Anomalies in target-controlled infusion: an analysis after 20 years of clinical use.

Although target-controlled infusion has been in use for more than two decades, its benefits are being obscured by anomalies in clinical practice caused by a number of important problems. These include: a variety of pharmacokinetic models available in open target-controlled infusion systems, which often confuse the user; the extrapolation of anthropomorphic data which provokes anomalous adjustments of dosing by such systems; and the uncertainty of regulatory requirements for the application of target-controlled infusion which causes uncontrolled exploitation of drugs and pharmacokinetic models in target-controlled infusion devices. Comparison of performance of pharmacokinetic models is complex and mostly inconclusive. However, a specific behaviour of a model in a target-controlled infusion system that is neither intended nor supported by scientific data can be considered an artefact or anomaly. Several of these anomalies can be identified in the current commercially available target-controlled infusion systems and are discussed in this review.

The influence of target concentration, equilibration rate constant (ke0 ) and pharmacokinetic model on the initial propofol dose delivered in effect-site target-controlled infusion.

One advantage of effect-site target-controlled infusion is the administration of a larger initial dose of propofol to speed up the induction of anaesthesia. This dose is determined by the combination of the pharmacokinetic model parameters, the target setting and the blood-effect time-constant, ke0 . With the help of computer simulation, we determined the ke0 values required to deliver a range of initial doses with three pharmacokinetic models for propofol. With an effect site target of 4 µg.ml(-1) , in a 35-year-old, 170-cm tall, 70-kg male subject, the ke0 values delivering a dose of 1.75 mg.kg(-1) with the Marsh, Schnider and Eleveld models were 0.59 min(-1) , 0.20 min(-1) and 0.26 min(-1) , respectively. These ke0 values have the attractive feature that, when used to simulate the administration schemes used in two previous studies, predicted effect site concentrations at loss of consciousness were close to those required for maintenance of anaesthesia.

Sevoflurane-enriched blood cardioplegia: the intramyocardial delivery of a volatile anesthetic.

Myocardial ischemia/reperfusion injury is a major problem in cardiac surgery, characterized by an enhanced inflammatory response postoperatively. Sevoflurane has anti-inflammatory effects and may attenuate this injury. This study describes a novel approach to using sevoflurane as a local anti-inflammatory drug and not as an anesthetic. Therefore, a pediatric oxygenator with a sevoflurane vaporizer was integrated into the blood cardioplegia system of an adult bypass system. In addition, a gas blender was implemented to regulate pO2 and pCO2 concentrations in the cardioplegia. This proof-of-principle study was tested in vivo and shows that it is feasible to deliver sevoflurane locally while regulating O2 and CO2 concentrations. Moreover, this set-up enables one to use only the specific cardioprotective features of sevoflurane. Inflammatory responses were attenuated, both locally (i.e. the heart) as well as systemically through intramyocardial delivery of sevoflurane.

A novel technique to determine an 'apparent ke0 ' value for use with the Marsh pharmacokinetic model for propofol.

Debate continues over the most appropriate blood-brain equilibration rate constant (ke0) for use with the Marsh pharmacokinetic model for propofol. We aimed to define the optimal ke0 value. Sixty-four patients were sedated with incremental increases in effect-site target concentration of propofol while using six different ke0 values within the range 0.2-1.2 min(-1). Depth of sedation was assessed by measuring visual reaction time. A median 'apparent ke0' value of 0.61 min(-1) (95% CI 0.37-0.78 min(-1)) led to the greatest probability of achieving a stable clinical effect when the effect-site target was fixed at the effect-site concentration displayed by the target-controlled infusion system, at the time when a desired depth of sedation had been reached. By utilising a clinically relevant endpoint to derive this value, inter-individual pharmacokinetic and pharmacodynamic variability may be accounted for.

Cardiospecific sevoflurane treatment quenches inflammation but does not attenuate myocardial cell damage markers: a proof-of-concept study in patients undergoing mitral valve repair.

BACKGROUND: Inflammation is considered a key mediator of complications after cardiac surgery. Sevoflurane has been shown to quench inflammation and to provide cardioprotection in preclinical studies. Clinical studies using sevoflurane confirm this effect on inflammation but do not consistently show clinical benefits. This paradox may indicate that the contribution of inflammation to postoperative sequalae is less than commonly thought or that systemic doses are too low in their local concentration. To test the latter, we evaluated the effects of intramyocardial sevoflurane delivery. METHODS: Selective myocardial sevoflurane delivery was performed during aortic cross-clamping in patients undergoing valve surgery (n=11). Results were compared with a control group not receiving sevoflurane (n=10). A reference group (n=5) was added to evaluate the effects of systemic sevoflurane delivery. Paired arterial and myocardial venous blood samples were collected at various time points post-reperfusion. Inflammatory mediators and myocardial cell damage were studied. RESULTS: Intramyocardial delivery was superior to systemic delivery in attenuation of interleukin-6 and interleukin-8 (-44% and -25%, respectively; both P=0.001). Myocardial and systemic sevoflurane delivery effectively suppressed surgery-related inflammatory responses including postoperative C-reactive protein levels when compared with controls [63 (47-99) (P=0.01) and 58 (56-81) (P=0.04) compared with 107 (79-144) mg litre(-1)]. Sevoflurane treatment did not reduce postoperative troponin T, creatine kinase, and creatine kinase-MB values. CONCLUSIONS: This proof-of-concept study suggests that intramyocardial delivery compared with the systemic delivery of sevoflurane more strongly attenuates the systemic inflammatory response after cardiopulmonary bypass without reducing postoperative markers of myocardial cell damage. CLINICAL TRIAL REGISTRATION: Nederlands Trial Register NTR2089.

Predictive performance of computer-controlled infusion of remifentanil during propofol/remifentanil anaesthesia.

BACKGROUND: The predictive performance of the available pharmacokinetic parameter sets for remifentanil, when used for target-controlled infusion (TCI) during total i.v. anaesthesia, has not been determined in a clinical setting. We studied the predictive performance of five parameter sets of remifentanil when used for TCI of remifentanil during propofol anaesthesia in surgical patients. METHODS: Remifentanil concentration-time data that had been collected during a previous pharmacodynamic interaction study in 30 female patients (ASA physical status I, aged 20-65 yr) who received a TCI of remifentanil and propofol during lower abdominal surgery were used in this evaluation. The remifentanil concentrations predicted by the five parameter sets were calculated on the basis of the TCI device record of the infusion rate-time profile that had actually been administered to each individual. The individual and pooled bias [median performance error (MDPE)], inaccuracy [median absolute performance error (MDAPE)], divergence and wobble of the remifentanil TCI device were determined from the pooled and intrasubject performance errors. RESULTS: A total of 444 remifentanil blood samples were analysed. Blood propofol and remifentanil concentrations ranged from 0.5 to 11 micro g ml(-1) and 0.1 to 19.6 ng ml(-1) respectively. Pooled MDPE and MDAPE of the remifentanil TCI device were -15 and 20% for the parameter set of Minto and colleagues (Anesthesiology 1997; 86: 10-23), 1 and 21%, -6 and 21%, and -6 and 19% for the three parameter sets described by Egan and colleagues (Anesthesiology 1996; 84: 821-33, Anesthesiology 1993; 79: 881-92, Anesthesiology 1998; 89: 562-73), and -24 and 30% for the parameter set described by Drover and Lemmens (Anesthesiology 1998; 89: 869-77). CONCLUSIONS: Remifentanil can be administered by TCI with acceptable bias and inaccuracy. The three pharmacokinetic parameter sets described by Egan and colleagues resulted in the least bias and best accuracy.

A comparison of the effects of propofol and midazolam on memory during two levels of sedation by using target-controlled infusion.

UNLABELLED: We examined memory during sedation with target-controlled infusions of propofol and midazolam in a double-blinded five-way, cross-over study in 10 volunteers. Each active drug infusion was targeted to sedation level 1 (asleep) and level 4 (lethargic) as determined with the Observer Assessment of Alertness/Sedation scale. At the target level of sedation, drug concentration was clamped for 30 min, during which time neutral words were presented. After 2 h, explicit memory was assessed by recall, and implicit memory by using a wordstem completion test. Venous drug concentrations (mean +/- SD) were 1350 ng/mL (+/-332 ng/mL) for propofol and 208 ng/mL (+/-112 ng/mL) for midazolam during Observer Assessment of Alertness/Sedation scale level 4; and 1620 ng/mL (+/-357 ng/mL) and 249 ng/mL (+/-82 ng/mL) respectively during level 1. The wordstem completion test frequencies at low level sedation were significantly higher than spontaneous frequencies (8.7% + 2.4%; P: < 0.05 in all cases), and lower than during placebo (33.6% + 23%) (P: < 0.05 in all cases, except P: = 0.076 for propofol at level 4). Clinically distinct levels of sedation were accompanied by small differences in venous propofol or midazolam concentrations. This indicates steep concentration-effect relationships. Neutral information is still memorized during low-level sedation with both drugs. The memory effect of propofol and midazolam did not differ significantly. IMPLICATIONS: Implicit memory can occur during different states of consciousness and might lead to psychological damage. In 10 volunteers, implicit memory was investigated during sedation with propofol and midazolam in a double-blinded, placebo-controlled study. To compare the effects of both drugs, they were titrated using a computer-controlled infusion system to produce similar high and low levels of sedation.

Target-controlled infusion systems: role in anaesthesia and analgesia.

Drug delivery by target-controlled infusion (TCI) allows automatic adjustments of the infusion rate of a drug to maintain a desired target concentration. Since drug effect is more closely related to blood concentration than to infusion rate, drug delivery via TCI is capable of creating stable blood concentrations of intravenous anaesthetics and analgesics. In this article the concept and history of TCI are described. The rational administration of TCI requires an appropriate pharmacokinetic data set and knowledge of the concentration-effect relationship; therefore, general pharmacokinetic and pharmacodynamic aspects of intravenous anaesthetics and analgesics are also addressed. Intraoperative investigations have demonstrated that TCI drug delivery allows rapid titration to a desired effect. The use of TCI for postoperative analgesia is still experimental, but TCI can, in part, overcome the disadvantages associated with continuous infusions and patient-controlled analgesia regimens in the postoperative period. Although TCI is capable of creating stable blood concentrations, when the target concentration is changed the resulting effect correlates better with a theoretical effect site concentration. The efficacy of TCI systems that can perform effect-site steering are still to be explored.

Target-controlled infusion of alfentanil for postoperative analgesia: contribution of plasma protein binding to intra-patient and inter-patient variability.

We have examined the influence of plasma protein binding on inter-individual and intra-individual variability in the effective postoperative analgesic concentration (EAC) of alfentanil and on the performance of the target-controlled infusion system used. Ten patients received standardized anaesthesia and target-controlled alfentanil for postoperative analgesia. Analgesia was assessed using a visual analogue scale (VAS). Plasma protein binding of alfentanil was assessed at four different times (on arrival in the recovery room, at 21:00 on the day of surgery and at 09:00 and 21:00 on the first postoperative day). Bias and inaccuracy were examined on the day of surgery and on the first postoperative day. Unbound fractions of alfentanil varied from 5 to 15% and varied in time. In general, the unbound fractions on the day of surgery were higher than those on the first postoperative day. Thirty-nine percent of inter-individual variability in the EAC of alfentanil (range 33-140 ng ml-1) at the onset of therapy could be explained by protein binding. At the other observation times, correlations between unbound fraction and EAC were only moderate. Bias on the day of surgery was -19% and 12% on the first postoperative day (ns). Inaccuracy was 23% and 18%, respectively (ns). We conclude that inter-individual variations in plasma protein binding can explain a significant portion of inter-individual variability in the EAC of alfentanil in the early postoperative phase.

Effect-site modelling of propofol using auditory evoked potentials.

Auditory evoked potentials (AEP) were used to monitor central nervous system effects during induction and recovery from anaesthesia produced by infusion of propofol 30 mg kg-1 h-1 in 22 healthy male patients. Non-parametric and parametric modelling techniques were used successfully to calculate the parameter keo which linked pharmacokinetic with pharmacodynamic aspects of drug action in only 15 of the study patients. In the non-parametric analysis, keo was found to have a mean value of 0.2 (range 0.1-0.36) min-1. Estimation of keo allowed calculation of the effect-site concentration (Ce50) associated with 50% of AEP effect for the population (2.08 micrograms ml-1; 95% confidence limits 1.7-2.45). There were no significant differences between keo values calculated by non-parametric and individual parametric modelling techniques. During recovery, 50% of patients demonstrated evidence of waking at an effect-site concentration of 2.28 micrograms ml-1.

Pharmacodynamic interaction of eltanolone and alfentanil during lower abdominal surgery in female patients.

We have studied the influence of eltanolone on intraoperative alfentanil requirements in 18 female patients undergoing lower abdominal surgery receiving target-controlled infusions of eltanolone and alfentanil. While target concentrations of eltanolone were maintained constant, target concentrations of alfentanil changed in response to the presence or absence of responses. With serum eltanolone concentrations increasing from 500 to 2000 ng ml-1, the EC50 of alfentanil for suppression of responses to surgical stimulation decreased from 233 to 9 ng ml-1. The findings suggest that the interaction between eltanolone and alfentanil is synergistic.

Pharmacokinetics of propofol during conscious sedation using target-controlled infusion in anxious patients undergoing dental treatment.

Infusion of propofol by a target-controlled infusion (TCI) system is effective in achieving conscious sedation for anxious patients presenting for dental surgery. It is a common clinical observation that anxious patients require more anaesthetic drugs than non-anxious individuals. In study 1 we have defined blood propofol concentrations necessary for conscious sedation in both anxious (n = 23) and non-anxious (n = 18) patients. The pump performance of the TCI system, using Gepts' pharmacokinetic model, was evaluated in these two patient groups. Subsequently, clearance of propofol was compared in the two groups. Mean measured venous serum propofol concentrations obtained between 20 and 35 min after the optimal sedation level was reached were 1.6 (SD 0.2) micrograms ml-1 in the anxious patients compared with 1.7 (0.3) micrograms ml-1 in the control group (study 1) and 1.4 (0.27) micrograms ml-1 in study 2. The pump systematically overpredicted measured propofol concentrations in both groups (study 1). There was no significant difference in propofol clearance between the two groups. In study 2, an optimized set of microconstants was derived which should more accurately predict the pharmacokinetic profile of the anxious population and this set was tested prospectively in another group of 12 anxious dental patients. Bias and precision with the optimized kinetic set were significantly less than the values obtained in study 1. We conclude that there was no significant pharmacokinetic differences between anxious and non-anxious subjects receiving subanaesthetic doses of propofol for conscious sedation.

Practical use of 'Diprifusor' systems.

This paper describes practical aspects relating to the safe use of target controlled infusion systems in anaesthesia. Consideration is given to the correct use of syringes and infusion lines for any target controlled infusion system. The importance of appropriate connections, minimising infusion line dead space and the avoidance of syphoning are emphasised. The first two commercially available infusion pumps to incorporate the 'Diprifusor' module for the administration of propofol by target controlled infusion are the Graseby 3500 (Graseby Medical Ltd) and the Vial Master TCI (Fresenius Vial SA, originally developed by Becton Dickinson). Particular features of these systems are discussed. Finally, the practical consequences of possible misuse of infusion systems incorporating pharmacokinetic models are considered.

Target-controlled infusion of alfentanil for postoperative analgesia: a feasibility study and pharmacodynamic evaluation in the early postoperative period.

We have examined the feasibility of target-controlled infusion of alfentanil (TCIA) and the pharmacodynamics of alfentanil in the early postoperative period. Patients were allocated randomly to one of the three groups to receive balanced anaesthesia with bolus injections of fentanyl (group F), sufentanil (group S) or alfentanil (group A). In the recovery room all patients received the same analgesic regimen, comprising TCIA. To evaluate the efficacy of postoperative analgesia, pain scores were measured on a visual analogue scale (VAS) and patients indicated a need for additional analgesia. EC50, the concentration at which, with a 50% probability, patients reported adequate analgesia, was estimated using logistic regression. Six patients did not complain of pain. The time from the last intraoperative bolus injection of opioid until patients complained of postoperative pain was shorter (P < 0.05) in group A (mean 68 min) than in group F (101 min) and group S (136 min). The time to onset of satisfactory analgesia was comparable in the three groups (median 18 min in group F, 15 min in group S and 14 min in group A). EC50 of alfentanil was determined in 28 patients; mean values were 26 ng ml-1 (group F), 39 ng ml-1 (group S) and 52 ng ml-1 (group A). We conclude that TCIA, under the conditions studied, resulted in a fast onset of adequate analgesia, irrespective of the opioid administered during operation. Also, there was no effect of opioids administered during operation on postoperative pharmacodynamics of alfentanil.