Midazolam and alfentanil are desirable anesthetic adjuncts for cardiac anesthesia. They are metabolized by cytochrome P450 3A (CYP3A) enzymes. These isozymes are inhibited by concurrent medications, including the calcium channel antagonist diltiazem, which may have an effect on recovery from anesthesia.
Thirty patients having coronary artery bypass grafting were randomly assigned to receive either diltiazem (60 mg orally 2 h before induction of anesthesia and an infusion of 0.1 mg.kg-1.h-1 started at induction and continued for 23 h) or placebo in a double-blind study. Anesthesia was induced with 0.1 mg/kg midazolam, 50 micrograms/kg alfentanil, and 20 to 80 mg propofol and maintained with infusions of 1 microgram.kg-1.min-1 of both midazolam and alfentanil supplemented with isoflurane. Plasma midazolam and alfentanil concentrations and areas under the plasma concentration-time curves were determined. The terminal half-life and the time for the drug plasma level to decrease 50% after cessation of the infusion (t50) were calculated for midazolam and alfentanil. Separation from mechanical ventilation and tracheal extubation were performed according to the study protocol.
Diltiazem increased the mean concentration-time curves (from end of anesthesia until 23 h) of midazolam by 24% (P < 0.05) and that of alfentanil by 40% (P < 0.05). The mean half-life of midazolam was 43% (P < 0.05) and that of alfentanil was 50% (P < 0.05) longer in patients receiving diltiazem. The mean t50 of alfentanil was 40% longer (P < 0.05) in patients receiving diltiazem, but the change in the mean t50 of midazolam (25%) was not statistically significant. In patients receiving diltiazem, tracheal extubation was performed on average 2.5 h later (P = 0.054) than in those receiving placebo.
Diltiazem slows elimination of midazolam and alfentanil and may delay tracheal extubation after large doses of these anesthetic adjuncts. CYP3A-mediated drug interactions should be considered as confounders when recovery from anesthesia with midazolam and alfentanil infusions is assessed.
Key words: Anesthetic, intravenous: alfentanil; midazolam. Drugs: calcium channel blockers; diltiazem. Enzyme: cytochrome P450 3A (CYP3A). Pharmacokinetics and pharmacodynamics: alfentanil; midazolam; drug-drug interactions; diltiazem. Surgery: coronary artery bypass graft.
Midazolam and alfentanil are intravenous anesthetic adjuncts with desirable pharmacokinetic characteristics, including fast recovery, for patients having cardiac surgery. [1-3]They are metabolized by cytochrome P450 3A (CYP3A) enzymes. [4-6]These isozymes are inhibited by many substances both in vitro [7,8]and in vivo. [9-12]In humans, the antibiotic erythromycin and some azole antimycotics, which inhibit CYP3A enzymes, reduce the clearance of intravenous midazolam by 50% to 70% [9,10]and that of alfentanil by 25% to 80%. [11,12]
The calcium antagonists diltiazem and verapamil are also effective inhibitors of CYP3A in vitro. [7,13]In humans, they increase the area under the oral midazolam plasma concentration time curve by 200% to 300%. The possible interaction of the calcium antagonists with intravenous midazolam or alfentanil has not been studied in humans. Because many patients scheduled for coronary artery bypass grafting (CABG) receive long-term medication with calcium antagonists, including diltiazem, we wanted to determine if diltiazem increases the plasma concentrations of midazolam and alfentanil and affects the time to awakening and tracheal extubation in patients having CABG.
Materials and Methods
We used a double-blind, randomized, placebo-controlled study design in parallel groups. Based on our experience of the time to tracheal extubation (which was regarded as the primary pharmacodynamic end point), we determined that 15 patients would be needed in each group to demonstrate a 50% difference in the time to tracheal extubation at a level of significance of P = 0.05 and power of 80%. We obtained institutional approval and written informed consent to study 30 patients having elective CABG. The patients were randomly assigned to receive either diltiazem or placebo. Because five of the initially enrolled 30 patients had to be excluded from the study because of violation of the study protocol, five additional patients were enrolled. Table 1shows the patient characteristics. All the staff in the ward, in the operating room, and in the intensive care unit was unaware of the patient randomization code.
Exclusion criteria were left ventricular ejection fraction less than 40%, significant valvular dysfunction, sick sinus syndrome or second- to third-degree atrioventricular block, liver insufficiency, treatment with any calcium antagonist during the previous month, treatment with any known inhibitor or inducer of CYP3A enzymes, and morbid obesity. All routine medications, including beta-blockers, long-acting nitrates, and angiotensin converting enzyme (ACE) inhibitors, were terminated the evening before the operation; aspirin was stopped 6 to 7 days before surgery. Two hours before induction of anesthesia, the patients received 40 micro gram/kg orally administered lorazepam (Temesta; Wyeth Laboratories, Munster, Germany). At the same time they received orally either 60 mg diltiazem (Dilzem; Orion, Espoo, Finland) or placebo.
Before anesthesia was induced, peripheral venous and radial arterial cannulas were inserted and an infusion of 0.1 mg [centered dot] kg sup -1 [centered dot] h sup -1 diltiazem (Dilzem; Godecke, Berlin, Germany) or the corresponding placebo (saline) was started and continued until the next morning. The rate of the diltiazem infusion was chosen according to a recent study in patients having CABG. Anesthesia was induced with 0.1 mg/kg midazolam (Dormicum; Hoffmann-La Roche, Basel, Switzerland), 50 micro gram/kg alfentanil (Rapifen; Janssen Pharma, Beerse, Belgium), and 20 to 80 mg propofol (Diprivan; Zeneca Pharma, Cheshire, UK). At the beginning of the induction, continuous infusions of 1 micro gram [centered dot] kg sup -1 [centered dot] min sup -1 midazolam and 1 micro gram [centered dot] kg sup -1 [centered dot] min sup -1 alfentanil were started and maintained unchanged until skin closure. Atracurium (Tracrium; Wellcome, London, UK), 0.5 mg/kg, was given for muscle paralysis, which was continued with a 0.5-mg [centered dot] kg sup -1 [centered dot] h sup -1 infusion. The dose of the atracurium infusion was halved during cardiopulmonary bypass (CPB), and the infusion was stopped at the beginning of the sternal closure, so that all four twitches in the train-of-four sequence were seen at the end of surgery. After endotracheal intubation, the lungs were ventilated with a mixture of oxygen in air. Isoflurane (Forene; Abbott, North Chicago, IL) administration was begun before the skin incision and increased to the end-tidal concentration of 1% according to the hemodynamic response. Perioperative medications are shown in Table 2.
All surgical procedures were performed under moderate hypothermia (nasopharyngeal temperature, 32 to 34 degrees Celsius). Cold crystalloid cardioplegic solution (Plegisol; Abbott) was used. Before separation from CPB, the patients were thoroughly rewarmed and an infusion of 0.04 micro gram [centered dot] kg sup -1 [centered dot] min sup -1 epinephrine (Adrenalin; Leiras, Turku, Finland) was started in all patients.
To determine the midazolam and alfentanil plasma concentrations, arterial blood samples were drawn before the anesthesia was induced, 10 min after the induction, every 30 min thereafter until the initiation of CPB, 15 min after the initiation of CPB (samples were drawn from the CPB circuit), at the end of CPB (samples from the CPB circuit), every 30 min thereafter until skin closure, at the end of midazolam and alfentanil infusions at skin closure, and at 3-h intervals until the next morning. Blood samples to determine the diltiazem plasma concentrations were drawn before anesthesia was induced, 15 min after the initiation of CPB, at the end of CPB, at skin closure, and at the end of the diltiazem infusion on the first postoperative morning.
The concentrations of midazolam and alfentanil were determined by gas chromatography. [16,17]The sensitivity of the method for midazolam was 1 ng/ml and for alfentanil it was 1 ng/ml. The coefficients of day-to-day variation for midazolam were 8.6% at 3.9 ng/ml (n = 18) and 5.1% at 206 ng/ml (n = 17). The coefficients of variation for alfentanil were 3.8% at 50 ng/ml (n = 7) and 2.1% at 252 ng/ml (n = 7). Diltiazem was quantified by high-performance liquid chromatography. [18,19]The sensitivity of the method was 1 ng/ml, and the coefficient of variation was 2.7% at 157 ng/ml (n = 6).
The pharmacokinetics of midazolam and alfentanil were characterized by three different areas under the drug plasma concentration-time curve: from the induction of anesthesia to the next morning (AUC0-23), from the induction of anesthesia to the initiation of CPB (AUC0-CPB), and from the end of anesthesia to the next morning (AUCEND-23). The AUC values were calculated using the logarithmic trapezoidal rule. Midazolam and alfentanil plasma concentrations just before CPB (CCPB), at the end of anesthesia (CEND), and during the elimination phase were noted. For each patient the terminal log-linear phase of the plasma concentration-time curve was identified visually. The elimination-rate constant (kel) was determined by regression analysis of the log-linear part of the curve. The terminal elimination half-life was calculated from t1/2 = ln2/kel. Furthermore, the time for the drug plasma concentration to decrease 50% after cessation of the infusion (t50) of midazolam and alfentanil was determined by interpolation using the logarithmic plasma concentration-time profile for each patient.
After the operation, the patients were actively warmed either under the Thermal Ceiling (Aragona, Stockholm, Sweden) or using the Bair Hugger forced-air warmer (Augustine Medical, Eden Prairie, MN) until they were awake. When patients gagged against the intubation tube or were extremely agitated (without contact with the patient), they were sedated with bolus doses of 10 to 30 mg propofol. If this was repeated four or five times in a short period, an infusion of propofol was initiated for 30 to 60 min. The time to return of appropriate responses to the command "Move your right and left arm and your legs" and to the question "Do you feel any pain?" was defined as the time of awakening. After awakening, analgesia was provided by intravenous bolus doses of 3 to 5 mg of the opioid oxycodone (Oxanest; Leiras, Turku, Finland).
Separation from mechanical ventilation was initiated when the patients were awake and calm. According to the end-tidal carbon dioxide concentration (ETCO2) and the arterial carbon dioxide tension (PaCO2), mandatory ventilations were reduced, allowing the ETCO2and the PaCO2to increase to 7% or 50 mmHg, respectively. Simultaneously, with increasing spontaneous ventilatory rate, the mandatory ventilations were reduced until the patients were breathing in the presence of 5 cm H2O of continuous positive airway pressure. Extubation criteria were as follows: the patient was breathing during continuous positive airway pressure (FiO2less or equal to 0.40) with the ventilation rate less than 20, the arterial oxygen tension was more than 75 mmHg, the PaCO2was less than 50 mmHg, and chest tube drainage was less than 100 ml/h. Every patient was observed at least until being awakened by the same anesthesiologist (J.A.), and in 24 of the 30 patients the trachea was extubated by him. In six patients the trachea was extubated by the intensive care unit physician strictly according to the study protocol.
Results are expressed as mean values +/- SD. Patient characteristics and pharmacokinetic and pharmacodynamic parameters between the groups were compared using the Student's t test or chi squared test, as appropriate. All the data were analyzed using the Systat for Windows program, version 5.0 (Systat, Evanston, IL).
The two groups were similar with respect to patient characteristics, severity of the coronary artery disease, duration of anesthesia, duration of aortic cross-clamping, duration of CPB, and perioperative medications (Table 1and Table 2).
Diltiazem increased the mean AUC0-23 of midazolam by 15% (P < 0.05) and that of alfentanil by 24% (P < 0.05; Table 3). In addition, diltiazem increased the mean AUCEND-23 of midazolam by 24% (P < 0.05) and that of alfentanil by 40% (P < 0.05). The mean values for the AUC0-CPB, CCPB, and the CENDof midazolam and alfentanil did not differ significantly between the groups. The mean plasma concentrations of midazolam at the last three time points and the mean plasma concentrations of alfentanil at all time points during the elimination phase were greater (P < 0.05) in patients receiving diltiazem (Figure 1, Table 3). The mean t1/2 of midazolam was 43% (P <0.05) and that of alfentanil was 50% (P < 0.05) longer in the diltiazem group. Furthermore, the mean t50of alfentanil was 40% longer (P <0.05) in patients receiving diltiazem, but the change in the mean t50of midazolam (25%) was not statistically significant. Between the groups, the difference in the time to awakening (mean 125 vs. 175 min) was not statistically significant. In patients receiving diltiazem, the trachea was extubated on average 2.5 h later (P = 0.054) than in those receiving placebo (Table 3). Between the end of anesthesia and tracheal extubation, the mean cumulative dose of propofol did not differ significantly between the groups. In the placebo group, 14 patients received propofol with the mean dose of 80 mg, and 12 patients in the diltiazem group received propofol with the mean dose of 100 mg (Table 2).
The mean concentrations of diltiazem were 62 +/- 32 ng/ml before anesthesia was induced; 70 +/- 13 ng/ml 15 min after the initiation of CPB; 65 +/- 16 ng/ml at the end of CPB; 72 +/- 16 ng/ml at the end of anesthesia; and 178 +/- 59 ng/ml at the end of the diltiazem infusion on the next morning.
Diltiazem significantly affected the pharmacokinetics of both midazolam and alfentanil. It increased the areas under the midazolam and alfentanil plasma concentration-time curves from the end of anesthesia until the first postoperative morning by about 25% and 40%, respectively. In patients receiving diltiazem, the time for the drug plasma level to decrease 50% after cessation of the alfentanil infusion was 40% longer, and the terminal elimination half-lives of midazolam and alfentanil were 40% and 50% longer, respectively. Slower elimination of midazolam and alfentanil in the patients receiving diltiazem is compatible with the observed delay of on average 2.5 h for tracheal extubation in these patients.
Based on previous studies, [9,10,14]we expected diltiazem to decrease the clearance of intravenous midazolam and alfentanil but that the pharmacokinetic interaction between diltiazem and these intravenous anesthetics would be identified only several hours after anesthesia was induced. After oral administration, midazolam undergoes extensive first-pass metabolism, with less than 50% bioavailability. Concurrent administration of erythromycin or diltiazem with oral midazolam results in markedly decreased first-pass metabolism and reduced clearance of this benzodiazepine. [9,14]When midazolam is given intravenously, the presystemic metabolism is bypassed and the pharmacokinetics of midazolam is not affected to the same extent as after oral administration. In our study, the total area under midazolam plasma concentration-time curve was increased only by 15% to 25% compared with the 200% to 300% increase after oral administration in healthy volunteers. We noted the differences in the plasma concentrations of midazolam and alfentanil between the patients receiving placebo or diltiazem only during the elimination phase after the end of the anesthesia. The times for the drug plasma level to decrease 50% after cessation of the midazolam (mean, 148 min) and alfentanil (mean, 107 min) infusions that we observed in patients receiving placebo were longer than the context-sensitive half-times modeled by Hughes and colleagues for an anesthetic of 4 to 6 h (70 and 60 min, respectively). However, in contrast to the modeled context-sensitive half-time, we determined the 50% decrease in the drug plasma concentration by interpolation using the logarithmic plasma concentration-time profile for each patient. In addition, the differences observed might, at least in part, be due to the complex effect of CPB on the pharmacokinetics of intravenous anesthetics. A further explanation could be a pharmacokinetic interaction between midazolam and alfentanil, both of which are substrates for CYP3A enzymes.
The perioperative infusion of diltiazem was recently shown to provide antiischemic and antidysrhythmic protection without any adverse effects on perioperative hemodynamics and systolic myocardial function in patients undergoing CABG surgery. Diltiazem plasma concentrations in these patients were not reported. However, the concentrations measured in the present study are comparable to those achieved with an oral sustained-release formulation of 300 mg a day and a standard formulation of 60 mg four times a day (mean peak concentrations of 150 ng/ml and mean trough concentrations of 65 to 80 ng/ml). In healthy volunteers, mean peak plasma concentrations of 70 ng/ml and trough concentrations of 50 ng/ml affected significantly the pharmacokinetics of concurrently administered oral midazolam. The oral bioavailability of midazolam was increased and the clearance was reduced. It is obvious, however, that higher diltiazem plasma concentrations are associated with stronger inhibition of CYP3A enzymes.
Theoretically, the observed differences in the time to awakening and tracheal extubation could be due to the changes in the pharmacokinetics of lorazepam, propofol, or isoflurane. Lorazepam, however, is metabolized by hepatic conjugation to the glucuronic acid, which is a nonmicrosomal reaction and is not affected by changes in cytochrome P450 activity. Isoflurane is eliminated by metabolism only to a clinically insignificant degree. This CYP2E1-mediated metabolism is not known to be affected by diltiazem. The doses of propofol used in our study were small; above all, there is no evidence that propofol metabolism was inhibited by diltiazem. In patients having no medications before surgery, the combination of midazolam and alfentanil was shown to have a supraadditive hypnotic interaction, and the addition of propofol did not change the relative hypnotic potency of the combination. Thus it is unlikely that the differences observed in the time to awakening or tracheal extubation were affected by lorazepam, propofol, or isoflurane. The combined use of midazolam and alfentanil limits our ability to make any conclusions about the relative importance of the diltiazem versus midazolam and diltiazem versus alfentanil interactions. The combined use of midazolam and opioids, however, is frequently applied during cardiac anesthesia, and our study design allows us to make conclusions in a clinically relevant setting.
The time to awakening and tracheal extubation of patients having CABG surgery also can be affected by factors not related to the anesthesia. Physiologic changes associated with the CPB and the surgical procedure itself can affect recovery. In our study, no major neurologic or renal complications occurred in any patient and no differences in the fluid balance or postoperative bleeding between the groups were detected. Furthermore, the patients were observed by the same anesthesiologist and the tracheal extubation followed a strictly defined study protocol.
In conclusion, regarding postoperative recovery of patients from CABG, diltiazem slows elimination of midazolam and alfentanil and may delay tracheal extubation after large doses of these anesthetic adjuncts. CYP3A-mediated drug interactions should be considered as confounders when recovery from anesthesia with midazolam and alfentanil infusions is assessed.
The authors thank all the staff of the operating room and the intensive care unit for their help during administration of anesthesia and patient recovery, and Jouko Laitila, Lisbet Partanen, Kerttu Martensson, and Eija Makinen-Pulli for the skillful determination of drug plasma concentrations.