Drugs 44 (Suppl. 5): 42-51. 1992 00 12-6667/92/0500-0042/$5.00/0 © Adis International Limited. All rights reserved.

DRSUP3465

Cardiovascular Risks and Benefits of Perioperative Nonsteroidal Anti-Inflammatory Drug Treatment Frederic Camu, Caroline Van Lersberghe and Marie Helene Lauwers Department of Anaesthesiology. Flemish Free University of Brussels Medical Centre, Brussels, Belgium

Summary Prostaglandins participate in the regulation of blood pressure in normotensive and hyperten-sive subjects; vascular tone is subject to the continuous relaxing influence of endogenous vaso­dilating prostaglandins. Prostaglandin 12 (PGI2; prostacyclin), probably the most important phys­iological modulator of vascular tone, decreases blood pressure together with a concomitant increase in cardiac output and a reduction in systemic vascular resistance secondary to peripheral vaso­dilation. In addition, vasodilation within the splanchnic, pulmonary and coronary vascular beds has been observed, with increased blood flow through the mesenteric, renal and coronary vascular beds. These changes in regional blood flow have been associated with the inhibition, by PGI2, of the vasoconstrictor response to sympathetic nervous stimulation and pressor hormones [nor­adrenaline (norepinephrine), angiotensin II]. However, other prostaglandins, such as prostaglan­din E2 (PGE2) and prostaglandin F2,,(PGF2,,), induce coronary vasoconstriction and have dif­ferent effects on pulmonary artery blood pressure because of their effect on pulmonary vascular resistance.

Nonsteroidal anti-inflammatory drugs (NSAIDs; e.g. indomethacin) have been reported to induce hypertension parallel to a fall in cardiac output, suggesting that the underlying mechanism is an increase in systemic vascular resistance. In animal models these agents reduced regional blood flow in the ischaemic myocardium, with a corresponding increase in infarct size. Ibuprofen, which inhibits prostaglandin synthesis to a lesser extent than indomethacin, did not exert systemic or coronary haemodynamic effects. NSAIDs also provide protection in shock models but may exacerbate haemodynamic derangements and decrease survival in acute hypovolaemic hypoten­sion.

To what extent do NSAIDs and opioids influence cardiovascular status during the postop­erative course and analgesic therapy? Continuous infusion of NSAIDs for analgesia had no major haemodynamic effects. Also, there were insignificant changes in indices of left heart function (cardiac output, stroke volume) and the systemic circulation (mean arterial pressure, systemic vascular resistance) following intravenous ketorolac injections, whereas cardiac output and mean arterial pressure decreased after administration of morphine. The pulmonary circulation was un­affected by ketorolac administration, whereas morphine administration induced an increase in pulmonary vascular resistance. Indices of right and left cardiac work were decreased by morphine. Thus, ketorolac produces fewer haemodynamic effects than morphine, although it is possible that some of the effects of morphine may result from morphine-induced histamine release.

NSAIDs may be seen as a worthwhile gain with respect to morphine in clinical situations when hypotension is disadvantageous or when reduction in afterload is not a specific therapeutic aim.

Cardiovascular Safety of NSAIDs

The haemodynamic safety of nonsteroidal anti­inflammatory drugs (NSAIDs) must be evaluated with regard to both specific cardiovascular effects and inhibition of prostaglandin synthesis, as wen as the haemodynamic effects of the administration of opioid analgesics during the postoperative course.

The major pharmacological effect of NSAIDs is the inhibition of cyclo-oxygenase, the enzyme that catalyses the incorporation of molecular oxygen into arachidonic acid to produce the cyclic endoperox­ides prostaglandin G2 (PGG2) and prostaglandin H2 (PGH2), which either spontaneously decom­pose to prostaglandin E2 (PGE2) and prostaglandin D2 (PGD2) or are enzymatically converted in some tissues to prostaglandin F2" (PGF2,,; in the uterus), thromboxane A2 (TXA2; in the platelets and lungs) and prostaglandin 12 (PGI2; prostacyclin; in the vascular endothelium). Figure I summarises the prostaglandin metabolism cascade. This review considers the cardiovascular and haemodynamic

43

effects of these prostaglandins and the influence of NSAIDs on pathological conditions, and compares the haemodynamic effects of NSAIDs and opioids when given for analgesic purposes.

1. Haemodynamic Effects of Prostaglandins

Prostaglandins have the capacity to affect local vascular tone in any tissue, either directly or through modulation of vascular responses to hor­mones, neuronal activity or changes in oxygen ten­sion. The action of prostaglandins on vascular smooth muscle is directly related to their ability to increase the concentration of cyclic adenosine mono phosphate (E series) or of cyclic guanosine monophosphate (F series) [Clyman et al. 1975; Dunham et al. 1974]. Prostaglandins are rapidly inactivated upon passage through the lungs and

Phospholipid

I Phospholipase A2

l --- Steroids

~ AdOO_ 'oM"" Lipo-oxygenase Cyclo-oxygenase

I Pro,ggl"dlo L::~ NSAIOs

Prostaglandin G2 Prostaglandin H2 12-Hydroxyperoxyeicosatetraenoic acid

! ~I~ 12-Hydroxyeicosatetraenoic

acid

Isomerase Prostacyclin Thromboxane I .,"-" prostLyclin

Prostaglandin E2

Prostaglandin 02 l 6-Keto-prostaglandin Flo

synthetase

! Thromboxane A2

! Thromboxane 62

Fig. 1. A schematic diagram showing the metabolism of phospholipid and arachidonic acid. Enzymes are indicated by italic type. NSAIDs inhibit cyclo-oxygenase and, thereby, suppress the synthesis of prostaglandin E2, prostacyclin and thromboxane A2 (Needleman & Isakson 1980).

44

liver (Ferreira & Vane 1967), or their effect is ter­minated by dilution and diffusion.

The endoperoxides PGG2 and PGH2 are potent vasoconstrictors and are 80 to 200 times more po­tent than PGE2 in constricting aortic wall in vitro (Hamberg et al. 1975). In platelets, PGG2 is con­verted to TXA2, which causes constriction in all vascular smooth muscle cells, including those in the aorta and the coronary, umbilical, coeliac and mesenteric arteries, and is the most potent natur­ally occurring vasoconstrictor agent. The biological activity of TXA2 is limited to its site of synthesis because of its extreme lability (it has a half-life of 30 seconds) and decomposition to its inactive me­tabolite thromboxane B2 (TXB2).

1.1 Prostaglandins and the Mesenteric Circulation

Perfused rabbit mesenteric blood vessels have been shown to release PGE2 when treated with angiotensin I, angiotensin II, arachidonic acid or bradykinin (Blumberg et al. 1977). The addition of indomethacin 10 ILg/minute to the perfusate abol­ished the ability of the vessels to release PGE2 in response to any of these agonists and produced an enhanced pressor response to angiotensin II, an in­hibition of vasodilation and an increase in the mesenteric perfusion pressure. Subsequent admin­istration of PGE2 caused a profound vasodilation and an abolition of the effects of angiotensin II.

In the cat and dog, PGF2a has vasoconstrictor properties and increases vascular resistance within the mesenteric and hindquarter vascular beds, whereas PGE2 and PGD2 dilate vascular smooth muscle in peripheral organs such as the kidney, and in skeletal muscle.

PGI2 exerts a marked vasodilator activity in the canine intestinal vascular bed, decreasing perfu­sion pressure by reducing regional vascular resist­ance. As the gastrointestinal tract has a larger PGI2-generating capacity than any other organ, en­hanced local synthesis of PGI2 would be expected to have a marked effect on the fraction of cardiac output flowing to the gastrointestinal tract. The changes in regional blood flow have been linked to

Drugs 44 (Suppl. 5) 1992

the inhibition by PGI2 of the vasoconstrictor re­sponse to sympathetic stimulation and pressor hor­mones [e.g. noradrenaline (norepinephrine), angio­tensin II] (Hedqvist 1977). The effects of PGI2 were concluded to be mainly post junctional in nature since the responses to nerve stimulation and nor­adrenaline were equally affected.

1.2 Coronary Arteries and Cardiac Effects

The intrinsic rate of prostaglandin biosynthesis in the coronary vasculature of the isolated rabbit heart was increased under conditions of reduced oxygen tension but was decreased during anoxia -findings that are consistent with the requirement for molecular oxygen during the cyclo-oxygenase reaction (Needleman et al. 1975). Surprisingly, al­though the primary prostaglandins (PGE2 and PGF2a) are constrictors of bovine and human cor­onary artery strips, arachidonic acid administra­tion produced relaxation of human and animal coronary arteries in vitro. This suggests that the coronary blood vessels generate a potent vasodi­lator substance, which has been identified as PGI2 (Dusting et al. 1977; Kulkarni et al. 1976).

PGI2 induced a dose-related decrease in mean aortic pressure, a reduction in systemic vascular resistance (SVR) and an increase of cardiac output when injected into the right or left atrium of an­aesthetised dogs. That the vasodepressor response to PGI2 0.25-3 ILg/kg was similar following both routes of administration suggests that PGI2 is not inactivated in the pulmonary vascular bed (Wald­man et al. 1978), in contrast to other prostaglan­dins that are metabolised in the lung such as pros­taglandin E) (PGE» and PGE2. Preload, afterload and coronary vascular resistance decreased in an­aesthetised dogs administered PGI2 in increasing dosages 0.25 to 2 ILg/kg' min (Mehta et al. 1981) while stroke volume increased, thus preserving left ventricular stroke work. Heart rate decreased dur­ing PGI2 administration, an effect not abolished by vagotomy. The vasodilating effect of intravenous PGI2 0.1 to 1 ILg/kg· min maintained but did not increase blood flow in the canine coronary bed (Dinerman et al. 1988), whereas blood flow in-

Cardiovascular Safety of NSAIDs

creased in the renal and mesenteric vascular beds. In addition, PGI2 inhibited platelet aggregation in­duced by collagen, thrombin, arachidonic acid and adrenaline (epinephrine). The effects of TXA2 on regional vascular resistances have not been fully elucidated because of the instability of this sub­stance under physiological conditions. Neverthe­less, coronary constriction and impairment ofven­tricular force have been reported to be associated with TXA2 administration in animals (Needleman et al. 1977).

PGE2, a prostaglandin used in obstetrics, exerts significant haemodynamic effects. Secher et al. (1982) noted a 31 % decrease in systemic blood pressure and a 33% decrease in SYR associated with an increase of 36% in cardiac output relative to baseline during a 30-minute intravenous infusion of PGE2 5 to 15 Ilg/min in 5 pregnant women undergoing suction abortion during the first trimester. Pulmonary vascular resistance (PYR) was unaffected by PGE2 administration in this study. In 5 similar patients, intravenous infusion ofPGF2" I 00 to 300 Ilg/min for 40 minutes caused a sig­nificant increase in cardiac output and systolic blood pressure, while pulmonary artery pressure increased by 125% and PYR doubled. Only a slight decrease in SYR was noted.

Intravenous PGE2 0.3 to 31lg had significant va­sodilator activity in the canine coronary vascular bed. In contrast, prostaglandin D2 (PGD2) at the same dosage had only half the vasodilating activity of PGE2, and PGF2" had no vasodilating effect (Hyman et al. 1978a).

PGE\, a coronary vasodilator, has been sug­gested as a treatment for coronary vasospasm. However, in a closed-chest dog model the vaso­dilator effect of this agent in doses of 0.1 to 2 Ilg/ min is limited to the coronary resistance vessels and does not affect the proximal coronary con­ductance vessels. Therefore, the lack of epicardial coronary vasodilation does not seem to support the usefulness of PGE\ for coronary spasm (Miller &

Bove 1989).

45

1.3 The Pulmonary Circulation

The lung is a major organ for the synthesis of TXA2 and PGI2, and the primary prostaglandins (PGE2 and PGF2,,) also influence the pulmonary circulation. In intact spontaneously breathing dogs, PGH2, PGE2, PGD2 and PGF2" (dose ranged from 0.1 to I Ilg/kg) increased lobar arterial pressure due to an increase in vascular resistance across the lung. PGF2" is a very potent vasoconstrictor (10 times as potent as PGE\), nearly doubling the PYR. PGE\ is a potent dilator of pulmonary veins and up­stream small arteries. In contrast to the effects of the primary prostaglandins, PGI2 had vasodilator activity in the pulmonary vascular bed, reducing lobar arterial pressure and acting equally on the arterial and venous pulmonary resistances. The re­sponse to PGI2 was greatly enhanced when vaso­motor tone and PYR were increased by a concom­itant infusion of PGH2, the precursor of TXA2 (Hyman et al. 1978b). Intravenous administration of NSAIDs (indomethacin 2.5 to 5 mg/kg; meclo­fenamate 2.5 to 5 mg/kg) resulted in a slow, grad­ual increase in PYR (maximum of 45%) in dogs (Kadowitz et al. 1975) but not in human subjects administered intravenous indomethacin 25mg (Wennmalm 1978). In this latter study, PYR was measured only once, upon completion of the in­domethacin injection, and so it is possible that a gradual rise in PYR may not have been recorded. As PGI2 is continuously released by the lung, this would indicate that the pulmonary circulation is maintained in a dilated state by a vasodilator prod­uct of the cyclo-oxygenase pathway.

2. Role of Prostaglandins in Pathological Situations 2.1 Anaesthesia and Surgery

During surgery, the levels of PGE2 and TXB2 in arterial plasma increased and remained elevated until the first postoperative day in patients anaes­thetised with halothane or enflurane (Kaukinen et al. 1987). On the contrary, in 30 patients receiving spinal or epidural anaesthesia for small operations ofthe lower extremities (e.g. knee, toe) no changes

46

in PGE2, PGI2 or 6-keto-PGFI" (the inactive me­tabolite of PGI2) were observed, although the con­centrations of TXB2 rose significantly (Kaukinen et al. 1989). Hudson et al. (1990) also observed a 24-fold increase in 6-keto-PGFI" after bowel ma­nipulation during aortic surgery. This huge release of PGI2, suggested by the increase in concentration of the metabolite, induced significant haemodyn­amic changes; these included a reduction of mean arterial pressure (MAP), a reduction of SVR, in­creased heart rate, and increased cardiac index. Pretreatment with oral ibuprofen 800mg com­pletely suppressed both the haemodynamic dis­turbances and the PGI2 release. A late rise in TXB2 was observed in the placebo-treated patients, and its temporal relationship to the PGI2 release might indicate that this was a compensatory response to the increased PGI2 synthesis and vasodilation. Seltzer et al. (1988) reported similar increases in PGh and the same beneficial effect of oral ibu­profen 12 mg/kg in aortic surgery. Elevations of TXA2, implied by the rise in its metabolite TXB2, may contribute to surgical morbidity, since TXA2 may promote embolus production and myocardial ischaemia through induction of platelet aggrega­tion and vasoconstriction.

2.2 Circulatory Deficiency and Shock

Elevated circulating levels of PGE2, PGF2", TXB2 and 6-keto-PGF I" were reported in various models of shock (endotoxic, traumatic, haemor­rhagic), and in patients with septic shock (Feuer­stein & Hallenbeck 1987). Administration of pros­taglandins of the E series and PGI2 enhanced recovery of haemodynamic, metabolic and endo­crine function, and generally improved survival following acute hypovolaemia in a rat model (Feuerstein & Hallenbeck 1987). In this model, in­fusions ofPGE2 (10 ~g/kg· min) and PGI2 (0.3 ~g/ kg· min) potentiated sympathetic nervous activity, as shown by increased plasma levels of noradren­aline (norepinephrine) and increased release of vasopressin. Inhibition of prostaglandin produc­tion by indomethacin, however, also increased sur­vival at least in the endotoxin model, but exacer-

Drugs 44 (Suppl. 5) 1992

bated the haemodynamic derangements in acute hypovolaemic hypotension (Perbeck & Hedqvist 1980). The elevated levels of TXB2 reported in these studies support a role for TXA2 as a media­tor of shock, in which potent vasoconstrictive and platelet aggregating properties may interfere with organ blood flow and induce ischaemia.

In 12 critically ill patients with circulatory de­ficiency, PGEI infusions (10 to 30 ng/kg· min for 24 to 72 hours) improved blood flow, cardiac out­put, oxygen delivery and oxygen consumption, possibly by overcoming the platelet aggregation and neutrophil migration stimulated by TXA2. The beneficial effects of PGEI administration were ob­served in the pulmonary and systemic circulations (Shoemaker & Appel 1986).

2.3 Pulmonary Hypertension and Acute Respiratory Distress Syndrome

Since PGEI and PGI2 induce pulmonary va­sodilation, their efficacy in the treatment of pul­monary hypertension (either primary or secondary to mitral stenosis), chronic lung disease or con­gestive heart failure has been evaluated. Most often, the intravenous administration of both prostaglan­dins improved cardiac output and systemic oxygen transport and significantly decreased PVR, al­though PVR was not changed in patients with chronic lung disease who were artificially venti­lated and who received PGEI 0.02 to 0.04 ~g/ kg· min (Naeije et al. 1992). PGI2 appeared to be more effective and better tolerated than PGEI (Long & Rubin 1987).

In critically ill patients with impaired gas ex­change, intravenous 15 to 30 ng/kg· min for 30 minutes PGEI increased the pulmonary shunt frac­tion in a study by Rademacher et al. (1989), al­though not in a study by Tokioka et al. (1985), and inhibited the hypoxic pulmonary vasoconstriction reflex. Endogenous PGI2 may have similar effects. Indeed, ibuprofen reinforced the hypoxic pulmo­nary vasoconstriction reflex after endotoxin injec­tions in dogs (lshibe et al. 1990). Ibuprofen also blocked the rise of 6-keto-PGF I" levels observed after endotoxin injection, a finding suggesting the

Cardiovascular Safety of NSAIDs

involvement of PGI2 in the inhibition of the hy­poxic pulmonary vasoconstriction response in­duced by endotoxin.

In patients with acute respiratory distress syn­drome (ARDS), plasma levels ofTXB2 and 6-keto­PGF 1" were elevated, with a significant correlation between the degree of ARDS and TXB2 levels (Lamy et al. 1985). However, there is no evidence that either PGI2 or TXA2 is implicated in the pathogenesis of ARDS in human subjects. The re­lease of these agents is probably a nonspecific re­sponse to injury, since inhibition of cyclo-oxygen­ase by indomethacin has prevented only acute and transient pulmonary vasoconstriction in experi­mental animals, and intravenous administration of dazoxiben (1.5 mg/kg), a selective thromboxane synthetase inhibitor, was ineffective in patients with established ARDS (Leeman et al. 1985; Reines et al. 1985).

2.4 Myocardial Ischaemia and Coronary Vasospasm

Coronary vasospasm may be implicated in the pathogenesis of myocardial infarction, angina pec­toris and coronary thrombosis. Although TXA2 synthesis has not been demonstrated in the cor­onary artery, this potent vasoconstrictor is syn­thesised by circulating platelets during thrombosis and aggregation. Thus, coronary tone and possibly spasm in ischaemic myocardial zones may be in­fluenced by the interplay between TXA2 released from platelets and endoperoxide products such as PGI2 synthesised within the endothelium. The cor­onary vasodilation observed as a consequence of hypoxia has been reported to be attenuated by the prior administration of indomethacin 150mg into the pulmonary artery of intact dogs (Afonso et al. 1974), although this has been questioned by other investigators, who observed that indomethacin had no effect on reactive hyperaemia, autoregulation of coronary blood flow or hypoxia-induced coronary dilation (Hintze & Kaley 1977; Rubio & Berne 1975).

Prostaglandins (PGE2, PGF2m PGI2 and TXA2) ate released during myocardial ischaemia and pos-

47

sibly during reperfusion. During such ischaemic episodes, prostaglandins may stimulate reflex car­dio-inhibitory and vasodepressor responses that re­duce myocardial oxygen demand and thereby limit the severity of ischaemia. In one study, inhibition of prostaglandin synthesis by intravenous indo­methacin 6.5 mg/kg and meclofenamate 4 mg/kg markedly attenuated the reflex inhibitory re­sponses to coronary occlusion and reperfusion in dogs (Thames & Minisi 1989). Moreover, Ogletree and Lefer (1978) demonstrated that prostaglandins of the E series protected ischaemic myocardium by stabilising cardiac lysosomal membranes in an open-chest cat model.

PGI2, in particular, appears to be the corner­stone of the defence mechanisms of cardiac cells in terms of restricting myocardial ischaemic dam­age and counteracting reperfusion damage (Araki & Lefer 1980). Berti et al. (1988) also reported that NSAIDs aggravated acute myocardial ischaemia in a perfused rabbit heart model.

NSAIDs have been used for limiting the extent of myocardial infarction, since ischaemic cell death initiates inflammatory processes. However, intra­venous indomethacin 10 mg/kg decreased regional blood flow in ischaemic myocardium and in­creased infarct size in conscious dogs (Jugdutt et al. 1979). These results indicate that the endog­enous synthesis of PGI2 is protective by means of eliciting coronary vasodilation, and could decrease infarct size. Ibuprofen, which is a less potent in­hibitor of prostaglandin synthesis than indometh­acin, has been shown to decrease infarct size in dogs while having no systemic or coronary haemo­dynamic effects (Jugdutt et al. 1980).

2.5 Hypertension

Prostaglandins may act on vascular smooth muscle as modulators of adrenergic neurotrans­mission, and play an important role in the regu­lation of vasomotor tone in the systemic vascular bed. Chronic blockade of cyclo-oxygenase could al­ter blood pressure by producing an imbalance in prostaglandin production and has been reported to aggravate hypertension in human subjects (Martin

48

et al. 1981). Watkins et al. (1980) reported that in­domethacin 100mg daily, given for 3 weeks, atten­uated the hypotensive action of both propanolol and thiazide diuretics. Recently, Gerber et al. (1990) clearly demonstrated that indomethacin did not af­fect the antihypertensive efficacy of hydrochloro­thiazide. Thus, diuretics do not seem to require an intact prostaglandin synthesis pathway to produce a hypotensive effect. In contrast, 13-adrenergic blocking agents stimulate the synthesis of PGI2 (Beckmann et al. 1988). Therefore, the use of NSAIDs in hypertensive patients treated with 13-blocking agents may be associated with loss of blood pressure control. Also, both piroxicam and indo­methacin acutely reversed the antihypertensive ef­fects of propanolol and atenolol without affecting noradrenaline release at the adrenergic synapse (Daniell et al. 1988). NSAIDs must, therefore, be used carefully in hypertensive patients.

3. Haemodynamic Effects of Analgesic Agents during Postoperative Pain Treatment

During the postoperative course of trauma or major surgery, patients often have an abnormal or unstable cardiovascular status as a result of factors such as fluid imbalance, increased catecholamine secretion and acute pain. The physiological changes seen in the postoperative period directly influence myocardial oxygen supply and demand and, thus, can precipitate myocardial ischaemia. Acute pain in particular contributes to tachycardia and elevated afterload, which are major determinants of increased myocardial oxygen consumption. Within this setting, the use of analgesic drugs is necessary and the absence of any significant car­diovascular effect of these agents is highly desira­ble. How well do NSAIDs compare with opioids with regard to their haemodynamic effects?

3.1 Opioid Analgesics

Opioid analgesics generally produce only minor cardiovascular changes at the doses generally re­quired to relieve pain. Morphine and fentanyl ex-

Drugs 44 (Suppl. 5) 1992

ert negative chronotropic effects but minimally af­fect the inotropy of the heart. In canine models of myocardial ischaemia, intravenous morphine 5mg, subcutaneous morphine 15mg and intravenous morphine 25 ~g/kg decreased left ventricular myo­cardial oxygen consumption by reducing heart rate and aortic blood pressure. It was postulated that this beneficial effect was a consequence of anal­gesia, in that it blunted the sympathetic response to coronary occlusion (Theroux et al. 1976; van der Vusse et al. 1979). In patients with coronary heart disease, the administration of morphine decreased oxygen consumption, cardiac work and left ven­tricular end-diastolic pressure (L VEDP) by means of peripheral arteriolar and venous dilation. It has been suggested that the release of histamine me­diates this peripheral vasodilatory response to morphine, pethidine and codeine (Philbin et al. 1981 ).

The peripheral vascular effects of morphine are beneficial in the treatment of pulmonary oedema; administration rapidly reduces pulmonary artery flow and pressure and LVEDP, and increases card­iac contractility. This improvement is a result of an increase in peripheral vascular capacitance, which decreases venous return. Caution is advised in treating hypovolaemic patients with morphine, since morphine-induced vasodilation can aggra­vate hypotension.

Pethidine has effects on the cardiovascular sys­tem that are similar to those of morphine, but may impair myocardial contractility at intravenous doses of 2 to 10 mg/kg. The fall in blood pressure may produce significant reflex tachycardia, and the peripheral vasodilating effect secondary to hist­amine release is more pronounced with pethidine than morphine (DeCastro et al. 1979). Buprenor­phine has minimal cardiovascular effects in nor­mal subjects and cardiac patients, producing only small reductions in heart rate and arterial pressure. The reduction in cardiac output is unlikely to be clinically significant, since myocardial contractility has been shown to be unaffected by administration of buprenorphine (Scott et al. 1980). In contrast, pentazocine may produce deleterious effects on the systemic and pulmonary circulation, including

Cardiovascular Safety of NSAIDs

hypertension and tachycardia. In critically ill patients administered pentazocine, systemic and pulmonary hypertension were accompanied by in­creases in left ventricular filling pressure as a result of peripheral vasoconstriction and negative inotro­pic effects (Alderman et al. 1972). Butorphanol has produced similar effects (Popio et al. 1978).

3.2 NSAIDs

There have been few studies investigating the haemodynamic effects of NSAIDs during the treat­ment of postoperative pain. Continuous infusions of indomethacin 0.1 mgjkgjh and diclofenac 5.4 mg/h induced no significant changes in heart rate or systolic and diastolic blood pressure (Claeys et al. 1992; Tigerstedt et al. 1991). These data are in sharp contrast to the findings ofWennmalm (1978), who reported a significant increase in systemic blood pressure with a concomitant fall in cardiac output and increased SVR in normal volunteers administered indomethacin 25mg into the superior vena cava.

The haemodynamic effects of ketorolac were compared with those of morphine in a double-blind study using right-sided heart catheterisation (Camu et al. 1990). Indices ofleft heart function (cardiac output, stroke volume) and the systemic circula­tion (MAP, SVR) showed insignificant changes with ketorolac 10mg and 90mg, while cardiac output and MAP decreased by 5 and 12%, respectively, with morphine 10mg. Ketorolac did not affect the pul­monary circulation but morphine increased PVR by 13%. Right and left cardiac work indices were decreased (13 and 19%, respectively) by morphine and slightly increased (8 and 4%, respectively) by ketorolac. Neither drug affected heart rate. Mor­phine caused a significant reduction in left ven­tricular stroke work index, consistent with reduc­tions in blood pressure and flow, whereas this parameter was unaffected by ketorolac administra­tion. Ketorolac thus seems to lack haemodynamic side effects despite its ability to inhibit endogenous prostaglandin production.

49

4. Conclusion

Thus, it would appear that, although exogenous prostaglandins have major effects on haemodyn­amic parameters, inhibition of endogenous pros­taglandin production seems to have generally mi­nor haemodynamic effects. This may reflect the balanced reduction of synthesis of prostaglandins with vasoconstrictor and vasodilator actions, with an almost zero net result. However, NSAIDs should be used with caution in clinical situations where PGI2 has proven therapeutic benefits, such as cir­culatory insufficiency, shock, myocardial ischae­mia, coronary vasospasm and hypertension.

References

Afonso S, Bandow GT, Rowe GG. Indomethacin and the pros­taglandin hypothesis of coronary blood flow regulation. Jour­nal of Physiology 241: 299-308, 1974

Alderman E, Barry W, Graham A, Harrison D. Hemodynamic effects of morphine and pentazocine differ in cardiac patients. New England Journal of Medicine 287: 623-627, 1972

Araki H, Lefer A. Role of prostacyclin in the preservation of is­chemic myocardial tissue in the perfused cat heart. Circulation Research 47: 757-763, 1980

Beckmann M, Gerber J, Byyny R, Loverde M, Nies A. Propan-0101 increases prostacyclin synthesis in patients with essential hypertension. Hypertension 12: 582-588, 1988

Berti F, Rossoni Q, Magni F, Omini C, Puglisi L. et al. Non­steroidal anti-inflammatory drugs aggravate acute myocardial ischemia in the perfused rabbit heart. Journal of Cardiovas­cular Pharmacology 12: 438-444, 1988

Blumberg A, Denny S, Marshall G, Needleman P. Blood vessel­hormone interactions: angiotensin, bradykinin and and pros­taglandins. American Journal of Physiology 232: H305-H31O. 1977

Camu F, Van Overberge L, Bullingham R, Lloyd J. Haemodyn­amic effects of two intravenous doses ofketorolac trometham­ine compared with morphine. Pharmacotherapy 10: 1225-127S, 1990

Claeys M, Camu F, Maes V. Prophylactic diclofenac infusions in major orthopaedic surgery: effects on analgesia and acute phase proteins. Acta Anaesthesiologica Scandinavica 36: 270-275, 1992

Clyman R, Sandler J, Manganiello V, Vaughan M. Guanosine 3', 5'-monophosphate content of human umbilical artery. Possi­ble role in perinatal arterial patency and closure. Journal of Clinical Investigation 55: 1020-1025, 1975

Daniell H, Webb J, Walle T. Oatis J, Gaffney T. Inhibitors of prostaglandin synthesis reverse the effects of chronic p-recep­tor blockade to attenuate adrenergic neurovascular transmis­sion in dogs. Journal of Cardiovascular Pharmacology 12: 300-307, 1988

DeCastro J, Vanderwater A, Wouters L, Xhonneux R, Reneman R, et al. Comparative study of cardiovascular, neurological and metabolic side effects of eight narcotics in dogs. Acta Anaes­thesiologica Belgica 30: 5-99, 1979

Dinerman J, Mehta J, Nichols w. Systemic and coronary hemo­dynamic effects of prostacyclin and nitroprusside in conscious dogs. Journal of Cardiovascular Pharmacology 12: 492-496, 1988

50

Dunham E. Haddox M. Goldberg N. Alteration of vein cyclic 3',5' nucleotide concentrations during changes in contractility. Proceedings of the National Academy of Sciences USA 71: 815-819.1974

Dusting G. Moncada S. Vane J. Prostacyclin (PGX) is the en­dogenous metabolite responsible for relaxation of coronary ar­teries induced by arachidonic acid. Prostaglandins 13: 3-16. 1977

Ferreira S. Vane J. Prostaglandins: their disappearance from and release into the circulation. Nature 216: 868-873. 1967

Feuerstein G. Hallenbeck J. Prostaglandins. leukotrienes and platelet activating factor in shock. Annual Review of Pharma­cology and Toxicology 27: 301-313. 1987

Gerber J. Loverde M. Byyny R. Nies A. The antihypertensive efficacy of hydrochlorothiazide is not prostacyclin dependent. Clinical Pharmacology and Therapeutics 48: 424-430. 1990

Hamberg M. Hedqvist P. Strandberg K. Svensson J. Samuelsson B. Prostaglandin endoperoxides. IV. Effects on smooth muscle. Life Sciences 16: 451-462. 1975

Hedqvist P. Basic methods of prostaglandin action on autonomic transmission. Annual Review of Pharmacology and Toxicol­ogy 17: 259-279. 1977

Hintze T. Kaley G. Prostaglandins and the control of blood flow in the canine myocardium. Circulation Research 40: 313-320. 1977

Hudson J. Wurm W. O'Donnell T, Kane F. Mackey W, et al. Ibuprofen pretreatment inhibits prostacyclin release during ab­dominal exploration in aortic surgery. Anesthesiology 72: 443-449. 1990

Hyman A. Kadowitz P. Lands W. Crawford C. Fried J. et al. Coronary vasodilator activity of 13. 14-dehydro-prostacyclin methylester: comparison with PGI2 and other prostanoids. Proceedings of the National Academy of Sciences of the United States of America 75: 3522-3526. 1978a

Hyman A. Spannhake E. Kadowitz P. Prostaglandins and the lung. A state of the art review. American Review of Respiratory Diseases 117: I 11-136. 1978b

Ishibe Y. Shiokawa Y. Umeda T. Izumi T. Nakamura M. et al. Effects of ibuprofen on the endotoxin-induced inhibition of the pulmonary pressure response to hypoxia in dogs. Anesthe­siology 73: A 1169. 1990

Jugdutt B. Hutchins G. Bulkley B. Pitt B. Becker L. Effect of indomethacin on collateral blood flow and infarct size in the conscious dog. Circulation 59: 734-743. 1979

Jugdutt B. Hutchins G. Bulkley B. Becker L. Salvage of ischemic myocardium by ibuprofen during infarction in the conscious dog. American Journal of Cardiology 46: 74-82, 1980

Kadowitz p. Chapnick B. Joiner P, Hyman A. Influence of in­hibitors of prostaglandin synthesis on the canine vascular bed. American Journal of Physiology 229: 941-946, 1975

Kaukinen S, Seppala E. Kaukinen L, Ojanen R, Vapaatalo H. Effects of halothane and enflurane on prostanoid concentra­tions in operation patients. Prostaglandins and Leukotrienes in Medicine 29: 85-94, 1987

Kaukinen S. Ojanen R, Kaukinen L, Seppala E, Vapaatalo H. Plasma thromboxane B2 levels and thromboxane B2 produc­tion by platelets are increased in patients during spinal and epidural anesthesia. Prostaglandins Leukotrienes and Essential Fatty Acids 37: 83-88, 1989

Kulkarni P, Roberts R, Needleman P. Paradoxical endogenous synthesis of a coronary dilating substance from arachidonate. Prostaglandins 12: 337-353. 1976

Lamy M. Deby-Dupont G. Pincemail J. Biochemical pathways of acute lung injury. Bulletin Europeen de Physiopathologie Respiratoire 21: 221-229, 1985

Leeman M, Boeynaems I, Degaute J. Vincent J, Kahn R. Admin­istration of dazoxiben, a selective thromboxane synthetase in­hibitor, in the adult respiratory distress syndrome. Chest 87: 726-730. 1985

Drugs 44 (Suppl. 5) 1992

Long W, Rubin L. Prostacyclin and PGE) treatment of pulmo­nary hypertension. American Review of Respiratory Disease 136: 773-776, 1987

Martin K. Zipser R, Horton R. Effect of prostaglandin inhibition on the hypertensive action of sodium retaining steroids. Hypertension 3: 622-628, 1981

Mehta J, Nichols W, Mehta P, Pepine C, Conti C. Effects of pros­tacyclin on systemic and coronary hemodynamics in the dog. American Heart Journal 102: 835-840, 1981

Miller W, Bove A. Prostaglandin E) and nitroglycerin effects on canine epicardial conductance and distal coronary resistance vessels. Journal of Cardiovascular Pharmacology 14: 260-267, 1989

Naeije R, Melot C, Mols P, Hallemans R. Reduction in pulmo­nary hypertension by prostaglandin E) in decompensated chronic obstructive pulmonary disease. American Review of Respiratory Diseases 25: 1-5, 1992

Needleman P, Isakson P. Intrinsic prostaglandin biosynthesis in blood vessels. In Bohr D, et al. (Eds) Handbook of physiology. The cardiovascular system, pp. 613-634, American Physiolog­ical Society, Bethesda, 1980

Needleman P, Key S, Isakson P, Kulkarni P. Relationship be­tween oxygen tension, coronary vasodilation and prostaglan­din biosynthesis in the isolated rabbit heart. Prostaglandins 9: 123-134,1975

Needleman P, Kulkarni P, Raz A. Coronary tone modulation: formation and action of prostaglandins, endoperoxides and thromboxanes. Science 195: 409-412, 1977

Ogletree M, Lefer AM. Prostaglandin induced preservation of is­chemic myocardium. Circulation Research 42: 218-222, 1978

Perbeck L, Hedqvist P. Prostaglandins E) and E2 antagonize in­domethacin induced decrease in survival rate of hemorrhagi­cally shocked rats. Acta Clinica Scandinavica 500: 91-94, 1980

Philbin DM, Moss J, Akins CW, Rosow CE, Kono K, et al. The use of H) and H2 histamine antagonists with morphine anes­thesia: a double blind study. Anesthesiology 55: 292-296, 1981

Popio K, Jackson D, Ross A, Schreiner B. Yu P. Hemodynamic and respiratory effects of morphine and butorphanol. Clinical Pharmacology and Therapeutics 23: 281-287, 1978

Rademacher P, Santak B. Becker H, Falke K. Prostaglandin E) and nitroglycerin reduce pulmonary capillary pressure but worsen ventilation-perfusion distributions in patients with adult respiratory distress syndrome. Anesthesiology 70: 601-606. 1989

Reines H. Halushka p. Olanoff L. Hunt P. Dazoxiben in human sepsis and adult respiratory distress syndrome. Clinical Pharmacology and Therapeutics 37: 391-395. 1985

Rubio R. Berne R. Regulation of coronary blood flow. Progress in Cardiovascular Diseases 18: 105-122. 1975

Shoemaker WC, Appel PL. Effects of prostaglandin E) in adult respiratory distress syndrome. Surgery 99: 275-282, 1986

Scott D, Arthur G. Scott DB. Haemodynamic changes following buprenorphine and morphine. Anaesthesia 35: 957-961, 1980

Secher N, Thayssen P, Arnsbo P, Olsen J. Effect of prostaglandin E2 and F2. on the systemic and pulmonary circulation in preg­nant anaesthetized women. Acta Obstetriccia et Gynecologica Scandinavica 61: 213-218, 1982

Seltzer J, Goldberg M, Larijani G. Ritter D. Starsnic M. et al. Prostacyclin mediation of vasodilation following mesenteric traction. Anesthesiology 68: 514-518, 1988

Thames M, Minisi A. Reflex responses to myocardial ischemia and reperfusion. Role of prostaglandins. Circulation 80: 1878-1885. 1989

Theroux P, Ross J, Franklin D. Regional myocardial function in the conscious dog during acute coronary occlusion and re­sponses to morphine. propanolol. nitroglycerine and lidocaine. Circulation 53: 302-314, 1976

Tigerstedt I, Tammisto T, Neuvonen P. The efficacy of intra­venous indomethacin in prevention of postoperative pain. Acta Anaesthesiologica Scandinavica 35: 535-540, 1991

Cardiovascular Safety of NSAIDs

Tokioka H. Kobayashi Y. Ohta T. Wakabayashi T. Kosaka F. The acute effects of prostaglandin E I on the pulmonary cir­culation and oxygen delivery in patients with adult respiratory distress syndrome. Intensive Care Medicine II: 61-64. 1985

van der Vusse GJ. van Belle W. van Gerven R. Kruger R. Rene­man RS. Acute effect of fentanyl on haemodynamics and myo­cardial carbohydrate utilization and phosphate release during ischaemia. British Journal of Anaesthesia 51: 927-935. 1979

Waldman H. Alter I. Kot P. Rose J. Ramwell P. Effect of lung transit on systemic depressor responses to arachidonic acid and prostacyclin in dogs. Journal of Pharmacology and Experi­mental Therapeutics 204: 289-293. 1978

Watkins J. Abbott E. Hensby C. Webster J. Dollery C. Atten­uation of hypotensive effect of propanolol and thiazide di­uretics by indomethacin. British Medical Journal 281: 702-705. 1980

Wennmalm A. Influence of indomethacin on the systemic and pulmonary vascular resistances in man. Clinical Science and Molecular Medicine 54: 141-145. 1978

Correspondence and reprints: Dr Fredertc Camu. Department of Anaesthesiology. Flemish Free University of Brussels Medical Centre. Laarbeeklaan 101. B-1090 Brussels. Belgium.

Discussion

Prof. H. Kehlet: The randomised trials of NSAIDs have not excluded patients with hyper­tension or a previous history of. myocardial isch­aemia, and severe cardiovascular adverse effects have not been reported. Can you give any definite contraindications for NSAIDs?

51

Prof. F. Camu: I would consider a history of coronary vasospasm to be an absolute contrain­dication to NSAID administration because pros­taglandin 12 (PGI2) has a major beneficial effect in this condition. Concerning hypertension, while patients with hypertension have not been excluded from clinical trials of NSAIDs, we know that chronic use of NSAIDs tends to precipitate the ac­cumulation of water and salt, which might induce hypertension or counteract the beneficial effects of antihypertensive therapies. To date, I am not aware of any studies in humans assessing the impact of NSAIDs on (j-blockade, but I am sure that this will be investigated in the near future. However, it has been suggested that PGI2 synthesis is increased at the noradrenergic synapse during (j-blockade, an effect that modulates the vasomotor tone at that particular site. If the increased synthesis of PGI2 is blocked by administration of a NSAID, then noradrenaline will be active at the synapse and there will be an increase in vasomotor tone. In pulmo­nary hypertension, I would not advocate the use of NSAIDs in patients receiving beneficial treat­ment with prostaglandin EI (PGE) or PGI2 in­fusions.