Retrospective comparison of Eleveld, Marsh, and Schnider propofol pharmacokinetic models in 50 patients
2019; Elsevier BV; Volume: 124; Issue: 2 Linguagem: Inglês
10.1016/j.bja.2019.10.019
ISSN1471-6771
AutoresTobias Hüppe, Felix Maurer, Daniel I. Sessler, Thomas Volk, Sascha Kreuer,
Tópico(s)Intensive Care Unit Cognitive Disorders
ResumoEditor—Numerous pharmacokinetic models have been developed to dose propofol administration accurately during anaesthesia.1Struys M.M. De Smet T. Glen J.I. Vereecke H.E. Absalom A.R. Schnider T.W. The history of target-controlled infusion.Anesth Analg. 2016; 122: 56-69Crossref PubMed Scopus (75) Google Scholar Pharmacokinetic models developed by Marsh and colleagues2Marsh B. White M. Morton N. Kenny G.N. Pharmacokinetic model driven infusion of propofol in children.Br J Anaesth. 1991; 67: 41-48Abstract Full Text PDF PubMed Scopus (939) Google Scholar (n=20) and Schnider and colleagues3Schnider T.W. Minto C.F. Gambus P.L. et al.The influence of method of administration and covariates on the pharmacokinetics of propofol in adult volunteers.Anesthesiology. 1998; 88: 1170-1182Crossref PubMed Scopus (897) Google Scholar (n=24) have been applied in commercial propofol target-controlled infusion (TCI) syringe pumps. Recently, Eleveld and colleagues4Eleveld D.J. Colin P. Absalom A.R. Struys M.M.R.F. Pharmacokinetic–pharmacodynamic model for propofol for broad application in anaesthesia and sedation.Br J Anaesth. 2018; 120: 942-959Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar used a large diverse population to develop a pharmacokinetic–pharmacodynamic model for predicting propofol concentrations and the bispectral index. Specifically, they used data from 1033 patients in the Open TCI Initiative which combines data from many investigators. However, the Eleveld model4Eleveld D.J. Colin P. Absalom A.R. Struys M.M.R.F. Pharmacokinetic–pharmacodynamic model for propofol for broad application in anaesthesia and sedation.Br J Anaesth. 2018; 120: 942-959Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar has yet to be validated in an independent patient population. It is unclear whether this pharmacokinetic model provides better predictive performance of measured plasma propofol concentrations than conventional models such as those of Marsh and colleagues2Marsh B. White M. Morton N. Kenny G.N. Pharmacokinetic model driven infusion of propofol in children.Br J Anaesth. 1991; 67: 41-48Abstract Full Text PDF PubMed Scopus (939) Google Scholar and Schnider and colleaguges.3Schnider T.W. Minto C.F. Gambus P.L. et al.The influence of method of administration and covariates on the pharmacokinetics of propofol in adult volunteers.Anesthesiology. 1998; 88: 1170-1182Crossref PubMed Scopus (897) Google Scholar We therefore evaluated plasma concentrations in 50 patients who were not included in the population used by Eleveld and colleagues4Eleveld D.J. Colin P. Absalom A.R. Struys M.M.R.F. Pharmacokinetic–pharmacodynamic model for propofol for broad application in anaesthesia and sedation.Br J Anaesth. 2018; 120: 942-959Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar for model development. Specifically, we tested the hypothesis that the Eleveld model provides better predictions (smaller median absolute performance error [MDAPE] and smaller median performance error [MDPE]) of arterial propofol plasma concentration than either the Marsh or Schnider models. After approval from the responsible ethics committee (Identification number 39/2016, March 22, 2016, Ärztekammer Saarland, Saarbrücken, Germany) and written informed consent, we enrolled 50 patients who were undergoing visceral or trauma surgery. All had total i.v. general anaesthesia and an arterial catheter. Data were obtained from a previous study related to development of an exhaled propofol monitor.5American Society of Anesthesiologists Annual meeting san Francisco.2018http://www.asaabstracts.com/strands/asaabstracts/abstract.htm?year=2018&index=8&absnum=4240Google Scholar Anaesthesia was provided with propofol using TCI with the Marsh model2Marsh B. White M. Morton N. Kenny G.N. Pharmacokinetic model driven infusion of propofol in children.Br J Anaesth. 1991; 67: 41-48Abstract Full Text PDF PubMed Scopus (939) Google Scholar and remifentanil using TCI with the Minto model.6Minto C.F. Schnider T.W. Egan T.D. et al.Influence of age and gender on the pharmacokinetics and pharmacodynamics of remifentanil. I. model development.Anesthesiology. 1997; 86: 10-23Crossref PubMed Scopus (963) Google Scholar,7Minto C.F. Schnider T.W. Shafer S.L. Pharmacokinetics and pharmacodynamics of remifentanil: II. Model application.Anesthesiology. 1997; 86: 24-33Crossref PubMed Google Scholar Quantification of propofol in plasma was performed by liquid chromatography (Agilent 1260 Infinity series, Agilent Technologies, Waldbronn, Germany). We calculated propofol plasma concentrations using automatically recorded perfusor flow rates with respective pharmacokinetic data sets of Eleveld and colleagues,4Eleveld D.J. Colin P. Absalom A.R. Struys M.M.R.F. Pharmacokinetic–pharmacodynamic model for propofol for broad application in anaesthesia and sedation.Br J Anaesth. 2018; 120: 942-959Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar Marsh and colleagues,2Marsh B. White M. Morton N. Kenny G.N. Pharmacokinetic model driven infusion of propofol in children.Br J Anaesth. 1991; 67: 41-48Abstract Full Text PDF PubMed Scopus (939) Google Scholar and Schnider and colleagues.3Schnider T.W. Minto C.F. Gambus P.L. et al.The influence of method of administration and covariates on the pharmacokinetics of propofol in adult volunteers.Anesthesiology. 1998; 88: 1170-1182Crossref PubMed Scopus (897) Google Scholar Specifically, we used transfer constants of the respective pharmacokinetic models and the propofol supply to the central distribution compartment (Microsoft Excel 2011, Microsoft Corporation, Redmond, WA, USA) for calculations. To compare measured and corresponding estimated plasma concentrations and the precision of different pharmacokinetic models, we calculated the MDAPE as a measure of accuracy. The MDAPE of individual patients were compared between different pharmacokinetic models using analysis of variance on ranks (anova). The MDPE was estimated as a measure of bias. Post hoc power analysis was performed using one-way anova with fixed effects in three groups with a total sample size of 150 and alpha-error 0.05. We analysed 488 arterial blood samples. Post hoc power analysis revealed an effect size of 0.11 with a power of 0.21. The MDAPE with the Eleveld model for individual patients was 22 (4–50)%; 25 (6–58)% with the Marsh model; and 26 (2–54)% with the Schnider model. Individual patient MDAPE values did not differ significantly among the three models, and did not depend on age, weight, or gender. MDPE values were −18% for the Eleveld model, 9% for the Marsh model, and −20% for the Schnider model. When using the Marsh model, measured propofol plasma concentrations were a mean of 0.25 μg ml−1 above estimated concentrations. The Schnider and Eleveld models overestimated measured concentrations by 0.64 and 0.53 μg ml−1, respectively. Using the Marsh model, individual ratios of measured-to-calculated concentrations indicated that measured propofol plasma concentrations were underestimated, especially after induction of anaesthesia. The Schnider and Eleveld models tended to overestimate measured plasma concentration throughout anaesthesia (Fig. 1). In our patients, prediction of actual measured propofol plasma concentration was slightly better with the Eleveld model than with either the Marsh or Schnider models. Presumably performance was better because Eveveld was able to incorporate more than 15 000 propofol concentrations from more than 1000 patients from 30 studies, making the results both robust and generalisable.4Eleveld D.J. Colin P. Absalom A.R. Struys M.M.R.F. Pharmacokinetic–pharmacodynamic model for propofol for broad application in anaesthesia and sedation.Br J Anaesth. 2018; 120: 942-959Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar The Eleveld model included covariates such as age, weight, height, and sex to predict arterial propofol concentrations. Nevertheless, the MDAPE of ∼22% was only slightly below those of the previous models. The clinical relevance of this small difference is likely to be low. Complexities specific to our patient population might possibly have led to increased MDAPE in our validation study. It seems unlikely that it will be possible to develop a pharmacokinetic model for propofol with MDAPE <20% using currently available covariates, largely because of time-varying distribution volumes and metabolic differences between individuals. The metabolism of propofol occurs predominantly via cytochrome P450 2B6, and is subject to individual genetic influence, alcohol consumption, and use of various medications, none of which is adequately included in current pharmacokinetic models. However, Eleveld's model is based on the largest data set and, as we have shown, has the lowest prediction error so far. Furthermore, Eleveld improved model performance at the extremes of age and in the obese.4Eleveld D.J. Colin P. Absalom A.R. Struys M.M.R.F. Pharmacokinetic–pharmacodynamic model for propofol for broad application in anaesthesia and sedation.Br J Anaesth. 2018; 120: 942-959Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar This model brings a significant gain in predicting plasma concentrations, especially in these patient groups. Thus, it should be integrated into commercial syringe pumps for TCI. The major limitation of our study is the low power with only a very small effect size. Small differences in the outcome parameter MDAPE between the three pharmacokinetic models would require a case number of nearly 700 patients to get a power of 0.8. In conclusion, the Eleveld pharmacokinetic model showed a slightly better predictive performance for measured propofol plasma concentrations compared with those of the Marsh and Schnider models. Further studies with significantly more patients must confirm these preliminary data to clarify whether this small difference from other pharmacokinetic models is clinically relevant. Performed the intraoperative measurements and wrote the manuscript: TH Measured and evaluated the blood plasma samples: FM Carried out the conceptualisation, formal analysis, data correction, and manuscript preparation: DIS Reviewed and edited the manuscript: TV Designed the manuscript and performed the pharmacokinetic calculations: SK The liquid chromatograph was a loan from B. Braun Melsungen (Melsungen, Germany). The authors declare that they have no conflicts of interest. The evolution of pharmacokineticsBritish Journal of AnaesthesiaVol. 124Issue 6PreviewA recent issue of the British Journal of Anaesthesia contains a landmark Letter to the Editor, 'Retrospective comparison of Eleveld, Marsh and Schnider propofol pharmacokinetic models in 50 patients'.1 A landmark Letter to the Editor? How can that be? Full-Text PDF Open Archive
Referência(s)