Pulmonary Pharmacology (1992) 5, 159-165 PULMONARY PHARMACOLOGY
Opiate Action in the Pulmonary Circulation T. S. Hakim*, M. M. Grunstein, R. P. Michel Department of Surgery, SUNY Health Science Center, Syracuse, New York, Children's Hospital of Philadelphia, University of Pennsylvania, Philadelphia, Pennsylvania, USA, and Department of Pathology, McGill University, Montreal, Quebec, Canada
SUMMARY: To gain insight into the mechanisms underlying the association between acute pulmonary edema and narcotic abuse, the direct action of morphine was examined in isolated, perfused left lower lobe (LLL) preparations in dogs and cats . Morphine sulphate injected (0 .6 mg/kg) into the pulmonary artery of the LLL increased the pulmonary vascular resistance (PVR) by about 100% in both species . The increase in PVR was primarily due to constriction of the veins, as determined with the arterial and venous occlusion technique . The increase in PVR with morphine injection was unaffected by a-adrenergic antagonists, but was reversed by chlorpheniramine, a histamine Ht -receptor antagonist . Pretreatment, but not post-treatment with the opiate antagonist, naloxone, blocked the effect of morphine on PVR . Thus, the rapid administration of morphine produces pulmonary venoconstriction via histamine release from the lung, and the latter may account for the well-documented association between acute pulmonary edema and narcotic abuse.
INTRODUCTION
avoided and changes in vascular resistance can be detected and localized precisely .
Heroin-induced pulmonary edema or `heroin lung' was first described by Osler in 1980 .' Although initially described after the intravenous injection of heroin, acute pulmonary edema has also been reported with the abuse of other narcotic substances including morphine, opium, methadone, and other agents .',' The pulmonary edema associated with these substances has been attributed to increased pulmonary capillary permeability,4'5 but this remains largely unconfirmed . Paradoxically, large doses of intravenous morphine (0 .5-3 .0 mg/kg) are being used in the clinical setting for anesthetic purposes,' and to treat acute pulmonary edema.' In the latter situation, the beneficial effects of morphine have been attributed to its peripheral (venous and arteriolar) vasodilatory properties ." Most studies on this subject have dealt primarily with the systemic effects of opiates . 6 " Indeed, there are no studies which have systematically characterized the direct effects of opiates on the pulmonary circulation, independent of systemic humoral and neural influences . To fill this gap in present knowledge, the direct effect of morphine on the pulmonary vasculature in isolated lungs was examined, where humoral and neural influences are
MATERIALS AND METHODS Thirteen dogs (20.8 ± 1 .3 kg) and 25 cats (3 .61 ± 0 .12 kg) were studied . The dogs were anesthetized with sodium pentobarbital (25 mg/kg i .v .) and the left lower lobe (LLL) was exposed through a left thoracotomy . After systemic heparinization (700 units/kg i .v .), the LLL was isolated and perfused in situ with heated autologous blood (37°C) from an extracorporeal perfusion system containing 200 ml of blood . This preparation has been fully described elsewhere . 12 The right lung and the left lower lobe were ventilated separately using two Harvard respirators, and the dog was kept alive by ventilating the right lung with room air. The LLL was ventilated with 5% C02/35% 02 /60% N2 to maintain blood gases in the perfusion system normal [PCO2 = 42 .3 ± 1 .7 (SE) mmHg, P0 2 = 220 ± 20 mmHg and pH = 7 .35 ± 0 .01] . The hematocrit in the perfusion system was 42 .3±2 .2% . The only communication between the dogs' systemic blood and the perfusion system was through the bronchial circulation of the LLL . The nerves to the LLL remain relatively intact, 13 and presumably the lymphatics are functional . In this preparation, substances injected into the LLL artery
* For correspondence and reprint requests at : Department of Surgery, SUNY Health Science Center, 750 E . Adams Street, Syracuse, New York 13210, USA . 0952-0600/92/030159 + 07 $08 .00/0
159
© 1992 Academic Press Limited
1 60
T . S. Hakim et al
do not reach the animal's circulation and vice versa . The blood flow rate to the LLL was set at 522 ± 48 ml/ min, to give a pulmonary arterial pressure (Pa) of 1015 mmHg . The venous pressure (Pv) was kept constant at 1 .3 ± 0 .2 mmHg relative to the top of the lobe . The lobe was ventilated with a constant tidal volume of 150 ml with an end expiratory airway pressure of 12 mmHg. Since the flow rate and venous pressure were kept constant, changes in Pa reflected alterations in pulmonary vascular resistance . The airway pressure of the LLL (Pals) was also monitored as an index of bronchomotor tone . Cats were also anesthetized with sodium pentobarbital (25 mg/kg i.p .) . The LLL was studied as in the dogs, but with slight modifications :14 the cats were exsanguinated to obtain an adequate volume to prime the perfusion system . Thus, the LLL was perfused in situ, but the animal was not alive . The hematocrit in the perfusion system was 31 .7± 1 .5% . Because of the small size of the animal, one tube was placed in the trachea and both the LLL and the right lung were ventilated with 5% C0 2/35% 02/60% N 2. The tidal volume was adjusted to keep peak inspiratory pressure below 10 mmHg, and end expiratory pressure set at 1-2 mmHg . As in the canine LLL, the venous pressure (Pv) was kept constant at 1 mmHg and flow rate was set at a constant value of 54 .4 ± 4 .4 ml/min, so that changes in Pa indicated changes in PVR . Changes in airway pressure were monitored but were not analysed because the LLL and the right lung were ventilated together . Perfusate blood gases and pH in the cats were : PCO 2 = 35 .8±0 .8 mmHg ; P02 =164±6 mmHg and pH=7 .38±0 .01 . Following a period of stabilization (30 min), the PVR in seven dogs and 13 cat LLLs was partitioned into three segments (arterial, middle, venous) using the arterial and venous occlusion technique ." The effect of various interventions on the partitioning of
PVR was examined to identify the site of change in vascular resistance . The substances and their dosages tested in this study included the opiate agonist morphine sulphate (0 .6 mg/kg ; Abbott, Montreal, Canada), the opiate antagonist, naloxone hydrochloride (0 .3 mg ; Narcan, Endo Labs, Montreal, Canada), the histamine H,-receptor antagonist, chlorpheniramine maleate (5 mg; Chlor-Tripolon, Schering, Kenilworth, NJ, USA), the a-adrenergic receptor antagonist, phentolamine mesylate (5 mg ; Rogitine CIBA, Summit, NJ, USA), and a non-specific vasodilator, papaverine (1 mg ; Sigma, St Louis, MO, USA) . These agents, abbreviated respectively as M, N, C, Ph and P, were injected in 0 .5 to 1 ml boluses over 3-5 s directly into the arterial inflow cannula . None of the antagonist drugs had any significant effect on baseline pulmonary vascular resistance in the isolated lung preparation when used alone (data not shown) . The changes in total pulmonary vascular resistance, and in the resistance of individual vascular segments were tested for statistical significance, by comparing the means during vasoconstriction with the means during baseline conditions using Student's paired ttest . Differences were considered significant for Pvalues of less than 0 .05. All values are expressed as mean ± SE .
RESULTS Figure 1 demonstrates the pulmonary vascular responses to morphine injection in six different cats . The administration of morphine produced a rise in pulmonary artery pressure, and although the magnitude of the response varied considerably between animals, it always peaked within about 2 min . The responses in canine LLL were very similar . The baseline and peak pulmonary artery pressure responses obtained in all
40
0
A M 0
A M
M
I min
Fig. 1 Typical pressor response tracings following injection of 0 .6 mg/kg morphine (M) in left lower lobes of seven different cats . Tracings in canine left lower lobes were very similar .
Opiate Action in Pulmonary Circulation
16 1
40 rn E E 30 v tR v n • 20 al v T
C C
E
10
a
0
Control
Morphine
Control
Morphine
Fig. 2 Pulmonary arterial pressure in the individual lobes of cat lungs (A) and dog lungs (B) during control and at peak pressor response following morphine injections. ( •) Mean value . 40 Arterial
Venous
o -rn v = • EE a-
a
Is 0
I
40 0
•
E E
A
W
• C O
q E o S a
n
I min
w 0
i
V . 7 S m E
•
E
Arterial
Venous
0
Fig. 3 Pressure tracings during the arterial and venous flow occlusion in one canine LLL just prior to morphine injection (bottom) and at near peak pressor response (top) . The middle trace shows the change in pulmonary artery pressure . The tracings allow one to partition the total vascular resistance (or pressure gradient) into arterial, middle, and venous components .
cat and dog LLL are shown in Figure 2 . The intensity and course of the response was quite variable . The lobes in which the response was small recovered spontaneously, and the Pa returned to baseline, while in those with a marked response, acute pulmonary edema with froth in the airways ensued within 20-
30 min . The mean pulmonary artery pressure rose from 11 .6 ± 0 .6 to 21 .9 ± 2 .0 mmHg in the feline lungs and from 12.5 ± 0.5 to 25 .3 ± 2 .0 mmHg in the canine lungs . The site of vasoconstriction was examined with the arterial and venous occlusion . Figure 3 shows the pulmonary artery pressure in one canine LLL (middle
1 62
T . S . Hakim et al
trace), and pressure tracings during the arterial and venous occlusion during baseline conditions (bottom tracings) and following vasoconstriction by morphine (top tracings) . The shape of the arterial pressure profile following arterial occlusion did not differ before (bottom tracing) and after (top tracing) vasoconstriction . In contrast the venous pressure tracings were markedly different, and suggested that a large increase in the resistance of the veins had occurred . The mean pressure gradients across the arterial, middle and venous segments under control conditions and following the injection of morphine are shown in Figure 4 . The arterial pressure gradient did not change significantly in either canine or feline lobes, whereas the venous pressure gradient rose markedly in both species (P < 0 .05) . There were also variable but not significant increases in the middle pressure gradient, most likely due to constriction of the small postcapillary veins, which are within the middle segment .
A
(n=7)
i E E
e
n=15)
The pressure profile along the three vascular segments in dog and cat lungs during control and following morphine injection are plotted in Figure 5 . Note that vasoconstriction causes the pressure in the capillaries within the middle segment to rise markedly in both species . Figure 6 illustrates examples of the effect of posttreatment with the opiate receptor antagonist naloxone (N), and the histamine H,-receptor antagonist, chlorpheniramine (C). Naloxone had minimal effect on the morphine-induced vasoconstriction, whereas chlorpheniramine invariably reversed it . In other experiments (not shown), pretreatment with naloxone attenuated the response to morphine : in three lungs the response was entirely abolished by naloxone, while in two other lungs, the response was not affected . Similarly pretreatment with chlorpheniramine (n = 3) also attenuated the response to morphine in dog lungs: the response was abolished in two lungs but there was a pressor response of 6 mmHg in the third lung . Phentolamine had no effect on the morphine induced vascular response, whether given before (n = 3) or after morphine (n = 3) . In three canine lungs, the effects of morphine were examined twice within 60 min . The injection of a second dosage of morphine (0 .6 mg/kg), following recovery from the first injection, produced no response even though the baseline vascular pressures were nearly the same . This suggests that tachyphylaxis or depletion of the involved mediators had occurred . Indeed in preliminary experiments, it was not possible to produce an acceptable dose-response relationship between the morphine and pulmonary artery pressure, starting at dosages under 0 .3 mg/kg . This lack of success may have been due to specific tachyphylaxis to morphine, since these lungs still responded to other vasoactive substances such as serotonin . To demonstrate that morphine acted independently of blood cells, two feline lungs were perfused with autologous plasma : the effect of morphine in these two lungs was indistinguishable from the blood perfused lung, and the Pa rose from 7 to 18 mmHg in one lung and from 11 to 13 mmHg in the other .
DISCUSSION
Control
Morphine
Fig. 4 Distribution of pressure gradient among the arterial (D), middle (0), and venous (®) segment during control and during morphine induced vasoconstriction in dog lungs (A) and cat lungs (B). Values are mean t SE. Note the large increase in venous pressure gradient signifying marked increase in venous resistance . (•) Significant differences from control; P
Opiate Action in the Pulmonary Circulation T. S. Hakim*, M. M. Grunstein, R. P. Michel Department of Surgery, SUNY Health Science Center, Syracuse, New York, Children's Hospital of Philadelphia, University of Pennsylvania, Philadelphia, Pennsylvania, USA, and Department of Pathology, McGill University, Montreal, Quebec, Canada
SUMMARY: To gain insight into the mechanisms underlying the association between acute pulmonary edema and narcotic abuse, the direct action of morphine was examined in isolated, perfused left lower lobe (LLL) preparations in dogs and cats . Morphine sulphate injected (0 .6 mg/kg) into the pulmonary artery of the LLL increased the pulmonary vascular resistance (PVR) by about 100% in both species . The increase in PVR was primarily due to constriction of the veins, as determined with the arterial and venous occlusion technique . The increase in PVR with morphine injection was unaffected by a-adrenergic antagonists, but was reversed by chlorpheniramine, a histamine Ht -receptor antagonist . Pretreatment, but not post-treatment with the opiate antagonist, naloxone, blocked the effect of morphine on PVR . Thus, the rapid administration of morphine produces pulmonary venoconstriction via histamine release from the lung, and the latter may account for the well-documented association between acute pulmonary edema and narcotic abuse.
INTRODUCTION
avoided and changes in vascular resistance can be detected and localized precisely .
Heroin-induced pulmonary edema or `heroin lung' was first described by Osler in 1980 .' Although initially described after the intravenous injection of heroin, acute pulmonary edema has also been reported with the abuse of other narcotic substances including morphine, opium, methadone, and other agents .',' The pulmonary edema associated with these substances has been attributed to increased pulmonary capillary permeability,4'5 but this remains largely unconfirmed . Paradoxically, large doses of intravenous morphine (0 .5-3 .0 mg/kg) are being used in the clinical setting for anesthetic purposes,' and to treat acute pulmonary edema.' In the latter situation, the beneficial effects of morphine have been attributed to its peripheral (venous and arteriolar) vasodilatory properties ." Most studies on this subject have dealt primarily with the systemic effects of opiates . 6 " Indeed, there are no studies which have systematically characterized the direct effects of opiates on the pulmonary circulation, independent of systemic humoral and neural influences . To fill this gap in present knowledge, the direct effect of morphine on the pulmonary vasculature in isolated lungs was examined, where humoral and neural influences are
MATERIALS AND METHODS Thirteen dogs (20.8 ± 1 .3 kg) and 25 cats (3 .61 ± 0 .12 kg) were studied . The dogs were anesthetized with sodium pentobarbital (25 mg/kg i .v .) and the left lower lobe (LLL) was exposed through a left thoracotomy . After systemic heparinization (700 units/kg i .v .), the LLL was isolated and perfused in situ with heated autologous blood (37°C) from an extracorporeal perfusion system containing 200 ml of blood . This preparation has been fully described elsewhere . 12 The right lung and the left lower lobe were ventilated separately using two Harvard respirators, and the dog was kept alive by ventilating the right lung with room air. The LLL was ventilated with 5% C02/35% 02 /60% N2 to maintain blood gases in the perfusion system normal [PCO2 = 42 .3 ± 1 .7 (SE) mmHg, P0 2 = 220 ± 20 mmHg and pH = 7 .35 ± 0 .01] . The hematocrit in the perfusion system was 42 .3±2 .2% . The only communication between the dogs' systemic blood and the perfusion system was through the bronchial circulation of the LLL . The nerves to the LLL remain relatively intact, 13 and presumably the lymphatics are functional . In this preparation, substances injected into the LLL artery
* For correspondence and reprint requests at : Department of Surgery, SUNY Health Science Center, 750 E . Adams Street, Syracuse, New York 13210, USA . 0952-0600/92/030159 + 07 $08 .00/0
159
© 1992 Academic Press Limited
1 60
T . S. Hakim et al
do not reach the animal's circulation and vice versa . The blood flow rate to the LLL was set at 522 ± 48 ml/ min, to give a pulmonary arterial pressure (Pa) of 1015 mmHg . The venous pressure (Pv) was kept constant at 1 .3 ± 0 .2 mmHg relative to the top of the lobe . The lobe was ventilated with a constant tidal volume of 150 ml with an end expiratory airway pressure of 12 mmHg. Since the flow rate and venous pressure were kept constant, changes in Pa reflected alterations in pulmonary vascular resistance . The airway pressure of the LLL (Pals) was also monitored as an index of bronchomotor tone . Cats were also anesthetized with sodium pentobarbital (25 mg/kg i.p .) . The LLL was studied as in the dogs, but with slight modifications :14 the cats were exsanguinated to obtain an adequate volume to prime the perfusion system . Thus, the LLL was perfused in situ, but the animal was not alive . The hematocrit in the perfusion system was 31 .7± 1 .5% . Because of the small size of the animal, one tube was placed in the trachea and both the LLL and the right lung were ventilated with 5% C0 2/35% 02/60% N 2. The tidal volume was adjusted to keep peak inspiratory pressure below 10 mmHg, and end expiratory pressure set at 1-2 mmHg . As in the canine LLL, the venous pressure (Pv) was kept constant at 1 mmHg and flow rate was set at a constant value of 54 .4 ± 4 .4 ml/min, so that changes in Pa indicated changes in PVR . Changes in airway pressure were monitored but were not analysed because the LLL and the right lung were ventilated together . Perfusate blood gases and pH in the cats were : PCO 2 = 35 .8±0 .8 mmHg ; P02 =164±6 mmHg and pH=7 .38±0 .01 . Following a period of stabilization (30 min), the PVR in seven dogs and 13 cat LLLs was partitioned into three segments (arterial, middle, venous) using the arterial and venous occlusion technique ." The effect of various interventions on the partitioning of
PVR was examined to identify the site of change in vascular resistance . The substances and their dosages tested in this study included the opiate agonist morphine sulphate (0 .6 mg/kg ; Abbott, Montreal, Canada), the opiate antagonist, naloxone hydrochloride (0 .3 mg ; Narcan, Endo Labs, Montreal, Canada), the histamine H,-receptor antagonist, chlorpheniramine maleate (5 mg; Chlor-Tripolon, Schering, Kenilworth, NJ, USA), the a-adrenergic receptor antagonist, phentolamine mesylate (5 mg ; Rogitine CIBA, Summit, NJ, USA), and a non-specific vasodilator, papaverine (1 mg ; Sigma, St Louis, MO, USA) . These agents, abbreviated respectively as M, N, C, Ph and P, were injected in 0 .5 to 1 ml boluses over 3-5 s directly into the arterial inflow cannula . None of the antagonist drugs had any significant effect on baseline pulmonary vascular resistance in the isolated lung preparation when used alone (data not shown) . The changes in total pulmonary vascular resistance, and in the resistance of individual vascular segments were tested for statistical significance, by comparing the means during vasoconstriction with the means during baseline conditions using Student's paired ttest . Differences were considered significant for Pvalues of less than 0 .05. All values are expressed as mean ± SE .
RESULTS Figure 1 demonstrates the pulmonary vascular responses to morphine injection in six different cats . The administration of morphine produced a rise in pulmonary artery pressure, and although the magnitude of the response varied considerably between animals, it always peaked within about 2 min . The responses in canine LLL were very similar . The baseline and peak pulmonary artery pressure responses obtained in all
40
0
A M 0
A M
M
I min
Fig. 1 Typical pressor response tracings following injection of 0 .6 mg/kg morphine (M) in left lower lobes of seven different cats . Tracings in canine left lower lobes were very similar .
Opiate Action in Pulmonary Circulation
16 1
40 rn E E 30 v tR v n • 20 al v T
C C
E
10
a
0
Control
Morphine
Control
Morphine
Fig. 2 Pulmonary arterial pressure in the individual lobes of cat lungs (A) and dog lungs (B) during control and at peak pressor response following morphine injections. ( •) Mean value . 40 Arterial
Venous
o -rn v = • EE a-
a
Is 0
I
40 0
•
E E
A
W
• C O
q E o S a
n
I min
w 0
i
V . 7 S m E
•
E
Arterial
Venous
0
Fig. 3 Pressure tracings during the arterial and venous flow occlusion in one canine LLL just prior to morphine injection (bottom) and at near peak pressor response (top) . The middle trace shows the change in pulmonary artery pressure . The tracings allow one to partition the total vascular resistance (or pressure gradient) into arterial, middle, and venous components .
cat and dog LLL are shown in Figure 2 . The intensity and course of the response was quite variable . The lobes in which the response was small recovered spontaneously, and the Pa returned to baseline, while in those with a marked response, acute pulmonary edema with froth in the airways ensued within 20-
30 min . The mean pulmonary artery pressure rose from 11 .6 ± 0 .6 to 21 .9 ± 2 .0 mmHg in the feline lungs and from 12.5 ± 0.5 to 25 .3 ± 2 .0 mmHg in the canine lungs . The site of vasoconstriction was examined with the arterial and venous occlusion . Figure 3 shows the pulmonary artery pressure in one canine LLL (middle
1 62
T . S . Hakim et al
trace), and pressure tracings during the arterial and venous occlusion during baseline conditions (bottom tracings) and following vasoconstriction by morphine (top tracings) . The shape of the arterial pressure profile following arterial occlusion did not differ before (bottom tracing) and after (top tracing) vasoconstriction . In contrast the venous pressure tracings were markedly different, and suggested that a large increase in the resistance of the veins had occurred . The mean pressure gradients across the arterial, middle and venous segments under control conditions and following the injection of morphine are shown in Figure 4 . The arterial pressure gradient did not change significantly in either canine or feline lobes, whereas the venous pressure gradient rose markedly in both species (P < 0 .05) . There were also variable but not significant increases in the middle pressure gradient, most likely due to constriction of the small postcapillary veins, which are within the middle segment .
A
(n=7)
i E E
e
n=15)
The pressure profile along the three vascular segments in dog and cat lungs during control and following morphine injection are plotted in Figure 5 . Note that vasoconstriction causes the pressure in the capillaries within the middle segment to rise markedly in both species . Figure 6 illustrates examples of the effect of posttreatment with the opiate receptor antagonist naloxone (N), and the histamine H,-receptor antagonist, chlorpheniramine (C). Naloxone had minimal effect on the morphine-induced vasoconstriction, whereas chlorpheniramine invariably reversed it . In other experiments (not shown), pretreatment with naloxone attenuated the response to morphine : in three lungs the response was entirely abolished by naloxone, while in two other lungs, the response was not affected . Similarly pretreatment with chlorpheniramine (n = 3) also attenuated the response to morphine in dog lungs: the response was abolished in two lungs but there was a pressor response of 6 mmHg in the third lung . Phentolamine had no effect on the morphine induced vascular response, whether given before (n = 3) or after morphine (n = 3) . In three canine lungs, the effects of morphine were examined twice within 60 min . The injection of a second dosage of morphine (0 .6 mg/kg), following recovery from the first injection, produced no response even though the baseline vascular pressures were nearly the same . This suggests that tachyphylaxis or depletion of the involved mediators had occurred . Indeed in preliminary experiments, it was not possible to produce an acceptable dose-response relationship between the morphine and pulmonary artery pressure, starting at dosages under 0 .3 mg/kg . This lack of success may have been due to specific tachyphylaxis to morphine, since these lungs still responded to other vasoactive substances such as serotonin . To demonstrate that morphine acted independently of blood cells, two feline lungs were perfused with autologous plasma : the effect of morphine in these two lungs was indistinguishable from the blood perfused lung, and the Pa rose from 7 to 18 mmHg in one lung and from 11 to 13 mmHg in the other .
DISCUSSION
Control
Morphine
Fig. 4 Distribution of pressure gradient among the arterial (D), middle (0), and venous (®) segment during control and during morphine induced vasoconstriction in dog lungs (A) and cat lungs (B). Values are mean t SE. Note the large increase in venous pressure gradient signifying marked increase in venous resistance . (•) Significant differences from control; P
