III-031 James Morse

Sevoflurane Pharmacokinetics and Pharmacodynamics in Adults During General Anaesthesia

James D Morse (1,2), Nick HG Holford (2), Brian J Anderson (1), Jacqueline A Hannam (2), Robin Kang (3), Simon J Mitchell (1,3), Doug Campbell (3), Timothy G Short (1,3)

1. Department of Anaesthesiology, University of Auckland, Auckland, New Zealand; 2. Department of Pharmacology and Clinical Pharmacology, University of Auckland, Auckland, New Zealand; 3. Department of Anaesthesia, Auckland City Hospital, Auckland, New Zealand.

Background Sevoflurane vapour is inhaled for the induction and maintenance of general anaesthesia. Sevoflurane vapour concentration is measured in real-time in the anaesthetic circuit and displayed to the anaesthetist as the end-tidal sevoflurane concentration (Vol%). The relationship between arterial sevoflurane concentration and depth of anaesthesia (e.g. electroencephalographic (EEG) activity) is poorly described.  Current models employed to describe sevoflurane pharmacokinetics often do not consider co-administered drugs on anaesthetic depth nor investigate covariates such as age, body size and composition.  

Objective The objective of this study was to derive a population pharmacokinetic pharmacodynamic (PKPD) model in adults quantifying the relationships between inspired sevoflurane, arterial plasma concentration and EEG activity measured by the bispectral index (BIS). We also sought to examine the effect of age and body composition on these relationships and assess the interaction of coadministered remifentanil and propofol with sevoflurane on BIS.

Methods Sevoflurane Vol%, gas flow rate and minute volume inspired were used to calculate sevoflurane input rate. Arterial sevoflurane plasma concentration and BIS were described using nonlinear mixed effects models. The PK parameters were estimated initially using sevoflurane concentration measurements then fixed in the PKPD model [1]. Sevoflurane input is confounded by an unknown bioavailability factor and is determined in part by physical (within the anaesthetic circuit) and physiological (within the lungs) deadspaces. Bioavailability was assumed to be 1 for this analysis. All participants received remifentanil target-controlled infusion (effect-site therapeutic range 3-5 µg/L) using the parameter set derived by Minto and colleagues [2]. Propofol boluses were given at the discretion of the anaesthetist for induction of anaesthesia. Zero-order input and first-order elimination were assumed and two and three-compartment models were evaluated for sevoflurane distribution. Models were parameterised in terms of clearances and volumes. Total body mass, fat-free mass and normal fat mass [3] were investigated to describe size differences scaled to a 70 kg individual using theory based allometry [4]. Fat-free mass was predicted for the individual patient based on sex, total body mass and height [5]. Age-related changes in sevoflurane clearance was investigated using a linear function standardised to a 40 year old individual.

The BIS monitor provided observations in the 0-100 range but processing of the internal signals at the extremes of this range was unknown. This is important when trying to estimate a baseline BIS value because the device is censored with a maximum value of 100. There were many BIS observations close to the censored value of 100 which made a statistical model for the residual error and baseline variability challenging. Consequently, a logit transform both sides method was applied to the observed BIS [6].

BIS was described using a sigmoidal maximal effect (EMAX) model. An effect-site compartment was used to relate sevoflurane concentrations to effect, linked using an equilibration half-life, T1/2 eq. The Greco [7] response-surface model described additive effects of propofol [8] and remifentanil [2] with sevoflurane on BIS. Population parameter variability (PPV) assumed a log-normal distribution for the random effects except for EMAX which assumed a logit distribution.

Results Participants were 60 adults (19-87 years, 44-160 kg, ASA grade 1-3) undergoing elective surgery. Data comprised 825 concentrations and 50,482 BIS observations. Sevoflurane PK were best described using a three-compartment structural model. Fat-free mass best described differences in sevoflurane clearance (CL) while normal fat mass best described differences in volume (FfatV=3.72). Parameter estimates (random effect shown as sqrt(variance) were CL 1050 (0.396) L/h/70 kg, Q2 636 (0.861) L/h/70 kg, Q3 405 (0.246) L/h/70 kg, V1 27 (0.716) L/70 kg, V2 83.5 (0.608) L/70 kg, V3 1380 L/70 kg. Sevoflurane CL decreased linearly with age (CL 1050 L/h/70 kg in a 40 year old versus 801 L/h/70 kg in an 80 year old).

The maximum effect on bispectral index (EMAX) was an 88 point reduction from a baseline bispectral index of 98. Using the original BIS scale produced a distribution of predicted values that were less than the observed distribution of baseline BIS responses. Two approaches were used to deal with this problem using the transform both sides (TBS) method. The first used a logit transformation of the observed BIS with expit values scaled between 0 and 102 to allow for a residual error that could in theory exceed the machine limitation of 100. The logit TBS method improved the ability to describe the distribution of baseline BIS measurements but it proved impossible to obtain plausible estimates of the random effect parameters in order to obtain an adequate visual predictive check (VPC) [9]. The second TBS method involved subtraction of the observed BIS value from the nominal machine baseline value of 100 which improved estimates of pharmacodynamic parameter variability and VPCs.

The effect-site sevoflurane concentration (Ce50 SEVO) producing half of EMAX was 4.4 mg/L. The Ce50 PROP for propofol was 1.78 mg/L and for remifentanil Ce50 REMI was 11.7 µg/L. The estimated sevoflurane T1/2, eq was 1.62 minutes. The T1/2, eq for propofol was 1.26 minutes and remifentanil 1.12 minutes.

Conclusion A three compartment model best described sevoflurane PK using fat-free mass for clearances and normal-fat mass for volumes. The sevoflurane PKPD model developed in this study could be used to implement a target-concentration strategy to determine a drug dose associated with the desired anaesthetic effect for sevoflurane alone and in combination with either propofol or remifentanil. If a target reduction of BIS to 50% of baseline is sought, then the target concentration will be equal to the Ce50 for propofol (1.78 mg/L) and remifentanil (11.7 µg/L) when combined with sevoflurane at a plasma concentration of 4.4 mg/L (end-tidal sevoflurane 0.35 Vol%). Modelling of BIS observations on their natural scale resulted in poor model predictions when assessed with a VPC. A transform both sides approach was employed to improve the fit of the model to the data. This PKPD model for sevoflurane could be used in commercially available target-controlled infusion devices to achieve target sevoflurane plasma or effect-site concentrations [10].

References

  1. Zhang L, Beal SL, Sheiner LB. Simultaneous vs. Sequential Analysis for Population PK/PD Data I: Best-Case Performance. Journal of Pharmacokinetics and Pharmacodynamics. 2003;30(6):387-404.
  2. Minto CF, Schnider TW, Egan TD, Youngs E, Lemmens HJ, Gambus PL, et al. Influence of age and gender on the pharmacokinetics and pharmacodynamics of remifentanil. I. Model development. Anesthesiology. 1997;86(1):10-23.
  3. Holford NHG, Anderson BJ. Allometric size: The scientific theory and extension to normal fat mass. European journal of pharmaceutical sciences : official journal of the European Federation for Pharmaceutical Sciences. 2017;109S:S59-S64.
  4. Anderson BJ, Holford NH. Mechanism-based concepts of size and maturity in pharmacokinetics. Annu Rev Pharmacol Toxicol. 2008;48:303-32.
  5. Janmahasatian S, Duffull SB, Ash S, Ward LC, Byrne NM, Green B. Quantification of lean bodyweight. Clinical pharmacokinetics. 2005;44(10):1051-65.
  6. Dosne AG, Bergstrand M, Karlsson MO. A strategy for residual error modeling incorporating scedasticity of variance and distribution shape. J Pharmacokinet Pharmacodyn. 2016;43(2):137-51.
  7. Greco WR, Park HS, Rustum YM. Application of a new approach for the quantitation of drug synergism to the combination of cis-diamminedichloroplatinum and 1-beta-D-arabinofuranosylcytosine. Cancer research. 1990;50(17):5318-27.
  8. Morse J, Hannam JA, Cortinez LI, Allegaert K, Anderson BJ. A manual propofol infusion regimen for neonates and infants. Paediatr Anaesth. 2019;29(9):907-14.
  9. Nguyen TH, Mouksassi MS, Holford N, Al-Huniti N, Freedman I, Hooker AC, et al. Model Evaluation of Continuous Data Pharmacometric Models: Metrics and Graphics. CPT Pharmacometrics Syst Pharmacol. 2017;6(2):87-109.
  10. Lerou JGC, Booij LHDJ. Model-based administration of inhalation anaesthesia. 1. Developing a system model. British Journal of Anaesthesia. 2001;86(1):12-28.

Reference: PAGE 32 (2024) Abstr 11084 [www.page-meeting.org/?abstract=11084]

Poster: Drug/Disease Modelling - Other Topics