Patients may receive more than one positive inotropic drug to improve myocardial function and cardiac output, with the assumption that the effects of two drugs are additive. The authors hypothesized that combinations of dobutamine and epinephrine would produce additive biochemical and hemodynamic effects.
The study was performed in two parts. Phase 1 used human lymphocytes in an in vitro model of cyclic adenosine monophosphate (cAMP) generation in response to dobutamine (10(-8) to 10(-4) M) or epinephrine (10(-9) M to 10(-5) M), and dobutamine and epinephrine together. Phase 2 was a clinical study in patients after aortocoronary artery bypass in which isobolographic analysis compared the cardiotonic effects of dobutamine (1.25, 2.5, or 5 microg x kg(-1) x min(-1)) or epinephrine (10, 20, or 40 ng x kg(-l) x min(-1)), alone or in combination.
In phase 1, dobutamine increased cAMP production 41%, whereas epinephrine increased cAMP concentration approximately 200%. However, when epinephrine (10(-6) M) and dobutamine were combined, dobutamine reduced cAMP production at concentrations between 10(-6) to 10(-4) M (P = 0.001). In patients, 1.25 to 5 microg x kg(-1) x min(-1) dobutamine increased the cardiac index (CI) 15-28%. Epinephrine also increased the CI with each increase in dose. However, combining epinephrine with the two larger doses of dobutamine (2.5 and 5microg x kg(-1) x mi(-1)) did not increase the CI beyond that achieved with epinephrine and the lowest dose of dobutamine (1.25 microg x kg(-1) x min(-1)). In addition, the isobolographic analysis for equieffective concentrations of dobutamine and epinephrine suggests subadditive effects.
Dobutamine inhibits epinephrine-induced production of cAMP in human lymphocytes and appears to be subadditive by clinical and isobolographic analyses of the cardiotonic effects. These findings suggest that combinations of dobutamine and epinephrine may be less than additive.
DRUG combinations can provide therapeutic advantages and improved outcome compared with monotherapy for the treatment of many disorders. Most drug combinations use agents with differing mechanisms of action to achieve additive or synergistic effects. Combinations of a catecholamine and a phosphodiesterase inhibitor have been used to treat chronic congestive heart failure and low cardiac output syndrome after cardiac surgery. [1,2]Two (or more) catecholamines are often infused simultaneously in critically ill patients. The theoretical rationale for confusing drugs in this manner is to combine the positive attributes of both drugs, limit the detrimental side effects of high doses of a single agent, and, perhaps, to achieve a greater therapeutic response than possible with either drug alone (pharmacologic synergism, also known as super-additivity). However, two agonist drugs may interact in an additive, synergistic, or pseudoantagonistic manner. This study was designed to determine whether the [Greek small letter beta]-receptor agonist dobutamine is additive to the effect of epinephrine.
The investigation was performed in two phases. In phase 1, using human lymphocytes studied in vitro, we measured cyclic adenosine monophoshate (cAMP) generation by the [Greek small letter beta]-adrenergic receptor-adenylyl cyclase-cAMP pathway in response to dobutamine and epinephrine stimulation. Phase 2 was a clinical investigation of patients who had aortocoronary bypass surgery to determine the cardiotonic effects of epinephrine and dobutamine, alone and in combination.
Materials and Methods
These studies were approved by the Clinical Research Practices Committee of The Bowman Gray School of Medicine, Wake Forest University, with informed consent obtained from all participants.
The lymphocyte cAMP method we used was similar to that previously published. [4–7]Samples of 50 ml venous blood were obtained from 10 healthy volunteers before 10: A.M. on the experimental day. Volunteers were excluded if any medication was taken 30 days before blood sampling and if [Greek small letter beta]-adrenergic antagonist medications had ever been taken. Blood from only one volunteer was used in any one experimental preparation. Blood was anticoagulated with heparin sodium and immediately diluted with an equal volume of phosphate-saline buffer (NaCl = 140 mM, KCl = 2.7 mM, KH2PO4= 1.5 mM, K2HPO4= 6 mM). Lymphocytes were isolated as described by Boyum by applying diluted blood to Ficoll-Paque (Pharmacia Laboratory Separation, Pharmacia, Piscataway, NJ) and subjecting it to density-gradient centrifugation at 1,300 g for 30 min to generate a lymphocyte-rich (70% T lymphocytes and 10% B lymphocytes) band of cells. This lymphocyte band was aspirated, washed twice in ice-cold phosphate-buffered saline solution, and divided into separate test tubes, each of which contained approximately 1 x 106lymphocytes/ml, quantified by Coulter Counter (Coulter Electronics, Hialeah, FL) set to recognize only mononuclear cells. 3-isobutyl-1-methylxanthine (IBMX, 0.01 M) was added to each test tube to block cAMP hydrolysis.
Adenylyl cyclase activity was determined in an incubation mixture containing 140 mM NaCl 1.5 mM CaCl2, and IBMX. The reaction was initiated by adding cells to the incubation mixture, for a final reaction volume of 500 [micro sign]l. Two tubes from each group were not exposed to any agonist drug to determine baseline cAMP production, whereas the remaining tubes of lymphocytes were mixed with graded concentrations of the [Greek small letter beta]-agonists epinephrine, dobutamine, or a combination of epinephrine plus dobutamine. The concentrations of each drug used in this study were determined from previous experience with the lymphocyte-cAMP experimental model [6,7]and ranged from 10-8M to 10-4M dobutamine, and from 10-9M to 10-5M epinephrine. The tubes were incubated for 10 min at 27 [degree sign]C.
Incubation was terminated with the addition of a solution of cold 3% perchloric acid and with subsequent addition of 2% KHCO3. All samples were chilled to 1–2 [degree sign]C, and cAMP was measured in duplicate using a radioimmunoassay method modified from Gilman (RIA, Diagnostic Products Corporation, Los Angeles, CA). Measured cAMP production was standardized to 1 x 106lymphocytes. The values for cAMP production for the two samples exposed to the same concentration of agonists were averaged, resulting in a single cAMP production value used for statistical analysis. Based on our experience with these procedures, we used an a priori quality control wherein cAMP concentrations from the two lymphocyte samples exposed to the same concentration of agonist that varied by <25% were averaged and considered valid. Our experience with this model finds that only 1–2% of samples deviate outside this range. The entire process was repeated on separate study days for each of the 10 volunteer blood samples.
After patients gave written informed consent, those undergoing elective aortocoronary bypass graft surgery with normal preoperative left-ventricular function (ejection fraction >or= to 0.50) were studied in the cardiac surgical intensive care unit (ICU) on their first postoperative day. Inclusion criteria required that the patient be extubated in the ICU, demonstrate respiratory and hemodynamic stability, and have manifest no need for any inotropic drug support during the entire postoperative period. Previous infusion of vasoactive drugs (such as phenylephrine, nitroglycerin, sodium nitroprusside, and so on) was permitted if such infusions were terminated before the study was begun. In addition, diuretic therapy was temporarily withheld during the clinical protocol. Only mild sedatives or analgesics were administered concomitant with study medications, if necessary.
Eight patients qualified for the study in the ICU. Baseline hemodynamic measurements obtained included heart rate (HR), mean arterial pressure, pulmonary artery pressure, pulmonary artery occlusion pressure, and cardiac output in duplicate (a third determination was done if the first two varied by more than 10%). Cardiac index (CI) and the percentage changes after therapy were calculated by standard formulas. All study inotrope medications were administered into a central catheter, with a secondary carrier solution of D5W infusing at a constant 100 ml/h to ensure rapid drug bioavailability. An epinephrine infusion at 10 ng [middle dot] kg-1[middle dot] min-1was infused for 8 min, and hemodynamic measurements were repeated. We previously determined that the clinical (cardiotonic) effects of these drugs reach steady state within 8 min, and that 15 min is adequate to allow for return to baseline. [2,8–10]The infusion was increased to 20 and then to 40 ng [middle dot] kg-1[middle dot] min-1with data recorded after each incremental dose was infused for 8 min. After a 15-min recovery period allowed for hemodynamic parameters to return to baseline, dobutamine was administered (1.25 [micro sign]g [middle dot] kg (-1)[middle dot] min-1) and data were recorded after 8 min. Then, while dobutamine was being infused, the epinephrine infusion was administered at 10, 20, and 40 ng [middle dot] kg-1[middle dot] min-1with hemodynamic data obtained 8 min after each incremental dose (see the flow diagram in Table 1). The epinephrine infusion was then discontinued and the dobutamine infusion was increased to 2.5 [micro sign]g [middle dot] kg-1[middle dot] min-1and the epinephrine infusion was added at the same three doses as previously described. The epinephrine infusion was again discontinued and the dobutamine infusion increased to 5 [micro sign]g [middle dot] kg-1[middle dot] min-1. After 15 min, hemodynamic measurements were performed and the epinephrine infusion was again added at the three doses, as previously described. No other intravenous fluids were administered during the study. Data were recorded with dobutamine alone and after each combination of dobutamine and epinephrine.
Basal cAMP production was compared using analysis of variance. To evaluate drug efficacy, the maximal cAMP production was used, divided by baseline cAMP production before the statistical analysis to adjust for baseline differences among lymphocyte groups. To evaluate drug potency, the minimum, maximum, and the median effective concentrations were determined by fitting the data to a sigmoidal distribution curve with a Hill coefficient of 1, defined by the Equation 1where E = the drug effect (cAMP production), Emin = the minimum cAMP concentration, Emax = the maximum cAMP concentration, EC50= the concentration producing 50% of the maximal effect, and conc = the concentration of the agonist. This Equation assumesthat the agonists bind freely with the receptor sites without negative or positive cooperativity. The slopes for change in absolute cAMP production between consecutive dobutamine concentrations along the epinephrine concentration curves were estimated and then tested for significant difference from 0. Statistical contrasts were performed to compare similar slopes at different dobutamine concentrations.
In phase 2, the actual recorded parameters (i.e., HR and CI) were analyzed to evaluate the interaction between epinephrine and dobutamine with a doubly repeated-measures mixed-effects analysis of variance. Within-subject factors included epinephrine, dobutamine, and their interaction. Corrections were made for multiple comparisons using Fisher's protected least significant difference test as appropriate. Statistical analyses were performed using software (Proc Mixed) in the SAS statistical package (version 6.08; SAS Institute, Cary, NC). In addition, a CI isobologram was constructed. This graphical technique uses equieffective concentrations of individual drugs and the combination of two drugs. [13–16]We plotted dose-response curves with dobutamine, epinephrine, and the dobutamine-epinephrine combinations and chose the dose from these curves that resulted in a 10% increase in CI. On the horizontal and vertical axes of the isobologram, the dose of dobutamine and epinephrine that would result in a 10% increase in CI is plotted. A line of additivity is drawn between these two points. It would be expected that any combination of drugs along this line would produce a 10% increase in CI. If a 10% increase in CI results from a drug dose combination below this line (i.e., smaller doses than predicted), drug synergism is present. If the dose combination producing a 10% increase in CI lies above this line (i.e., larger doses than predicted), then pseudoantagonism is present. Data are presented as mean +/- SD, and alpha levels of 0.05 were used throughout.
Dobutamine alone resulted in a small but significant increase in cAMP production of 41%(3.2 +/- 0.9 to 4.6 +/- 0.9 pmole [middle dot] 10-6lymphocytes [middle dot] 10 min-1; P = 0.005) in concentrations varying from 10-8M to 10-4M dobutamine (Figure 1). Previously we showed that concentrations of dobutamine <or= to 10-8M are below the stimulatory threshold of cAMP production in this model and not significantly different from a zero drug concentration. Epinephrine concentrations from 10-9M to 10-5M increased cAMP above baseline by [approximately] 200%(3.6 +/- 0.9 to 10.6 +/- 2.5 pmole [middle dot] 10-6lymphocytes [middle dot] 10 min-1; P < 0.0001;Figure 1).
(Figure 2) shows cyclic AMP production stimulated by varying concentrations of epinephrine plus dobutamine. For epinephrine at 10-6M, cAMP production decreased significantly when dobutamine was increased from 10-6M to 10-5M (P = 0.0011), and also from 10-5M to 10-4M (P = 0.0009). Similarly, an epinephrine concentration of 10-7M was significantly inhibited by dobutamine concentrations from 10-6M to 10 (-5) M (P = 0.024). Cyclic AMP production did not decrease when epinephrine was mixed with more dilute concentrations of dobutamine (i.e., dobutamine = 10-8M to 10-6M). Furthermore, concentrations of epinephrine > 10 (-8) M showed no evidence for inhibition by dobutamine (Figure 2).
In a subset of lymphocytes, blood samples (n = 5) were examined to determine whether the [Greek small letter beta]-agonist inhibition demonstrated above is competitive. If so, then very high doses of the full agonist epinephrine should overcome the pseudoantagonism produced by the partial agonist dobutamine. In this subset analysis, cAMP production produced by 10-5M epinephrine was not significantly inhibited by dobutamine concentrations in the range of 10-6M to 10-5M.
Eight patients (of 37 who gave initial, preoperative consent) qualified for the study in the ICU. Twenty-three patients were excluded because of use of concomitant medications (epinephrine, amrinone, dopamine, nitroglycerin, or nitroprusside), one patient could not be studied because no investigator was available, one patient was still intubated, one patient was having gastrointestinal distress, and three patients had their surgery canceled. Figure 3and Figure 4show the HR and CI responses for both drugs alone and for the various combinations. Dobutamine significantly increased HR with each incremental increase in dose (overall dobutamine dose effect on HR; P <or= to 0.0001). Epinephrine significantly increased HR from baseline only at the highest dose (40 ng [middle dot] kg-1[middle dot] min-1; P = 0.0081). Combining epinephrine and dobutamine resulted in the same incremental increase in HR as that produced by doses of dobutamine alone.
Dobutamine doses of 1.25 and 2.5 [micro sign]g [middle dot] kg-1[middle dot] min-1produced equivalent increases in CI (Figure 4), whereas 5 [micro sign]g [middle dot] kg-1[middle dot] min-1of dobutamine produced the greatest increase in CI (28%). Epinephrine alone increased CI (compared with baseline) with each incremental increase in dose (10, 20, and 40 ng [middle dot] kg-1[middle dot] min-1). However, the combination of epinephrine with the two larger doses of dobutamine (2.5 and 5 [micro sign]g [middle dot] kg-1[middle dot] min-1) was no better at increasing CI than that achieved with epinephrine plus the lowest dose of dobutamine (1.25 [micro sign]g [middle dot] kg-1[middle dot] min-1; P > 0.05).
(Figure 5) illustrates the isobologram for dobutamine and epinephrine on CI. Using dose-response curves with dobutamine, epinephrine, and the dobutamine-epinephrine combinations, the dose that resulted in a 10% increase in CI was determined. On the horizontal and vertical axes of the isobologram, the dose of dobutamine and epinephrine that would result in a 10% increase in CI is plotted. A line of additivity is drawn between these two points. If the effects were additive, it would be expected that any combination of drugs that achieved a 10% increase in CI would result in doses of drugs taken from this line. If the 10% increase in CI results from drug dose combinations below this line (smaller doses than predicted), drug synergism is present. If the dose combination results in drug doses above this line (larger doses than predicted), then pseudoantagonism is present. Visual inspection reveals that the observed data points fall above and to the right of the line of additivity, suggesting clinical pseudoantagonism between these two [Greek small letter beta]-agonists.
This study shows pseudoantagonism of biochemical and clinical endpoints when the two [Greek small letter beta]-agonists dobutamine and epinephrine are combined. Phase 1 tested the intracellular molecular signaling via the [Greek small letter beta]-adrenergic receptor-adenylyl cyclase-cAMP pathway in human lymphocytes. We found that dobutamine (>or= to 10-6M), combined with epinephrine (concentrations of 10-6M and 10 (-7) M), significantly reduced cAMP production in the in vitro model (compared with epinephrine alone). In addition, the cardiotonic effects of epinephrine and dobutamine were prospectively analyzed in postoperative cardiac surgery patients via isobolographic analysis of CI. The combination of dobutamine with epinephrine requires more drug (than predicted) to produce a 10% increase in CI, based on the epinephrine-dobutamine isobologram. We conclude that the combination of these two [Greek small letter beta]-adrenergic agonists (at these concentrations) results in inotropic drug pseudoantagonism.
The classic drug-receptor interaction model postulates an agonist binding to a cellular receptor, activating it, and producing a maximal physiologic response (usually defined as drug efficacy), whereas a partial agonist binds to the same receptor but generates only a fraction of this maximal response. By these definitions, epinephrine is a full agonist in our model (maximal cAMP response = 10.6 pmole [middle dot] 10-6lymphocytes in 10 min) and dobutamine is a partial agonist (maximal cAMP response = 4.6 pmole [middle dot] 10-6lymphocytes in 10 min). We previously identified similar drug efficacy and potency while describing six [Greek small letter beta]-adrenergic receptor agonists in the human lymphocyte model. This classic model of drug-receptor interaction implies that partial agonists should have “pseudoantagonistic properties”; that is, they occupy receptors while producing weaker effects. This pharmacologic phenomenon has been well characterized previously with other drugs. 
However, a new model for receptor-mediated response suggests that the key step in transmembrane signal transduction is the transition of the [Greek small letter beta]-receptor from the inactive (R) to the active (R*) state (Figure 6). [17,19,20]In the traditional model, the occupation of these receptors by an agonist results in activation (R*). However, evidence now suggests that hormones may not actually be involved in the conversion from the inactive to the fully active receptor. Rather, the receptor is in a natural state of flux between R and R* conformations, and agonists preferentially select the active form, thereby shifting the equilibrium toward the active (R*) conformation. [17,19,20]Thus, in this model, agonists do not activate receptors but rather bind selectively to the already active conformation, favoring movement of the equilibrium in that direction. An antagonist binds with equal affinity to both active and inactive forms but does not alter the equilibrium while occupying receptors. Finally, a partial agonist has only slightly greater affinity for R* compared with R, which means that even at saturating concentrations, the magnitude of the observed effect will be less (i.e., less cAMP production by dobutamine compared with epinephrine). Many drugs used clinically are partial agonists and are predicted to produce subadditive effects when combined with a full agonist. 
From our previous experience and characterization of this human lymphocyte model, epinephrine is about 10 times more potent (EC50= 10 (-7).335 M) than dobutamine (EC50= 10-6.124 M) at the [Greek small letter beta]2-adrenergicreceptor. This would be consistent with inspection of the data shown in Figure 1and Figure 2. That is, pseudoantagonism of 10-7M epinephrine requires 10 to 100 times greater concentrations of dobutamine, concentrations that presumably are sufficient to account for the differences in molecular affinity at the [Greek small letter beta]-receptor. Similarly, for 10-6M epinephrine, it requires 10 to 100 times greater concentrations of dobutamine for subadditive effects to be observed. Conversely, however, higher concentrations of epinephrine (10 (-5) M) overcome the inhibition of dobutamine in the range of 10-6M to 10-4M.
Are the concentrations and doses of epinephrine and dobutamine used in this study relevant to clinical practice? Most patients receiving infusions of catecholamines exhibit highly variable serum concentrations of these drugs. [21–23]Standard dobutamine infusions in patients in the ICU produce a 50-fold range of serum concentrations (0.05 x 10-6M to 2.5 x 10-6M). We found that postoperative epinephrine infusions of 0.03 [micro sign]g [middle dot] kg-1[middle dot] min-1patients having coronary artery bypass grafting produced blood concentrations of about 6 x 10-6M. Thus the drug concentrations and doses we used in the in vitro phase of this study appear to be within, or near, the concentrations achieved clinically with typical infusion rates for these two catecholamines in critically ill patients.
The effects of dobutamine and epinephrine on CI are displayed in Figure 4and in the isobologram of Figure 5. The low dose of dobutamine (1.25 [micro sign]g [middle dot] kg-1[middle dot] min-1) increases CI as it is added to low doses of epinephrine, and this is clinically what we would expect, and indeed demonstrated. However, the combination of epinephrine with the two larger doses of dobutamine (2.5 and 5 [micro sign]g [middle dot] kg-1[middle dot] min-1) was no better at increasing CI than that achieved with epinephrine plus the lowest dose of dobutamine (1.25 [micro sign]g [middle dot] kg-1[middle dot] min-1). This is a less-than-additive effect between a partial agonist (dobutamine) and a full agonist (epinephrine). We would not expect to see an actual decrease in the CI by the addition of dobutamine unless higher doses of epinephrine were being used. Obviously, we had to restrict the clinical doses of these inotropes to reasonable, safe, and prudent ones. However, the clinical effects on CI are less than additive (less than the predicted amount), as seen by isobolographic analysis. Although we wanted to enroll additional patients into the clinical investigation to increase the statistical power, the clinical pathway for cardiac surgery commonly called “fast tracking” significantly altered the recovery paradigm for these patients (early removal of pulmonary artery catheter, aggressive diuresis, early mobilization, and ICU discharge) which restricted our opportunity to study these patients under the same clinical conditions we describe here.
Additional limitations are evident. First, our findings may be applicable only to epinephrine and dobutamine used in postoperative cardiac patients. However, we specifically selected these patients, and these drugs, because of the frequency that these drugs are used after this common surgical procedure. Second, acute [Greek small letter beta]-receptor desensitization could occur during repeated catecholamine infusions during the study protocol (Table 1). However, we have not seen acute, demonstrable alterations in the clinical response to repetitive catecholamine infusions in this patient population in our past investigations. [2,8–10,24]Nonetheless, others have identified acute receptor desensitization of both [Greek small letter beta]1and [Greek small letter beta]2receptors after cardiopulmonary bypass. [25,26]This pattern may be reversed by 30 min after extracorporeal circulation (although the time course of [Greek small letter beta] receptor downregulation is more variable). Because our cardiac surgery patients were tested 24 h after operation, we assumed that many of these alterations abated at the time of this study. A third potential limitation is the in vitro model of human lymphocytes, which contains a homogenous population of [Greek small letter beta]2receptors. Despite some continued debate, most investigators acknowledge that the [Greek small letter beta]2-adrenoreceptoron circulating lymphocytes accurately predicts adrenergic function in general, [27,28]and [Greek small letter beta]2receptor function specifically in the healthy and septic heart. We have extensive experienced with this model that shows the interaction of local anesthetics and sepsis with harvested lymphocytes. Although only 40%(range, 14–59%) of [Greek small letter beta]-adrenergic receptor subtype, the [Greek small letter beta]2receptor subtype is particularly critical for intracellular cAMP generation, because it may account for as much as 10 times more cAMP production per activated receptor than the [Greek small letter beta]1receptor. Thus we believe this combined evidence supports the validity of the circulating lymphocyte-cAMP model as an acceptable in vitro model to evaluate human [Greek small letter beta]2adrenoreceptor function, and that the advantages of studying actual human tissue offset potential limitations. Finally, we recognize that it would be ideal to test the isobole interaction at several CI endpoints, such as drug doses required to produce increases in CI of 10%, 30%, and 50%. However, clinical limitations, patient safety, and patient tolerance did not allow for such an ideal drug titration.
Although many physician believe that all inotropic drug combinations are additive or even synergistic, we found that specific concentrations of dobutamine and epinephrine are, in fact, less than additive. It is unclear how these findings might influence the management of critically ill patients, but we believe clinicians should be aware of the complexity of drug interactions when drugs have an affinity for the same receptor. Decreased responsiveness to catecholamine infusions is recognized in a wide variety of clinical scenarios, including congestive heart failure, after thoracic and abdominal surgical operations, [36,37]after reversible myocardial ischemia, and during sepsis. [39–42]Although receptor desensitization and downregulation certainly play a role in the decreased responsiveness to catecholamine infusions, we speculate that drug pseudoantagonism also may occur during concurrent administration of two or more [Greek small letter beta]-adrenergic inotropic drugs. We would advise close monitoring of the hemodynamic and cardiovascular system to observe for possible drug interactions.
The authors thank Jian P. Leith and Yonggu A. Lin for their help in performing the cAMP assays.