Predictive accuracy of alfentanil infusion in coronary artery surgery: A prebypass study in middle-aged and elderly patients

Predictive Accuracy A Prebypass

of Alfentanil Infusion Study in Middle-Aged

L. Barvais, MD, A. D’Hollander, F. Cantraine,

MD, PhD, D. Schmartz,

T

HE PHARMACOKINETIC profile of alfentanil appears particularly suited for cardiac anesthesia because its effector site concentration decreases more rapidly than the other opioids after a long-lasting infusion.’ The use of high doses of alfentanil as a continuous infusion associated with lorazepam premeditation provides satisfactory anesthetic conditions in patients undergoing myocardial revascularization.? Moreover, computer-assisted continuous infusion (CACI) of an opioid drug during cardiac surgery was demonstrated to give greater hemodynamic stability, to require fewer adjuvant drug interventions, and to experience fewer hypotensive and hypertensive episodes than the manual bolus method.” The predictive accuracy of alfentanil in patients anesthetized using a CACl system has already been extensively studied in adult-” and elderly patients’ undergoing general surgery; however. data are lacking in patients undergoing cardiac surgery. The aim of the present study was to evaluate the predictive accuracy of alfentanil using either, prospectively, the population pharmacokinetic parameters of Maitrc et aI8 and the pharmacokinetic set of Helmets ct al” derived from elderly patients, or retrospectively, using conventional pharmacokinetic data sets published in the literature”‘-‘” in order to select the most appropriate pharmacokinetic set for subsequent USC in middle-aged and elderly patients scheduled for CABG surgery.

From the Depurtment of Anesthesiology, Erasmus Hospital, Brussels, and the Department qf Computer Science, ULB (Free Universir) of Brussels), Brussels, Belgium. Address correspondence und reprint requests to L. Barvais, MD, Depurtment of Anesthesiology, Erusmus Hospital. 808 route de Lennick, 1070, Brussels, Belgium. Copyright o 1994 by W.B. Saunders Company 1053-0770/9410803-0004$03.00/0

Artery Surgery: Patients

MD, C. Hendrice,

MS, PhD, and E. Coussaert,

Twenty-three informed and consenting patients scheduled for CABG were anesthetized using computer-controlled infusions of alfentanil, midazolam, and pancuronium. Thirteen middle-aged patients received a preprogrammed infusion scheme of alfentanil, simulated using the population pharmacokinetic set of Maitre et al (Group M), and 10 elderly patients received a preprogrammed infusion scheme simulated using the model of Helmers et al (Group H). The target alfentanil concentrations in groups M and H for tracheal intubation were: 300-500 ng/mL and for sternotomy: 500700 ng/mL. Blood alfentanil concentrations were measured at tracheal intubation, skin incision, sternotomy, and aortic cannulation. The bias, inaccuracy, and precision of each pharmacokinetic set were assessed by the median performance error (MDPE), the median absolute performance error (MDAPE), and the 10th and 90th percentiles of the performance errors (PlO, P90), respectively. The predictive accu-

278

in Coronary and Elderly

MD,

MS

racy of seven other alfentanil pharmacokinetic sets selected from the literature was also evaluated retrospectively. The measured alfentanil concentrations were underpredicted when using all the pharmacokinetic sets, except the set of Scott et al (MDPE: -15.9%). The sets of Maitre et al and Helmers et al were found not to be accurate (MDAPE > 40%) in both groups M and H. The set of Scott et al with the lowest clearance (2.4 mL/ kg/ min) shows the best accuracy (MDAPE: 19.5%) and precision (PlO: -40%. P90: 16%). In conclusion, the set of Scott et al should preferably be selected to predict prebypass alfentanil infusion accurately in either middle aged or elderly patients who have normal myocardial function (LVEF > 50%) and are scheduled for CABG. Copyright (f‘1994 by W.B. Saunders Company KEY WORDS: analgesics,

alfentanil,

MATERIALS

anesthetic,

intravenous

AND METHODS

The study was approved by the ethics committee of this institution and all patients gave informed consent. Inclusion criteria for the 23 selected patients were subjects within 25% of ideal body weight, normal renal and hepatic function based on preoperative routine blood chemical analysis, normal blood pressure for age and sex, no cimetidine treatment, and a left ventricular ejection fraction equal or greater than 50% (estimated by preoperative ventriculography and/or echocardiography). Preoperative medications were continued the day and morning before, except for the P-blockers. which were stopped on the morning of surgery, if the heart rate was less than 60 beatsimin. For premeditation, the patients received bromazepam, 6-12 mg, orally the night before and 3 to 6 mg orally 1 hour prior to surgery. Anesthesia consisted of a computer-controlled infusion of alfentanil and midazolam.‘7~‘x During the placement of the arterial and the flow-directed pulmonary artery catheters, all the patients were sedated using a preprogrammed infusion regimen of midazolam in of 100 ng/mL, order to maintain a stable target concentration

based on the pharmacokinetic set of Persson et al.]’ Thereafter, anesthesia was induced with an alfentanil infusion. Pancuronium (0.12 mgikg) was given to obtain muscle relaxation. Tracheal intubation was performed and the patients were ventilated with 50% 02 in air to maintain the PaCOz within normal limits. Thirteen patients younger than 65 years of age received a preprogrammed alfentanil infusion regimen using the population pharmacokinetic set of Maitre et als (Group M). The target concentration was 500 ng/mL for endotracheal intubation and 700 ng/mL for sternotomy and the prebypass period. Ten patients older than 65 years of age received a preprogrammed alfentanil infusion regimen using the pharmacokinetic set of Helmers et al, studied in elderly patients (9) (Group H). The target concentration was 300 ng/mL for endotracheal intubation and 500 ng/mLfor sternotomy and the prebypass period. Arterial blood samples were drawn at intubation, skin incision, sternotomy. and aortic cannulation. Plasma concentrations of alfentanil were determined by a specific capillary gas chromatographic assay.“’ The mean coefficient of intra-day and inter-day variation was less than 5%. Based on the same infusion rates as those given to each patient,

Journal of Cardiothoracic and Vascular Anesthesia, Vol8, No 3 (June), 1994: pp 278-283

279

PREDICTIVE ACCURACY OF ALFENTANIL IN CABG SURGERY

Teble 1. Llet of the Atfentanil Pharmacokinetic

Author

Dete Sate Selected in the Liierature

ClWranCe

Vl

KlO

K12

K21

K13

K31

(mL/kglmin)

(mLlkg1

(min-‘)

(min-‘1

(min-‘)

(min-‘)

(min-‘)

0.047

0.018

0.024

0.02

0.0221

0.00997

0.113

0.017

BovilPo

5.078

0.063

0.362

0.271

Bowerrr

3.434

154

80.6

0.0223

0.055

0.0357

Camur2

8.360

220

0.038

0.057

0.135

Fragenr3

3.601

130

0.0277

0.0522

0.0303

Helmem

4.431

211

0.021

0.031

0.02

Maitre*

5.084

111

0.0458

0.117

0.0775

Schuttlerr4

5.924

160

0.037

0.104

0.052

Scottrs

2.339

0.091

0.656

0.214

ShaferI

4.400

0.022

0.07

0.1

25.7 200

Abbreviations: Vl, volume of the central compartment:

KlO, elimination time constant from the central compartment;

K12, equilibration time

constant from the central compartment to the superficial peripheral compartment; K21, equilibration time constant from the superficial peripheral compartment to the central compartment; K13, equilibration time constant from the central compartment to the deep peripheral compartment: K31, equilibration time constant from the deep peripheral compartment to the central compartment.

theoretical

predicted

central

compartment

concentrations

were

using the selected average phannacokinetic data sets taken from the literature and included in Table 1. To evaluate the predictive accuracy of these different pharmacokinetic sets, the performance error (PE) and the absolute performance error (APE) were calculated for all the samples of both groups M and H.*ry” The PE measures the difference between the measured concentration and the predicted concentration as a fraction of the predicted concentration.21*22 To assess the predictive accuracy of each pharmacokinetic set, the median performance error (MDPE), the median absolute prediction error (MDAPE), and the dispersion between the percentiles 10 (PlO) and 90 (P90) of the PEs were calculated. MDPE, MDAPE, PlO, and P90 are expressed as percentages to eliminate the influence of high or low predicted alfentanil concentrations. For each pharmacokinetic set, MDPE represents the bias that is an estimate of the overprediction or underprediction of the measured concentrations. MDAPE is a measure of the inaccuracy, and the difference between PlO and P90 provides information on the dispersion within the studied population. Using a single set of pharmacokinetic parameters to describe a group of patients, a 20% to 40% MDAPE and a dispersion of the 10th to 90th percentiles of -35% to +35% have been proposed as an acceptable predictive accuracy,” reflecting the fundamental interindividual pharmacokineticvariability. Linear regression analyses were used to study the relation between the predictive accuracy values (MDPE and MDAPE) of the studied data sets and their pharmacokinetic parameters. calculated

dieted when using all the pharmacokinetic sets, except the set of Scott et a1,15with a negative bias of -18.6% and -9.6% in both groups M and H, respectively (Fig 2). When the 23 patients are pooled, the bias of the set of Scott et al is -15.9%15 (Fig 2). For both groups M and H, MDAPE is low ( < 20%) when using the set of Scott et al, l5 between 20 and 40% using the sets of Fragen et all3 or Bower et al,” and greater than 40% using the six other studied alfentanil pharmacokinetic setss-10,12J4,16 (Fig 2). The MDAPE of the nine pharmacokinetic sets selected from the literature shows almost the same ascending order in both groups M and H (Fig 2). The dispersion between PlO and P90 was also lower using the set of Scott et al (-40% to 16%) compared with the other data sets (Fig 2). A good linear regression analysis between the clearance of the pharmacokinetic sets listed in Table 1 and their MDPE and MDAPE was found. The coefficient of correlation was greater than 0.88 in both groups M and H (Fig 3 A and B). No relation was found with either the volume of the central compartment or the intercompartment rate constant values. The set of Scott et al, which has the lowest clearance (2.4 mL/kg/min), has the best predictive accuTable 2. Sex, Age, Weight, Height, and Concomitant Drug Therapy of Both Groupe of Patients

RESULTS

Nineteen male and 4 female patients participated in the study. The average age was 54.9 years (range 38 to 62) in group M and 70.5 years (range 65 to 84) in group H (Table 2). Table 2 contains the weight, height, and associated drug therapy for each patient. Figure 1 A and B shows the alfentanil measured and target concentrations using the pharmacokinetic parameters of Maitre et al for group M and Helmers et al for group H, respectively. In both groups, the target alfentanil concentrations clearly underpredict the measured concentrations. This is associated with a large bias (49.4% and 89.2%, respectively), a very poor MDAPE (49.4% and 89.2%, respectively), and a large dispersion between PlO and P90 (Fig 2). The measured alfentanil concentrations were underpre-

N

Sex (M/F)

Group M

Group H

Mean 2 SD

Mean ? SD

13

10

12/l

713

54.9 2 6.5

70.5 2 4.8

Age (Year) Weight (kg)

70.7 + 9.7

68 f 9.4

Height (cm)

171.5 + 6.3

164 ? 7.1

Beta blocker Calcium antagonist Nitrates

11

8

5

6

13

9

NOTE. Group M = 13 patients, younger than 65, receiving a preprogrammed alfentenil infusion regimen based on the population pharmacokinetic set of Maitre et al.B Group H = 10 patients older than 65 receiving a preprogrammed alfentanil infusion regimen based on the pharmacokinetic set of Helmers et al,Ostudied in elderly patients.

BARVAIS ET AL

1*5 .'--_~.._._ 1.4 _I 1.3 4 1.2 *, 1.1 4 1' 0.9 J 0.8 4 0.7 ;

INTUSATION

1.6 I 1.5 1.4 1.3 1.2 1.1 l0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 -

01

INCISION STERNOTOW

CANNULATION

1

,: I

I

I

I

lNTUSATlON

INCWON

STERNOTOW

racy, whereas all the sets with a clearance greater than 3 ml/kg/mm systematically underevaluate the measured concentrations. DISCUSSION

The predictive accuracy of alfentanil using either a conventional or a population pharrnacokinetic model has already been studied in general surgery. Using the conventional average alfentanil pharmacokinetic data of Schuttler et a1,14Ausems et al4 showed that there was no systematic overprediction or underprediction of the alfentanil concentrations in 3.5 patients receiving alfentanil by bolus and

CANNULATION

Fig 1. (A) Measured affentanil concentratfons (+) in the 13 patients of group M, receiving the preprogrammed infurion scheme simulated using the population pharmacokfnetk set of Maitre et al. The data points are linked to identify the patients. (B) Measured alfentanil concentrations (+) in the 10 patients in group H, receiving the preprogrammed infusion scheme simulated using the pharmacokinetic set of Helmers et al. The data points are linked to identify the patients. W, mean measured alfentanil concentration;---, target affentanil concentration.

continuous infusion for lower and upper abdominal surgery. Maitre et al5 demonstrated that their population pharmacokinetic parameters appeared to be “robust” in a different group of surgical patients, given IV boluses and a variable-rate infusion of alfentanil for induction and maintenance of anesthesia. The use of alfentanil pharmacokinetic parameters derived from bolus administration in elderly patients resulted in a general underprediction of measured alfentanil concentrations in the elderly surgical patients.’ Recently, Raemer et al6 demonstrated that the pharmacokinetic parameters of Scott et all5 were more suitable than the population pharmacokinetic parameters

281

PREDICTIVE ACCURACY OF ALFENTANIL IN CABG SURGERY

250-1GROUP M: 13 PATIENTS / 52 SAMPLES

-50 -100 r

d

- P50

* MDAPE

p P90

- PlO

BCOTTFRKlEN-BO~~~LLMMREWAFERHELMEWBCHUTTLERCAMU

%300 250

Fii 2. Predictive accuracy of the nine selected alfentanil phwmacokinetlc sets for the 23 prtients pooled, for the 13 patients of group M and the 10 patients of group H, respectively. The median of the performance emw (MDPE) represents the bias, ?? . The median of the absolute performance error (MDAPE) w the inacwracy, +. The dispersion is represented bv the difference between the percentiles 10 and 90.

-5o-100 ’

of Maitre et al6 for use in a CACI system. These authors found no explanation for the better suitability of the set of Scott et all5 The influence of factors such as the assay method, the accuracy of the CACI system used, the patients’ selection, the data analysis method, and the site of sampling were excluded.6 Recently, Barvais et al” have showed that pharmacokinetic sets of alfentanil derived from a large bolus must not be selected to predict continuous infusion accurately. Among the study conditions of the pharmacokinetic sets included in Table 1, the differences between the dose regimen used could partially explain the variable predictive accuracy. The model of Scott et all5 is derived from a bolus injected slowly, whereas the models of Maitre et al8 and Helmers et al9 are derived from a large alfentanil bolus injected over a short period. In the present sample of coronary artery patients undergoing myocardial revascularization, the set of Scott et all5 had the lowest bias and inaccuracy and the best precision in

GROUP H: 10 PATIENTS / 39 SAMPLES /

/

- P50

-MDAPE

-P90

- PlO

8aJl-rFwQENBowERl3mlu~R~mLMERBscHIIITLEqcAMU

both groups M and H, such as in the study of Raemer et a1.6 However, the sets of Fragen et all3 and Bower et al” also had an acceptable predictive accuracy whereas the sets of Bovill et a1,t0 Maitre et al,8 Helmers et aL9 Schuttler et a1,14 and Camu et all2 demonstrated a poor predictive accuracy. Does any factor influence the suitability of the set of Scott et alI5 for patients receiving alfentanil by a CACI system? During continuous alfentanil infusion, the plasma concentration is mainly influenced by its clearance because the equilibration time constants for the fast and slow compartments are relatively brief for alfentanil compared with the other phenylpiperidine opioid drugs.’ In this study, a strong linear regression analysis between the clearance of the different alfentanil data sets and their respective bias (MDPE) or inaccuracy (MDAPE) was evidenced. The set of Scott et all5 with the lowest clearance value of 2.4 mL/kg/min shows the best predictive accuracy (Fig 3 A and B), whereas the other sets whose clearance rate values were greater than 3 mL/kg/min systematically underpredicted

282

A

BARVAIS ETAC

-.___-_

‘40T-----;z

Y = - 55.02

__..__...

+ (22.98

- 31.77

110 1001

+ (19.025

* cbarance)

i

’ //'

.

.

. . /

:

R2 = 0.89

/

180.

Y=4.86

Y=-34.56

lao-

+(26.12*ckuance)

160-

+(23.44’dearance)

loo -

"MO#

120.

': 40 80

. ..

-

.

$:I

.

I

e0

. =w-

140. 20 0 0

___i

R'=

.

.

I

/

0.88

3

2clearak (mvlkglmin)” ’

Fig 3. (A) Group M: 13 patients/S models. Rogrossion ~MI@s botwoon MDfE and MDAPE with the clrronco of the nine phwmacokinetic sete selected in the literature. (B) Group H: 10 patients/B models. Regression analysis betwoan MOAPE and MDAPE wfth the clearancr of the nine phrrmacokinrtic sets selectrd in the Ilterature.

the measured concentrations. In this sample of 23 studied patients, a low clearance of alfentanil could be expected because of their age.9 Moreover, the P-blockers by their effect on the hepatic blood flow will decrease the metabolism of the opioids,2S whereas the interaction between opioids and calcium entry blockers appears to be mild. In conclusion, the pharrnacokinetic set of Scott et all5 with a low alfentanil clearance was found to be accurate to predict prebypass alfentanil concentrations, in either adult

or elderly patients with normal myocardial function (LVEF > 50%) undergoing myocardial revascularization. ACKNOWLEDGMENT

The authors are grateful to Dr Peter SA Glass of Duke University, Durham, NC, for the advice he provided during the preparation of this article. The authors also acknowledge Dr D. Gangji, PhD, for his support in the assay samples of alfentanil by GLC, and Mrs Geeraerts for her secretarial assistance.

REFERENCES

1. Shafer SL, Varvel JR: Pharmacokinetics, pharmacodynamics, and rational opioid selection. Anesthesiology 7453~63,199l 2. Mantz J, Abi-Jaoude F, Ceddaha A, et al: High-dose alfentanil for myocardial revascularization: A hemodynamic and pharmacokinetic study. J Cardiothorac Vast Anesth 5:2:107-110,199l 3. Alvis M, Reves JG, Govier AV, et al: Computer-assisted continuous infusions of fentanyl during cardiac anesthesia: Comparison with a manual method. Anesthesiology 63:41-49,1985 4. Ausems ME, Stanski DR, Hug CC: An evaluation of the accuracy of pharmacokinetic data for the computer assisted infusion of alfentanil. Br J Anaesth 57:1217-1225,1985 5. Maitre PO, Ausems ME, Vozeh S, Stanski DR: Evaluating the accuracy of using the population pharmacokinetic data to predict plasma concentrations of alfentanil. Anesthesiology 68:5967,1988

6. Raemer DB, Buschman A, Varvel JR, et al: The prospective use of population pharmacokinetics in a computer-driven infusion system for alfentanil. Anesthesiology 73:66-72,199O 7. Lemmens HJM, Bovill JG, Burm AGL, Hennis PJ: Alfentanil infusion in the elderly. Prolonged computer-assisted infusion of alfentanil in the elderly surgical patient. Anaesthesia 43:850-856,1988 8. Maitre OP, Vozeh S, Heykants J, et al: Population pharmacokinetics of alfentanil: The average dose-plasma concentration relationship, and interindividual variability in patients. Anesthesiology 66:3-12,1987 9. Helmers JH, Van Peer A, Woestenborghs R, et al: Alfentanil kinetics in the elderly. Clin Pharmacol Ther 36239-243, 1984 10. Bovill JG, Sebell PS, Blackburn CL, Heykants J: The pharmacokinetics of alfentanil, a new opioid analgesic. Anesthesiology 57~439~443,1982

PREDICTIVE ACCURACY

OF ALFENTANIL

IN CABG SURGERY

11. Bower S, Hull CJ: Comparative pharmacokinetics of fentarry1and alfentanil. Br J Anaesth 54:871-877, 1982 12. Camu F, Gepts E, Rucquoi M, Heykants J: Pharmacokinetits of alfentanil in man. Anesth Analg 61:657-661, 1982 13. Fragen RJ, Booij L, Braak G, et al: Pharmacokinetics of the infusion of alfentanil in man. Br J Anaesth 55:1077-1081,1983 14. Schuttler J, Stoeckel H: Alfentanil [R 392091: Ein neues kurzwirkendes Opioid. Anaesthesist 31:10-14, 1982 15. Scott JC, Ponganis KV, Stanski DR: EEG quantification of narcotic effect: The comparative pharmacodynamics of fentanyl and alfentanil. Anesthesiology 62:234-241,1985 16. Shafer A, Sung M, White PF: Pharmacokinetics and pharmacodynamics of alfentanil infusions during general anesthesia. Anesth Analg 65:1021-1028, 1986 17. Cantraine F, Coussaert E, d’Hollander A, Barvais L: MINA: Monitoring of intravenous anaesthesia. A system to monitor the infusion of 4 intravenous drugs under the control of a personal computer. Int J Clin Monit Comput 6:109-117, 1989 18. Barvais L, Coussaert E, Cantraine F, d’Hollander A: The pharmacokinetics of intravenous anaesthetic drugs given by infusion: SPINA a software program. Eur J Anaesth 6:435-447,1989 19. Persson P, Nilsson A, Hartvig P, Tamsen A: Pharmacokinet-

283

its of midazolam in total IV anesthesia. Br J Anaesthesia 59:548556,1987 20. Bjorkman S, Aziz N, Stein DA: Determination of alfentanil in serum by radioimmunoassay or capillary column gas-liquid chromatography: A comparison of two assays. Acta Pharm Nordica 1:211-220,1989 21. Shafer SL, Varvel JR, Aziz N, Scott JC: Pharmacokinetics of fentanyl administered by computer-controlled infusion pump. Anesthesiology 73:1091-1102, 1990 22. Glass PSA, Jacobs JR, Smith LR, et al: Pharmacokinetic model-driven infusion of fentanyl: Assessment of accuracy. Anesthesiology 73:1082-1090,199O 23. Glass PSA, Jacobs JR, Reves JG: Intravenous anesthetic delivery in Miller R (ed): Anesthesia, vol 1 (ed 3). 1990, pp 367-388, sect 2, chap 11 24. Barvais L, Cantraine F, d’Hollander A, Coussaert E: Predictive accuracy of continuous alfentanil infusion in volunteers: Variability of different pharmacokinetic sets. Anesth Analg 77:801810,1993 25. Riddell JG, Harron DWG, Shanks RG: Clinical pharmacokinetics of beta-adrenoreceptor antagonists and update. Clin Pharmacokinet 12:305-310, 1987