Pharmacokinetics Of Busulfan In Hematopoietic Stem Cell Transplantation Patients And Application Of Limited Sampling Strategy In Therapeutic Drug Monitoring Of Busulfan

Objectives1. To establish a high-performance liquid chromatography–mass spectrometry(LC-MS/MS) method for determining concentration of busulfan in human plasma. 2. To evaluate the intra-patient variability of Bu’s pharmacokinetic parameters in HSCT patients such as area under the concentration-time curve(AUC), maximal busulfan concentration(Cmax) and clearance rate(CL). 3. Analysis the correlation between AUC and clinical efficacy, and obtain the AUC range of Bu for HSCT patients. 4. To establish a limited sampling strategy model for HSCT patients who have received Bu intravenous infusion.Method1. The LC-MS/MS method was used to measure the blood concentration of busulfan. Chromatographic conditions: Agilent ZORBA SB-C18 chromatographic column(5μm, 150mm×4.6mm), 10mmol·L-1 ammonium acetate: methanol(60:40, v/v), flow rate of 0.5m L·min-1, column temperature of 35℃, sample size of 10μL, busulfan-d8 as the internal standard Mass spectrum conditions: electrospray ion source(ESI), multiple reaction monitoring(MRM), selected ion monitoring(SIM) for detecting positive ions, busulfan( [M+H]+,m/z151)and busulfan- d8( [M+H]+,m/z159)for detection of objects, atomization pressure of 45 PSI, dry gas flow rate of 10L·min-1, capillary voltage of 3.5k V, decomposition voltage of 70 V, collision energy of 5V. 2. Blood samples were collected from HSCT patients at specific time points,namely, 0h before the first dose of Bu-based regimen, 1h, 2h, 3h, 4h, 5h, 6h after the first dose. The proposed LC-MS/MS method was used to measure the blood concentration of busulfan. DAS 2.0 was used to do the statistical analysis, fit the blood concentration-time curve, determine the compartment model, calculate AUC,t1/2α,V1, Cmax, CL and other pharmacokinetic parameters and their variability. 3. SPSS20.0 was used to examine the influence of age, gender, disease type and other factors on pharmacokinetic parameters of Bu. Locally weighted scatterplot smoothing(LOWESS) was used to evaluate the correlation between AUC and clinical efficacy. Logistic regression was used to assess the correlation between AUC and the incidence of toxicity and side effects, and thus obtain the most suitable AUC range of Bu for Chinese patients. 4. SPSS20.0 was used to build a multiple regression model based on data points about the blood concentration of Bu. Internal validation was performed over the model using the Jackknife and Bootstrap methods. The consistency between results produced by the traditional method and the proposed LSS respectively was assessed by the intraclass correlation coefficient(ICC) and Bland-Altman(BA) analysis.Results1. In the LC-MS/MS method to measure blood concentration of busulfan, the standard curve equation was y=1.902x+0.010, correlation coefficient r=0.996; linear range 0.021-10.435μmol·L-1; intra-day Bu RSD≤7.1%, inter-day Bu RSD≤3.8%; methodological recovery of 98.0%-104.8%; extraction recovery of 105.1±2.2%, 106.8±2.4% and 99.0±2.0%. Endogenous substances in blood plasma did not affect the measuring of busulfan. 2. The plasma concentration-time curves after the first dose intravenous busulfan in 30 patients were fitted to a two-compartment model. AUC(0-t) and AUC(0-∞) of Bu in HSCT patients vary in the range 356.80-946.91μmol·L-1 and 457.04-1379.83μmol·L-1, respectively. t1/2α, V1, K10, K12, and K21 had a large variability. 3. This study found significant difference in Cmax between male and female patients;interclass t1/2β among "AML/ALL", "CML" and "MDS"; t1/2α, K12 and K21 among "Arac、Flu、Bu、Cy", "Arac、Bu、Cy", “Flu、Bu” and "Arac、Flu、Cy、Bu". Meanwhile, with the recovery time of hematopoietic function in post-transplant patients being the indicator of short-term clinical efficacy, LOWESS was used to analyze the correlation between AUC and the recovery time. In consideration of stability, the optimal AUC range 550-850μmol·min·L-1 was obtained. The results of logistic regression analysis showed that there was no significant correlation between AUC and the incidence of toxicity and side effects while the incidence of GVHD was related to age(P=0.034). 4. The pharmacokinetic parameters regression model based on two points(C3, C1) had a good linear relativity, with average prediction error of 0.25%, root mean square error of 4.83%, 95% confidence interval of ICC being 0.932-0.984, the limit of agreement in the BA-9.7-9.4. Results of the proposed LSS method were consistent with those produced by the classic method.Conclusions1. This LC-MS/MS method to measure blood concentration of busulfan is fast, convenient, accurate and applicable in clinical routine tests. 2. Pharmacokinetic parameters of the fixed dosage regimen have a large variability, which is affected by gender, disease type, and preparative regimen. The optimal AUC range for HSCT patients obtained by using short-term clinical efficacy and LOWESS needs further verification. 3. The LSS model established in this study is highly reliable which has improved therapeutic drug monitoring and provided HSCT patients with valuable clinical reference.