TY - JOUR
T1 - Drug clearance in neonates
T2 - a combination of population pharmacokinetic modelling and machine learning approaches to improve individual prediction
AU - Tang, Bo Hao
AU - Guan, Zheng
AU - Allegaert, Karel
AU - Wu, Yue-E.
AU - Manolis, Efthymios
AU - Leroux, Stephanie
AU - Yao, Bu-Fan
AU - Shi, Hai-Yan
AU - Li, Xiao
AU - Huang, Xin
AU - Wang, Wen-Qi
AU - Shen, A.-Dong
AU - Wang, Xiao-Ling
AU - Wang, Tian-You
AU - Kou, Chen
AU - Xu, Hai-Yan
AU - Zhou, Yue
AU - Zheng, Yi
AU - Hao, Guo-Xiang
AU - Xu, Bao-Ping
AU - Thomson, Alison H.
AU - Capparelli, Edmund V.
AU - Biran, Valerie
AU - Simon, Nicolas
AU - Meibohm, Bernd
AU - Lo, Yoke-Lin
AU - Marques, Remedios
AU - Peris, Jose-Esteban
AU - Lutsar, Irja
AU - Saito, Jumpei
AU - Burggraaf, Jacobus
AU - Jacqz-Aigrain, Evelyne
AU - van den Anker, John
AU - Zhao, Wei
PY - 2021/5/27
Y1 - 2021/5/27
N2 - Background: Population pharmacokinetic evaluations have been widely used in neonatal pharmacokinetic studies, while machine learning has become a popular approach to solving complex problems in the current era of big data. Objective: The aim of this proof-of-concept study was to evaluate whether combining population pharmacokinetic and machine learning approaches could provide a more accurate prediction of the clearance of renally eliminated drugs in individual neonates. Methods: Six drugs that are primarily eliminated by the kidneys were selected (vancomycin, latamoxef, cefepime, azlocillin, ceftazidime, and amoxicillin) as 'proof of concept' compounds. Individual estimates of clearance obtained from population pharmacokinetic models were used as reference clearances, and diverse machine learning methods and nested cross-validation were adopted and evaluated against these reference clearances. The predictive performance of these combined methods was compared with the performance of two other predictive methods: a covariate-based maturation model and a postmenstrual age and body weight scaling model. Relative error was used to evaluate the different methods. Results: The extra tree regressor was selected as the best-fit machine learning method. Using the combined method, more than 95% of predictions for all six drugs had a relative error of < 50% and the mean relative error was reduced by an average of 44.3% and 71.3% compared with the other two predictive methods. Conclusion: A combined population pharmacokinetic and machine learning approach provided improved predictions of individual clearances of renally cleared drugs in neonates. For a new patient treated in clinical practice, individual clearance can be predicted a priori using our model code combined with demographic data.
AB - Background: Population pharmacokinetic evaluations have been widely used in neonatal pharmacokinetic studies, while machine learning has become a popular approach to solving complex problems in the current era of big data. Objective: The aim of this proof-of-concept study was to evaluate whether combining population pharmacokinetic and machine learning approaches could provide a more accurate prediction of the clearance of renally eliminated drugs in individual neonates. Methods: Six drugs that are primarily eliminated by the kidneys were selected (vancomycin, latamoxef, cefepime, azlocillin, ceftazidime, and amoxicillin) as 'proof of concept' compounds. Individual estimates of clearance obtained from population pharmacokinetic models were used as reference clearances, and diverse machine learning methods and nested cross-validation were adopted and evaluated against these reference clearances. The predictive performance of these combined methods was compared with the performance of two other predictive methods: a covariate-based maturation model and a postmenstrual age and body weight scaling model. Relative error was used to evaluate the different methods. Results: The extra tree regressor was selected as the best-fit machine learning method. Using the combined method, more than 95% of predictions for all six drugs had a relative error of < 50% and the mean relative error was reduced by an average of 44.3% and 71.3% compared with the other two predictive methods. Conclusion: A combined population pharmacokinetic and machine learning approach provided improved predictions of individual clearances of renally cleared drugs in neonates. For a new patient treated in clinical practice, individual clearance can be predicted a priori using our model code combined with demographic data.
KW - drug clearance
KW - neonates
KW - combination
KW - population pharmacokinetic modelling
KW - machine learning
KW - individual prediction
UR - http://www.scopus.com/inward/record.url?scp=85106464544&partnerID=8YFLogxK
U2 - 10.1007/s40262-021-01033-x
DO - 10.1007/s40262-021-01033-x
M3 - Article
AN - SCOPUS:85106464544
SN - 0312-5963
VL - 60
SP - 1435
EP - 1448
JO - Clinical Pharmacokinetics
JF - Clinical Pharmacokinetics
IS - 11
ER -