Remember that all models are wrong; the practical question is how wrong do they have to be to not be useful
George Box (1987) 
As the population of morbidly obese patients increases worldwide, anaesthesiologists are increasingly confronted with the dilemma of finding appropriate drug dosing regimens for these difficult-to-manage patients. The altered physiological changes that accompany obesity include increases in blood volume, body water, muscle mass and cardiac output; factors that inter-alia affect volumes of distribution and total body clearance [2,3]. Although dose regimens based on lean body mass (LBM) or ideal body weight (IBW) may result in inadequate anaesthesia, concern arises that dosing according to total body weight (TBW) may give rise to adverse haemodynamic effects. Unfortunately, there is a paucity of scientific studies that addresses these issues, and therefore many practitioners compromise by estimating a dosing weight that is between IBW and TBW .
Commercially available devices, approved for delivery of target-controlled infusions (TCIs), have been accepted readily into routine practice in Europe, Britain and several other countries, apart from the United States. With regard to propofol, only two pharmacokinetic parameter sets for adults have been programmed into these syringe pumps: those of Marsh et al.  and Schnider et al. . These were derived from patients of normal body mass index (BMI), and it is likely that applying them to patients who are vastly different from the original experimental group is unsound. There are published pharmacokinetic parameter sets that have been derived from special classes of patients (e.g. intensive care ), but pump manufacturers are understandably hesitant to incorporate them, considering the cost of meeting stringent regulatory requirements. The morbidly obese represents a special class in whom the pharmacokinetics of propofol has been the subject of only one formal study . In that investigation, propofol was administered by a manually controlled infusion that was based on a dosing weight calculated to be IBW + 40% of the difference between TBW and IBW. Although that regimen provided adequate anaesthesia, the authors concluded that it was not supported by the pharmacokinetic findings. They suggested that propofol infusions should perhaps be based on TBW.
In a previous case series , propofol TCI was administered to 20 morbidly obese patients targeting the electroencephalographic bispectral index (BIS) as the endpoint. The ‘Marsh’  pharmacokinetic parameter set was used with the weight adjustment described as above . Comparisons between measured and predicted concentrations revealed that, in all patients, the TCI system overpredicted the propofol concentration on average by 32%. This means that, for most of the time, blood propofol concentrations were considerably less than the predicted values. In this issue, La Colla and coworkers, from the same institution, report on a sequel to their previous study in which they administered total intravenous (i.v.) anaesthesia to morbidly obese patients using propofol and remifentanil. For propofol, either TBW or the aforementioned adjusted body weight was input to the ‘Marsh’ model . As before, they fine-tuned the targeted concentrations to maintain BIS values between 40 and 50, while remifentanil-targeted concentrations were titrated according to haemodynamic responses to surgery. Although stable propofol blood concentrations were maintained in both groups, the measured concentrations were generally less than those predicted by the TCI system. Furthermore, the errors between the measured and predicted concentrations were similar in the two groups. These results contrast with the findings in studies of propofol–remifentanil anaesthesia for patients of normal body weight in which measured concentrations were consistently greater than predicted [10–12].
In spite of using an inherently unsuitable TCI device, the authors were able to achieve and maintain stable anaesthesia in morbidly obese patients by titrating to pharmacodynamic effects, namely two of the main components of balanced anaesthesia, hypnosis and suppression of haemodynamic responses to surgical stimulation. Using this empirical approach, the investigators could probably have achieved equally satisfactory anaesthesia had they used any other pharmacokinetic parameter set, or even a manually adjusted infusion regimen. This study illustrates an important principle, namely, that it is wrong to extrapolate pharmacokinetic models beyond the attributes of the patients who participated in the original research. The ‘Marsh’ pharmacokinetic model was developed and tested in young patients of normal weight [13–15]. It would, therefore, have been most surprising if this study of morbidly obese patients had resulted in good agreement between predicted and measured propofol concentrations. Indeed, it is likely that if they had attempted to base their dosages only on the simulated plasma concentrations displayed by the TCI system instead of electroencephalograph (EEG) monitoring, the levels of anaesthesia as well as the arterial concentrations would have fluctuated extensively.
La Colla and coworkers have addressed an important clinical problem, namely, how to administer total i.v. anaesthesia to morbidly obese patients using an inappropriate tool, because no other TCI system is available. Their study indicates that the ‘Marsh’ model is inappropriate, with as well as without a weight adjustment. If the available TCI systems are to be used in these patients, little faith can be placed in the displayed concentrations, and therefore it is important to use some form of EEG monitoring in these patients. Furthermore, considering that there is no correlation between BIS values and measured propofol concentrations during propofol–remifentanil anaesthesia [10,11] (this is probably due to the synergistic interaction that the two drugs exert on the BIS ), we can conclude that the ‘Marsh’ pharmacokinetic parameter set is ‘wrong enough’ not to be useful in morbidly obese patients.
The ‘Marsh’ pharmacokinetic parameter set  adjusts only the central volume of distribution according to weight. The question arises as to whether the other available pharmacokinetic parameter set in the commercially available TCI systems, the ‘Schnider’ model, would be more appropriate, considering that several pharmacokinetic parameters are adjusted according to age, sex and weight . However, the parameter adjustments according to weight are based on LBM and, in morbidly obese patients, these calculations have an anomaly. For each height, LBM increases with TBW to a certain maximum, after which the calculated LBM decreases. This aberration is due to the fact that the original publication by James  probably did not include morbidly obese patients. The practical result is that if TBW of obese patients is input into the ‘Schnider’ model in the available TCI pumps, one make of pump does not allow propofol administration to patients whose BMI exceeds 42 kg m−2 in males or 35 kg m−2 in females. The precautionary software of another pump does not permit usage in patients whose TBW exceeds a certain weight determined by the peak value of the James equations. A recent study by Janmahasatian et al.  derived a new formula for LBM that incorporated morbidly obese patients, and the calculations do not have the anomaly of the James equation (Fig. 1) [17,18]. We may conjecture that introducing the Janmahasatian equation for LBM to the ‘Schnider’ pharmacokinetic parameter set would result in greater accuracy of predicted propofol concentrations during TCI. However, among the 25 patients from whom the model was derived, only one had a BMI greater than 30 kg m−2, and, therefore, it is likely that extrapolating the model to morbidly obese patients would also result in misspecification.
In order to develop propofol TCI dosing regimens that are better suited to morbidly obese patients, we need formal studies based on concentration–time data. It is surprising that the parameter set derived by Servin et al. in 1993 has never been subjected to evaluation of its predictive performance during TCI. La Colla and coworkers have demonstrated that clinicians can compensate for the deficiencies of the available clinical tools by intelligent application of electroencephalographic and haemodynamic monitoring. In the group to whom propofol was dosed according to TBW, more than a few measured propofol concentrations were greater than those predicted. Therefore, it is probably prudent to employ a dosing weight that lies somewhere between IBW and TBW and to proceed cautiously, following the advice of Struys et al. to ‘give some drug and observe your patient to see what happens’.
In general, dosing guidelines for obese patients that appear in the literature are, to say the least, confusing. Recommendations include nonlinear adjustments to TBW for all drugs  as well as different suggestions for each drug that vary from calculations based on LBM as well as IBW and TBW [2,3,20,21]. Predicted normal weight , another commonly used correction for dosing-weight, is also inappropriate, because the calculation of LBM is included in the formula. Bouillon and Shafer  suggested that a reasonable approach to obese patients would be to base dosing weight on IBW increased by a factor determined by the ratio of LBM to the LBM of a person of IBW. They have provided nomograms from which a dosing weight could be determined from height and weight. These nomograms, which have not been subjected to formal evaluation, could possibly be applicable to obese or even some morbidly obese patients. However, for the superobese patients of La Colla and coworkers, the corrections by Bouillon and Shafer would be inadequate as they have the same anomaly as the calculation of LBM: that is, decreasing dosing weight after peaking at a certain TBW. If the LBM calculations of Janmahasatian are used, this does not occur, and the calculated dosing weights are roughly similar to those of the empiric formula used by La Colla. Figure 2[4,18,22] illustrates various calculations of dosing weight that have been obtained for a person of 1.70 m height. It is interesting to note that the adjusted dosing weights are consistently less than TBW. Considering the findings of La Colla in this issue for propofol, it is unlikely that applying any weight adjustments to the two available pharmacokinetic models for propofol will result in improved accuracy of predicted concentrations. These dosing weight calculations may, however, be appropriate to other drugs.
1 Box GEP, Draper NR. Empirical model-building and response surfaces
. New York: John Wiley & Sons; 1987, p. 424.
2 Adams JP, Murphy PG. Obesity in anaesthesia and intensive care. Br J Anaesth 2000; 85:91–108.
3 Shenkman Z, Shir Y, Brodsky JB. Perioperative management of the obese patient. Br J Anaesth 1993; 70:349–359.
4 Bouillon T, Shafer SL. Does size matter. Anesthesiology 1998; 89:557–558.
5 Marsh B, White M, Morton N, Kenny GN. Pharmacokinetic model driven infusion of propofol in children. Br J Anaesth 1991; 67:41–48.
6 Schnider TW, Minto CF, Gambus PL, et al
. The influence of method of administration and covariates on the pharmacokinetics of propofol in adult volunteers. Anesthesiology 1998; 88:1170–1182.
7 Barr J, Egan TD, Sandoval NF, et al
. Propofol dosing regimens for ICU sedation based upon an integrated pharmacokinetic-pharmacodynamic model. Anesthesiology 2001; 95:324–333.
8 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; 78:657–665.
9 Albertin A, Poli D, La Colla L, et al
. 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.
10 Hoymork SC, Raeder J, Grimsmo B, Steen PA. Bispectral index, predicted and measured drug levels of target-controlled infusions of remifentanil and propofol during laparoscopic cholecystectomy and emergence. Acta Anaesthesiol Scand 2000; 44:1138–1144.
11 Hoymork SC, Raeder J, Grimsmo B, Steen PA. Bispectral index, serum drug concentrations and emergence associated with individually adjusted target-controlled infusions of remifentanil and propofol for laparoscopic surgery. Br J Anaesth 2003; 91:773–780.
12 Wietasch JK, Scholz M, Zinserling J, et al
. The performance of a target-controlled infusion of propofol in combination with remifentanil: a clinical investigation with two propofol formulations. Anesth Analg 2006; 102:430–437.
13 Coetzee JF, Glen JB, Wium CA, Boshoff L. Pharmacokinetic model selection for target controlled infusions of propofol. Assessment of three parameter sets. Anesthesiology 1995; 82:1328–1345.
14 Schraag S, Flaschar J. Delivery performance of commercial target-controlled infusion devices with Diprifusor module. Eur J Anaesthesiol 2002; 19:357–360.
15 Swinhoe CF, Peacock JE, Glen JB, Reilly CS. Evaluation of the predictive performance of a ‘Diprifusor’ TCI system. Anaesthesia 1998; 53(Suppl 1):61–67.
16 Ropcke H, Konen-Bergmann M, Cuhls M, et al
. Propofol and remifentanil pharmacodynamic interaction during orthopedic surgical procedures as measured by effects on bispectral index. J Clin Anesth 2001; 13:198–207.
17 James WPT. Research on obesity. London: Her Majesty's Stationery Office; 1976.
18 Janmahasatian S, Duffull SB, Ash S, et al
. Quantification of lean bodyweight. Clin Pharmacokinet 2005; 44:1051–1065.
19 Struys M, Mortier E, Versichelen L. Strategies to prevent awareness. In: Vuyk J, Engbers F, Groen-Mulder S, editors. On the study and practice of intravenous anaesthesia. Dordrecht, The Netherlands: Kluwer Academic Publishers; 2000. pp. 113–127
20 Egan TD, Huizinga B, Gupta SK, et al
. Remifentanil pharmacokinetics in obese versus lean patients. Anesthesiology 1998; 89:562–573.
21 Slepchenko G, Simon N, Goubaux B, et al
. Performance of target-controlled sufentanil infusion in obese patients. Anesthesiology 2003; 98:65–73.
22 Duffull SB, Dooley MJ, Green B, et al
. A standard weight descriptor for dose adjustment in the obese patient. Clin Pharmacokinet 2004; 43:1167–1178.