Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Natural opiates, opioid peptides, and their synthetic derivatives belong to the most potent analgesics known so far and are used in almost all fields of medicine. In the human body, activation of µ-opioid receptors is considered to cause not only analgesic effects, but also respiratory depression and euphoria (McQuay, 1991; Reisine and Pasternak, 1996). After i.v. administration of opioids, the predominant side effects consist of nausea, vomiting, constipation, as well as cardiovascular depression and subsequent hypotension (McQuay, 1991; Reisine and Pasternak, 1996). The hemodynamic effects of opioids differ significantly. For example, morphine is well known to release histamine, which results in venous pooling. In contrast, fentanyl and its potent analog sufentanil do not release histamine, but have probably more effect on vagal tone. However, the exact mechanism of these responses are not yet fully understood.

Studies in conscious animals and in healthy young men revealed an increase of endogenous catecholamine serum levels after application of opioid agonists such as morphine (May et al., 1988) or fentanyl (Hoehe and Duka, 1993). This effect was attributed to the liberation of endogenous catecholamines mediated by opioid agonists within the central nervous system. An increase of circulating catecholamines after morphine administration was observed in conscious rabbits, thus leading to arterial hypertension (May et al., 1988). In contrast, when given to patients, a hypotensive effect of opioids is usually observed. This effect is particularly pronounced if the patient stands up or if the patient's circulation is compromised (McQuay, 1991; Reisine and Pasternak, 1996).

Moreover, circulating endogenous opioids are known to be involved in the regulation of blood pressure; in patients suffering from septic shock, -endorphin concentrations in blood are elevated. Administration of the opioid antagonist naloxone increases blood pressure (Dirksen et al., 1981; Editorial, 1981; Canady et al., 1989). In addition, systemically administered -endorphin has been shown to produce a naloxone-reversible hypotension (Lemaire et al., 1978).

To achieve adequate sedation and analgesia, high doses of opioid agonists are administered and combinations of benzodiazepines and opioid agonists are often used (Murray and Plevak, 1994). On the other hand, many critically ill patients depend on catecholamines to maintain hemodynamic stability. The aim of the present investigation was to evaluate the pharmacodynamic and pharmacokinetic interactions between the potent opioid derivative fentanyl and the adrenergic agonist dobutamine in critically ill patients. Our study was carried out in sedated patients who received dobutamine and were subsequently treated with fentanyl. We determined the hemodynamic response to this treatment and measured the plasma kinetics of circulating dobutamine and endogenous catecholamines. After an increase in plasma dobutamine levels (after administration of fentanyl) was established, we investigated the interaction between fentanyl and -adrenergic receptors on rat myocardium. We hypothesized a direct replacement of dobutamine from the receptor or an effect on the -adrenergic receptor G protein coupling.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Clinical Investigation

The study protocol was approved by the institutional review board (Ethical Committee of the University Hospital of Vienna); written informed consent was waived.

Patients. Six consecutive hemodynamically stable patients, elected from the local Emergency Department, were enrolled in this nonrandomized prospective trial. All patients were unconscious/comatose before initiation of the study (demographic data see Table 1). No baseline sedation was necessary. Exclusion criteria were defined as hemodynamic instability, prior continuous catecholamine application, or opioid medication within the past 24 h. All patients were intubated and subjected to controlled mechanical ventilation. The respiratory settings were not changed during the study period. None of the patients received any medication known to interact with catecholamines or opioids, namely monoamine oxidase inhibitors; - or -adrenergic blockers; -adrenergic agonists other than dobutamine; amphetamine and related compounds; angiotensin-converting enzyme inhibitors; methylxanthines; or antihistaminics.

Study Protocol. After insertion of an Edwards Swan-Ganz catheter (Baxter Healthcare Corporation, Edwards Critical-Care Division, Irvine, CA) and an arterial line for continuous invasive blood pressure monitoring, patients first received continuous sedation with i.v. midazolam (Dormicum, Hoffmann-La Roche AG, Basel, Switzerland). This benzodiazepine derivative was administered at a dose of 0.4 mg · kg¹ · h¹ with a loading dose of 15 mg as a bolus injection, paralleled by a continuous infusion of dobutamine (Dobutrex, Eli Lilly, Indianapolis, IN) at a dose of 10 µg · kg¹ · min¹. Loading doses of midazolam (and thereafter fentanyl) were calculated according to the formula of Ritschel (1986). Due to the manufacturer's recommendation and because of feared severe hypotension, the loading dose of fentanyl was reduced to 1.0 mg. Then, the protocol differed for the first two (A, B) versus the final 4 patients (C-F). In patients A and B, i.v. fentanyl (Janssen Pharmaceutica, Beerse, Belgium) was added at a dose of 4 µg · kg¹ · h¹ (loading dose 1 mg as a bolus injection) 60 min after initiation of the study and continuously administered at its fixed dose until the end of the observation period. Arterial blood samples were drawn and hemodynamic measurements were evaluated before initiation of the study, and 15, 30, 45, 60, 75, 90, 105, 120, 150, 180, and 240 min thereafter. In patients C to F, the combined treatment with midazolam and dobutamine was extended up to 90 min, and the observation period including the additional administration of fentanyl up to 510 min. Hemodynamic measurements and catecholamine plasma levels were again determined before the initiation of the study, and 30, 60, 90, 150, 210, 270, 330, 420, and 510 min thereafter.

Hemodynamic Measurements. The following hemodynamic parameters were evaluated at time points when plasma samples were drawn by use of a HP-CMS-M1166A monitor (Hewlett-Packard-GesmbH, Böblingen, Germany): systolic arterial blood pressure (ABP), diastolic ABP, mean ABP (ABPm), pulmonary ABP, pulmonary arterial wedge pressure (PAWP), and central venous pressure (CVP). Cardiac output (CO) was evaluated by using the thermodilution technique as described previously (Ganz and Swan, 1972). Cardiac index (CI), systemic vascular resistance index (SVRI), pulmonary vascular resistance index (PVRI), stroke volume index (SVI), left and right ventricular stroke work index (LVSWI, RVSWI), as well as oxygen delivery index (DO₂-I), oxygen consumption index (VO₂-I), and oxygen extraction rate were calculated according to standard formulae (Shoemaker et al., 1974). All indices refer to the patients' individual body surface area (m²).

Laboratory Investigations

Materials. Epinephrine, norepinephrine, dopamine, and epinine (deoxyepinephrine) were obtained from Sigma (St. Louis, MO). Water (Rathburn, Walkerburn, UK), acetonitrile, and methanol (both from Promochem, Wesel, Germany) were of HPLC grade. All other reagents were supplied by Merck (Darmstadt, Germany) and were of analytical grade. The -adrenergic antagonistic radioligand ()-[¹²⁵I]iodocyanopindolol ([¹²⁵I]CYP; specific activity 2200 Ci/mmol)) was obtained from NEN (Boston, MA); GTPS and HEPES were purchased from Boehringer Mannheim (Mannheim, Germany), and ()-isoproterenol and d,l-propranolol were obtained from Sigma (St. Louis, MO); all other reagents (analytical grade) were obtained from Merck. Male Rats (200-250 g body weight; Sprague-Dawley) were obtained from the Institute for Animal Breeding (Himberg, Austria).

Determination of Plasma Catecholamine Levels. Plasma samples were collected in chilled glass tubes containing glutathione and EGTA (Amersham, Buckinghamshire, UK), and were stored at 70°C until analysis after centrifugation.

1

Preparation of Rat Cardiac Ventricular Membranes. The project was approved by the local Animal Care Committee and was performed at the Department of Pharmacology of the University of Vienna. The preparation of cardiac microsomes from rat ventricles was carried out as described previously (Freissmuth et al., 1986). In brief, ten rats were sacrificed by cervical dislocation; the hearts were rapidly removed and perfused with isotonic saline in a retrograde fashion via Langendorff column until the effluent was clear. Thereafter, the cardiac ventricles were minced in ice-cold buffer (composition in mmol · l¹: 20 HEPES-NaOH, pH 7.4, 1 EDTA, 2 MgCl₂, 250 sucrose) and subsequently homogenized by means of an Ultra-Turrax at a tissue-to-volume-ratio of 1:4 at 2 × 20 s at half-maximum speed and 1 × 2 s at maximum speed. The resulting homogenate was filtered over four layers of cheesecloth. A microsomal fraction was obtained by differential centrifugation (10 min at 10,000g followed by 20 min at 50,000g). The resulting pellet was washed three times in sucrose-free buffer, taken up at a protein concentration of 3 mg · ml¹, snap frozen in liquid nitrogen, and stored at 80°C. The protein concentration was determined by dye binding with Coomassie Blue using an assay kit provided by BioRad (Richmond, CA).

Radioligand Binding Experiments. In saturation experiments, cardiac ventricular membranes (2.5 µg membrane protein/assay) were incubated in 200 µl reaction buffer (composition in mmol · liter¹ 50 Tris.HCl, pH 7.4, 1 EDTA, 5 MgCl₂, 1 ascorbate as antioxidant) in the absence and presence of 0.5 µM fentanyl. Increasing concentrations of [¹²⁵I]CYP (from ~3-300 pM, cf. Fig. 3) were added to the reaction mixture. Nonspecific binding was determined in the presence of 3 µM propranolol. Preliminary experiments verified that binding equilibrium was reached within 90 min at 30°C even at the low concentrations of [¹²⁵I]CYP; binding was stable for at least 2 h (data not shown). After 90 min at 30°C, the reaction was stopped by filtration over glass fiber filters using a Skatron cell harvester (Skatron, Lier, Norway). The filters were rinsed with 10 ml of ice-cold wash buffer (composition in mmol · l¹: 10 Tris · HCl, pH 7.4, 1 MgCl₂). The radioactivity trapped on the filters was determined in a gamma counter (Minaxi- 5000, Packard) with a counting efficiency of 75%.

View larger version (30K): [in this window] [in a new window]   Fig. 1.   Hemodynamic changes during different study periods. Graphics reveal the changes in heart rate (A), CI (B), DO2-I (C), and VO2-I (D) in the four patients with identical study protocol (patients C-F, see Experimental Procedures) during the different study periods. Values are expressed as the mean ± S.D.

View larger version (20K): [in this window] [in a new window]   Fig. 2.   Dobutamine plasma levels during different study periods. Patients were first treated with midazolam and dobutamine and then subsequently with the combination of midazolam, dobutamine, and fentanyl (fentanyl was added 60 min after initiation of the study in patients A and B, and 90 min after initiation of the study in patients C to F). The solid lines were drawn by subjecting the data points to nonlinear least-squares curve fitting (see Experimental Procedures). Values are expressed as the mean ± S.D.

Statistical Analysis

Statistical calculations were performed using the Statistical Analysis Software package (SAS Institute, Cary, NC). The first two patients were investigated for a shorter time period than the subsequent four patients. However, the pharmacodynamic effects of the compounds occurred rapidly and the dobutamine levels had invariably approached steady-state levels. Hence, all data from the six patients were combined and analyzed within the framework of General Linear Model's handling classification variables, which have discrete levels (i.e., the three different study periods: before therapy; during administration of midazolam and dobutamine; and during combined administration of midazolam, dobutamine, and fentanyl) as well as continuous variables, which measure quantities. We used the General Linear Model procedure as repeated measures ANOVA (for unbalanced data). We computed means for the classification variable and performed Tukey's studentized range test on the means. The four patients subjected to the longer study protocol were also analyzed separately; this analysis yielded statistical differences that were similar to those obtained with all six patients (data not shown). The p values in Results therefore represent the whole study population. Values are expressed as the mean ± S.D. A p value less than .05 was considered statistically significant.

The data points obtained in the binding experiments were subjected to nonlinear least-squares curve fitting to the appropriate equations (rectangular hyperbola, Hill equation, displacement from one of two sites) using a Gauss-Newton or Marquart-Levenberg algorithm. Similarly, the increase in dobutamine concentration upon continuous infusion was analyzed by fitting the data to the equation describing a monoexponential rise yielding estimates for the elimination half-life and the steady-state concentration.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Clinical Investigation. None of the patients had to be withdrawn from the study because of severe hemodynamic instability during the study. Both loading doses, midazolam as well as fentanyl, only led to transient hypotension. We compared hemodynamic parameters before initiation of the study and during i.v. infusions of dobutamine and midazolam. The following changes were observed (see Table 2 and Fig. 1, A-C): heart rate, CI, SVI, and DO₂-I increased significantly (p < .05), whereas SVRI and PVRI significantly decreased (see Table 2). During continuous i.v. application, dobutamine plasma levels achieved a steady state within 60 min in patients A and B. This steady state could be confirmed in the longer observation period (90 min) in patients C to F. 

Effect of Fentanyl on Pharmacokinetics of Dobutamine. In five of the six patients, the continuous infusion of dobutamine resulted in a rapid increase in the plasma concentration (rate constant ± S.D. = 0.098 ± 0.043 min¹). The steady-state plasma levels varied only modestly in these patients (see Fig. 2, A and B; mean steady-state concentration ± S.D. = 177.3 ± 23.3 ng · ml¹). The total body clearance was derived from the constant infusion rate and the steady-state concentration and calculated as 57.3 ± 8.7 ml · min¹ · kg¹. In one patient (patient F, Fig. 2C), however, the increase in dobutamine plasma levels was delayed (rate constant ± S.E. of the estimate = 0.038 ± 0.03 min¹) with a much higher steady-state level (steady-state concentration ± S.E. of the estimate = 534.7 ± 15.6 ng · ml¹). Accordingly, the total clearance of this patient was considerably lower (18.7 ml · min¹ · kg¹) than determined for the other five patients.

1

1

-Adrenergic Receptor Binding Studies. After the administration of fentanyl, the cardiac effects of dobutamine were blunted. In spite of an increase in circulating epinephrine and dobutamine, the CO was reduced (cf. Tables 2 and 3, and Fig. 1). This indicated that the initial stimulation of cardiac -adrenergic receptors was counteracted by fentanyl. We therefore used radioligand binding experiments to investigate whether fentanyl impairs the coupling of cardiac -adrenergic receptors to its cognate G protein G_s in rat cardiac ventricular membranes. The -adrenergic receptors were labeled with the antagonist radioligand [¹²⁵I]CYP. Saturation experiments were carried out in the absence and presence of fentanyl to confirm that fentanyl does not interact with the ligand binding pocket of the -adrenergic receptor (Fig. 3); fentanyl decreased the total binding of [¹²⁵I]CYP (Fig. 3A, ). However, this effect was attributed to a decrease in nonspecific binding (Fig. 3A, ) such that the specific binding calculated from the difference of total and nonspecific binding remained unaffected by fentanyl (Fig. 3B, ). Therefore, the binding parameters calculated for the saturation isotherms in the absence and presence of fentanyl were virtually identical (see Fig. 3B; K_D = 18.1 ± 2.2 and 17.9 ± 1.8 pM; B_max = 281.8 ± 14.5 and 281.8 ± 14.5 fmol · mg¹ in the absence and presence of fentanyl respectively).

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Binding of the -adrenergic radioligand [125I]CYP to rat cardiac ventricular membranes in the absence and presence of fentanyl. Membranes (2.5 µg · assay) were incubated in the absence (, , ) and in the presence (, , ) of 0.5 µM fentanyl for 90 min at 30°C as described in Experimental Procedures. A, total binding (, ) and nonspecific binding defined by the addition of 3 µM propranolol (, ). B, specific binding (, ), i.e., the difference between total binding and nonspecific binding. Data are means of duplicate determinations in a single experiment carried out in duplicate; a second experiment gave similar results. The solid lines were drawn by fitting the data to the appropriate equations.

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Displacement of specific [125I]CYP binding to rat cardiac ventricular membranes by isoproterenol in the absence and presence of fentanyl. Membranes (2.5 µg · assay) were incubated with 60 pM [125I]CYP and the indicated concentrations of isoproterenol in the absence (, ) and in the presence of 0.5 µM fentanyl (, ) for 90 min at 30°C as outlined in Experimental Procedures; (, ), binding in the presence of 10 µM GTPS. Specific binding determined in the absence of isoproterenol was set 100%. Data are means from duplicate determinations in a single experiment which was reproduced twice; the parameter estimates (IC50H and IC50L = the concentration of the competitor resulting in half-maximal displacement at the high- and low-affinity sites, respectively; %H and %L = percentage of high- and low-affinity sites, respectively) derived from these experiments were calculated by nonlinear least-squares curve fitting (in the absence of GTPS: %H = 44 ± 7 and 47 ± 3, in the absence and presence of fentanyl; IC50H = 2.8 ± 1.1 nM and 1.9 ± 0.3 nM; IC50L = 0.58 ± 0.2 and 0.70 ± 0.43 µM in the absence and presence of fentanyl; in the presence of GTPS, a single population of low-affinity sites were detected with IC50 values = 0.80 ± 0.29 and 0.85 ± 0.34 µM in the absence and presence of fentanyl, respectively).

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

The present study suggests that fentanyl electively antagonizes, at least in part, various effects that dobutamine exerts on the myocardium by stimulation of cardiac ₁-adrenergic receptors (Tuttle and Mils, 1975; Leier and Unverferth, 1983). The actions of dobutamine on the vasculature, i.e., decreases in SVRI and PVRI, attributable to stimulation of ₂-receptors (Leier and Unverferth, 1983) and/or blockade of  receptors (Ruffolo et al., 1981), remained unaffected.

CI, DO₂-I, and VO₂-I are relevant parameters in the management of critically ill patients, established to predict the risk of patients to develop potential lethal complications. In patients suffering from shock, the elevation of these parameters up to "supranormal" values might be important, because normal or decreased values reflect a net cumulative oxygen deficit accompanied with higher mortality (Shoemaker, 1995).

The administration of opioid derivatives such as fentanyl can result in hypotension. Several mechanisms contribute to this response, including peripheral arterial vasodilation, reduced responsiveness to the baroreceptor reflex, and venous dilatation. However, these effects were not prominent in our patients, because ABPm, SVRI, PAWP, and CVP remained unchanged under fentanyl.

Fentanyl exerts direct effects on the heart. When reviewing the literature, clinical studies on cardiovascular effects of fentanyl reveal different and confusing results: a decrease in heart rate (Liu et al., 1977) was as well observed as an unchanged heart rate (Stanley and Webster, 1978; Bennett and Stanley, 1979; Thomson et al., 1988). Inotropic effects on CO are described as positive or negative (Liu et al., 1977; Bennett and Stanley, 1979). SVRI may decrease (Liu et al., 1977; Bennett and Stanley, 1979) or remain unchanged (Bennett and Stanley, 1979). Accordingly, the exact effects of fentanyl on the heart and vasculature remain obscure.

Benzodiazepines are known to alter hemodynamics: midazolam decreases heart rate, ABPm, PAWP, CI, and SVRI (Massaut et al., 1983). Moreover, negative inotropic effects of fentanyl seem to be enforced by concomitant administration of benzodiazepines (Stanley and Webster, 1978; Heikkila et al., 1984; Thomson et al., 1988). SVRI decreases during combined treatment with benzodiazepines and fentanyl (Stanley and Webster, 1978; Tomichek et al., 1983). Interestingly, no cardiovascular depression was noticed when the sympathetic and the parasympathetic tone was eliminated (Flacke et al., 1985).

Our analysis suggests that the change in CI is mainly due to the negative chronotropic effect of fentanyl. The parameters used indirectly to assess the force of ventricular contraction (i.e., SVI, LVSWI, and RVSWI) remained unaffected. In contrast to previous studies, SVRI did not further decrease when fentanyl was administered in addition to midazolam. We therefore attempted to evidenciate mechanisms contributing to the observed antagonism between dobutamine and fentanyl. In isolated perfused hearts, morphine and other agonists decreased CO predominantly via depression of the heart rate. This negative chronotropic effect was not relieved by the concomitant application of atropine but was blocked by naloxone (Vargish et al., 1987a,b). The existence of opioid receptors on the myocardium was further substantiated by experiments on cardiac myocytes of rats. Activation of opioid receptors resulted in G_i-dependent inhibition of contraction frequency (Ela et al., 1993). As a consequence, the morphine-induced depression of cardiac function was attenuated by naloxone (Vargish et al., 1987b). A direct cross talk between G_s- and G_i-coupled receptors has been demonstrated to occur in the absence of second messenger synthesis via a membrane-delimited pathway (Ferré et al., 1991) and may involve an intermediate protein that is distinct from G protein  subunits and dimers (Nanoff et al., 1995).

However, our observations clearly show that fentanyl does not affect the ability of -adrenergic receptors to couple to their cognate G protein G_s. As demonstrated by saturation experiments, a direct replacement on the receptor can be ruled out either.

Noteworthy, the negative chronotropic effect of fentanyl was observed although the levels of circulating dobutamine and epinephrine actually increased. Previous studies have reported clear cut increases in the plasma concentrations of catecholamines after the administration of opioid agonists in human volunteers (Hoehe and Duka, 1993) and in animals (May et al., 1988). These changes have been attributed to an action of opioids within the central nervous system. The changes detected in the present study are modest when considering epinephrine and undetectable for norepinephrine. We believe that this difference may be attributable to the fact that we administered fentanyl to sedated patients. It is reasonable to assume that activation of the sympathetic nervous system via a central site of action may be impaired under these conditions. Moreover, all patients have been comatose, which might have further altered hormonal and hemodynamic response by a reduced sympathetic outflow (Flacke et al., 1985).

The steady-state levels of dobutamine (~180 ng · ml¹) and the corresponding total body clearance (~57 ml · min¹ · kg¹) determined in five patients are in excellent agreement with the parameters reported in literature (~170 ng · ml¹ and ~60 ml · min¹ · kg¹; Kates and Leier, 1978). In one of our patients, the clearance of dobutamine was lower and, accordingly, the steady-state level was higher. This interindividual variability is not unexpected. A previous study indicated that the values determined for the clearance of dobutamine differed up to 5-fold in healthy young volunteers (Berg et al., 1993) and even larger variations have been reported for critically ill pediatric patients (Schwartz et al., 1991). The elimination of dobutamine is rapid; half-lives in the range of 2.4 min were originally calculated from the decay of the plasma concentration after discontinuation of the dobutamine infusion (Kates and Leier, 1978). However, a study on the pharmacokinetics of dobutamine indicated that its elimination is governed by a biexponential function, i.e., the sum of an initial rapid process (T_1/2 = 1.7 min) and a second phase (T_1/2 = 26 min); this is consistent with elimination of dobutamine from two compartments (Schwartz et al., 1991). Our blood sampling protocol was not designed to address this issue but to confirm that steady-state values had been reached before the initiation of the fentanyl infusion. Nevertheless, if the half-life is calculated from the rate constant describing the approach to steady state, intermediate values are obtained (mean ± S.D. = 9.9 ± 5 min; range 4.7-18 min). It is likely that these half-lives reflect, in fact, an average estimate of the two phases. The infusion of fentanyl increased the plasma concentrations of dobutamine. To the best of our knowledge, drug interactions based on pharmacokinetic interferences have not been documented for fentanyl and dobutamine. The pharmacokinetics of dobutamine are known to be affected by the concomitant administration of dopamine (Schwartz et al., 1991; Eldadah et al., 1991). However, administration of fentanyl did not alter the plasma concentration of endogenous dopamine. Thus, the observed increase might, at least in part, be due to the changes in CO, which can per se account for a reduction in the clearance of dobutamine. However, correlation analysis between the increase of dobutamine plasma concentrations and the decrease in heart rate did not reveal any statistical significance (data not shown). Irrespective of the nature of the underlying process, we stress that the changes in the plasma concentrations of dobutamine induced by fentanyl were modest. Most importantly, these alterations were not accompanied by any increased pharmacodynamic effect of dobutamine.

Summarizing our results, fentanyl in addition to midazolam and dobutamine neither exerted negative inotropic effects on the heart nor decreased SVRI, as expected. We only could observe negative chronotropic effects, which seem to be responsible for wrecking the beneficial effects of dobutamine on oxygen metabolism. Because these hemodynamic alterations were not similar to those usually observed during combined administration of fentanyl and benzodiazepines (Liu et al., 1977; Tomichek et al., 1983; Heikkila et al., 1984), we assume that first and foremost fentanyl alone was responsible for the observed sympatholytic effects.

Both dobutamine and potent opioid analgesics such as fentanyl are widely used for hemodynamic support and analgesia as well as sedation in the management of critically ill patients. However, based on our results we conclude that the combined treatment with dobutamine, benzodiazepines, and fentanyl or other opioids should be carefully monitored. This is particularly relevant in patients who depend on the salutary effect of dobutamine on oxygen delivery and consumption (Shoemaker et al., 1974; Shoemaker, 1995), especially those suffering from septic shock, patients with severe cardiac failure (Shoemaker et al., 1991), and patients affected by an acute respiratory distress syndrome (Steltzer et al., 1994). When using dobutamine, the clinician who resorts to the concomitant administration of fentanyl and midazolam should therefore be aware of the functional antagonism exerted by this opioid. A substitution with other agents such as ketamine, which does not decrease CI (Schwartz and Horwitz, 1975) and is discussed to be the anesthetic agent of choice in clinical situations when O₂ availability is reduced (Berman et al., 1990), might be preferable. Thus, the leading role of opioids as analgesic agents should be redefined in patients depending on high oxygen delivery and consumption. In contrast, the pharmacokinetic interaction described in the present work, which results in an altered metabolism of dobutamine, is presumably of little clinical relevance in the short-term management of patients.