mechanism of morphine-induced venodilation. The dorsal hand vein technique was used because local responsiveness, independent of systemic and reflex responses, can be measured by infusion of very small quantities of drug. The results show that morphine acts as a venodilator in human hand veins in vivo, likely in- directly mediated by histamine release.

All subjects reported itching over and around the hand vein under investigation, which was subjec- tively augmented with increasing doses of morphine. This finding is in agreement with the observation of other workers,14 as well as the clinical observation that pruritus may develop in patients receiving morphine.22 There is also a growing body of evidence that supports the hypothesis that endogenous opioid agonists contribute to the pruritus of cholestasis and

Naloxone, an opiate receptor antagonist with high affinity for the p,-receptor, caused a small but statis- tically significant reduction in morphine-induced venodilation. Therefore some involvement of the p-opiate receptor in the morphine venodilator response seems to be present. There is evidence that p-opiate receptors can promote histamine release in some ce11s,5Y25 whereas in other cells this response appears independent of p-opiate receptors.8P13T15 In- terestingly, fentanyl, a p-opiate receptor agonist, does not cause vasodilation or histamine release from skin mast cells in patients undergoing elective coronary bypass surgery.8915 Fentanyl is known to have twentyfold to eightyfold higher affinity for k-opiate receptors and is more potent in terms of analgesic effect than morphine.8,26 Fentanyl produces little change in peripheral vascular tone and systemic concentrations of histamine in humans.7’8 In view of the much higher affinity of fentanyl for

(;I,INICAL PHARMACOLOGY &THERAPEUTICS VOLUME 60. NUMBEK 5

k-opiate receptors, the fentanyl infusion rate in our study, 30 times lower than the maximal morphine infusion rate, should produce approximately equipotent k-opiate receptor stimulation. Fentanyl can cause wheal and flare responses in the skin without any evidence of mast cell degranulation.14 Levy et a1.14 demonstrated that intradermal injection of morphine, fentanyl, meperidine (pethidine), and sufentanil produced variable wheal and flare responses. Electron microscopy of the biopsies from the wheals produced by fentanyl showed normal mast cell architecture with no evidence of mast cell degranulation. However, in our experiments fentanyl did not evoke any vasodilatation in the hand vein and it did not produce localized itching or urticaria at the maximum dose infused. These findings are consistent with those of Hermens et al.,” who found that fentanyl failed to release histamine at any concentration in human skin mast cells.

We found no evidence that naloxone infusion by itself produces any effect on the venous tone. Con- sistent with our findings, in skin mast cell preparations, naloxone (5 X 10e4 mol/L) neither released histamine nor inhibited morphine-induced histamine release.15 Considering the heterogeneity of human mast cells, it is reasonable to use skin mast cell as the closest model to vascular mast cells.‘6*27

Our findings, along with earlier literature data, show that mechanism of histamine release by morphine in hand veins is unlikely to be mediated by k-opiate receptors. There has been speculation about the differing capacities of opiates to release histamine.*,15 Opiates from the same chemical class differ in their ability to induce histamine release from mast cells, even at very high doses.” An im- portant characteristic of histamine-releasing opiates seems to relate to positive charge and hydrophobic- ity.i3 Recent data have suggested that other histamine-releasing compounds may directly acti- vate G-proteins, leading to an increase in histamine release through increased phosphoinositol metabo- lism and an increase in intracellular calcium concentrations.28 Dynorphin has been shown to enhance G-protein activity in reconstituted vesicles.29 Al- though it remains speculative, some opiates may act directly on G-proteins without stimulating opiate receptors to promote histamine release from mast cells.

Our study shows that two opiate agonists, morphine and fentanyl, differ in their ability to relax human hand veins in vivo. The results suggest that morphine induces venodilation through release of

Grossmann et al. 559

histamine because histamine receptor blockade abolishes the venodilation completely. Fentanyl had no venodilatory activity at equipotent p-opiate receptor doses. The mechanism of morphine-induced histamine release remains unknown, and it remains speculative whether morphine interacts directly with a G-protein or whether another opiate receptor not fully antagonized by naloxone is involved.

We thank the nurses and staff of the Aging Study Unit at the Veterans AII’airs Medical Center for their assis- tance.

References 1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

Gazes P. Treatment of acute pulmonary edema. Heart Dis Stroke 1994;3:205-9. Zelis R, Mansour E, Capone R, Mason D, KIeckner R. The cardiovascular effects of morphine. .I Clin Invest 1974;54:1247-58. Lowenstein E, Whiting R, Bittar D, Sanders C, Powell W Jr. Local and neurally mediated effects of morphine on skeletal muscle vascular resistance. J Phar- macol Exp Ther 1972;180:359-67. Evans A, Nasmyth P, Stewart H. The fall of blood pressure caused by intravenous morphine in the rat and the cat. Br J Pharmacol 1952;7:542-52. El Sharkawy T, AI Shireida M, Pilcher C. Vascular effects of some opioid receptor agonists. Can J Physiol Pharmacol 1991;69:846-51. Henney R, Vasko J, Brawley R, Oldham H, Morrow A. The effects of morphine on the resistance and capacitance vessels of the peripheral circulation. Am Heart J 1966;72:242-50. Popio K, Jackson D, Ross A, Schreiner B, Yu P. Hemodynamic and respiratory effects of morphine and butorphanol. Clin Phamacol Ther 1978;23:281-7. Rosow C, Moss J, Philbin D, Savarese J. Histamine release during morphine and fentanyl anesthesia. An- esthesiology 1982;56:93-6. Withington D, Patrick J, Reynolds F. Histamine release by morphine and diamorphine in man. Anaes- thesia 1993;48:26-9. Moss J, Rosow C. Histamine release by narcotics and muscle relaxants in humans. Anesthesiology 1983;59: 330-9. Fahmy N, Sunder N, Soter N. Role of histamine in the hemodynamic response and plasma catecholamine response to morphine. Clin Pharmacol Ther 1983;33: 615-20. Doenicke A, Moss J, Lorenz W, Hoernecke R. Intra- venous morphine and nalbuphine increase histamine and catecholamine release without accompanying hemodynamic changes. Clin Pharmacol Ther 1995;58: 81-9.

13. Barke K, Hough LB. Opiates, mast cells and histamine release. Clin Sci 1993;53:1391-9.

560 Grossmann et al. CLINICAL PHARMACOLOGY &THERAPEUTICS

NOVEMBER 1996

14. Levy J, Brister N, Shearin A, Ziegler J, Hug C, Ad- elson DM, et al. Wheal and flare responses to opioids in humans. Anesthesiology 1989;70:756-60.

15. Hermens J, Ebertz J, Hanifin J, Hirshman C. Com- parison of histamine release in skin mast cells induced by morphine, fentanyl, and oxymorphone. Anesthesi- ology 1985;62:124-9.

16. Stellato C, Cirillo R, dePaulis A, Casolaro V, Patella V, Mastronardi P, et al. Human basophihmast cell releasability; IX: heterogeneity of the effects of opioids on mediator release. Anesthesiology 1992;77: 932-40.

17. Pan H, Hoffmann B, Pershe RA, Blaschke TF. De- cline in beta adrenergic receptor-mediated relaxation with aging in man. J Pharmacol Exp Ther 1986;239: 802-7.

18. Kenakin TP. Pharmacologic analysis of drug-receptor interactions. New York: Raven Press, 1987:205-44.

19. Hustvet 0, Setekleix J, Oye I. Naloxone counteracts fast development of tolerance to morphine in guinea pig ileum. Pharmacol Toxic01 1992;71:224-8.

20. Dachman W, Bedarida G, Blaschke TF, Hoffman BB. Histamine-induced venodilation in human be- ings involves both Hl and H2 receptor subtypes. J Allergy Clin Immunol 1994;93:606-14.

21. De Lean A, Munson P, Rodbard D. Simultaneous analysis of families of sigmoidal curves: application to bioassay, radioligand assay and physiological dose-response curves. Am J Physiol 1978;235:E97- E102.

22. Ballantyne J, Loach A, Carr D. Itching after epidural and spinal opiates. Pain 1988;33:149-60.

23. Bergasa N, Talbot T, Alling D, Schmitt JM, Walker EC, Baker BL, et al. A controlled clinical trial of naloxone infusions for the pruritus of chronic cholestasis. Gastroenterology 1992;102:544-9.

24. Jones E, Bergasa N. The pruritus of cholestasis and the opioid system [clinical conference]. JAMA 1992; 268~3359-67.

25. Chikai T, Oishi R, Saeki K. Increase in the extracel- lular histamine concentration in the rat striatum by mu-opioid receptor activation. J Neurochem 1994;62: 724-9.

26. Jaffe J, Martin W. Opioid analgesics and antagonists. Goodman & Gilman’s The pharmacological basis of therapeutics. 8th edition. New York: Pergamon Press, 1984:485-521.

27. Patella V, de Crescendo G, Ciccarelli A, Marino I, Adt M, Marone G. Human heart mast cells: a defin- itive case of mast cell heterogeneity. Int Arch Allergy Immunol 1995;106:386-93.

28. Mousli M, Bronner C, Bueb J, Landry Y. Evidence for the interaction of mast cell-degranulating peptide with pertussis toxin-sensitive G proteins in mast cells. Eur J Pharmacol 1991;207:249-55.

29. Higashijima T, Burnier J, Ross E. Regulation of Gi and Go by mastoparan, related amphophilic peptides, and hydrophobic amines. Mechanism and structural determinants of activity. J Biol Chem 1990;265:14176- 86.

morphine-induced venodilation in humans*

Documents