(Almost) Everything You Learned About Pharmacokinetics Was (Somewhat) Wrong!

The Performance of an Artificial Neural Network Model in Predicting the Early Distribution Kinetics of Propofol in Morbidly Obese and Lean Subjects

BACKGROUND

Induction of anesthesia is a phase characterized by rapid changes in both drug concentration and drug effect. Conventional mammillary compartmental models are limited in their ability to accurately describe the early drug distribution kinetics. Recirculatory models have been used to account for intravascular mixing after drug administration. However, these models themselves may be prone to misspecification. Artificial neural networks offer an advantage in that they are flexible and not limited to a specific structure and, therefore, may be superior in modeling complex nonlinear systems. They have been used successfully in the past to model steady-state or near steady-state kinetics, but never have they been used to model induction-phase kinetics using a high-resolution pharmacokinetic dataset. This study is the first to use an artificial neural network to model early- and late-phase kinetics of a drug.

METHODS

Twenty morbidly obese and 10 lean subjects were each administered propofol for induction of anesthesia at a rate of 100 mg/kg/h based on lean body weight and total body weight for obese and lean subjects, respectively. High-resolution plasma samples were collected during the induction phase of anesthesia, with the last sample taken at 16 hours after propofol administration for a total of 47 samples per subject. Traditional mammillary compartment models, recirculatory models, and a gated recurrent unit neural network were constructed to model the propofol pharmacokinetics. Model performance was compared.

RESULTS

A 4-compartment model, a recirculatory model, and a gated recurrent unit neural network were assessed. The final recirculatory model (mean prediction error: 0.348; mean square error: 23.92) and gated recurrent unit neural network that incorporated ensemble learning (mean prediction error: 0.161; mean square error: 20.83) had similar performance. Each of these models overpredicted propofol concentrations during the induction and elimination phases. Both models had superior performance compared to the 4-compartment model (mean prediction error: 0.108; mean square error: 31.61), which suffered from overprediction bias during the first 5 minutes followed by under-prediction bias after 5 minutes.

CONCLUSIONS

A recirculatory model and gated recurrent unit artificial neural network that incorporated ensemble learning both had similar performance and were both superior to a compartmental model in describing our high-resolution pharmacokinetic data of propofol. The potential of neural networks in pharmacokinetic modeling is encouraging but may be limited by the amount of training data available for these models.

Predictive ability of propofol effect-site concentrations during fast and slow infusion rates

BACKGROUND

The performance of propofol effect-site pharmacokinetic models during target-controlled infusion (TCI) might be affected by propofol administration rate. This study compares the predictive ability of three effect-site pharmacokinetic models during fast and slow infusion rates, utilizing the cerebral state index (CSI) as a monitor of consciousness.

METHODS

Sixteen healthy volunteers, 21-45 years of age, were randomly assigned to receive either a bolus dose of propofol 1.8 mg/kg at a rate of 1200 ml/h or an infusion of 12 mg/kg/h until 3-5 min after loss of consciousness (LOC). After spontaneous recovery of the CSI, the bolus was administered to patients who had first received the infusion and vice versa. The study was completed after spontaneous recovery of CSI following the second dose scheme. LOC was assessed and recorded when it occurred. Adequacies of model predictions during both administration schemes were assessed by comparing the effect-site concentrations estimated at the time of LOC during the bolus dose and during the infusion scheme.

RESULTS

LOC occurred 0.97 +/- 0.29 min after the bolus dose and 6.77 +/- 3.82 min after beginning the infusion scheme (P<0.05). The Ce estimated with Schnider (ke0=0.45/min), Marsh (ke0=1.21/min) and Marsh (ke0=0.26/min) at LOC were 4.40 +/- 1.45, 3.55 +/- 0.64 and 1.28 +/- 0.44 microg/ml during the bolus dose and 2.81 +/- 0.61, 2.50 +/- 0.39 and 1.72 +/- 0.41 microg/ml, during the infusion scheme (P<0.05). The CSI values observed at LOC were 70 +/- 4 during the bolus dose and 71 +/- 2 during the infusion scheme (NS).

CONCLUSION

Speed of infusion, within the ranges allowed by TCI pumps, significantly affects the accuracy of Ce predictions. The CSI monitor was shown to be a useful tool to predict LOC in both rapid and slow infusion schemes.

Contributions of PK/PD modeling to intravenous anesthesia

Pharmacokinetic (PK)/pharmacodynamic (PD) modeling has made an enormous contribution to intravenous anesthesia. PK/PD models have provided us with insight into the factors affecting the onset and offset of drug effect. For example, we are now able to describe the influence of cardiac output on the disposition of intravenous drugs within the first few minutes after administration of the drug. We are able to calculate intravenous loading doses that allow for the delay between the concentration of the drug in the plasma and the rising concentration at the site of drug effect. We are able to achieve and maintain a stable level of anesthetic effect using computerized infusion pumps that target the site of drug effect rather than the plasma. Importantly, on the basis of models of drug interaction and an understanding of how drug offset varies with duration of administration, we are now able to rationally combine hypnotics and opioids.

The two-compartment recirculatory pharmacokinetic model'an introduction to recirculatory pharmacokinetic concepts

BACKGROUND

Some limitations of traditional ("mamillary") compartmental pharmacokinetic models of anaesthetic related drugs arise from representing the blood as a central compartment. Recirculatory pharmacokinetic models overcome these limitations. It is proposed that the simplest recirculatory model has only two compartments, and that understanding the properties of this model is a useful introduction to recirculatory pharmacokinetic concepts.

METHODS

The compartments of the model are the lungs and the remainder of the body. The traditional rate constants (e.g. k12 and k21) are replaced by terms that include cardiac output. Drug infusion is into the lung compartment, and drug clearance is from the "body" compartment. The "total" drug concentrations can be thought of as the sum of the first-pass and recirculated drug concentrations at any time. Equations for both first-pass and total drug concentrations in arterial and mixed venous blood are presented. The effects of cardiac output and injection time on these concentrations were analysed.

RESULTS

The first-pass arterial concentrations were shown to make a significant contribution to the total concentrations for high-clearance drugs and/or bolus drug administration. There was an inverse relationship between these first-pass concentrations and cardiac output, and a direct relationship with bolus injection rate. Thus, the total arterial concentrations are affected by these factors in these circumstances.

CONCLUSIONS

The two-compartment recirculatory model is the simplest tool available for elaborating recirculatory pharmacokinetic concepts. The recirculatory approach may provide a conceptual framework of drug disposition that better matches the clinical experience of anaesthetists.

Pharmacokinetic studies of neuromuscular blocking agents: good clinical research practice (GCRP)

In September 1997, an international consensus conference on standardization of studies of neuromuscular blocking agents was held in Copenhagen, Denmark. Based on the conference, a set of guidelines for good clinical research practice (GCRP) in pharmacokinetic studies of neuromuscular blocking agents is presented. Guidelines include: design of the study; relevant patient groups to investigate; test drug administration, sampling and analysis; pharmacokinetic analysis; pharmacokinetic/pharmacodynamic modeling; population pharmacokinetics; statistics; and presentation of pharmacokinetic data. The guidelines are intended to aid those working in this research area; it is hoped that they will assist researchers, editors of scientific papers, and pharmaceutical companies in improving the quality of pharmacokinetic studies.

Cumulation characteristics of cisatracurium and rocuronium during continuous infusion

PURPOSE

The dissimilar pharmacokinetic properties of cisatracurium (CIS) and rocuronium (ROC) predict different potential for drug cumulation when these drugs are administered by continuous infusion. A study was therefore undertaken to compare cumulation potential of CIS and ROC during surgical procedures of relatively long duration (2-4 hr).

METHODS

Sufentanil/propofol-N2O anesthesia was administered to 40 ASA I and II adults. In a double-blind protocol, patients were randomly allocated to receive a continuous i.v. infusion of either CIS or ROC, titrated in progressive increments or decrements as required to achieve and maintain 95 +/- 5% depression of the T1 response of the adductor pollicis muscle, using a Datex NMT-100 Relaxograph EMG monitor applied at the wrist. At the end of surgery, 60 microg x kg(-1) neostigmine plus 15 microg x kg(-1) atropine were administered for reversal.

RESULTS

The duration of infusion was 104 +/- 33 min in group CIS and 110 +/- 23 min in group ROC (P=NS). In both groups, a progressive decrease in potency-adjusted infusion rates was observed after 30 min, then stabilized beyond 60 min. When allowing for an initial period of stabilization, mean potency-adjusted infusion requirements were: CIS 0.81 +/- 0.02 microg x kg(-1) x min(-1) and ROC 5.58 +/- 1.94 microg x kg(-1) x min(-1). There were no differences between groups at any time with regard to potency-adjusted infusion requirements necessary to maintain 90-99% block (P=NS). However, drug costs/hr for maintenance of neuromuscular block were less with CIS ($3.57 +/- 0.09) than with ROC ($6.03 +/- 0.27), P < 0.001.

CONCLUSION

When adjusted to equipotency, infusion requirements of CIS and ROC vary at similar rates during general anesthesia. Despite pharmacokinetic differences, neither drug demonstrates cumulation for infusion lasting up to 3.5 hr.

Is the recovery profile of mivacurium independent of the rate of decay of its plasma concentration in patients with normal plasma cholinesterase activity?

BACKGROUND

The duration of action of muscle relaxants is poorly correlated to the rate of decay of their plasma concentration. The plasma concentration of mivacurium may rapidly decrease below its active concentration because of the extensive hydrolysis of mivacurium. By inflating a tourniquet on one upper limb for 3 min after the administration of atracurium, mivacurium or vecuronium, we studied the influence of the initial decline of their plasma concentration on their effect.

METHODS

In 50 patients anaesthetised with thiopental, isoflurane and fentanyl, the effect of bolus doses of 0.15 or 0.25 mg.kg-1 mivacurium (MIV 15, MIV 25), 0.3 or 0.5 mg.kg-1 atracurium (ATR 30, ATR 50) and 0.06 or 0.1 mg.kg-1 vecuronium (VEC 06, VEC 10) were measured on both arms (evoked response of the adductor pollicis to train-of-four stimulation every 12 s), a tourniquet being applied on one arm just before and during 3 min after the muscle relaxant bolus.

RESULTS

Tourniquet inflation of 3 min almost abolished the neuromuscular effect of mivacurium. In the vecuronium groups and in the ATR 50 group, tourniquet inflation did not modify the maximum degree of depression of the twitch response. Also, the duration of action of vecuronium was unaffected by the tourniquet. In the ATR 30 group, times to return of the twitch response to 25% (duration 25%) and 75% (duration 75%) of control response were significantly shorter in the cuffed arm, 23 min vs 27 min, and 41 min vs 45 min, respectively. In the ATR 50 group, only duration 25% was significantly shorter in the cuffed arm (41 min vs 45 min).

CONCLUSION

The results suggest that the rate of decline of the plasma concentration of mivacurium is so rapid, that a very low and almost clinically ineffective concentration is present as soon as 3 min after its administration. The results also indicate that the recovery from a mivacurium-induced neuromuscular blockade is not influenced by the rate of decay of its plasma concentration in patients with genotypically normal plasma cholinesterase.

Dose proportionality of cisatracurium

The dose proportionality of cisatracurium pharmacokinetics was assessed using a population approach by incorporating the collection of sparse blood samples from patients in clinical trials. Plasma concentration-time data from 131 patients with limited concentration-time data and 38 patients with full sampling were pooled and analyzed using nonlinear mixed-effects modeling (NONMEM). Dose proportionality was assessed using dichotomous parameterization and a linear model. The population pharmacokinetic approach revealed that the pharmacokinetics of cisatracurium are independent of dose between 0.1 mg/kg and 0.4 mg/kg, as was expected based on the importance of Hofmann elimination, a chemical process dependent on pH and temperature.