Determination Of Caffeine And Its Three Primary Metabolites In Premature Infants And Its Population Pharmacokinetics Study

Objective:To develop an ultra-high performance liquid chromatography-tandem mass spectrometry(UHPLC-MSIMS)method for the determination of caffeine and its three primary metabolites in human serum,and then use this method to analyse the clinic scavenged blood samples from routine biochemical tests of apnea of prematurity(AOP)treated with caffeine citrate.Concentration data of caffeine,administration information,blood sampling time,the pathological and physiological data and genotype of CYP1A2 were collected.Then select the appropriate covariate to establish for a population pharmacokinetic(PPK)model of caffeine in premature infants by the nonlinear mixed effects modeling(NONMEM),which provides the reference for safe and reasonable clinical medication.Methods:?The establishment and validation of UHPLC-MS/MS for the determination of caffeine and its three primary metabolites in human serum.Caffeine-d9 was selected as the internal standard.Aliquot of 50?L serum was placed in a 1.5 mL centrifuge tube.Methanol solution of 200 ?L containing caffeine-d9 was added precisely.The protein was precipitated by vortex oscillation for 1 min and standing for 5 min.The mixture was centrifuged at 13000 rpm for 10 min,and the supernatant was analyzed by UHPLC-MS/MS.The pretreatment process of standard curve and quality control(QC)was same as the samples.Chromatographic column:Agilent C18 column(2.1 mm×50 mm.2.2?m);Mobile phase:gradient elution of aqueous solutioin(including 0.1%formic acid.A)and methanol solution(including 0.1%formic acid,B);Flow rate:0.3 mL/min.Using electrospray ionization(ESI),positive ion mode,multiple reaction monitoring(MRM)scanning detection.?Establishment and validation of PPK model of caffeine in premature infants.Accurate administration time and dose,blood sampling time and pathological and physiological data were collected.At the same time,the concentrations of caffeine and its three primary metabolites from scavenged blood samples were determined by UHPLC-MS/MS.First order conditional estimation with interaction(FOCEI)method was used to estimate pharmacokinetic parameters and their variability by NONMEM program.Inter-individual variability and residual variability were both described using appropriate model.The potential covariates were investigated by a forward and backward selection process.The likelihood ratio test(LRT)was used to test the effect of each covariate on model parameters.The final PPK model was established and model validation was based on graphical and statistical criteria.Results:?The determination method of UHPLC-MS/MS was established.The linear relationships of caffeine,theophylline,1,7-dimethylxanthine and theobromine were good in the range of 200-75000 ng/mL,10-25000 ng/mL,10-10000 ng/mL and 10-2500 ng/mL,respectively.The relative standard deviations(RSDs)of intra-day and inter-day precision were both less than 7.9%.The average recoveries were 89.2%-109.4%.The analysis method showed good selectivity and high sensitivity.The matrix effects,extraction recoveries and process efficiencies were all in accordance with the requirements.The QC samples placed at room temperature for 24 h,4? for 2 d,-80?for 16 d and 3 times freezing and thawing cycles were all stable.?It showed that one-compartment model with first-order elimination was optimal for pharmacokinetic data.Inter-individual variability and residual variability were both best described using exponential model in basic model.The allometric size approach was used by incorporating a priori the current weight(CW)into the basic model(allometric coefficients of 0.75 for CL,1 for Vd),which caused a significant drop in the objective function value(OFV)of 117.88 units(P<0.05).Postmenstrual age(PMA)was associated with a drop in the OFV of 8.49 units(P<0.05)on CL.A further decrease in the OFV of 5.99 units(P<0.05)was achieved by implementing serum creatinine concentration(CREA)on CL.By backward selection,it was shown that both PMA and CREA had significant effects on OFV(P<0.05).Genotype of CYP1A2 and other covariates were implemented but did not result in a further decrease in the OFV.The final PPK model was described as follows:CL(L/h)=?1 ×(CW/1280)0.75×(PMA/31.1)?3 × 1/(CREA/68)?4 Vd(L)=?2x(CW/128O) CL:clearance;Vd:apparent volume of distribution;CW:current weight in gram;PMA:postmenstrual age in weeks;CREA:serum creatinine concentration in ?mol/L.In our population,1280 gram,31.1 weeks and 68 ?mol/L are the median values of CW,PMA and CREA,respectively.?1 and ?2 are the population typical values of CL and Vd,respectively.?3 and ?4 are the fixed effect influencing factors of PMA and CREA,respectively.Conclusion:A rapid and sensitive analysis method of UHPLC-MS/MS was developed.It showed that the method was suitable for the determination of caffeine and its three primary metabolites in scavenged blood samples from routine biochemical tests of premature infants.A population pharmacokinetic model of caffeine was established in premature infants by NONMEM program.In the final PPK model,the clearance(CL)of caffeine was correlated with postmenstrual age(PMA),current weight(CW)and serum creatinine concentration(CREA),and the apparent volume of distribution(Vd)was only correlated with current weight(CW).Genotype of CYP1A2 and other covariates were tested with no effect to population pharmacokinetic parameters.The final PPK model shows a good accuracy,applicability and stability by model validation based on graphical and statistical criteria.It could provide a reference for the safe and reasonable medication of premature infants.