Increasing the utility of target-controlled infusions: one model to rule them all
2018; Elsevier BV; Volume: 120; Issue: 5 Linguagem: Inglês
10.1016/j.bja.2018.02.012
ISSN1471-6771
AutoresTimothy G. Short, Douglas Campbell, Talmage D. Egan,
Tópico(s)Antibiotics Pharmacokinetics and Efficacy
Resumo‘Essentially, all models are wrong, but some are useful’.1Box G.E.P. Draper N.R. Empirical model-building and response surfaces. Wiley, New York1987: 424Google Scholar Target-controlled infusion (TCI) for anaesthesia became available for widespread clinical use in 1996 with the commercial launch of Diprifusor™ (AstraZeneca UK Ltd, Macclesfield, Cheshire, England).2Glen J.B. The development and regulation of commercial devices for target-controlled drug infusion.in: Absalom A.R. Mason K.P. Total intravenous anaesthesia and target controlled infusions. Springer, Basel, Switzerland2017: 9-29Crossref Google Scholar The pharmacokinetic model incorporated into the syringe pump was a modified version of a three-compartment model derived from a study of propofol infusions in 16 ASA Physical Status I and II patients, 12 of whom had also received spinal anaesthesia for their procedure.3Marsh B. White M. Morton N. Kenny G.N.C. Pharmacokinetic model driven infusion of propofol in children.Br J Anaesth. 1991; 67: 41-48Abstract Full Text PDF PubMed Scopus (942) Google Scholar, 4Gepts E. Camu F. Cockshott I.D. Douglas E.J. Disposition of propofol administered as constant rate intravenous infusions in humans.Anesth Analg. 1987; 66: 1256-1263Crossref PubMed Scopus (397) Google Scholar Known as the Marsh model (from Marsh and colleagues3Marsh B. White M. Morton N. Kenny G.N.C. Pharmacokinetic model driven infusion of propofol in children.Br J Anaesth. 1991; 67: 41-48Abstract Full Text PDF PubMed Scopus (942) Google Scholar), it included an adjustment of the central volume of distribution for weight (Table 1). It had been prospectively validated in 46 patients aged 18–80 yr undergoing non-cardiac surgery and 20 cardiac-surgery patients, was approved for use in patients more than 16 yr of age, and found to have acceptable accuracy. When used for propofol infusion in healthy adults, the Diprifusor pump rapidly produces a pseudo-steady-state infusion that mimics the simplicity of using a vapouriser.Table 1Number and diversity of volunteers and patients included in some target-controlled-infusion models for propofol and remifentanilModelNumber of patientsAge range (yr)Weight range (kg)Number of blood samplesNumber of model parametersPropofol Marsh3Marsh B. White M. Morton N. Kenny G.N.C. Pharmacokinetic model driven infusion of propofol in children.Br J Anaesth. 1991; 67: 41-48Abstract Full Text PDF PubMed Scopus (942) Google Scholar, 4Gepts E. Camu F. Cockshott I.D. Douglas E.J. Disposition of propofol administered as constant rate intravenous infusions in humans.Anesth Analg. 1987; 66: 1256-1263Crossref PubMed Scopus (397) Google Scholar1625–6548–844817 Schnider5Schnider T. Minto C. Gambus P. 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 (899) Google Scholar2425–8144–123100611 Eleveld6Eleveld 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 (117) Google Scholar10330.5–820.68–16015 43318Remifentanil Minto7Minto 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 (965) Google Scholar6020–8545–106199211 Eleveld8Eleveld D.J. Proost J.H. Vereecke H. et al.An allometric model of remifentanil pharmacokinetics and pharmacodynamics.Anesthesiology. 2017; 126: 1005-1018Crossref PubMed Scopus (58) Google Scholar1315 days–852–106263412 Kim9Kim T.K. Obara S. Egan T.D. et al.Disposition of remifentanil in obesity: a new pharmacokinetic model incorporating the influence of body mass.Anesthesiology. 2017; 126: 1019-1032Crossref PubMed Scopus (24) Google Scholar22920–8545–21544559 Open table in a new tab In 2003, a new generation of programmable syringe pumps became available. Alternative models for infusion of propofol could now be used, such as the more complex Schnider model (from Schnider and colleagues5Schnider T. Minto C. Gambus P. 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 (899) Google Scholar) for use in adults and the Kataria model (from Kataria and colleagues10Kataria B.K. Ved S.A. Nicodemus H.F. et al.The pharmacokinetics of propofol in children using three different data analysis approaches.Anesthesiology. 1994; 80: 104-122Crossref PubMed Scopus (280) Google Scholar) for use in children. The Schnider model was derived from 24 volunteers, aged 25–81 yr, who were intensively studied with both a bolus dose and an infusion on two occasions, and includes adjustments for age, weight, height, and sex. The model reflected the significant reduction in clearance of drugs from the blood that occurs with age. Most validation studies have found the Schnider model to be more accurate than the Marsh model in clinical practice.6Eleveld 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 (117) Google Scholar When we create a mathematical model for a drug, we create a description of the typical disposition or effects of the drug in the population of patients we have studied. Numerous studies have been published describing the use of these models in various subsets of patients that we meet in our daily practice that are distinct from the population they were derived in. Not surprisingly, the models have often been found to be less accurate when used in these different patient populations.11Sahinovic M.M. Eleveld D.J. Miyabe-Nishiwaki T. Struys M.M.R.F. Absalom A.R. Pharmacokinetics and pharmacodynamics of propofol: changes in patients with frontal brain tumours.Br J Anaesth. 2017; 118: 901-909Abstract Full Text Full Text PDF PubMed Scopus (11) Google Scholar, 12Lee Y.H. Choi G.H. Jung K.W. et al.Predictive performance of the modified Marsh and Schnider models for propofol in underweight patients undergoing general anaesthesia using target-controlled infusion.Br J Anaesth. 2017; 118: 883-891Abstract Full Text Full Text PDF PubMed Scopus (18) Google Scholar, 13Hara M. Masui K. Eleveld D.J. Struys M.M.R.F. Uchida O. Predictive performance of eleven pharmacokinetic models for propofol infusion in children for long-duration anaesthesia.Br J Anaesth. 2017; 118: 415-423Abstract Full Text Full Text PDF PubMed Scopus (17) Google Scholar In this edition of the British Journal of Anaesthesia, a new general model for propofol is presented by Eleveld and colleagues6Eleveld 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 (117) Google Scholar (hence known as the Eleveld model). It is based on a pooled analysis of 30 pharmacokinetic studies and five pharmacodynamic studies that had also used the bispectral index as a measure of hypnosis. The study includes data from 1033 patients aged 0.5–88 yr, and attempts to provide one model for propofol that is derived from a diverse group of patients more representative of everyday practice. It includes adjustments for age, weight, height, and sex, and for the presence or absence of opioids. It has been specifically calibrated to be useful for providing both anaesthetic and sedative concentrations of propofol. Most readers will probably find the pharmacokinetic analysis arcane, and indeed, the model contains 18 parameters. However, the logic of systematically adding parameters and seeing whether they significantly improve model performance is valid and well presented for the sceptical reader to review. A conservative approach to adding new model parameters has been taken, ensuring that there is value to the additional complexity in the model. Extensive data have been provided on the internal validity of the model in the patient cohorts from which they were derived. Indeed, the greatest strength of this study is that the derivation cohort is much larger, has more varied patient groups, and is geographically diverse. Internal validation has found the model to work as well as, or better than, the models derived from niche studies in all the common subgroups of patients included in it. The study represents a milestone in pharmacokinetic and pharmacodynamic analysis in the anaesthetic literature for propofol. Adjustments in drug dosage required for age, weight, and obesity are characterised. Because of its complexity and the diverse population of patients that it embraces, external validation in clinically diverse groups of patients is still absolutely warranted, as predictive models can demonstrate poor calibration when exported to geographically or ethnically distinct populations.14Liu J. Hong Y. D'Agostino R.B. et al.Predictive value for the Chinese population of the Framingham CHD risk assessment tool compared with the Chinese multi-provincial cohort study.JAMA. 2004; 291: 2591-2599Crossref PubMed Scopus (511) Google Scholar However, experience with validation studies of pharmacological models suggests this new model is very unlikely to perform worse than current models, although only a prospective study can conclusively demonstrate this. Also, it is fair to say, when anaesthetising a typical adult, the model may perform no better than the current models in terms of the pharmacokinetic or clinical outcome. By being more generalisable than other propofol TCI models, the model proposed by Eleveld and colleagues6Eleveld 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 (117) Google Scholar should fulfil the requirements of a busy anaesthetist faced with a diverse group of patients to sedate or anaesthetise. One model that can be trusted to deliver stable propofol concentrations to most patients will be simpler and more useful. The model is more complicated than previous models, but the additional parameters have all been justified by the authors based on statistical properties and physiological and pharmacological plausibility. Whilst we have instincts to trust simple models that are more easily understood, this should not be at the expense of accuracy. When using computer-controlled pumps, this complexity is hidden from the clinician, and the Eleveld model contains only one more input parameter than the Schnider model, namely, whether opioids are being used, so should be similarly easy to use in practice. The Schnider model is certainly deeply entrenched clinically and has been well received amongst TCI users. Because total i.v. anaesthesia (TIVA) practitioners have been achieving excellent results with the Schnider model for more than a decade, there may be some hesitation to switch to a new model. Until there are some clinical experience and validation studies describing its performance, the Eleveld model usage may be limited to clinical situations, in which the Schnider model clearly has limitations, such as in children and obese patients. Interestingly, a prior pooled analysis of a propofol model with some similarities to the Eleveld study was never widely embraced in the clinical domain for reasons that are unclear.15Schüttler J. Ihmsen H. Population pharmacokinetics of propofol: a multicenter study.Anesthesiology. 2000; 92: 727-738Crossref PubMed Scopus (354) Google Scholar The study is similar to the two recently published pooled analyses of remifentanil pharmacokinetics. These analyses both addressed the demographic limitations of the currently used Minto model (from Minto and colleagues7Minto 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 (965) Google Scholar) for remifentanil. The study by Eleveld and colleagues8Eleveld D.J. Proost J.H. Vereecke H. et al.An allometric model of remifentanil pharmacokinetics and pharmacodynamics.Anesthesiology. 2017; 126: 1005-1018Crossref PubMed Scopus (58) Google Scholar essentially addressed the issue of age-related changes in remifentanil pharmacokinetics, creating a model for use in all age groups and with improved performance in obese patients. The second study, by Kim and colleagues9Kim T.K. Obara S. Egan T.D. et al.Disposition of remifentanil in obesity: a new pharmacokinetic model incorporating the influence of body mass.Anesthesiology. 2017; 126: 1019-1032Crossref PubMed Scopus (24) Google Scholar (hence known as the Kim model), specifically addressed the issue of accounting for obesity in adult patients. Disappointingly, there is currently no pooled analysis that embraces all the major studies of remifentanil to enable one model to be used confidently for remifentanil in all patients. Having said that, there are also very little pharmacokinetic data available on remifentanil in paediatric patients, and both these new models will likely perform well in adult patients of all ages and sizes. Like the Eleveld model for propofol, we expect that these new remifentanil models will also eventually be available in TCI pumps. What is lacking from these TCI models for propofol and remifentanil? All the volunteers and most of the patients studied were relatively healthy. Care must be applied when using them in the presence of significant renal, liver, cardiovascular, or brain disease, or acute trauma. The models also do not account fully for drug interactions, particularly interactions between remifentanil and propofol.16Bouillon T.W. Bruhn J. Radulescu L. et al.Pharmacodynamic interaction between propofol and remifentanil regarding hypnosis, tolerance of laryngoscopy, bispectral index, and electroencephalographic approximate entropy.Anesthesiology. 2004; 100: 1353-1372Crossref PubMed Scopus (282) Google Scholar The disposition of propofol is known to be altered by the presence of opioids, but the simple calculation used in the new Eleveld model is based on typical ‘low-dose’ opioid anaesthesia. What are the implications of the Eleveld model for non-TCI users? This is an important question, because there are still countries, most notably the USA, where TCI is not yet approved by regulatory agencies for clinical use.17Shafer S.L. Egan T. Target-controlled infusions: surfing USA redux.Anesth Analg. 2016; 122: 1-3Crossref PubMed Scopus (15) Google Scholar, 18Egan T.D. Shafer S.L. Target-controlled infusions for intravenous anesthetics: surfing USA not!.Anesthesiology. 2003; 99: 1039-1041Crossref PubMed Scopus (55) Google Scholar In countries where TCI is approved, the technology is presumably not available at every anaesthetising location in every hospital or ambulatory surgery centre, and some TIVA practitioners may still prefer to use standard calculator pumps without TCI capability. It is true that the full clinical utility of the Eleveld model can only be realised using a TCI pump. Without the aid of a computer model, it is impossible for a clinician to implement in real time the complex covariate effects incorporated into advanced pharmacokinetic models. When covariates, such as body weight or age, are related to a pharmacokinetic model parameter, such as clearance in a mathematical equation, the covariate effect typically influences the dosing scheme in a complex way that varies over time.19Egan T.D. Target-controlled drug delivery: progress toward an intravenous “vaporizer” and automated anesthetic administration.Anesthesiology. 2003; 99: 1214-1219Crossref PubMed Scopus (79) Google Scholar The clinician is no match for the computer when it comes to solving polyexponential equations. Nonetheless, many of the essential findings of complex pharmacokinetic models can be applied by the practitioner using standard infusion pumps. In the case of the Eleveld model, for example, the practitioner can formulate a maintenance infusion dosing strategy knowing that women require slightly higher doses than men, that children require substantially higher doses than the elderly, that very obese patients require lower doses on a per-kilogram basis than lean patients, amongst others. Thinking of the lean, middle-aged adult as the standard patient, dosing adjustments for the young, the elderly, or the obese can be made in a semi-quantitative way leveraging the Eleveld and colleagues findings6Eleveld 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 (117) Google Scholar. Interestingly, such an approach requires perhaps a more careful reading of the study by Eleveld and colleagues by non-TCI users6Eleveld 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 (117) Google Scholar. For the TCI enthusiasts, the technology simplifies an otherwise very complex situation, as technology should. Indeed, practically speaking, the trusting TCI user does not need to read the paper, although we do not recommend that approach. We expect these new models for propofol and remifentanil to be available in our programmable syringe pumps for clinical use in the future. It will be important that anaesthetists read these articles carefully and understand the models and their limitations before using them. We must also thank all the researchers who have freely made their data available so that these pooled analyses could be done, thereby improving the utility of our pharmacokinetic models. Proposal for writing: T.S. Writing of manuscript and approval of final version: all authors. T.D.E. is a member of the associate editorial board of the British Journal of Anaesthesia. Pharmacokinetic–pharmacodynamic model for propofol for broad application in anaesthesia and sedationBritish Journal of AnaesthesiaVol. 120Issue 5PreviewPharmacokinetic (PK) and pharmacodynamic (PD) models are used in target-controlled-infusion (TCI) systems to determine the optimal drug administration to achieve a desired target concentration in a central or effect-site compartment. Our aim was to develop a PK–PD model for propofol that can predict the bispectral index (BIS) for a broad population, suitable for TCI applications. Full-Text PDF Open Archive
Referência(s)