Free
Education  |   December 2002
Pharmacokinetics of Propofol Infusions in Critically Ill Neonates, Infants, and Children in an Intensive Care Unit
Author Affiliations & Notes
  • Ann E. Rigby-Jones, B.Sc.
    *
  • Judith A. Nolan, M.R.C.P., F.R.C.A.
  • Melanie J. Priston, Ph.D.
  • Peter M. C. Wright, M.D., Ph.D., F.C.A., R.C.S.I.
    §
  • J. Robert Sneyd, M.D., F.R.C.A.
  • Andrew R. Wolf, M.D., F.R.C.A.
    #
  • *Postgraduate Research Student, ‡Postdoctoral Research Scientist, Plymouth Postgraduate Medical School and Department of Pharmacy, Derriford Hospital. †Consultant Anaesthetist, Bristol Royal Children's Hospital. §Senior Lecturer, Department of Anaesthesia, University of Newcastle upon Tyne; current position: Professor of Anesthesia, University of California, San Francisco, California. ∥Professor of Anaesthesia and Associate Dean, Peninsula Medical School. #Professor, Department of Anaesthesia and Critical Care, University of Bristol, and Consultant Paediatric Cardiac Anaesthetist/Intensivist, Bristol Royal Children's Hospital.
  • Received from the Paediatric Intensive Care Unit, Bristol Royal Children's Hospital, Bristol, United Kingdom; the Department of Anaesthesia, University of Newcastle upon Tyne, Newcastle upon Tyne, United Kingdom; the Department of Pharmacy, Derriford Hospital, Plymouth, United Kingdom; the Plymouth Postgraduate Medical School, Plymouth, United Kingdom; and the Peninsula Medical School, Plymouth, United Kingdom.
Article Information
Education
Education   |   December 2002
Pharmacokinetics of Propofol Infusions in Critically Ill Neonates, Infants, and Children in an Intensive Care Unit
Anesthesiology 12 2002, Vol.97, 1393-1400. doi:
Anesthesiology 12 2002, Vol.97, 1393-1400. doi:
OPIOIDS, benzodiazepines, and chloral hydrate are commonly used for the sedation of critically ill children on the pediatric intensive care unit, but all have side effects, such as respiratory depression, delayed recovery from relative overdose, drug tolerance, and withdrawal phenomena. 1,2 Propofol has been used to provide smooth and predictable sedation in children, 3,4 but recently its use has been contraindicated because of concerns that its use may be associated with increased mortality 5 and that it can cause a syndrome characterized by bradycardia, rhabdomyolisis, metabolic acidosis, hypotension, and death. 6–8 While there are limited data on the kinetics of propofol in well children, 9–11 even less is known of the kinetics in critically ill neonates and infants. 12 We hypothesized that the pharmacokinetics of propofol in neonates and young babies might be substantially altered compared with older infants and children and that these differences may be clinically important. We therefore wished to accurately describe the pharmacokinetics of propofol when given as a sedative infusion to very young critically ill children, including those with low cardiac outputs. We also wished to relate these data to factors such as age, weight, gender, infusion duration, and clinical diagnosis.
Materials and Methods
After obtaining local ethics committee approval and written informed parental consent, we studied 21 neonates and children up to the age of 12 yr requiring sedation and ventilation following cardiac surgery or for single organ failure. Cardiac surgery patients were excluded from the study if prolonged postoperative ventilation or major inotropic support was anticipated. Sedation was achieved with an infusion of 2% propofol combined with a background infusion of morphine. The aim was to provide a constant morphine infusion rate while an individualized infusion rate of propofol was delivered to achieve target sedation scores. Sedation scoring was performed hourly using an observational pain scale 13 modified for intensive care (table 1), 14 with a range of scores from 0 to 8. Adequate sedation was considered to be a score of 2–4, which is consistent with the degree of sedation normally achieved in the pediatric intensive care unit.
Table 1. Sedation Score for Ventilated Non-Paralyzed Children
Image not available
Table 1. Sedation Score for Ventilated Non-Paralyzed Children
×
Patients undergoing cardiac surgery were anesthetized with isoflurane and fentanyl (50 μg/kg). Morphine sulfate (0.5–1 mg/kg) was added to the cardiomyotomy reservoir before commencing cardiopulmonary bypass, and isoflurane was administered via  the sweep gases. Propofol infusion commenced at 4 mg · kg−1· h−1without an initial bolus, after cardiopulmonary bypass had been discontinued or on returning to the pediatric intensive care unit. In all other children, propofol was introduced as an infusion, either after induction of anesthesia-sedation with another agent or to replace a previous sedative agent that had been considered unsatisfactory. Morphine was commenced at 20 μg · kg−1· h−1. Sedation was maintained with 4 mg · kg−1· h−1propofol and 20 μg · kg−1· h−1morphine if sedation scores remained within the target range. Undersedation was treated with a bolus injection of 20 μg/kg morphine and the morphine infusion rate was increased to 40 μg · kg−1· h−1if sedation remained unsatisfactory. Morphine boluses were also given prior to tracheal suctioning and physiotherapy. Propofol was increased to 5 mg · kg−1· h−1and then to a maximum of 6 mg · kg−1· h−1if target sedation scores were not achieved with 40 μg · kg−1· h−1morphine. No propofol boluses were given.
Arterial blood pressure was monitored invasively in all cases. Hypertension was treated in the same way as undersedation, or by adjusting inotropes, according to clinical impression. Hypotension was defined as a persistent reduction in mean blood pressure by more than 20%. If judged to be caused by poor cardiac performance (increased arterial lactate and decreased venous saturations), it was treated with appropriate inotropes. Initial support was with dopamine infused at 5–10 μg · kg−1· min−1with epinephrine as a second agent. Hypotension associated with low central venous pressure was treated with volume replacement, initially 20 ml/kg crystalloid. If hypotension was considered to result from oversedation, the rate of propofol infusion was reduced in increments of 1 mg · kg−1· h−1to a minimum of 2 mg · kg−1· h−1. Propofol was discontinued for weaning or if sedation was required for more than 24 h when it was replaced by another sedative agent. Triglyceride and cholesterol levels were determined before commencing and immediately on stopping the infusion. No child received parenteral nutrition during the study. Laboratory investigations were routinely performed for urea, electrolytes, liver function tests, lactate, and acid-base status before, during, and after the propofol infusion.
Propofol Analysis
Propofol was infused at constant rate for 4 h or more in each patient before the infusion was withdrawn. Once target sedation scores were achieved, arterial blood samples were obtained hourly. The purpose of these samples was to establish steady state blood propofol concentrations during optimal sedation, as well as to contribute to the pharmacokinetic model fitting. After the propofol infusion was withdrawn, arterial blood samples were taken immediately and at 5, 10, 15, and 30 min and at 1, 2, 3, 6, 12, and 24 h, and 48 h when possible. Blood samples were collected in oxalate tubes and stored at 4°C until analysis.
Propofol was extracted from whole blood using a solid phase extraction procedure 15 and analyzed by high-performance liquid chromatography. 16 The high-performance liquid chromatography assay was stability indicating and had proven linearity. Intraday precision was 6.3% and 11.8% at 100 and 1,000 ng/ml, respectively (n = 5). Interday precision at 100, 500, 1,000, and 2,000 ng/ml was less than 8% (n = 5). The limit of quantification was 2 ng/ml.
Pharmacokinetic Model
Mixed-effects population models were fitted to the propofol concentration data. The program NONMEM V was used, running on a SUN Enterprise computer with a Solaris operating system. 17 The mixed-effects approach defines a single basic model of typical values (population means) for the pharmacokinetic parameters. Variations in each individual from the basic model were defined by the use of a variable number of additional, user-defined “interindividual variability parameters,” each defining a degree of variability in one or more of the basic parameters. For instance, clearance was modeled as:MATHwhere Cl is the value for an individual, Cltypicalis the typical value for the population, and η is a normally distributed random variable with a mean zero. Both the basic model and the interindividual variability can also be wholly or partially modeled as functions of physiologic covariates, the aim being to reduce the residual degree of interindividual variability.
The basic parameters of the models used here were volume of the central compartment (V1), volume of the peripheral compartments (V2and V3), clearance (Cl, elimination clearance equal to V1·k  10) and distribution clearances (Q1equal to V1·k  12and Q2, equal to V1·k  13). Volume of distribution at steady state (Vss) was equal to V1plus V2plus V3. Models were fitted using NONMEM's first-order conditional estimates with the “centered” option. A model building approach was used, and improvements in three criteria were used to determine if additional parameters should be incorporated into the model. These criteria were goodness of fit (−2 log likelihood) evaluated against a chi-square distribution, determinable precision for all parameters, and visual acceptability.
We first tested models with two and three compartments. When these indicated that three compartments were justified, we subsequently used only three-compartment models. The population pharmacokinetic model was developed by adding interindividual variation parameters until no further model variation could be justified. Next, guided by visual plots, we evaluated models that permitted structural parameters (i.e.  , clearances and volumes) to differ with covariates. We systematically attempted to model each structural pharmacokinetic parameter as a simple or complex function of age or weight and as a function of gender or type of operation. The justification for each additional effect added to the model was for it to improve the goodness-of-fit statistic (−2 log likelihood) by more than 3.7 (evaluated against the chi-square distribution, this is equivalent to significance at the 0.05 level) and to result in a visual improvement in the goodness of fit. When all justified additional effects had been added to the model, the necessity for each was tested by removing it from the model and evaluating the resultant fit.
Simulations
To investigate our pharmacokinetic findings, simulations were performed using our optimal model. Concurrently, we performed simulations using the propofol pharmacokinetic model developed by Schuttler and Ihmsen, 18 to allow comparison of our model with a model developed from older healthy children and adults. To demonstrate the influence of weight in both models, and age in the Schuttler model, profiles were simulated for children of different weights and ages. The assignment of age to weight was based on our study population. We simulated 12-h infusions (our median propofol infusion duration) at a constant rate of 4 mg · kg−1· h−1.
Results
Twenty-one children were recruited to the study. Median age was 16 months (range, 1 week to 12 yr), and median weight was 8.9 kg (range, 3.1–33 kg). Details of the patient population and propofol delivery are shown in table 2. Duration of propofol infusion ranged from 4.5 to 28 h (median, 12 h). In three patients, propofol infusion was extended beyond 24 h because planned extubation was delayed and it was considered inappropriate to change to another sedative agent.
Table 2. Study Population
Image not available
Table 2. Study Population
×
Sedation scoring was performed in 20 children (1 child required paralysis, and sedation scores were not performed). Fifteen of these 20 completed the study with 20 μg · kg−1· h−1morphine, while 5 children (2 postoperative cardiac and 3 noncardiac) required dose increases. At 4 mg · kg−1· h−1propofol, target sedation scores were achieved in 17 of 20 children. Two of the 17 required a reduction in the infusion rate of propofol because of hypotension. No child had arrhythmias during the infusion. No neonate required more than 20 μg · kg−1· h−1morphine or 4 mg · kg−1· h−1propofol. Plasma concentrations of triglyceride and LDL and HDL cholesterol were unaffected by 2% propofol. Urea and electrolytes and liver function test results were not significantly different from baseline.
Pharmacokinetics
A three-compartment model with interindividual variation modeled in clearance, slow and fast distributional clearances, and V1was accepted. This model had a median prediction error of −1.5% and a median absolute prediction error of 29.7%. Some visual representations of the fit are in figure 1. An abbreviated summary of the model-building process is given in table 3. The optimal model (i.e.  , the one that fit the data best and in which no parameter could be removed without significantly worsening the fit) was one with three rather than two compartments. The structural parameters of the model (along with 95% confidence intervals for the “typical values” and the associated degree of interindividual variability) are shown in table 4. The structural parameters Cl, Q2, Q3, V1, and V2were all proportional to weight, while the largest of the three compartments (V3) was related to weight in a complex way with a constant component in addition to the weight-related component. In addition, children recovering from cardiac surgery had significantly reduced propofol clearance. Concentration-versus  -time profiles for a typical prediction and the most extreme underprediction and overprediction, respectively, are shown in figure 2.
Fig. 1. Plots allowing the evaluation of the optimal model to the data. (A  ) Observed Cb versus  population model-predicted Cb. (B  ) Observed Cb versus  individual model-predicted Cb. (C  ) Weighted residuals versus  time for the optimal population model. (D  ) Weighted residuals versus  time for the optimal individualized models. (E  ) Population model-predicted/observed Cb for each of the 21 subjects. (F  ) Individualized model predicted/observed Cb for each of the 21 subjects. Cb = Concentration of propofol in whole blood. The plots of the predicted versus  observed concentrations (population model [A  ] and individualized models [B  ]) demonstrate the overall goodness of fit. A plot of the weighted residuals (or SD units) versus  time (population model [C  ] and individualized models [D  ]) will show whether the pattern of the residuals is dependent on time. Plots of the population model-predicted Cb (E  ) or individualized model-predicted Cb (F  ) versus  observed Cb for each subject will demonstrate if any individual data set is an outlier.
Fig. 1. Plots allowing the evaluation of the optimal model to the data. (A 
	) Observed Cb versus 
	population model-predicted Cb. (B 
	) Observed Cb versus 
	individual model-predicted Cb. (C 
	) Weighted residuals versus 
	time for the optimal population model. (D 
	) Weighted residuals versus 
	time for the optimal individualized models. (E 
	) Population model-predicted/observed Cb for each of the 21 subjects. (F 
	) Individualized model predicted/observed Cb for each of the 21 subjects. Cb = Concentration of propofol in whole blood. The plots of the predicted versus 
	observed concentrations (population model [A 
	] and individualized models [B 
	]) demonstrate the overall goodness of fit. A plot of the weighted residuals (or SD units) versus 
	time (population model [C 
	] and individualized models [D 
	]) will show whether the pattern of the residuals is dependent on time. Plots of the population model-predicted Cb (E 
	) or individualized model-predicted Cb (F 
	) versus 
	observed Cb for each subject will demonstrate if any individual data set is an outlier.
Fig. 1. Plots allowing the evaluation of the optimal model to the data. (A  ) Observed Cb versus  population model-predicted Cb. (B  ) Observed Cb versus  individual model-predicted Cb. (C  ) Weighted residuals versus  time for the optimal population model. (D  ) Weighted residuals versus  time for the optimal individualized models. (E  ) Population model-predicted/observed Cb for each of the 21 subjects. (F  ) Individualized model predicted/observed Cb for each of the 21 subjects. Cb = Concentration of propofol in whole blood. The plots of the predicted versus  observed concentrations (population model [A  ] and individualized models [B  ]) demonstrate the overall goodness of fit. A plot of the weighted residuals (or SD units) versus  time (population model [C  ] and individualized models [D  ]) will show whether the pattern of the residuals is dependent on time. Plots of the population model-predicted Cb (E  ) or individualized model-predicted Cb (F  ) versus  observed Cb for each subject will demonstrate if any individual data set is an outlier.
×
Table 3. Abbreviated Summary of the Model Building Process
Image not available
Table 3. Abbreviated Summary of the Model Building Process
×
Table 4. Magnitude of Parameters for the Optimal Model
Image not available
Table 4. Magnitude of Parameters for the Optimal Model
×
Fig. 2. Examples of model fits to three individuals’ propofol concentration data. The dashed line represents the prediction of the “typical value” model. The solid line is the individualized model. (Top  ) Typical fit. (Middle  ) The most extreme overprediction. (Bottom  ) The most extreme underprediction.
Fig. 2. Examples of model fits to three individuals’ propofol concentration data. The dashed line represents the prediction of the “typical value” model. The solid line is the individualized model. (Top 
	) Typical fit. (Middle 
	) The most extreme overprediction. (Bottom 
	) The most extreme underprediction.
Fig. 2. Examples of model fits to three individuals’ propofol concentration data. The dashed line represents the prediction of the “typical value” model. The solid line is the individualized model. (Top  ) Typical fit. (Middle  ) The most extreme overprediction. (Bottom  ) The most extreme underprediction.
×
Many combinations of covariate interactions were examined during the course of the model-building process, in particular weight, age, gender, type of operation, and duration of propofol infusion. Age, gender, and duration of propofol infusion were not supported as covariates. The intercompartmental rate constants were calculated from the typical clearance and volume values for cardiac surgery patients. The intercompartmental rate constants were used to construct the context sensitive half-time profiles (time required for a 50% decrement in the blood propofol concentration as a function of infusion duration) for children of different weights, using the computer software package RECOV (fig. 3). RECOV was developed by Steven L. Shafer, MD (Department of Anesthesia, Stanford University, Palo Alto, CA), and is freely available ().
Fig. 3. Context-sensitive half-time for post-cardiac surgery children of different weights.
Fig. 3. Context-sensitive half-time for post-cardiac surgery children of different weights.
Fig. 3. Context-sensitive half-time for post-cardiac surgery children of different weights.
×
Simulations
The concentration-versus  -time profiles for children of different weights (and ages) simulated using our final pharmacokinetic model and the pharmacokinetic model developed by Schuttler and Ihmsen 18 are shown in figure 4. Compared with our data, the Schuttler model significantly underpredicts the propofol blood concentration resulting from a 12-h infusion at 4 mg · kg−1· h−1, administered to critically ill children after cardiac surgery.
Fig. 4. Pharmacokinetic simulations. (Left  ) Simulated concentration-versus  -time profiles resulting from 12-h propofol infusions (4 mg · kg−1· h−1) administered to children of different weights undergoing cardiac surgery using the pharmacokinetic parameters determined in this study. (Right  ) Simulated concentration-versus  -time profiles resulting from 12-h propofol infusions (4 mg · kg−1· h−1) administered to children of different weights and ages, using the pharmacokinetic parameter estimates reported by Schuttler and Ihmsen. 18 
Fig. 4. Pharmacokinetic simulations. (Left 
	) Simulated concentration-versus 
	-time profiles resulting from 12-h propofol infusions (4 mg · kg−1· h−1) administered to children of different weights undergoing cardiac surgery using the pharmacokinetic parameters determined in this study. (Right 
	) Simulated concentration-versus 
	-time profiles resulting from 12-h propofol infusions (4 mg · kg−1· h−1) administered to children of different weights and ages, using the pharmacokinetic parameter estimates reported by Schuttler and Ihmsen. 18
Fig. 4. Pharmacokinetic simulations. (Left  ) Simulated concentration-versus  -time profiles resulting from 12-h propofol infusions (4 mg · kg−1· h−1) administered to children of different weights undergoing cardiac surgery using the pharmacokinetic parameters determined in this study. (Right  ) Simulated concentration-versus  -time profiles resulting from 12-h propofol infusions (4 mg · kg−1· h−1) administered to children of different weights and ages, using the pharmacokinetic parameter estimates reported by Schuttler and Ihmsen. 18 
×
Complications
One child developed persistent hypotension and metabolic acidosis after 5 h of propofol infusion at a constant infusion rate of 4 mg · kg−1· h−1. This child had a mitral valve atresia and total anomalous pulmonary venous drainage and had undergone a Fontan procedure. Metabolic acidosis was apparent at the start of the propofol infusion and persisted after propofol was discontinued. It was a clinical decision to discontinue the propofol, and it was considered that the acidosis was related primarily to poor cardiac output. Midazolam was used as a replacement infusion. The hypotension and acidosis responded over the following 8 h to intravenous fluid and vasoconstrictors. There were no arrhythmias, and the child did not develop bradycardia. The blood steady state concentrations of propofol in this patient were similar to those of the other patients in the study, and the elimination curve was unremarkable. Triglyceride concentrations were normal. The lowest blood pressure recorded for this child was 70/45 mmHg, and pulse rate ranged from 150 to 165 beats/min. No other patients developed an acidosis. No other major complications were observed using propofol in this series.
Discussion
Our model demonstrates that critically ill children and infants have a pharmacokinetic profile for propofol that is broadly similar to previously reported studies in well adults and children 9–11,19–22 and in critically ill children. 12 However, we found altered kinetics in very small babies and in children recovering from cardiac surgery.
In neonates, our model indicated proportionately increased distribution of propofol into slowly equilibrating tissues compared with older children. This is evidenced by the large constant component of V3. This will have more significance in smaller children as the volume of the deep compartment becomes proportionally larger as body weight decreases. The redistribution rate constant from this compartment, k  31, is also highly weight dependent with smaller children having a slower rate of drug movement out of V3than larger children. In neonates and infants, the combination of a large slow peripheral compartment and slow redistribution rate has relevant effects late after the discontinuation of the infusion in that residual concentrations of propofol are detectable for longer. However, the clinically relevant early context sensitive decrease in blood propofol concentration (context sensitive half-time) is shorter in smaller children after prolonged infusion. The proportionally larger deep compartment allows drug distribution from the central compartment to occur rapidly even after prolonged infusion. The kinetic model therefore indicates that when a propofol infusion is stopped in a young infant, the initial decrease in blood concentration is more rapid, while the later decline is slower than in an older child. This fits with our clinical impression that neonates can emerge from sedation infusions rapidly, but full recovery can be considerably delayed.
Our typical parameter estimates, with the exception of V3and K  31, are within the ranges reported by Reed and colleagues. 12 Reed et al.  reported a median V3/V1ratio of 20 and a median V3/V2ratio of 11 for children aged 0.02 to 3.2 yr (personal communication, Michael D. Reed, Pharm.D., Professor of Pediatrics, School of Medicine, Case Western Reserve University, Cleveland, OH, March 2002). This is similar to our parameter estimations in larger children. In 10-kg and 15-kg children (corresponding to children aged 1–4 yr in our study), our V3/V1ratios are 27 and 21, respectively, while our V3/V2ratios are 12 and 9. However, the relative increase in peripheral distribution in smaller babies is demonstrated by our volume ratios for a 3-kg baby, where V3/V1is approximately 68 and V2/V1is 29. The major difference between our study protocol and that performed by Reed and colleagues 12 is that our patients received concomitant morphine infusions, while those of Reed et al.  received ketorolac. The interactions between propofol and the synthetic opioids are well documented, 23 and it has been reported that alfentanil reduces propofol elimination clearance and increases the deep volume of distribution. 24 The interaction between propofol and morphine is less clear, but it is possible that the administration of morphine may have influenced the pharmacokinetics of propofol in our study and may have contributed to our increased apparent volume of distribution. We did not quantify morphine blood concentrations, and morphine administration was not evaluated as a model covariate.
The increased peripheral drug distribution in the smaller babies may be explained by their altered body composition. Total body water, extracellular fluid, and blood volume are considerably larger in neonates and young infants than in older children, when expressed as a percentage of total body weight. Also, reduced plasma protein binding caused by the state of critical illness can have the effect of increasing the apparent distribution volume because more free drug is available for tissue binding. 25 In previous pharmacokinetic studies in children, propofol has been administered as a single bolus 10,11,19 or as a short infusion. 9,12 The duration of propofol infusion in this study (up to 28 h) was significantly longer, and this will have aided our ability to fully characterize late propofol pharmacokinetics.
As with other studies of propofol pharmacokinetics in children, 12,19 age was not found to be a significant covariate for our model. The association of weight but not age as a covariate in the model was interesting. The infants and children recruited for the study were not from a normal population. Specifically, some of the infants who underwent cardiac surgery were below the 10th centile for weight compared with age. Hence, there was little correlation of age with weight. The pharmacokinetic analysis of propofol in children by Kataria and colleagues 9 found age to be a statistically significant covariate on V2but was not thought to be clinically relevant as the actual improvement to the model was very small.
The propofol pharmacokinetic model developed by Schuttler and Ihmsen 18 was based on data from healthy children and adults aged 2–88 yr. Age and weight were included as model covariates. Our simulations of propofol infusions administered to children of different weights and ages describe the differences between our pharmacokinetic parameter values (clearance based on cardiac surgery patients) and those derived by the analysis of Schuttler and Ihmsen (fig. 4). The parameter estimates of Schuttler and Ihmsen demonstrate increased metabolic and distributional clearance, particularly in the smaller babies. This results in an underprediction of the propofol blood concentration compared with the simulations produced using our model. Our simulated age-weight relations were based on the very underweight children seen in our study, and it is therefore not surprising that these simulations demonstrate significant kinetic differences between the two models.
Elimination of morphine is prolonged in children after cardiac surgery. 26 This is in keeping with our findings of propofol pharmacokinetics on the pediatric intensive care unit. Our optimal pharmacokinetic model also indicates that patients undergoing cardiac surgery had reduced values for metabolic clearance. Mild liver impairment is common following cardiopulmonary bypass in children and may continue into the postoperative period. 27 This could effect the hepatic clearance of propofol. Cardiac surgery patients also demonstrate reduced cardiac output, which may affect propofol elimination. As postcardiac surgery patients provided the majority of our data, this may potentially limit of the applicability of our kinetic parameters to noncardiac surgery, critically ill pediatric patients. However, despite the low number of nonsurgical patients in our study, we were able to detect a statistically significant effect of surgery on clearance.
Concerns about propofol infusion syndrome in children have limited the use of this drug in intensive care, and it is now contraindicated in both the United States and United Kingdom for sedation of children younger than 16 yr. Our study in this patient group demonstrates that the pharmacokinetics, although different, did not result in excessively high blood concentrations of propofol. Current data seem to indicate that the cause of propofol infusion syndrome is an inhibition of mitochondrial function leading to an increase in short and medium chain fatty acids. 6,8,28 This study was completed within the guidelines recommended at the time for propofol infusion in children, and we saw no indications of propofol infusion syndrome in this series. Our data set included a neonate following a “switch” procedure for transposition of the great arteries, a Blalock Taussig shunt, a repair of Fallots tetralogy, and a Fontan procedure. The results from this study showed that it was feasible to use short-term propofol infusions for the critically ill child and neonate. However, because of the results of a recent clinical trial (unpublished) 5 that demonstrated significantly higher mortality in children sedated with propofol compared with other sedative agents, propofol has now been withdrawn from use as a sedative agent in critically ill children aged 16 yr or younger. Whether it still should continue to be used as an anesthetic infusion in children who require a brief period of additional anesthesia in the critical care unit after surgery remains debatable.
References
Marx CM, Smith PG, Lowrie LH, Hamlett KW, Ambuel B, Yamashita TS, Blumer JL: Optimal sedation of mechanically ventilated pediatric critical care patients. Crit Care Med 1994; 22: 163–70Marx, CM Smith, PG Lowrie, LH Hamlett, KW Ambuel, B Yamashita, TS Blumer, JL
Keyes WG, Holbrook PR: Appropriate use of analgesics, sedatives and neuromuscular blocking agents in the pediatric intensive care unit. Curr Opin Pediatr 1990; 2: 538–42Keyes, WG Holbrook, PR
Newman LH, McDonald JC, Wallace PG, Ledingham IM: Propofol infusion for sedation in intensive care. Anaesthesia 1987; 42: 929–37Newman, LH McDonald, JC Wallace, PG Ledingham, IM
Norreslet J, Wahlgreen C: Propofol infusion for sedation of children. Crit Care Med 1990; 18: 890–2Norreslet, J Wahlgreen, C
Propofol (Diprivan) infusion: Sedation in children aged 16 years or younger contraindicated. Curr Problems Pharmacovigilance 2001; 27: 10NA,
Cray SH, Robinson BH, Cox PN: Lactic acidemia and bradyarrhythmia in a child sedated with propofol. Crit Care Med 1998; 26: 2087–92Cray, SH Robinson, BH Cox, PN
Mehta N, DeMunter C, Habibi P, Nadel S, Britto J: Short-term propofol infusions in children. Lancet 1999; 354: 866–7Mehta, N DeMunter, C Habibi, P Nadel, S Britto, J
Wolf A, Weir P, Segar P, Stone J, Shield J: Impaired fatty acid oxidation in propofol infusion syndrome. Lancet 2001; 357: 606–7Wolf, A Weir, P Segar, P Stone, J Shield, J
Kataria BK, Ved SA, Nicodemus HF, Hoy GR, Lea D, Dubois MY, Mandema JW, Shafer SL: The pharmacokinetics of propofol in children using three different data analysis approaches. A nesthesiology 1994; 80: 104–22Kataria, BK Ved, SA Nicodemus, HF Hoy, GR Lea, D Dubois, MY Mandema, JW Shafer, SL
Saint-Maurice C, Cockshott ID, Douglas EJ, Richard MO, Harmey JL: Pharmacokinetics of propofol in young children after a single dose. Br J Anaesth 1989; 63: 667–70Saint-Maurice, C Cockshott, ID Douglas, EJ Richard, MO Harmey, JL
Jones RD, Chan K, Andrew LJ: Pharmacokinetics of propofol in children. Br J Anaesth 1990; 65: 661–7Jones, RD Chan, K Andrew, LJ
Reed MD, Yamashita TS, Marx CM, Myers CM, Blumer JL: A pharmacokinetically based propofol dosing strategy for sedation of the critically ill, mechanically ventilated pediatric patient. Crit Care Med 1996; 24: 1473–81Reed, MD Yamashita, TS Marx, CM Myers, CM Blumer, JL
Hannallah RS, Broadman LM, Belman AB, Abramowitz MD, Epstein BS: Comparison of caudal and ilioinguinal/iliohypogastric nerve blocks for control of post-orchiopexy pain in pediatric ambulatory surgery. A nesthesiology 1987; 66: 832–4Hannallah, RS Broadman, LM Belman, AB Abramowitz, MD Epstein, BS
Wolf AR: Analgesia in the neonate, Textbook of Meeting Abstracts, 3rd edition. Edited by Rennie JM, Roberton NRC. Edinburgh, Churchill Livingstone, 1999, pp 435–41
Ragno G, Cicinelli E, Schonauer S, Vetuschi C: Propofol assay in biological fluids in pregnant women. J Pharm Biomed Anal 1997; 15: 1633–40Ragno, G Cicinelli, E Schonauer, S Vetuschi, C
Yeganeh MH, Ramzan I: Determination of propofol in rat whole blood and plasma by high-performance liquid chromatography. J Chromatogr B Biomed Sci Appl 1997; 691: 478–82Yeganeh, MH Ramzan, I
Boeckmann AJ, Beal SL, Sheiner LB: NONMEM Users Guide. San Francisco, University of California-San Francisco, 1999
Schuttler J, Ihmsen H: Population pharmacokinetics of propofol: A multicenter study. A nesthesiology 2000; 92: 727–38Schuttler, J Ihmsen, H
Murat I, Billard V, Vernois J, Zaouter M, Marsol P, Souron R, Farinotti R: Pharmacokinetics of propofol after a single dose in children aged 1–3 years with minor burns: Comparison of three data analysis approaches. A nesthesiology 1996; 84: 526–32Murat, I Billard, V Vernois, J Zaouter, M Marsol, P Souron, R Farinotti, R
Cockshott ID, Briggs LP, Douglas EJ, White M: Pharmacokinetics of propofol in female patients: Studies using single bolus injections. Br J Anaesth 1987; 59: 1103–10Cockshott, ID Briggs, LP Douglas, EJ White, M
Shafer A, Doze VA, Shafer SL, White PF: Pharmacokinetics and pharmacodynamics of propofol infusions during general anesthesia. A nesthesiology 1988; 69: 348–56Shafer, A Doze, VA Shafer, SL White, PF
Morgan DJ, Campbell GA, Crankshaw DP: Pharmacokinetics of propofol when given by intravenous infusion. Br J Clin Pharmacol 1990; 30: 144–8Morgan, DJ Campbell, GA Crankshaw, DP
Vuyk J: Pharmacokinetic and pharmacodynamic interactions between opioids and propofol. J Clin Anesth 1997; 9: 23S–6SVuyk, J
Vuyk J: Clinical interpretation of pharmacokinetic and pharmacodynamic propofol-opioid interactions. Acta Anaesthesiol Belg 2001; 52: 445–51Vuyk, J
Bodenham A, Shelly MP, Park GR: The altered pharmacokinetics and pharmacodynamics of drugs commonly used in critically ill patients. Clin Pharmacokinet 1988; 14: 347–73Bodenham, A Shelly, MP Park, GR
Lynn A, Nespeca MK, Bratton SL, Strauss SG, Shen DD: Clearance of morphine in postoperative infants during intravenous infusion: The influence of age and surgery. Anesth Analg 1998; 86: 958–63Lynn, A Nespeca, MK Bratton, SL Strauss, SG Shen, DD
Mitchell IM: Liver function after cardiopulmonary bypass in children. J Thorac Cardiovasc Surg 1995; 110: 284–6Mitchell, IM
Schenkman KA, Yan S: Propofol impairment of mitochondrial respiration in isolated perfused guinea pig hearts determined by reflectance spectroscopy. Crit Care Med 2000; 28: 172–7Schenkman, KA Yan, S
Fig. 1. Plots allowing the evaluation of the optimal model to the data. (A  ) Observed Cb versus  population model-predicted Cb. (B  ) Observed Cb versus  individual model-predicted Cb. (C  ) Weighted residuals versus  time for the optimal population model. (D  ) Weighted residuals versus  time for the optimal individualized models. (E  ) Population model-predicted/observed Cb for each of the 21 subjects. (F  ) Individualized model predicted/observed Cb for each of the 21 subjects. Cb = Concentration of propofol in whole blood. The plots of the predicted versus  observed concentrations (population model [A  ] and individualized models [B  ]) demonstrate the overall goodness of fit. A plot of the weighted residuals (or SD units) versus  time (population model [C  ] and individualized models [D  ]) will show whether the pattern of the residuals is dependent on time. Plots of the population model-predicted Cb (E  ) or individualized model-predicted Cb (F  ) versus  observed Cb for each subject will demonstrate if any individual data set is an outlier.
Fig. 1. Plots allowing the evaluation of the optimal model to the data. (A 
	) Observed Cb versus 
	population model-predicted Cb. (B 
	) Observed Cb versus 
	individual model-predicted Cb. (C 
	) Weighted residuals versus 
	time for the optimal population model. (D 
	) Weighted residuals versus 
	time for the optimal individualized models. (E 
	) Population model-predicted/observed Cb for each of the 21 subjects. (F 
	) Individualized model predicted/observed Cb for each of the 21 subjects. Cb = Concentration of propofol in whole blood. The plots of the predicted versus 
	observed concentrations (population model [A 
	] and individualized models [B 
	]) demonstrate the overall goodness of fit. A plot of the weighted residuals (or SD units) versus 
	time (population model [C 
	] and individualized models [D 
	]) will show whether the pattern of the residuals is dependent on time. Plots of the population model-predicted Cb (E 
	) or individualized model-predicted Cb (F 
	) versus 
	observed Cb for each subject will demonstrate if any individual data set is an outlier.
Fig. 1. Plots allowing the evaluation of the optimal model to the data. (A  ) Observed Cb versus  population model-predicted Cb. (B  ) Observed Cb versus  individual model-predicted Cb. (C  ) Weighted residuals versus  time for the optimal population model. (D  ) Weighted residuals versus  time for the optimal individualized models. (E  ) Population model-predicted/observed Cb for each of the 21 subjects. (F  ) Individualized model predicted/observed Cb for each of the 21 subjects. Cb = Concentration of propofol in whole blood. The plots of the predicted versus  observed concentrations (population model [A  ] and individualized models [B  ]) demonstrate the overall goodness of fit. A plot of the weighted residuals (or SD units) versus  time (population model [C  ] and individualized models [D  ]) will show whether the pattern of the residuals is dependent on time. Plots of the population model-predicted Cb (E  ) or individualized model-predicted Cb (F  ) versus  observed Cb for each subject will demonstrate if any individual data set is an outlier.
×
Fig. 2. Examples of model fits to three individuals’ propofol concentration data. The dashed line represents the prediction of the “typical value” model. The solid line is the individualized model. (Top  ) Typical fit. (Middle  ) The most extreme overprediction. (Bottom  ) The most extreme underprediction.
Fig. 2. Examples of model fits to three individuals’ propofol concentration data. The dashed line represents the prediction of the “typical value” model. The solid line is the individualized model. (Top 
	) Typical fit. (Middle 
	) The most extreme overprediction. (Bottom 
	) The most extreme underprediction.
Fig. 2. Examples of model fits to three individuals’ propofol concentration data. The dashed line represents the prediction of the “typical value” model. The solid line is the individualized model. (Top  ) Typical fit. (Middle  ) The most extreme overprediction. (Bottom  ) The most extreme underprediction.
×
Fig. 3. Context-sensitive half-time for post-cardiac surgery children of different weights.
Fig. 3. Context-sensitive half-time for post-cardiac surgery children of different weights.
Fig. 3. Context-sensitive half-time for post-cardiac surgery children of different weights.
×
Fig. 4. Pharmacokinetic simulations. (Left  ) Simulated concentration-versus  -time profiles resulting from 12-h propofol infusions (4 mg · kg−1· h−1) administered to children of different weights undergoing cardiac surgery using the pharmacokinetic parameters determined in this study. (Right  ) Simulated concentration-versus  -time profiles resulting from 12-h propofol infusions (4 mg · kg−1· h−1) administered to children of different weights and ages, using the pharmacokinetic parameter estimates reported by Schuttler and Ihmsen. 18 
Fig. 4. Pharmacokinetic simulations. (Left 
	) Simulated concentration-versus 
	-time profiles resulting from 12-h propofol infusions (4 mg · kg−1· h−1) administered to children of different weights undergoing cardiac surgery using the pharmacokinetic parameters determined in this study. (Right 
	) Simulated concentration-versus 
	-time profiles resulting from 12-h propofol infusions (4 mg · kg−1· h−1) administered to children of different weights and ages, using the pharmacokinetic parameter estimates reported by Schuttler and Ihmsen. 18
Fig. 4. Pharmacokinetic simulations. (Left  ) Simulated concentration-versus  -time profiles resulting from 12-h propofol infusions (4 mg · kg−1· h−1) administered to children of different weights undergoing cardiac surgery using the pharmacokinetic parameters determined in this study. (Right  ) Simulated concentration-versus  -time profiles resulting from 12-h propofol infusions (4 mg · kg−1· h−1) administered to children of different weights and ages, using the pharmacokinetic parameter estimates reported by Schuttler and Ihmsen. 18 
×
Table 1. Sedation Score for Ventilated Non-Paralyzed Children
Image not available
Table 1. Sedation Score for Ventilated Non-Paralyzed Children
×
Table 2. Study Population
Image not available
Table 2. Study Population
×
Table 3. Abbreviated Summary of the Model Building Process
Image not available
Table 3. Abbreviated Summary of the Model Building Process
×
Table 4. Magnitude of Parameters for the Optimal Model
Image not available
Table 4. Magnitude of Parameters for the Optimal Model
×