Secondary Logo

Journal Logo


Prospective randomised comparison of Marsh and Schnider pharmacokinetic models for propofol during induction of anaesthesia in elective cardiac surgery

Viterbo, João F.; Lourenço, André P.; Leite-Moreira, Adelino F.; Pinho, Paulo; Barros, Fernanda

Author Information
European Journal of Anaesthesiology: October 2012 - Volume 29 - Issue 10 - p 477-483
doi: 10.1097/EJA.0b013e3283542421



Target-controlled infusion (TCI) has become an established technique for administration of intravenous anaesthetics.1 It has been shown to be superior to manual infusion in cardiac anaesthesia.2,3 Although TCI with propofol has not shown advantages in relation to time to extubation, drug consumption or haemodynamics, it is still widely used in cardiac anaesthesia, requiring fewer physician interventions and allowing a timely schedule of extubation.4–9 Two models are generally employed and commercially available.10 Based on venous blood sampling after steady-state infusion, Marsh's model derives from the pharmacokinetic variables described by Gepts et al.11 and sets compartmental volumes proportional to weight.12 In contrast, Schnider's model, based on arterial blood sampling after an intravenous bolus followed by continuous infusion, was obtained from a larger sample of volunteers with age, body weight, height and lean body mass (estimated from weight, height and sex) as covariates.13 Although modern TCI syringe pumps incorporate both pharmacokinetic models, the clinical differences between the two remain poorly investigated.10 Schnider's model may be of added value in special populations, for example target concentrations must be reduced in the elderly.14 Nevertheless, Marsh's model has been reported to correlate better with bispectral index (BIS) monitoring and sedation scores.15

Haemodynamic stability and smoothness of induction have long been cornerstones of cardiac anaesthesia. Induction with propofol by TCI might be particularly useful. Indeed, induction of anaesthesia with manual infusion usually leads to an overshoot in effect-site concentration, and lower infusion rates avoid hypotension in the elderly.16 According to the authors’ knowledge, differences between Marsh and Schnider models have not been investigated during induction of anaesthesia in cardiac surgery. The aim of our study was to compare the two models in respect of haemodynamic stability, as assessed by the decrease in mean blood pressure (mBP) during induction of anaesthesia in patients undergoing elective cardiac surgery. Time to achieve induction of anaesthesia, total dose of propofol required and predicted effect-site concentrations were also compared as secondary end-points.


Ninety adult patients scheduled for elective cardiac surgery at the Department of Cardiothoracic Surgery in the Hospital São João between July and December 2010 were enrolled consecutively (by J.F.V. and A.P.L.). Preoperative echocardiography and cardiac catheterisation were used to establish left ventricular ejection fraction. Serum creatinine concentration was determined on preoperative biochemical evaluation. Medical information was collected during the preanaesthetic visit as well as from the clinical records. Logistic European System for Cardiac Operative Risk Evaluation (EuroSCORE) was employed as a mortality risk predictor. The study was approved by the Hospital São João Ethics Committee (Chairperson Professor F. Almeida) on 26 February 2010, and conforms to the Helsinki declaration. All patients gave their informed consent.

Preoperative cardiac medication was continued until the day of surgery. Patients were premedicated with oral diazepam 0.15 mg kg−1 on the night before and on the morning of surgery. On arrival in the operating room, patients were randomly assigned (by J.F.V.) in blocks of 10, by a sealed envelope technique, to groups Marsh (n = 45) or Schnider (n = 45). Participants were blinded to the intervention. Patients were sedated with intravenous midazolam 0.05 mg kg−1 and monitoring devices were connected. Invasive blood pressure, heart rate (HR), pulse oximetry and BIS were continuously recorded (Anesthesia Manager 7.1, Picis inc., Wakefield, Massachusetts, USA). Induction of anaesthesia by TCI (Orchestra Base Primea Fresenius, Brezins, France) began with an effect-site-targeted remifentanil infusion, at a maximum rate of 1200 ml h−1, until a stable estimated effect-site concentration of 2.5 ng ml−1 was achieved, using Minto's model. In both groups, propofol was then immediately started according to the respective pharmacokinetic model, also in effect-site-targeting mode, at a maximum rate of 1200 ml h−1 initially to attain a stable estimated effect-site concentration of 1.5 μg ml−1, and then titrated upward by means of successive 0.5 μg ml−1 increments in the effect-site concentration until induction of anaesthesia was achieved, defined by the reliable cutoff value of BIS less than 50.17 At that time, the cumulative dose of infused propofol was recorded and muscle relaxation was induced with intravenous rocuronium 0.6 mg kg−1. Before tracheal intubation, intravenous ephedrine 5 mg was given through a peripheral vein whenever mBP fell below 60 mmHg. After tracheal intubation, anaesthetic and surgical procedures continued with no further intervention. The study design is summarised in Fig. 1. Data from the anaesthesia records and the TCI system were exported to Excel (Microsoft Office Excel 2007, Microsoft corp., Redmond, Washinton, USA).

Fig. 1
Fig. 1:
No captions available.

Our TCI system (Orchestra Base Primea Fresenius, Brezins, France) incorporates the option of employing effect-site TCI using the Marsh model, the so-called ‘modified Marsh’ model,18 by including a plasma/effect-site equilibration constant (keo) of 1.2 min−1, which has a time to peak effect similar to that of Schnider's model and better predicts the time course of clinical effects as assessed by BIS compared with first generation TCI systems (keo of 0.26 min−1).

All recruited patients were randomised to the intervention and none was excluded from the intervention or statistical analysis thereafter.

Statistical analysis

Sample size was estimated in order to detect, at least, a 10% difference between groups in decrease in mBP with a power of 0.8 and an α of 0.05. Marsh and Schnider groups were compared by a blinded researcher (A.P.L.) using either Student's t-test or the Mann–Whitney Rank Sum test for continuous variables and χ2 or Fisher's exact test for proportions. Variables monitored during induction of anaesthesia were compared with two-way Analysis of variance (ANOVA) for repeated measures. Two-way ANOVA was used to compare Marsh and Schnider groups regarding haemodynamic and TCI variables at induction, with age and weight as cofactors. Data were analysed with the aid of software (PAWS Statistics 18.0, International Business Machines Corp., NY, USA) and are presented as value (%) and mean ± SD for qualitative and quantitative variables, respectively. Statistical significance was set at two-tailed P value less than 0.05.


Patients underwent valve replacement or repair procedures (33%), coronary artery bypass grafting (44%), both (11%) or other interventions such as atrial septal defect correction (five patients), epicardial pacemaker insertion (one patient), removal of infected pacemaker electrodes (one patient), Bentall procedure (two patients) and subaortic membrane removal (one patient). One patient underwent synchronous carotid endarterectomy.

Marsh and Schnider groups did not differ in respect of type of surgery, personal characteristics, logistic EuroSCORE, ASA physical status, major comorbidities or current medication (Table 1).

Table 1
Table 1:
Patient characteristics

Both mBP and HR decreased significantly during induction of anaesthesia, but no significant differences were observed between Marsh and Schnider groups (Fig. 2). There were no differences in the need for vasopressor support (20 and 24% in groups Marsh and Schnider, respectively; P = 0.800). Regarding the influence of covariates of Schnider's pharmacokinetic model, when both age and BMI were included along with pharmacokinetic model as independent variables in a stepwise multiple linear regression model, only age was a significant predictor of a decrease in mBP (P = 0.028). Patients older than 60 years of age (51 and 49% of patients from Marsh and Schnider groups, respectively) showed a greater decrease in mBP compared with younger patients, but this could not be attributed to the pharmacokinetic model (Fig. 3). Obese patients, as defined by a BMI of 30 kg m−2 or more (16 and 20% of patients from Marsh and Schnider groups, respectively) did not differ from the other patients regarding the decrease in mBP, irrespective of the pharmacokinetic model (Fig. 3).

Fig. 2
Fig. 2:
No captions available.
Fig. 3
Fig. 3:
No captions available.

Although the total dose of propofol, normalised for body weight at induction, did not differ between groups, patients in group Schnider needed more time and required higher predicted propofol effect-site concentrations to achieve induction (BIS <50) compared with those in group Marsh (Fig. 4). When age and weight covariates, as defined above, were analysed, we found that the total dose of propofol per kilogram body weight in the obese subgroup of group Schnider was significantly lower compared to the nonobese patients or to patients in group Marsh (Table 2).

Fig. 4
Fig. 4:
No captions available.
Table 2
Table 2:
Target-controlled infusion variables in subgroups of patients defined by age and weight at the time of induction


Failing to appreciate differences between pharmacokinetic models in TCI systems may result in administration of inadequate doses of propofol and potentially harmful effects.10 However, these models remain poorly investigated in many clinical settings. Although the choice of model does not influence maintenance infusion rates for most patients, it may have a significant impact on initial drug dosage during induction.10 This may be particularly important in patients at cardiovascular risk. We compared the two most widely used pharmacokinetic models for propofol during induction of anaesthesia for elective cardiac surgery and found no significant differences in mBP and HR decrease between Marsh and Schnider pharmacokinetic models, when used in effect-site TCI and when guided by BIS values. No differences were observed between pharmacokinetic models even when subgroups of elderly and obese patients were analysed.

Patients in both Marsh and Schnider groups had significant and similar decreases in HR and mBP. A decrease in HR and cardiac index has been described during induction of anaesthesia using Minto and Schnider pharmacokinetic models for remifentanil and propofol, respectively, in patients scheduled for myocardial revascularisation, but no significant change in mBP.7 This discrepancy could be due to differences in patient selection, drug dosing and criteria used to define a decrease in mBP.

The decrease in mBP was greater in older patients, probably due to increased sensitivity to propofol. Surprisingly, although age is one of Schnider's pharmacokinetic model covariates,10 the magnitude of the decrease in mBP was similar in groups Schnider and Marsh, even in the elderly. In addition, older and younger patients behaved similarly in groups Marsh and Schnider in respect of TCI variables. This disagrees with the differences reported by Ouattara et al.14 between elderly and young patients during induction of anaesthesia for cardiac surgery with Marsh's original model in plasma-targeting mode (Diprifusor, AstraZeneca, Manchester, England). This may have been due to the fact that we used a ‘modified effect-site Marsh model’ with a keo of 1.2 min−1. With effect-site targeting, just enough propofol is administered in order to maximally alter plasma concentration without overshoot, and clinical effect is reached rapidly upon equilibration with the effect-site. The time delay is related to the equilibration constant; the higher the keo the faster the equilibration and, therefore, the smaller the plasma overshoot. The keo value of 1.2 min−1 not only predicts the time course of clinical effect more accurately but also allows more subtle manipulations of plasma concentration.10,18 Our results confirm the observations of Absalom et al.10 regarding the safety of using the modified effect-site Marsh model with a keo of 1.2 min−1 in cardiac surgery patients. Indeed, within group Marsh, we found no differences between age subcategories.

Although no differences were observed in the decrease in mBP or the mean total dose of propofol per kg weight to achieve induction when group Schnider was compared to group Marsh, the predicted effect-site concentrations and the time until induction (BIS <50) were significantly higher in group Schnider, independently of age and weight. The absence of differences in dose required for induction and decrease in mBP could also be partly explained by the similarity between the Schnider model and the ‘modified Marsh model’ when used in effect-site targeting mode. The shorter time to achieve induction and the lower predicted effect-site concentrations in group Marsh may be due to the fact that Base Primea uses a keo of 1.2 min−1 for Marsh's model compared with 0.456 min−1 for Schnider's model. This keo, in conjunction with the pharmacokinetic variables, predicts an average time to peak effect of 1.69 min,after a bolus dose, in group Schnider compared with a time to peak effect of 1.6 min in group Marsh.19 Interestingly, Barakat et al. found a closer correlation between predicted effect-site concentrations and sedation scores and BIS with the Marsh model compared with the Schnider model.15 Differences in total propofol requirements per kilogram weight to achieve induction between these two models only became relevant for the obese. However, this may not be clinically relevant because no haemodynamic differences were observed during induction. No other differences were observed between groups Marsh and Schnider in obese patients.

The study was intentionally conducted with minimal modifications to clinical practice, and the infusion of propofol was guided by BIS values. In a purely experimental setting, an alternative would have been to perform induction with a planned anaesthetic effect-site target concentration in both models. Although this would have not allowed the titration of dose to effect in individual patients and would deviate from standard clinical practice, it would certainly have been more accurate in comparing haemodynamic consequences. Nevertheless, because infusion was started with the same effect-site target concentration followed by carefully titrated stepwise increases with both models, we believe that this approach fairly approximates the experimental setting. However, we must draw attention to the fact that, as infusion was guided by BIS values, we were comparing the haemodynamic effects of different predicted effect-site concentrations for the two models, whereas if it had been determined by a planned anaesthetic effect-site concentration, we would have expected differences in propofol dosage and, perhaps, enhanced haemodynamic disturbances.


When used in effect-site targeting mode with a Base Primea TCI computer (Orchestra Base Primea Fresenius, Brezins, France) and titrated to BIS values, there were no clinically significant differences between the Marsh and Schnider pharmacokinetic models for induction of anaesthesia in elective cardiac surgery patients, with the exception of slightly longer times to achieve induction with Schnider's model. It seems that, although there can be a rationale for choosing the Schnider model for the elderly and the obese, the most relevant issue is that the anaesthesiologist uses the model that he or she is most familiar with and titrates it to the individual patient.


Assistance with the study: none declared.

Sources of funding: none declared.

Conflicts of interest: none declared.


1. Ihmsen H, Schraag S, Kreuer S, et al. Target-controlled infusion. Clinical relevance and special features when using pharmacokinetic models. Anaesthesist 2009; 58:708–715.
2. Alvis JM, Reves JG, Govier AV, et al. Computer-assisted continuous infusions of fentanyl during cardiac anesthesia: comparison with a manual method. Anesthesiology 1985; 63:41–49.
3. Theil DR, Stanley TE 3rd, White WD, et al. Midazolam and fentanyl continuous infusion anesthesia for cardiac surgery: a comparison of computer-assisted versus manual infusion systems. J Cardiothorac Vasc Anesth 1993; 7:300–306.
4. Leslie K, Clavisi O, Hargrove J. Target-controlled infusion versus manually-controlled infusion of propofol for general anaesthesia or sedation in adults. Cochrane Database Syst Rev 2008(3):CD006059.
5. Olivier P, Sirieix D, Dassier P, et al. Continuous infusion of remifentanil and target-controlled infusion of propofol for patients undergoing cardiac surgery: a new approach for scheduled early extubation. J Cardiothorac Vasc Anesth 2000; 14:29–35.
6. Lehmann A, Zeitler C, Thaler E, et al. Comparison of two different anesthesia regimens in patients undergoing aortocoronary bypass grafting surgery: sufentanil–midazolam versus remifentanil-propofol. J Cardiothorac Vasc Anesth 2000; 14:416–420.
7. Guarracino F, Penzo D, De Cosmo D, et al. Pharmacokinetic-based total intravenous anaesthesia using remifentanil and propofol for surgical myocardial revascularization. Eur J Anaesthesiol 2003; 20:385–390.
8. Russell D, Wilkes MP, Hunter SC, et al. Manual compared with target-controlled infusion of propofol. Br J Anaesth 1995; 75:562–566.
9. Lehmann A, Boldt J, Rompert R, et al. Target-cntrolled infusion or manually controlled infusion of propofol in high-risk patients with severely reduced left ventricular function. J Cardiothorac Vasc Anesth 2001; 15:445–450.
10. 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.
11. Gepts E, Camu F, Cockshott ID, Douglas EJ. Disposition of propofol administered as constant rate intravenous infusions in humans. Anesth Analg 1987; 66:1256–1263.
12. Marsh B, White M, Morton N, Kenny GN. Pharmacokinetic model driven infusion of propofol in children. Br J Anaesth 1991; 67:41–48.
13. 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.
14. Ouattara A, Boccara G, Lemaire S, et al. Target-controlled infusion of propofol and remifentanil in cardiac anaesthesia: influence of age on predicted effect-site concentrations. Br J Anaesth 2003; 90:617–622.
15. Barakat AR, Sutcliffe N, Schwab M. Effect site concentration during propofol TCI sedation: a comparison of sedation score with two pharmacokinetic models. Anaesthesia 2007; 62:661–666.
16. Peacock JE, Lewis RP, Reilly CS, Nimmo WS. Effect of different rates of infusion of propofol for induction of anaesthesia in elderly patients. Br J Anaesth 1990; 65:346–352.
17. Hirschi M, Meistelman C, Longrois D. Effects of normothermic cardiopulmonary bypass on bispectral index. Eur J Anaesthesiol 2000; 17:499–505.
18. Struys MM, De Smet T, Depoorter B, et al. Comparison of plasma compartment versus two methods for effect compartment: controlled target-controlled infusion for propofol. Anesthesiology 2000; 92:399–406.
19. Schnider TW, Minto CF, Shafer SL, et al. The influence of age on propofol pharmacodynamics. Anesthesiology 1999; 90:1502–1516.

age; blood pressure; body weight; cardiac surgery; pharmacology; propofol; target-controlled infusion

© 2012 European Society of Anaesthesiology