Alfentanil and fentanyl are widely used and extensively studied opioids. The potency ratio of these two drugs has been estimated previously by different methods, the ratio varying somewhat depending on the pharmacological end point used. The alfentanil:fentanyl analgesic concentration potency ratio for skin incision was estimated as 1:57 (1,2) and for experimental pain 1:48 (3). The respiratory depressant effects of alfentanil and fentanyl have been studied separately (4–7), but their relative potencies have not been quantified properly. However, there are clinical studies suggesting that for a given level of analgesia alfentanil might cause less respiratory depression than fentanyl (8–10).
In many of the previous studies, responsiveness to carbon dioxide (4,6) or hypoxia (5) was used to measure respiratory depression. Responsiveness to carbon dioxide is a well-documented and sensitive test for drug-induced respiratory depression. However, this method is associated with some problems (11). In this study, respiratory inductive plethysmography (12,13) and indirect calorimetry (14) were used for monitoring of resting ventilation to obtain continuous data and to not disturb the drug effect by the monitoring system (11). To obtain stable target plasma concentrations, a computer-driven infusion pump with escalating drug plasma concentrations was used for drug delivery. Thus, the aim of this study was to determine the potency ratio for respiratory depression induced by alfentanil and fentanyl using a continuous, noninvasive monitoring technique.
Eight healthy male volunteers were enrolled in the study. Subjects ranged in age from 22 to 28 yr and were within normal weight limits (range, 65–94 kg). None had a history of alcohol and/or drug abuse and none was currently using any medication. A screening test for drugs (cocaine, amphetamine, cannabis, opioids, and benzodiazepines) was performed for each volunteer before participating in the study. This study was approved by the Institutional Ethics Committee of Turku University Hospital and written informed consent from each volunteer was obtained before recruitment.
Each subject was randomly allocated to receive two different treatments according to a balanced, cross-over and double-blinded design with 5–7 days washout period. The order of drug administration was assigned using random permutation. The volunteers received fentanyl and alfentanil as a continuous infusion delivered with a computer-driven (STANPUMP; developed by Steven L. Shafer, MD; Palo Alto, CA) infusion pump (Harward 22 Basic Syringe Pump; Harward Apparatus, South Natick, MA). Nonweight-adjusted pharmacokinetic variables were used (15,16). Ascending logarithmic escalations of pseudo steady-state plasma concentrations of 5, 8, 12.5, 20, 31.5, 50, 80, 125, 200, 315, and 500 ng/mL for alfentanil and 0.1, 0.16, 0.25, 0.4, 0.63, 1, 1.6, 2.5, 4, 6.3, and 10 ng/mL for fentanyl were accomplished at 10-min intervals (i.e., 10 concentrations over a 100-fold concentration range). The study drugs were prepared and administered by an independent person not participating in the study by diluting them with 50 mL physiological saline.
Before each session, food, caffeinated drinks, and chocolate were not allowed for 12 h preceding the study. Both sessions for each subject were performed at the same time of day to avoid possible diurnal variation. After admission to the study site, a radial artery and a forearm vein were cannulated. The plethysmograph belts were fitted around the chest wall and abdomen, and the subjects were asked to breathe through the pneumotachometer for at least 2 min to have a volume reference for the plethysmograph. Then the canopy of the indirect calorimetry was placed over each subjects’ head. After a 15-min stabilization period, baseline values for all variables were recorded. The drug infusion was started according to the method described above. The infusions were terminated when respiratory rate decreased to 2/min or oxygen saturation (Spo2) decreased below 85%. After terminating the drug infusion, the volunteers were asked to take a few deep breaths and were given IV naloxone 0.4 mg and/or, in case of emesis, IV tropisetron 2 mg. The subjects were observed at the study site until no adverse drug effect was seen according to standard outpatient criteria.
Respiratory inductive plethysmography (RIP), (RESPITRACE; NIMS, Miami Beach, FL) (12,13) was used for respiratory monitoring. Changes in rib cage (RC) and abdominal circumferences (AB) were simultaneously measured using two differential linear transformers connected to tight fitting belts around the chest at nipple level and around the abdomen at umbilical level. The RC and AB signals were summed electronically to provide a signal equivalent to tidal volume (Vt). The calibration was done semiquantitatively before each measurement in each subject by first collecting baseline breathing data for 5 min and thereafter by calibrating against a known volume reference (17). A heated pneumotachometer (Hans Rudolph Inc., Kansas City, MO) connected to a differential pressure transducer (VALIDYNE MP 45, ± 2.0 cm H2O; Validyne, Northridge, CA) was used as the volume reference. Flow, volume, and analog RC+AB signals were amplified and recorded using a physiologic recording system (DIREC; Raytech Instruments, Vancouver, BC). At least five breaths were sampled for calibration. The calibrated RC+AB signal was analyzed with the software program (Respievents) provided with the RIP. Minute volume (V̇e) was calculated from Vt and respiratory rate. Mean inspiratory flow (Vt/Ti), an indicator of respiratory drive, was calculated by dividing Vt by inspiratory time (Ti). The proportion of total ventilation that can be attributed to expansion of rib cage (RC%) was also measured with RIP. Both Vt/Ti and RC% are standardized and validated variables produced by the Respievents and are described in detail elsewhere (13,18,19). The data were analyzed by including every breath from a 2-min period.
Indirect calorimetry (DELTATRAC; Datex, Helsinki, Finland) is an open system device designed for measuring oxygen consumption (V̇o2) and carbon dioxide production (V̇co2) in both spontaneously breathing and mechanically ventilated subjects, with a relative error of V̇o2 and V̇co2 of ± 5%(14). The Deltatrac flow generator entrains room air through the canopy at a rate of 40 L/min. The sampling is done from a side-stream port at the proximal side of the canopy. The difference between the inspired and expired oxygen fractions (Fio2 and Feo2), analyzed from the mixed expiratory gas flowing through the monitor, was measured with a fast-response paramagnetic differential oxygen sensor. The expired CO2 fraction (FeCO2), analyzed from the mixed expiratory gas, was measured with an infrared CO2 sensor. The gas exchange monitor was calibrated before each study session with a standardized, certified gas mixture (95% O2, 5%CO2). The values of V̇o2 were produced every 60 s from the Deltatrac and collected online to a computer for further analysis. The V̇co2 values were used for calculation of alveolar ventilation (VA).
In summary, the following respiratory variables were recorded during the study: Pao2(kPa), Paco2(kPa), Spo2(%), VA(L/min), V̇e(L/min), Vt(mL), RR(/min), RC%(%), Vt/Ti(mL/s), and V̇o2(mL/min).
Mean arterial blood pressure and heart rate were directly measured from the radial artery and monitored along with Spo2 using a Cardiocap monitor (Datex). Data from these variables were also recorded online to a computer every 10 s for further analysis.
Data from minutes 7–9 at each concentration level were used for analysis. RIP data were analyzed from a representative 120-s period within this time period and Deltatrac data as mean value of the two successive 1-min values from minutes 7–9. Data from the Cardiocap monitor were analyzed as a mean value of all values produced during minutes 8–9.
Arterial blood samples were drawn at the end of each concentration level from the radial artery for measurements of Pao2 and Paco2 as well as for plasma drug concentration determinations. Alveolar ventilation was calculated as VA (L/min) = 0.115 × V̇co2 (mL/min)/Paco2 (kPa) (18,20).
Plasma concentrations of fentanyl and alfentanil were determined using radioimmunoassay methods (21,22). The intra- and interassay coefficients of variations were 3.7% and 3.3% for alfentanil and 6.0% and 6.9% for fentanyl. The measured, not predicted, concentrations of alfentanil and fentanyl were used for pharmacodynamic modeling and for determining the cutoff values.
To determine the potency of respiratory depression induced by the opioids, pharmacodynamic modeling was performed separately for each subject using the nonlinear regression program PCNONLIN (Version 4.2; Scientific Consulting Inc., Apex, NC) without weighting. The concentrations producing 50% of the maximal decrease in V̇e and respiratory rate (i.e., apparent EC50 values) were estimated by fitting the fractional Emax and sigmoidal Emax models to the data according to Equations 1 and 2, respectively, where E = effect, E0 = baseline effect (maximal drug effect, Emax is equal to E0), C = drug concentration, EC50 = concentration at 50% of Emax and γ = sigmoidicity (Hill) factor (7). The goodness of the fit was determined by Akaike’s information criterion and by assessment of scatter of the observed data points in relation to the fitted function. In addition, the increase in Paco2 was fitted to a linear model, i.e.;, where k = the slope of the line relating effect to concentration. Potency ratios were calculated by dividing the respective EC50 values.
Data are presented as mean ± sd. The different pharmacodynamic responses were analyzed using analysis of variance for crossover design with repeated measures within periods. In this analysis, a significant step effect was indicative for concentration-dependent effects, and the two factor interaction sequence*period or the three factor interaction sequence*period*step indicative for differences between the two study drugs. To assess the sensitivity of the various respiratory variables to detect drug-induced respiratory depression, the analysis was continued in case of a significant concentration effect by calculating paired contrasts vs baseline and by correcting corresponding P values for multiple comparisons (Bonferroni correction). A P value <0.05 was considered as statistically significant. In addition, 95% confidence intervals were calculated for the estimated EC50 values and potency ratios. Statistical computing was performed using the SAS system for Windows (Release 6.12, 1996; SAS Institute Inc., Cary, NC).
The cutoff for respiratory depression was reached and the infusion terminated at targeted concentration levels of 200–500 ng/mL during alfentanil sessions and at concentration levels of 2.5–6.3 ng/mL during fentanyl sessions. Mean measured cutoff concentrations were 254 ± 88 ng/mL and 5.1 ± 1.7 ng/mL for alfentanil and fentanyl, respectively. The measured alfentanil concentrations were slightly below and fentanyl concentrations slightly above the targeted concentrations (Fig. 1). Mean ± sd total alfentanil dose was 4.48 ± 1.30 mg and fentanyl dose 0.47 ± 0.13 mg. Spo2 was used as a cutoff variable in 15 sessions and respiratory rate in one session.
In subject 8, the alfentanil infusion was terminated at a concentration level of 200 ng/mL because of nausea and vomiting. Data from this individual are included in the statistical analysis until the occurrence of vomiting. All other subjects completed both sessions as planned. One subject vomited after stopping the fentanyl infusion and two after stopping the alfentanil infusion. Other side effects reported were pruritus in four subjects during both infusions and dryness of mouth in three subjects after fentanyl and in two subjects during alfentanil infusion. One subject reported hallucinations during both drug infusions. The analyses of Paco2 failed in subject 1 during the fentanyl session because of a laboratory error. In addition, respiratory rate data could not be reliably modeled in subject 6 during fentanyl (low respiratory rate at baseline) and V̇e data in subject 8 during alfentanil (because of premature termination).
Both drugs depressed ventilation in a concentration-dependent manner in a nonlinear way (Fig. 2, Table 1). Spo2 decreased from 97% ± 1% (baseline) to 83% ± 4% (at cessation of the infusion) during alfentanil treatment and from 97% ± 1% to 84% + 3% during fentanyl treatment. V̇e decreased from 8.9 ± 1.5 L/min to 4.4 ± 1.6 L/min and from 9.9 ± 1.7 L/min to 4.9 ± 1.2 L/min during alfentanil and fentanyl, respectively. Vt increased slightly and RC% more clearly toward the end of both infusions. There were no clear changes in Vt/Ti, but statistical analysis revealed a significant time (concentration) effect in analysis of variance (P = 0.037). Changes in all other respiratory variables were highly significant for the concentration effect (P < 0.001). However, there were no significant differences between the study drugs except for VA (P = 0.029) and Spo2 (P = 0.003), which were both decreased slightly earlier during fentanyl treatment (Table 1). V̇o2 decreased during alfentanil infusion by 8% ± 9% and during fentanyl by 15% ± 14% (Table 1). No differences were found between the treatments. Heart rate remained at baseline level throughout both sessions. There were no significant differences between the treatments in either heart rate or mean arterial blood pressure, but statistical analysis revealed a significant concentration effect in mean arterial blood pressure (Table 1).
The modeled apparent EC50 values using the fractional Emax model (Equation 1) for V̇e and respiratory rate data and the slopes for PAco2 increases (Equation 3) are presented in Table 2. The derived potency ratios were between 1:39 and 1:51. The sigmoidal inhibitory Emax model (Equation 2) did not improve the fit (i.e., estimates for the sigmoidicity factor were close to unity).
In this study, the potency ratio for respiratory depression of alfentanil to fentanyl was estimated to range between 1:39 to 1:51. Fentanyl was 39 times more potent than alfentanil in decreasing minute ventilation and 51 times more potent in decreasing respiratory rate. These estimates are in agreement with the analgesic potency ratio of the two drugs reported earlier (Table 3).
As expected, both opioids have concentration-dependent respiratory effects. However, some previous studies have suggested that for a given level of analgesia, alfentanil might cause less respiratory depression than fentanyl (8–10). These results can at least partly be explained by using doses instead of concentrations in calculating the potency ratio. Misinterpretations may occur because the distribution volume of fentanyl is larger than that of alfentanil and ideally concentration-effect relationships for respiratory depression and analgesia should be determined at steady state. Although there is a minimal amount of information on the potency ratio for respiratory depression caused by alfentanil and fentanyl, these drugs have been extensively studied with respect to relative potencies (Table 3). The relative potency ratio of fentanyl and alfentanil measured with different forms of pain is in close correlation with the potency ratios for respiratory depression measured in our study, thus indicating a similar respiratory depressant effect with equianalgesic plasma concentrations. Also, the potency ratio of alfentanil and fentanyl measured as slowing of electroencephalogram resembles our findings closely (Table 3).
Fentanyl and alfentanil caused respiratory depression in an almost similar fashion. However, the apparent EC50 values for alfentanil obtained in the present study are strikingly higher than EC50 values reported by other groups. In the study by Glass et al. (6), alfentanil-induced respiratory depression was studied with healthy volunteers using the minute ventilation response to 7.5% CO2, administered via “bag in the box” system and the EC50 was 49.4 ng/mL. In the study of Bouillon et al. (7), the increase of Paco2 was used as a determinant of respiratory depression in a sophisticated indirect response model and the EC50 for VA “corrected to normocapnia” was 60.3 ng/mL. The discrepancy in the derived EC50 values was a result of different end points. In unstimulated resting ventilation, which was used in this study, both hypercarbia and hypoxia can considerably stimulate ventilation and affect the potency estimates. Thus, our (apparent) EC50 values can even be considered as biased estimates of the true potency of the opioids (7). However, our primary goal was to determine the potency ratio of the two drugs. The current results clearly show that the potency of ventilatory depression of opioids (i.e., EC50 values) is highly dependent on the physiological conditions in which the measurements are made, and different estimates can be obtained with concomitantly measuring and modeling or standardizing CO2 levels in the body (7). However, both kind of studies, i.e., ventilatory challenges where Paco2 is controlled and measurements of resting ventilation, are needed to extend the knowledge of respiratory depression induced by different opioids (11).
Fentanyl is not as widely studied as alfentanil in respect to the EC50 values for respiratory depression. In the study of Cartwright et al. (4), the EC50 for fentanyl-induced respiratory depression in surgical patients was estimated to range between 1.5 and 3.1 ng/mL measured with responsiveness for CO2. These values are lower than the “apparent” EC50 values obtained from our study; the reason is probably the same as with alfentanil. In the study of Fung and Eisele (25), where a single bolus of fentanyl was given to healthy volunteers, the EC50 for respiratory depression was 4.6 ng/mL. However, the use of a single bolus, instead of a steady-state infusion, makes direct comparisons difficult.
The most sensitive variable was VA, which was significantly decreased even at quite small plasma drug concentrations. A significant decrease in respiratory rate and an increase of Paco2 were also detected at small plasma drug concentrations that were clearly below anesthetic and even analgesic drug concentrations (1). The decrease of V̇e was largely accounted for by the reduced respiratory rate because Vt increased toward the end of the study. In previous studies, the decrease in V̇e was the result of either decrease in Vt or respiratory rate, depending on the dose of opioid administered (5,21,26). In our study, wide concentration ranges were used; the result was a decrease of respiratory rate. The mode of the monitoring system, e.g., resting ventilation, can account for this decrease in respiratory rate (27).
Despite the large amount of opioids infused, no chest wall rigidity was detected during either of the infusions as seen by the continuous increase of RC%. This is in accordance with previous studies in which chest wall rigidity was mostly seen with rapid bolus doses rather than infusions (9,28). No paradoxical chest movement was detected in any of the subjects, which indicates patent airway. The developing respiratory depression in this study was thus a result of central apnea, not upper airway obstruction. In our study, Vt/Ti increased slightly but significantly during both drug infusions, suggesting maintained respiratory drive (19). However, this increase was most probably caused by the stimulatory effect of increased Paco2, which masks the opioid-induced depression in respiratory drive. The increase of mean inspiratory flow has been associated with increasing activation of respiratory muscle and progressive increase in lung volume (29). This is in line with our observation of increasing Vt and RC%.
The respiratory effects were assessed by indirect calorimetry and respiratory inductive plethysmography because they were validated, noninvasive monitoring systems for continuous respiratory monitoring (12–14,17). The use of a nondisturbing system to measure respiratory pattern and ventilation creates a setting close to the clinical situation. This monitoring system allows natural breathing without any external stimuli from the equipment. In addition, use of these methods enabled the measurement of VA, determined by the rate of V̇co2 and the alveolar partial pressure of CO2, which is closely related to Paco2. VA thus depicts the actual gas exchange in lungs and the efficacy of breathing. Tobin et al. (19) have compared Vt/Ti, measured with RIP to standard indices of respiratory center outputP measured with mouth occlusion pressure and demonstrated a close correlation between the two measurements. We did not measure sedation, as it is difficult to do properly without disturbing these respiratory measurements.
A computer-driven infusion pump (STANPUMP-Harvard 22) was used in the present study to rapidly achieve and maintain the desired pseudo steady-state plasma concentrations of the opioids. The 10-minute time period we used in the present study was, however, not quite sufficient to achieve equilibration with the effect site during fentanyl infusion. Based on a STANPUMP simulation the effect site concentration at pharmacologically relevant concentrations was approximately 90% of the plasma concentrations at the end of the 10-minute period. Thus the true EC50 values of fentanyl may be slightly less than those reported and may underestimate the respiratory depression caused by fentanyl. A 10-minute interval was chosen to reduce the total dose.
In conclusion, both alfentanil and fentanyl produced respiratory depression in a similar fashion with escalating drug plasma concentrations. Using a continuous measurement of resting ventilation, the potency ratio for respiratory depression was between 1:39 and 1:51 (alfentanil:fentanyl). This is analogous to the potency ratio for pain stimulation studied elsewhere. Thus, the respiratory depressant effects of alfentanil and fentanyl are identical at equianalgesic drug plasma concentrations.
We would like to thank Elina Kahra for excellent technical assistance, Hans Helenius and Lauri Sillanmäki for performing the statistical analyses, and Janssen Pharma for analyzing plasma drug concentrations.
1. Glass PSA, Shafer SL, Jacobs JR, Reves JG. Intravenous drug delivery systems. In: Miller RD, ed. Anesthesia. 4th edition. New York: Churchill Livingstone, 1994: 389–416.
2. Westmoreland CL, Sebel PS, Gropper A. Fentanyl or alfentanil decreases the minimum alveolar anesthetic concentration of isoflurane in surgical patients. Anesth Analg 1994; 78: 23–8.
3. Hill H, Chapman C, Saeger L, et al. Steady-state infusions of opioids in human. II. Concentration-effect relationships and therapeutic margins. Pain 1990; 43: 69–79.
4. Cartwright P, Prys-Roberts C, Gill K, et al. Ventilatory depression related to fentanyl plasma concentrations during and after anesthesia in humans. Anesth Analg 1983; 62: 966–74.
5. Cartwright CR, Henson LC, Ward DS. Effects of alfentanil on the ventilatory response to sustained hypoxia. Anesthesiology 1998; 89: 612–9.
6. Glass PSA, Iselin-Chaves IA, Goodmand D, et al. Determination of the potency of remifentanil compared with alfentanil using ventilatory depression as the measure of opioid effect. Anesthesiology 1999; 90: 1556–63.
7. Bouillon T, Schmidt C, Gartska G, et al. Pharmacokinetic-pharmacodynamic modeling of the respiratory depressant effect of alfentanil. Anesthesiology 1999; 91: 144–55.
8. Scamman FL, Ghoneim MM, Korttila K. Ventilatory and mental effects of alfentanil and fentanyl. Acta Anaesthesiol Scand 1984; 28: 63–7.
9. White PF, Coe V, Shafer A, Sung M-L. Comparison of alfentanil with fentanyl for outpatient anesthesia. Anesthesiology 1986; 64: 99–106.
10. Andrews CJH, Sinclair M, Prys-Roberts C, Dye A. Ventilatory effects during and after continuous infusion of fentanyl and alfentanil. Br J Anaesth 1983; 55: S211–6.
11. Jordan C. Assessment of the effects of drugs on respiration. Br J Anaesth 1982; 54: 763–82.
12. Watson HL, Poole DA, Sackner MA. Accuracy of respiratory inductive plethysmographic cross-sectional areas. J Appl Physiol 1988; 65: 306–8.
13. Valta P, Takala J, Foster R, et al. Evaluation of respiratory inductive plethysmography in the measurement of breathing pattern and PEEP induced changes in lung volume. Chest 1992; 102: 234–8.
14. Takala J, Keinänen O, Väisänen P, Kari A. Measurement of gas exchange in intensive care: laboratory and clinical validation of a new device. Crit Care Med 1989; 17: 1041–7.
15. Scott JC, Stanski DR. Decreased fentanyl and alfentanil dose requirements with age: a simultaneous pharmacokinetic and pharmacodynamic evaluation. J Pharm Exp Ther 1987; 240: 159–66.
16. Scott JC, Ponganis KV, Stanski DR. EEG quantitation of narcotic effect: the comparative pharmacodynamics of fentanyl and alfentanil. Anesthesiology 1985; 62: 234–41.
17. Sackner MA, Watson H, Belsito AS, et al. Calibration of respiratory inductive plethysmography during natural breathing. J Appl Physiol 1989; 66: 410–20.
18. Tulla H, Takala J, Alhava E, et al. Breathing pattern and gas exchange in emergency and elective abdominal surgical patients. Int Care Med 1995; 21: 319–25.
19. Tobin MJ, Chadha TS, Jenouri G, et al. Breathing patterns. 1. Normal subjects. Chest 1983; 84: 202–5.
20. Kiiski R, Takala J, Eissa T. Measurement of alveolar ventilation and changes in deadspace by indirect calorimetry during mechanical ventilation: a laboratory and clinical validation. Crit Care Med 1991; 19: 1303–9.
21. Michiels M, Hendriks R, Heykants J. Radioimmunoassay of the new opiate analgesic alfentanil and sufentanil: preliminary pharmacokinetic profile in man. J Pharm Pharm 1983; 35: 86–93.
22. Woestenborgh R, Stanski D, Scott J, Heykants J. Assay methods for fentanyl in serum: gas-liquid chromatography versus radioimmunoassay. Anesthesiology 1987; 67: 85–90.
23. Ebling WF, Lee EN, Stanski DR. Understanding pharmacokinetics and pharmacodynamics through computer simulation: the comparative clinical profiles of fentanyl and alfentanil. Anesthesiology 1990; 72: 650–8.
24. Gambus PL, Gregg KM, Shaefer SL. Validation of the alfentanil canonical univariate parameter as a measure of opioid effect on the electroencephalogram. Anesthesiology 1995; 83: 747–56.
25. Fung DL, Eisele JH. Fentanyl pharmacokinetics in awake volunteers. J Clin Pharm 1980; 20: 652–8.
26. Santiago TV, Edelman NH. Opioids and breathing. J Appl Physiol 1985; 59: 1675–85.
27. McClain DA, Hug DD. Pharmacodynamics of opiates. Int Anesthesiol Clin 1984; 4: 75–94.
28. McDonnell TE, Bartkowski RR, Williams JJ. ED50
of alfentanil for induction of anesthesia in unpremedicated young adults. Anesthesiology 1984; 60: 136–40.
© 2001 International Anesthesia Research Society
29. Drummond GB. Comparison of decreases in ventilation caused by enflurane and fentanyl during anesthesia. Br J Anaesth 1983; 55: 825–35.