Behavioral and Physiological Effects of Remifentanil and Alfentanil in Healthy Volunteers 

The study was approved by the local institutional review board. Informed written consent was obtained from each participant before the the study began. The participants were told on the consent form that the intravenous drugs being studied were used commonly in medical settings and may come from one of six classes: sedative-tranquilizer, stimulant, opiate, general anesthetic at subanesthetic doses, alcohol, or placebo. Before the first session, the participants attended a screening interview in which psychiatric and medical assessments were made to determine whether contraindications existed to their participation in the study. Eight men and two women completed the study (mean age [+/− SD], 24.9 +/− 3 yr; mean weight, 72 +/− 6.1 kg). All participants had at least a high school education (or its equivalent). Payment for study participation was made during a debriefing session held after the experiment was completed.

A randomized, placebo-controlled, double-blinded, crossover trial was conducted. Participants received, in a forearm vein, infusions of saline (the vehicle), remifentanil, or alfentanil during a 2-h period, when three target plasma levels were achieved for 40 min each. The target plasma levels for remifentanil and alfentanil were 0.75, 1.5, and 3 ng/ml, and 16, 32, and 64 ng/ml, respectively. These levels were chosen from a study that detected a 19-fold difference in the potency between the two drugs using the electroencephalogram as the pharmacodynamic end point. [4]The values of doses that produced similar minimum alveolar concentration conditions for each of the opioids were desired, thus eliminating the potential for undesired side effects at higher doses. The corresponding infusion rates were all within the maximum recommended doses for minimum alveolar concentrations. They were less than 0.2 [micro sign]g [middle dot] kg^-1[middle dot] min (^-1) for remifentanil and less than 1 [micro sign]g [middle dot] kg^-1[middle dot] min^-1for alfentanil. By using each participant's age, sex, and weight, the infusion rates for the target plasma levels were determined using a computer pharmacokinetic modeling program, STANPUMP (copyright S. Shafer 1986-1995; revision 5/11/96;steven.shafer@stanford.edu;Palo Alto, CA). The STANPUMP program was used in simulation mode to develop theoretical weight-adjusted plasma concentrations and pump rates at 10-s intervals. The kinetic sets used were weight adjusted for alfentanil and remifentanil. The rates were averaged over five intervals during each 40-min block. During the 2-h infusion, 15 pump rates were entered manually by the anesthetist on the syringe infusion pump (model AS206H-2; Baxter Healthcare, Hookset, NH). The accuracy of the settings was confirmed by reentering the rates on STANPUMP, which were within 5% of the targeted infusion values. The time points were 0, 1, 3, 7, and 15 min for each of the 40-min blocks. A typical range of infusion rates would be from 20 ml/h to 1.4 ml/h for remifentanil at a concentration of 50 [micro sign]g/ml; and 14.1 ml/h to 2.6 ml/h for alfentanil at a concentration of 250 [micro sign]g/ml. The total dose for each participant was weight adjusted. For example, a 30-yr-old man weighing 70 kg would receive, during a 120-min period, total doses of 2,060 [micro sign]g alfentanil and 652 [micro sign]g remifentanil. During and after the infusion, mood, psychomotor performance, and physiologic status were assessed. In addition, the analgesic effects of the study drugs were assessed during the infusion period by asking the participants to report their pain levels while they had their nondominant forearm immersed in 2 +/− 1 [degree sign]C water.

The volunteers participated in three 5-h sessions spaced at least 1 week apart. They were instructed not to eat food or drink any nonclear liquids for 4 h, not to drink clear liquids for 2 h, and not to use any drugs (including alcohol, but excluding normal amounts of caffeine and nicotine) 24 h before the study session. A toxicology screening was required before the start of each session, as was a pregnancy test for all female participants. Blood alcohol levels (as measured by a breath intoximeter) also were measured. While each patient lay in a hospital bed in a semirecumbent position, an anesthetic inserted an angiocatheter into the participants' forearm vein. Noninvasive monitoring of heart rate and arterial oxygen saturation was initiated. An anesthetist began the 2-h infusion into the angiocatheter via the syringe infusion pump. Before the infusion, the participant was told, “the infusion you are about to receive may or may not contain a drug.” The drug was drawn previously by one anesthetist and administered by another to preserve the double-blinded nature of the study. However, the anesthetist who was monitoring the infusion was aware of the possible drugs involved, so appropriate measures could be taken to ensure the participants' safety and well-being if an adverse event occurred. At periodic intervals during and after the infusion, mood, psychomotor performance, and physiologic status of the participants were assessed. In addition, at three times during the infusion, a 2-min cold-pressor test was administered. The cold-pressor test was instituted after a given dose had been infused for 35 min and, presumably, when steady state plasma levels had been achieved. During the cold-water immersion, the participants reported their pain status.

After the study sessions were complete, the participants were transported home via a livery service with instructions not to engage in certain activities for the next 12 h (cooking, driving, caring for children, drinking alcohol).

The entire battery of subjective and psychomotor tests took approximately 10 min for the participants to complete. Subjective-effects testing was performed first, followed by psychomotor testing. The order of each test within the battery was held constant and was the same for all participants. Testing was initiated 20 min into each infusion period; that is, 20, 60, and 100 min after the onset of the first infusion, and 15, 30, 60, 120, 150, and 180 min after the infusion was discontinued. The cold-pressor test was initiated 35 min into each infusion period; that is, 35, 75, and 115 min after the onset of the first infusion.

1. The Addiction Research Center Inventory is a true-or-false questionnaire designed to differentiate among different classes of psychoactive drugs. [10]A paper-and-pencil short form of the Addiction Research Center Inventory was used [11]that had 29 items and yielded scores for two different scales: lysergic acid diethylamide, sensitive to somatic and dysphoric effects; and morphine-benzedrine group, often described as euphoria.

2. A locally developed visual analog scale consisted of 23 100-mm lines, each labeled with an adjective (e.g., floating, heavy or sluggish feeling, high [a drug high], lightheaded, nauseous, sleepy [drowsy, tired]). A mark at the leftmost point of the line indicated a response of “not at all,” whereas a mark at the rightmost point indicated a response of “extremely.”

3. A locally developed adjective checklist was constructed using items from the opiate adjective checklist (derived from the Single Dose Questionnaire [12]) and a list reported as sensitive to the somatic and subjective effects of opiates from the [micro sign]- and mixed-agonist-antagonist class. [13]The checklist consisted of 13 items that the participant rated on a five-point scale from 0 ("not at all") to 4 ("extremely"). The items were as follows: carefree, depressed, drive (motivated), dry mouth, flushing, good mood, headache, nodding, numb, skin itchy, sleepy, sweating, turning of stomach, and vomiting.

4. The Drug Effects-Liking Questionnaire assessed the extent to which participants felt a drug effect, on a scale of 1 to 5 (1 =“I feel no effect from it at all”; 2 =“I think I feel a mild effect, but I'm not sure”; 3 =“I definitely feel an effect, but it is not real strong”; 4 =“I feel a strong effect”; 5 =“I feel a very strong effect”) and assessed the extent that the participants liked the drug effect on a 100-mm line (0 = dislike a lot; 50 = neutral; 100 = like a lot).

1. The Maddox Wing test measures the relative position of the eyes in prism diopters. Some drugs cause extraocular muscles of the eye to diverge (exophoria), and this divergence is considered an indicator of psychomotor impairment. [14] 

2. The digit-symbol substitution test (DSST) was a 1-min paper-and-pencil test that required the participants to replace digits with corresponding symbols according to a digit-symbol code listed on the top of the paper. [15]The dependent measure was the number of symbols the participant drew correctly. Different forms of the test (e.g., different symbol-digit codes) were used each time the test was presented to the participant. The DSST evaluated changes in information processing performance and the ability to concentrate. [16] 

3. An auditory reaction test measured the time it took for participants to react to an auditory stimulus. [17]Ten 50-dBA (decibels auditory) computer-generated tones were delivered at random time intervals (between 1 and 10 s) in a 1-min period. The tone remained on until the participants depressed the computer keyboard space bar, or until 2 s had elapsed, whichever occurred first. The mean reaction time (in seconds) was the dependent measure.

4. The Backward Digit Span [15]is a computerized memory test that presents digits to participants, which they are to reproduce in backward order; successful completion of one trial initiates a succeeding trial in which the number of digits is increased by one. The first trial consisted of one digit, and the dependent measure was the number of trials completed successfully.

Five physiologic measures were assessed: heart rate, blood pressure, arterial oxygen saturation, respiration rate, and miosis. Heart rate, blood pressure, and arterial oxygen saturation were measured noninvasively using a Merlin model 54 monitor (Hewlett Packard, Andover, MA). Respiration rate was the number of breaths the participants took in 30 s (multiplied by 2 to yield breaths/min). This was assessed by counting the number of times the participant's chest or stomach rose and fell, and it was measured by one of the experimenters (J.L.H.) who was blinded to the dose and drug administered. Miosis, or pupil constriction, is a physiologic marker of opiate effects. The diameter of the pupil was measured by photographing the participant's right eye with a 2X magnification instant flash-equipped camera in a dimly lit room. [13] 

Participants were asked 10, 30, 70, and 110 s into each cold-water immersion, “On a scale from 0 to 10, how painful is it?” and “How much does it bother you?” in which 0 =“not painful/bothersome at all, and 10 =“the most painful/bothersome feeling imaginable.” The sensory and affective dimensions of pain were assessed using a paper-and-pencil version of the short-form McGill Pain Questionnaire, [18]which was completed 75 s after the immersion began. Fifteen descriptors (e.g., “throbbing,”“aching,”“gnawing”) were rated on a scale of 0 to 3 (0 =“none,” 1 =“mild,” 2 =“moderate,” and 3 =“severe”).

Participants were given an adjective rating checklist to complete at home 24 h later, noting whether they had experienced any of the symptoms listed (confusion, dry mouth, headache, nausea, skin itching) during the 24 h after the session. Each symptom was rated on a five-point scale ranging from 0 =“not at all” to 4 =“extremely.”

Repeated-measures analysis of variance was used to evaluate the data. Three sets of analyses of variance were performed. The first set evaluated dependent measures during the infusion period. Factors in the first analysis were drug (three levels: saline, remifentanil, alfentanil) and dose (three for each narcotic: 0.75, 1.5, 3 ng/ml for remifentanil and 16, 32, 64 ng/ml for alfentanil). For pain measures, factors were drug, dose, and time (five levels: 10, 30, 50, 70, and 110 s). F values were considered significant for P <or= to 0.05 with adjustments of within-factor degrees of freedom (Huynh-Feldt) to protect against violations of symmetry. The second set of analyses evaluated dependent measures during the postinfusion period. Factors in the second analysis were drug (three levels: saline, remifentanil, alfentanil) and time (four to seven levels). Tukey post hoc testing was performed on significant effects for each analysis of variance. The third set of analyses of variance evaluated postsession adjective checklist ratings, using drug as a factor.

The Addiction Research Center Inventory. Significant drug effects were obtained on the lysergic acid diethylamide scale (P < 0.05). Remifentanil significantly increased lysergic acid diethylamide scores, and the increase tended to be dose related but was not significant. The morphine-benzedrine group scores were not increased by either drug.

Visual Analog Scale. Table 1shows those visual analog scale adjectives in which significant drug x dose or drug effects were obtained. Generally, although both drugs tended to increase visual analog scale ratings in a dose-related manner, remifentanil had larger effects at each infusion level than did alfentanil. In all instances in which there was a significant drug effect, remifentanil, but not alfentanil, differed significantly from saline ratings.

Adjective Checklist. Table 1shows those opiate adjective checklist ratings in which significant drug x dose or drug effects were obtained. Remifentanil, unlike alfentanil, increased ratings of dry mouth, nodding, numb, and skin itching in a dose-related manner. Significant drug effects were obtained on the ratings of flushing, sweating, and turning of the stomach; post hoc tests revealed that remifentanil, but not alfentanil, significantly increased these ratings.

Drug Effect-Liking Questionnaire. As Table 1and Figure 1(left frame) show, ratings of “feel drug effect” increased in a dose-related manner for alfentanil and remifentanil. However, ratings were greater at each remifentanil infusion level, relative to alfentanil. Liking ratings showed much variability among participants, with some reporting liking remifentanil and alfentanil effects and others reporting either dislike or neutrality (drug x dose: F(4,36)= 0.7, not significant).

Psychomotor Performance. Significant drug x dose, drug effects, or both were obtained using the Maddox Wing test, DSST, and backward digit span test. Table 2shows ratings as a function of drug and infusion level. Remifentanil and alfentanil both induced exophoria on the Maddox Wing test and impaired DSST performance, but post hoc testing revealed greater effects with remifentanil. In addition, remifentanil, but not alfentanil, impaired performance on the backward digit span test.

Physiologic Measures. Remifentanil and alfentanil induced miosis in a dose-related manner (Table 2), but the effects of remifentanil were significantly greater than those of alfentanil. Remifentanil, but not alfentanil, decreased arterial oxygen saturation and respiration rates. The arterial oxygen saturation decrease was dose related.

Pain Measures. Significant drug x dose effects were obtained on ratings of pain intensity (P < 0.001) and bothersomeness of pain (P < 0.001). Figure 2shows that alfentanil and remifentanil both decreased pain ratings and that reductions were greater with remifentanil than with alfentanil. Significant drug x dose effects were obtained on the short-form McGill Pain Questionnaire ratings of fearfulness, sharpness, and throbbing. Post hoc test revealed that 1.5 and 3 ng/ml remifentanil significantly decreased ratings of fearfulness and that these doses and the 64 ng/ml alfentanil infusion reduced ratings of sharpness and throbbing. Significant drug effects were obtained on ratings of aching, hot-burning, shooting, splitting, and stabbing. Remifentanil and alfentanil significantly decreased ratings of shooting, but only remifentanil significantly decreased the other four ratings.

Significant drug or drug x time effects were obtained for nine visual analog scale ratings. Table 3shows that remifentanil had effects that lasted as long as 30 min after the infusion was stopped. Some of the effects were dysphoric in nature: Participants reported disliking the drug effects, having unpleasant bodily sensations, and feeling nauseated. Alfentanil had effects that were limited to the first 30 min after the infusion was stopped. No significant effects were obtained on the opiate adjective checklist ratings. The number of symbols drawn correctly on the DSST was reduced significantly for as long as 15 and 60 min, respectively, after alfentanil and remifentanil infusions were stopped (Figure 1, right). Miosis was evident for as long as 120 and 15 min, respectively, after the alfentanil and remifentanil infusions were discontinued (Figure 3).

No significant effects were obtained for any of the ratings from the postsession checklist, indicating that the effects of remifentanil and alfentanil were limited to the experimental session.

The major findings of this study were as follows: The analgesia potency ratio of remifentanil to alfentanil, using theoretical target plasma concentrations, appears to be closer to 40:1 than to the 20:1 ratio used in this study; remifentanil had subjective effects similar to those of other full [micro sign]-opioid agonists; and a dissociation existed between recovery from remifentanil, as measured by psychomotor performance and subjective effects.

When the analgesic effects of the two drugs were evaluated, it is apparent that the middle dose of remifentanil tested produced an amount of analgesia that was similar to that of the high dose of alfentanil. We chose a 20:1 ratio based on two studies that found remifentanil: alfentanil potency ratios ranging from 20:1 to 30:1, using the end points of pressure algesimetry, respiratory depression, and electric activity of the brain. [4,5]However, at least one other study suggested that a 40:1 potency ratio may be more accurate when characterizing the potency ratio of remifentanil to alfentanil. In that study, 1 ng/ml remifentanil and 40 ng/ml alfentanil produced similar minute ventilation rates when participants breathed steady state end-tidal 7.5% carbon dioxide. [1] 

Determining and defining potency is a complex issue. In this study, the theoretical plasma concentrations were predicted using a computer simulation. Plasma drug concentrations were not verified in this study. Therefore, it is conceivable, given the potential inaccuracies of the computer modeling, that the potency ratio was actually 20:1 rather than 40:1.

The spectrum of remifentanil subjective effects and side effects obtained in the current study corresponds with mood effects and side effects obtained in other studies of healthy volunteers that characterized effects of full and partial [micro sign]-agonist opiods, such as morphine, fentanyl, dezocine, and buprenorphine. [19-22]The typical subjective effects reported include difficulty in concentrating, a feeling of heaviness or sluggishness, dizziness, and drowsiness. Typical side effects reported include pruritus, dry mouth, and nausea. For 30 min after the remifentanil infusion was discontinued, drug liking ratings were significantly lower than saline liking ratings, perhaps because of the degree of nausea that occurred at approximately the same period. Approximately 50% of participants reported nausea during or after the remifentanil infusion, which is within the range of that reported in several laboratory and clinical studies. [4,9,23,24]and four participants vomited at some time during the session. Alfentanil produced some changes in mood, but no side effects, which again suggests that the remifentanil:alfentanil potency ratio is greater than 20:1.

Finally, the time course of recovery from remifentanil as measured by different indices is noteworthy. According to participants' self-reports, “feel drug effect” ratings and other mood measures were similar to that of placebo (saline) 60 min after the infusion was stopped. The miotic effects of remifentanil dissipated by 30 min after the infusion was discontinued. However, DSST performance was still impaired 60 min after infusions were stopped. On average, eight fewer symbols were drawn 60 min after the remifentanil infusion than at the same time after the saline infusion. This impairment is by no means trivial: In a previous study [25]that evaluated DSST performance after participants ingested a placebo beverage or an alcoholic beverage that produced a blood alcohol level of 0.11%, nine fewer symbols were drawn in the latter condition. The substantial, and lingering, impairment was a surprising finding given that the context-sensitive half-time is 3-6 min. [1-3]The clinical relevance of this finding is that patients receiving remifentanil may feel “normal” soon after the infusion is discontinued and their physiologic parameters may return to baseline levels, but they may still be cognitively impaired from the drug. Indeed, studies have documented decreased awareness of psychomotor impairment with several different drugs, including opioids, [26]barbiturates, [27,28]and benzodiazepines. [29,30]The DSST has proved in many studies to be exquisitely sensitive to different drugs and different doses. This study emphasizes the importance of using objective measures of impairment, such as the DSST, and subjective measures or recovery, and its suggests that an overt behavioral measure may be more sensitive than mood measures. Finally, the notion that the pharmacodynamic effects of remifentanil are extremely short-lived after the drug is no longer administered must be called into question given our findings that psychomotor effects were still apparent 1 h after infusions were stopped.

The authors thank Robert Shaughnessy, C.R.N.A., Mary Maurer, C.R.N.A., and Nada Williamson, C.R.N.A., for their assistance in administering the agents and monitoring the physiologic status of the participants.