The Effective Effect-Site Propofol Concentration for Induction and Intubation with Two Pharmacokinetic Models in Morbidly Obese Patients Using Total Body Weight

Echevarría, Ghislaine C. MD, MSc; Elgueta, María F. MD; Donoso, María T. MD; Bugedo, Diego A. MD; Cortínez, Luis I. MD; Muñoz, Hernán R. MD, MSc

Anesthesia & Analgesia:
doi: 10.1213/ANE.0b013e31825d6254
Anesthetic Pharmacology: Research Reports

BACKGROUND: Most pharmacokinetic (PK) models used for propofol administration are based on studies in normal-weight patients. Extrapolation of these models for morbidly obese patients is controversial. Using 2 PK models and a target-controlled infusion system, we determined the predicted propofol effect-site concentration (Ce) needed for induction of anesthesia in morbidly obese subjects using total body weight.

METHODS: Sixty-six morbidly obese subjects from 18 to 50 years of age were randomized to receive propofol to reach and maintain a predetermined propofol Ce, based on the PK models of either Marsh or Schnider. All patients were monitored with a Bispectral Index electroencephalographic monitor. Fentanyl 3 μg/kg total body weight was administered before starting the propofol infusion. After loss of consciousness, vecuronium was administered to facilitate endotracheal intubation. Groups of 6 patients each received propofol at a different, predetermined target propofol Ce. An “effective Ce” (ECe) was defined as the propofol Ce that provided adequate hypnosis (Bispectral Index <60) during the complete induction period (45 seconds after reaching the predetermined target Ce until 5 minutes after tracheal intubation). Heart rate and arterial blood pressure were measured every 1 minute throughout the study period. Probit regression analysis was performed to calculate the effective propofol Ce values to induce hypnosis in 50% (ECe50) and 95% (ECe95) of patients with 95% confidence intervals (CIs).

RESULTS: Patient characteristics were similar between models and across the propofol target concentration groups. The ECe50 of propofol was 3.4 μg/mL (95% CI: 2.9, 3.7 μg/mL) with the Marsh model and 4.5 μg/mL (95% CI: 4.1, 4.8 μg/mL) with the Schnider model (P < 0.001). The ECe95 values were 4.2 μg/mL (95% CI: 3.8, 6.2 μg/mL) and 5.5 μg/mL (95% CI: 5.0, 7.2 μg/mL) with Marsh and Schnider models, respectively. At the ECe95, hemodynamic effects were similar with the 2 PK models.

CONCLUSION: Different propofol target concentrations for each PK model must be used for induction when using total body weight in morbidly obese patients.

In Brief

Published ahead of print September 5, 2012

Author Information

From the Departamento de Anestesiología, Escuela de Medicina, Pontificia Universidad Católica de Chile, Santiago, Chile.

Supported by departmental funding.

The authors declare no conflicts of interest.

Reprints will not be available from the authors.

Address correspondence to Hernán R. Muñoz, MD, MSc, Departmento de Anestesiología, Hospital Clínico U.C., Marcoleta 367, PO Box 114-D, Santiago, Chile. Address e-mail to

Accepted April 30, 2012

Published ahead of print September 5, 2012

Article Outline

The prevalence of morbidly obese (MO) patients (body mass index [BMI] ≥40 kg/m2) undergoing surgery is increasing in both developed and developing countries.1,2

Propofol, a short-acting IV anesthetic drug, seems an alternative for providing anesthesia in this population because of its rapid recovery profile.3,4 This anesthetic drug can be administered using target-controlled infusion (TCI) systems that allow achieving and maintaining predetermined predicted propofol concentrations, either in the plasma (Cp) or at the site of effect (Ce).

TCI systems, however, require an accurate PK model to perform in a clinically acceptable manner in any given population. In adults, the most widely used PK models for propofol are those developed by Marsh et al.5 and Schnider et al.6 These models were developed in normal-weight individuals, and their use in MO patients (as the weight to be entered into the TCI system) is controversial. Several alternatives, such as the use of total body weight (TBW), lean body weight (LBW), ideal body weight, or Servin's weight correction formula7 have been suggested for MO patients.810 During maintenance of general anesthesia, however, and at least for the Marsh model, the performance of TCI may be improved by the use of TBW in MO patients.7,9,11

For the induction of anesthesia with the TCI device, the situation is more complicated. Central volume (V1) is a critical determinant of the initial loading dose of propofol and this parameter is very different between the 2 PK models. In Marsh's model, the central volume is proportional to weight,5 approaching 22.8 L in a patient of 100 kg. In Schnider's model, the central volume is 4.27 L for all patients.6 Thus, major differences in the initial bolus dose of propofol to reach the same predicted Cp or Ce are observed with these models, particularly in MO patients when TBW is used.9

Because each patient needs a unique amount of drug to reach a given propofol Ce (i.e., the administration of a different mass of drug should result in a different Ce), this suggests that one of the models (or both) may perform poorly during induction of anesthesia. In addition, this difference in central volume between models is potentially risky for patients. For example, one concern with the Marsh's model in MO subjects is the adverse hemodynamic consequences of a potentially excessive loading dose of propofol.9 However, the use of a fixed central volume with the Schnider model might result in inadequate dosing to induce anesthesia in MO individuals.

If using TBW is a good alternative for maintenance of anesthesia in MO patients, one approach to circumvent this problem and reduce adverse effects during induction is to select an appropriate and different target propofol concentration with each of these 2 PK models. Our hypothesis is that the appropriate target propofol Ce for the Marsh model is lower than that for the Schnider model.

The aim of this study was to determine the target propofol Ce needed to provide adequate hypnosis for induction and tracheal intubation for each PK model using TBW in MO patients. In addition, we wanted to define the hemodynamic adverse effects associated with these propofol target concentrations.

Back to Top | Article Outline


After Institutional Ethics Committee approval (School of Medicine, Pontificia Universidad Católica de Chile, Santiago, Chile) and written informed consent, 66 MO patients (BMI ≥40 kg/m2), aged 18 to 50 years and ASA physical status III, scheduled for elective bariatric surgery were studied. Exclusion criteria included suspect of difficult airway management, drug or alcohol abuse, chronic or acute (within 48 hours before surgery) intake of any drug acting on the central nervous system and any known adverse effect to the study drug. Patients fasted for at least 8 hours before surgery and no premedication was given.

In the operating room, standard monitoring was placed, including electrocardiogram, noninvasive arterial blood pressure, and pulse oximetry. Bispectral Index (BIS) values (A2000 BIS® XP monitor, version 3.2; Aspect Medical Systems Inc., Newton, MA) were obtained and recorded. The BIS smoothing time period was set at 15 seconds.

Using a computer-generated random-number sequence, patients were allocated to 1 of 8 groups of predefined propofol target Ce for induction, namely, 3.5, 4.0, 4.5, or 5.0 μg/mL, as predicted by either Marsh's or Schnider's PK model based on TBW (i.e., 8 groups of 6 patients each). The TCI system consisted of a Pilot anesthesia 2-syringe infusion pump (Fresenius Vial S.A., Brézins, France) controlled by a Toshiba Satellite 2450-SP300 laptop computer using the AnestFusor® Serie II Pro a software (School of Medicine, Universidad de Chile) and allowing a maximum infusion rate of 1500 mL/h. The Marsh model was implemented with a keo of 1.2 min−1,12 and the Schnider model with a keo of 0.456 min−1.13 These keo values were chosen for the respective PK models because they both predict a time to peak effect of approximately 1.6 minutes after a bolus dose.9 Constraints to the weight entered to avoid the paradoxical decrease in LBW that occurs with the James formula used in the Schnider model8,9 were not implemented.

After the administration of 5 mL/kg lactated Ringer solution in 5 to 10 minutes and oxygen administration, anesthesia was induced with fentanyl 3 μg/kg TBW. Three minutes later, propofol was administered by the TCI device to achieve the predetermined target propofol Ce. Loss of consciousness was defined by the disappearance of the eyelash reflex that was assessed every 5 seconds after patients spontaneously closed their eyes. During this period, no stimulus other than a facemask supplying O2 100% close to the nose was allowed. After loss of consciousness, vecuronium 0.1 mg/kg TBW was administered and tracheal intubation was performed 3 minutes later only if the BIS value was <60 and target propofol Ce had been reached. After intubation, patients were connected to mechanical ventilation adjusted to obtain an end-tidal CO2 of 30 to 35 mm Hg and the propofol target concentration was reduced to 2.5 μg/mL. Anesthesiologists performing loss of consciousness assessment and intubation were blinded to the target propofol Ce. Laryngoscopy and tracheal intubation had to be accomplished within <15 seconds for the patient to be included in the data analysis. In case of an intubation attempt lasting >15 seconds, the patient was excluded and the next patient was assigned the same randomization number.

Arterial blood pressure and heart rate (HR) were recorded at baseline (before fentanyl administration) and then every 1 minute until 5 minutes after tracheal intubation when the study finished.

As a consequence of the potential imprecision in the initial selected target propofol Ce, a preliminary analysis was planned in advance at 50% recruitment. On the basis of this analysis, dose-response curves were seen to be inadequately defined in the lower dose range for the Marsh model and in the upper dose range for the Schnider model. Accordingly, 3 extra groups were added, namely, Marsh target propofol concentration of 3.0 μg/mL, and Schnider target propofol concentrations of 5.5 and 6.0 μg/mL. These 18 extra patients, along with the remaining 24 of the original randomization, were randomized in a new group of 42 patients until the end of the study.

Normality was tested using the Shapiro-Wilk test. The dose-response relation for propofol Ce within each PK model was determined using probit analysis.14 Probit regression was performed to calculate values for the effective Ce (ECe) to induce hypnosis in 50% (ECe50) and 95% (ECe95) of patients with 95% confidence intervals (CIs). An effective Ce, or success, was defined as the Ce that provided adequate hypnosis (BIS <60) within 45 seconds after reaching the predetermined Ce and until the Ce was reset to 2.5 μg/mL after intubation. The model parameters were estimated by maximum likelihood, and the Pearson χ2 goodness-of-fit test was used to evaluate the fitness of the model.

Statistical analysis comparing data between models was performed using Wilcoxon rank sum test and across models using the Kruskal-Wallis test or Jonckheere-Terpstra trend test as appropriate.15 The hemodynamic analysis was restricted to patients with doses within the 95% CI of ECe95 and is expressed as a percentage change from baseline. Pearson χ2 test with simulated P value (based on 2000 replicates) was used for inferences on proportions. Data are expressed as median and interquartile range. A P value <0.05 was considered significant. Analyses were performed using R statistical programming language, version 2.11 (R Project for Statistical Computing: and SAS 9.2 software (SAS Institute Inc., Cary, NC).

Back to Top | Article Outline


Patient characteristics were similar between models and across target propofol Ce groups (Table 1). Sixteen patients with well-controlled chronic arterial hypertension were included in this study: 13 were using angiotensin-converting enzyme inhibitors, 4 were using calcium antagonists, 3 were using β-blockers, and 2 were using diuretics. These drugs were continued until the day of surgery. No statistically significant differences in the percentage of hypertensive patients among groups were found (Table 1). All initially selected patients were tracheally intubated within 15 seconds. As expected, initial bolus doses of propofol for the same predicted target Ce were significantly higher with the Marsh model (Table 2, Fig. 1).

Figure 2 shows the dose-response curve fitted to the propofol Ce predicted by both PK models. For the Marsh model, the calculated ECe50 of propofol was 3.4 μg/mL (95% CI: 2.9, 3.7 μg/mL), and the ECe95 was 4.2 μg/mL (95% CI: 3.8, 6.2 μg/mL). For the Schnider model, the calculated ECe50 was 4.5 μg/mL (95% CI: 4.1, 4.8 μg/mL), and the ECe95 was 5.5 μg/mL (95% CI: 5.0, 7.2 μg/mL). The ECe50 determined with the Marsh model was significantly lower than with the Schnider model (P < 0.001).

Baseline mean arterial blood pressure (MAP) and HR did not differ between PK models and target propofol concentration groups (Table 2). After propofol TCI was started, a significant decrease in MAP and HR from baseline was observed in all groups. For the Marsh model, the median maximum decrease for MAP in the different target concentration groups before tracheal intubation ranged from −26% to −35%, and for HR from −16% to −21% (not significantly different among target Ce groups). After tracheal intubation, MAP was still lower than baseline, ranging the median maximum MAP from −9% to −20%, and the HR from 3% to 26% (not significantly different among target Ce groups) (Fig. 3). A similar hemodynamic profile was observed for the Schnider model, with a MAP median maximum decrease from −26% to −39%, and −10% to −15% for HR before intubation, and a median maximum increase after intubation up to −7% to −10% from baseline in MAP and 2% to 5% in HR (not significantly different among target Ce groups in both cases) (Fig. 4).

Back to Top | Article Outline


This study has determined the appropriate target propofol Ce for induction and tracheal intubation in MO patients using TBW with 2 PK models. As stated in the hypothesis, our findings show that the appropriate target propofol Ce for the Marsh model is lower than the appropriate predicted target propofol Ce for use with the Schnider model.

There are several practical problems when using TCIs to induce anesthesia in MO patients. The first one is the choice of the weight entered into the TCI system. The second is the choice of the target propofol Ce to use for anesthesia induction. There is no consensus on the best way to calculate drug dosage in MO patients.4,8,16 Suggestions for the appropriate “pharmacologically active weight” include ideal body weight, TBW, and LBW, among others.710 Several clinical studies have attempted to define the appropriate weight scalar for the induction dose of hypnotics in MO subjects.16,17 In one of them, 200 mg or 350 mg propofol was administered over 60 seconds to patients with a BMI >38 kg/m2 followed by fentanyl 250 μg, atracurium, and tracheal intubation.17 These doses of propofol were chosen because 350 mg corresponds to 2.5 mg/kg for a TBW of 140 kg (the average weight of the obese population in the authors' clinic); in turn, 200 mg is the authors' usual induction dose of propofol for non-obese subjects.17 Based on BIS values and the fact that 60% of patients receiving 200 mg had systolic arterial blood pressure >160 mm Hg before tracheal intubation, the authors recommended the use of TBW to dose propofol at least in patients up to 140 kg.17 In a more recent study, propofol was administered to MO subjects at a constant rate (on a milligrams-per-kilogram basis) until loss of consciousness based either on LBW or TBW. The coefficient of determination (R2) between LBW and dose was 0.74 compared with 0.65 between TBW and dose of propofol. According to these results, the authors suggest that LBW is more appropriate than TBW to dose propofol for induction in MO patients.16 Therefore, there is no consensus on the “weight” to be used for dosing propofol for induction when injected “manually” (i.e., without a TCI system supported by a PK model) in this population. During maintenance of anesthesia, however, clinical studies with the Marsh model and propofol TCI anesthesia in morbid obesity suggest that the use of TBW results in a better performance than the use of other surrogates of patient mass.7,9,11 In addition, during this phase, there are time and tools, such as the depth of anesthesia monitors, to titrate propofol administration, and the availability of any of these 2 variables might compensate for the use of an inaccurate PK model.18 The situation is different during the induction phase when the amount of propofol (either in milligrams or target Ce) cannot be titrated, such as in a rapid-sequence induction that some advocate for MO patients.16 This is more relevant, and potentially more dangerous, with the use of a predetermined target propofol concentration predicted with a PK model in which the volume of the central compartment varies linearly with the weight compared with other PK models, such as Schnider's,6 in which the central volume is constant, independent of the weight of the patient.

The TCI technique might improve the manner of administering drugs. Targeting the propofol Ce, instead of propofol Cp, is a more logical approach. However, the effective propofol Ce for induction and intubation when using the TCI technique in MO patients has not been defined. We decided to define the ECe50 and ECe95 of propofol for induction with 2 PK models using TBW in MO patients. In particular, we selected these models because they are probably the most frequently used for propofol TCI in adults and because of their large differences in the estimation of the volume of the central compartment in MO subjects. As expected, significant differences in the predicted ECe values in similarly overweight patients receiving fentanyl 3 μg/kg TBW were found. Whereas a propofol Ce of 4.5 to 5.0 μg/mL predicted with the Marsh model seems to be adequate, using the Schnider model, an approximately 25% to 30% higher propofol Ce (i.e., 5.5–6.0 μg/mL) is needed for induction and intubation in these patients. Interestingly, although to reach its propofol ECe the Marsh model injected as a bolus approximately 50% more propofol than the Schnider model to reach a propofol ECe, immediately after intubation, both models administered essentially the same amount of propofol (Table 2). This last fact can also be the explanation for the similar hemodynamic effects, averaging a 32% decrease in MAP before intubation, observed with the 2 PK models.

The appropriate dose or Ce of any induction drug critically depends on the dose and characteristics of any other coadministered drug and our results, expressed as ECe values, might vary with the use of a different opioid or when given in a different dose. In turn, to determine accurate Ce values for induction and intubation, we also need accurate PK models for this stage of anesthesia. Because the development of accurate PK models for some patients (i.e., MO), and particularly for the induction phase,19,20 has proven to be a difficult task, one alternative is the “adjustment” of the existing models by the clinical determination of the effective predicted Ce for different stimuli and patients with each PK model. Our study did not aim to define the best weight scalar to be used as the input of the TCI system, but to define an adequate propofol Ce for induction using TBW in MO patients with the currently used PK models. With high certainty, different target propofol concentrations from the ones estimated in this study must be used when other surrogates of body mass are used in the TCI system.

Finally, whereas by using a specific predicted propofol Ce it is possible to compensate for differences between these PK models during induction of anesthesia in MO patients, other differences between the models will become apparent during prolonged infusions, such as during the maintenance phase of anesthesia, particularly in this population. For instance, Marsh's only covariate is the weight and while compartmental volumes are proportional to it, rate constants for slow and fast redistribution and elimination are (k10) fixed.5 In contrast, the more complex Schnider model includes the weight, height, age, and lean body mass as covariates. In particular, k10, which determines the infusion rate of propofol, is critically dependent on lean body mass.6 This variable, however, is calculated by the James formula, which has a paradoxical behavior because for every height, lean body mass increases with weight until a peak and then decreases to negative values (Fig. 5A).8 Given the mathematical relationship between k10 and weight and lean body mass in the Schnider model, after this peak lean body mass, k10 increases dramatically with a high risk of overdosing patients (Fig. 5B). Although there are other equations to calculate lean body mass without this odd behavior,21,22 the peak lean body mass with the James formula occurs with a BMI of approximately 42 kg/m2 in males and 37 kg/m2 in females,9 both values well within those found in clinics; therefore, the Schnider model should not be used in this population for maintenance of anesthesia.

In conclusion, given the inaccuracies of the PK models of propofol, particularly when used in MO subjects and during induction, specific propofol Ce values determined with each model must be used during this stage of anesthesia when using TBW, at least in this population.

Back to Top | Article Outline


Name: Ghislaine C. Echevarría, MD, MSc.

Contribution: This author helped in the development of the original idea, design of the study, the collection of the original study data, performed the analysis of data, prepared part of the manuscript, and approved the final manuscript.

Name: María F. Elgueta, MD.

Contribution: This author has helped in the collection of the original study data, reviewed the analysis of the data, and reviewed and approved the final manuscript.

Name: María T. Donoso, MD.

Contribution: This author has helped in the collection of the original study data, reviewed the analysis of the data, and reviewed and approved the final manuscript.

Name: Diego A. Bugedo, MD.

Contribution: This author has helped in the collection of the original study data, reviewed the analysis of the data, and reviewed and approved the final manuscript.

Name: Luis I. Cortínez, MD.

Contribution: This author has helped in the collection of the original study data, reviewed the analysis of the data, prepared the manuscript, and reviewed and approved the final manuscript.

Name: Hernán R. Muñoz, MD, MSc.

Contribution: This author helped in the development of the original idea, design of the study, has seen the original study data, reviewed the analysis of the data, and wrote part of the manuscript, reviewed and approved its final form.

This manuscript was handled by: Steven L. Shafer, MD.

Back to Top | Article Outline


1. Fontana MA, Wohlgemuth SD. The surgical treatment of metabolic disease and morbid obesity. Gastroenterol Clin North Am 2010;39:125–33
2. James WP. WHO recognition of the global obesity epidemic. Int J Obes (Lond) 2008;32:S120–6
3. De Baerdemaeker LE, Mortier EP, Struys MM. Pharmacokinetics in obese patients. Contin Educ Anaesth Crit Care Pain 2004;4:152–5
4. Casati A, Putzu M. Anesthesia in the obese patient: pharmacokinetic considerations. J Clin Anesth 2005;17:134–45
5. Marsh B, White M, Morton N, Kenny GN. Pharmacokinetic model driven infusion of propofol in children. Br J Anaesth 1991;67:41–8
6. Schnider TW, Minto CF, Gambus PL, Andresen C, Goodale DB, Shafer SL, Youngs EJ. The influence of method of administration and covariates on the pharmacokinetics of propofol in adult volunteers. Anesthesiology 1998;88:1170–82
7. Servin F, Farinotti R, Haberer JP, Desmonts JM. Propofol infusion for maintenance of anesthesia in morbidly obese patients receiving nitrous oxide: a clinical and pharmacokinetic study. Anesthesiology 1993;78:657–65
8. Bouillon T, Shafer SL. Does size matter? Anesthesiology 1998;89:557–60
9. Absalom AR, Mani V, De Smet T, Struys MM. Pharmacokinetic models for propofol: defining and illuminating the devil in the detail. Br J Anaesth 2009;103:26–37
10. Coetzee JF. Total intravenous anaesthesia to obese patients: largely guesswork? Eur J Anaesthesiol 2009;26:359–61
11. Albertin A, Poli D, La Colla L, Gonfalini M, Turi S, Pasculli N, La Colla G, Bergonzi PC, Dedola E, Fermo I. Predictive performance of ‘Servin's formula' during BIS-guided propofol-remifentanil target-controlled infusion in morbidly obese patients. Br J Anaesth 2007;98:66–75
12. Struys MM, De Smet T, Depoorter B, Versichelen LF, Mortier EP, Dumortier FJ, Shafer SL, Rolly G. Comparison of plasma compartment versus two methods for effect compartment– controlled target-controlled infusion for propofol. Anesthesiology 2000;92:399–406
13. Schnider TW, Minto CF, Shafer SL, Gambus PL, Andresen C, Goodale DB, Youngs EJ. The influence of age on propofol pharmacodynamics. Anesthesiology 1999;90:1502–16
14. Tallarida RJ. Quantal dose-response data: probit and logit analysis. In: Tallarida RJ. Drug Synergism and Dose-Effect Data Analysis. Boca Raton, FL: Chapman & Hall/CRC, 2000: 91–118
15. Higgins JJ. An Introduction to Modern Nonparametric Statistics. Pacific Grove, CA: Brooks/Cole, 2004
16. Ingrande J, Brodsky JB, Lemmens HJ. Lean body weight scalar for the anesthetic induction dose of propofol in morbidly obese subjects. Anesth Analg 2011;113:57–62
17. van Kralingen S, Diepstraten J, van de Garde EM, van der Lely AJ, van Dongen EP, van Ramshorst B, Knibbe CA. Comparative evaluation of propofol 350 and 200 mg for induction of anaesthesia in morbidly obese patients: a randomized double-blind pilot study. Eur J Anaesthesiol 2010;27:572–4
18. Coppens MJ, Eleveld DJ, Proost JH, Marks LA, Van Bocxlaer JF, Vereecke H, Absalom AR, Struys MM. An evaluation of using population pharmacokinetic models to estimate pharmacodynamic parameters for propofol and Bispectral Index in children. Anesthesiology 2011;115:83–93
19. Struys MM, Coppens MJ, De Neve N, Mortier EP, Doufas AG, Van Bocxlaer JF, Shafer SL. Influence of administration rate on propofol plasma-effect site equilibration. Anesthesiology 2007;107:386–96
20. Masui K, Kira M, Kazama T, Hagihira S, Mortier EP, Struys MM. Early phase pharmacokinetics but not pharmacodynamics are influenced by propofol infusion rate. Anesthesiology 2009;111:805–17
21. Hume R. Prediction of lean body mass from height and weight. J Clin Pathol 1966;19:389–91
22. Janmahasatian S, Duffull SB, Ash S, Ward LC, Byrne NM, Green B. Quantification of lean bodyweight. Clin Pharmacokinet 2005;44:1051–65
© 2012 International Anesthesia Research Society