【作者】 张菁；

【导师】 张婴元； 施耀国； 芮建中； 赵耐青； 吴菊芳；

【作者基本信息】 复旦大学 ， 内科学， 2005， 博士

【摘要】 当前，在感染性疾病诊治领域中，耐药菌感染的治疗已成为难题，高度耐药的甲氧西林耐药金葡菌（methicillin-resistant Staphylococcus aureus，MRSA）等革兰阳性球菌感染即是其中之一。去甲万古霉素、万古霉素等糖肽类抗生素是治疗MRSA感染的重要治疗药物，然而该类药物治疗指数低，治疗窗狭窄，有肾、耳毒性，在一些特殊生理、病理情况下患者（老年患者、肾功能减退患者）中尤易发生毒性反应，因此有必要在感染患者中进行去甲万古霉素药物动力学研究以了解不同的患者群体中因生理、病理及基础疾病不同导致的去甲万古霉素体内过程的差异，即开展群体药物动力学（population pharmacokinetics，PPK）研究。近年来在临床药物动力学基础上发展起来的群体药物动力学已在发达国家中开始应用于临床治疗学领域的研究。我们研究所在对去甲万古霉素等数十种抗菌药开展临床药物动力学研究的基础上计划引进国外技术，建立去甲万古霉素群体药物动力学模型，并据此制订去甲万古霉素应用于不同患者群体的安全而有效的治疗方案。群体药物动力学研究设计在耐药革兰阳性菌感染，需应用去甲万古霉素进行治疗的患者群体中进行。为开展PPK研究尚需建立去甲万古霉素血药浓度测定方法，由于此次PPK研究系在患者中进行，不仅样本量大，患者又具有各种不同的原发病，感染病情亦常较严重，又常合并用其他治疗药物，因此对血药浓度测定的要求是快速、准确、特异性高，取血点少。为此我们建立了去甲万古霉素快速血药浓度测定方法—荧光偏振免疫法（fluorescence polarization immunoassay，FPIA法），并对该方法进行了方法学考核及验证。在获得PPK参数后，将结合去甲万古霉素对患者病原菌MIC测定的资料（药效学pharmacodynamics，PD）进行PPK／PD研究，以制订用于耐药革兰阳性菌不同菌种的各患者群体的优化给药方案，最终达到提高抗感染治疗疗效以及安全性的目的。本课题研究内容包括以下三部分。第一部分去甲万古霉素快速血药浓度监测方法的建立及验证去甲万古霉素临床应用至今，缺乏快速血药浓度测定方法以监测感染患者的治疗药物浓度，现有的去甲万古霉素微生物测定法虽灵敏度和准确性较高，但耗时长和特异性差，易受合并用药的影响，因此不适合用于该药的治疗药物浓度监测（therapeutic drug monitoring，TDM）。依据去甲万古霉素与万古霉素结构相似的特点，选择在已有的万古霉素FPIA测定方法的基础上，利用现有的万古霉素试剂盒和荧光偏振免疫仪（TDx），建立去甲万古霉素血药浓度FPIA方法，方法的建立在20名健康志愿者中进行，并与传统的去甲万古霉素微生物测定法和高效液相色谱法（high pressure liquid chromatography，HPLC法）比较该测定方法的准确性、重复性、灵敏度、特异性以及所建立的去甲万古霉素FPIA方法与微生物法和HPLC法的相关性。建立步骤为测定10名健康年轻志愿者和10名健康老年志愿者静脉滴注去甲万古霉素800mg后于不同时间点采集的239份血清样本中的药物浓度，同时以微生物法测定上述血样，并以其测得结果作测定标准，建立微生物法与FPIA法测定结果间线性回归方程：Y=0.7534X—0.5948，其中X为FPIA法测得值，Y为微生物法测得值，R²=0.9703，即可采用FPIA法测定结果借助回归方程间接测定去甲万古霉素的血药浓度。去甲万古霉素FPIA法的日内和日间的变异系数（RSD）≤6.08％和≤4.75％，方法回收率为（87.74～114.34）％，并且不受联合应用β内酰胺类和氨基糖苷类等药物的影响，以该法与HPLC方法同时测定371份去甲万古霉素治疗浓度监测的血样，结果显示FPIA法经上述公式折算后的浓度与HPLC法测得的血药浓度结果相近，两种方法的线性回归方程：Y=1.016X+0.0041（R²=0.9782），其中X为FPIA法折算后的浓度，Y为HPLC法测得浓度。因此去甲万古霉素FPIA法可作为治疗药物监测的主要手段，并应用于去甲万古霉素PPK研究中感染患者及健康志愿者的血药浓度测定。第二部分去甲万古霉素群体药物动力学研究药物的群体药物动力学研究可全面地反映治疗药物在不同群体患者中个体过程的差异，此为制定治疗药物用于不同群体的合理治疗方案提供可靠的依据。目前采用NONMEM（nonlinear mixed effect model）法进行群体药物动力学参数的分析，NONMEM法优点为虽要求病例数多，但每名患者取血点少，适合于群体药物动力学模型的研究，同时NONMEM软件中POSTHOC子程序可估算患者个体药物动力学参数，因此可制定合理的个体化给药方案。为此我们采用NONMEM法建立感染患者中去甲万古霉素群体药物动力学模型。共收集了146例华山医院各临床科室诊断或拟诊为金葡菌等耐药革兰阳性菌感染患者因临床治疗使用去甲万古霉素后的血药浓度监测结果和临床资料，同时也将10名男性年轻健康志愿者和10名老年健康志愿者单剂静脉滴注去甲万古霉素800mg后于不同时间测得的血药浓度数据用于PPK研究，将上述两部分数据共166例作为去甲万古霉素群体药物动力学研究的数据集，其中110例数据用作建立去甲万古霉素群体药物动力学模型，研究患者或健康志愿者的性别、年龄、体重、肝、肾功能（Ccr值）及合并用药（氨基糖苷类药物和利尿药）等固定效应对该药药物动力学参数的影响以及患者个体间、个体内药物动力学参数的差异，筛选出对药物动力学参数有影响的固定效应，建立去甲万古霉素群体药物动力学模型，将其余56例数据用于对该模型验证。在判断模型的稳定性和有效性后，166例数据合并拟合，获得去甲万古霉素最终群体药物动力学参数。结果显示NONMEM法拟合去甲万古霉素基础药物动力学模型为线性二房室模型，最终群体药物动力学参数为：CL（Ccr≦85ml／min）、CL（Ccr＞85ml／min）、V₁、Q和V₂分别为2.54L／h、5.66L／h、23.89L、8.50L／h和21.70L，CL、V₁、Q和V₂的个体间变异分别为35.92％、11.40％、0、79.75％，残差误差为3.05mg／L。群体药物动力学研究结果显示：①肾功能减退对去甲万古霉素PK参数有显著影响，146例感染患者中14例肾功能重度减退者平均药物清除率（CL）为（0.23±0.09）L／h，平均消除半衰期(T_1／2β为（154.26±74.28）h；16例肾功能中度减退者平均CL为（2.17±0.95）L／h，平均T_1／2β为（22.86±20.73）h；45例肾功能轻度减退者平均CL为（4.01±1.27）L／h，平均T_1／2β为（9.57±4.34）h。结果提示肾功能呈轻、中、重度减退时，去甲万古霉素自体内消除明显减缓。②年龄对去甲万古霉素PK参数亦有显著影响，59例老年感染患者（≥65岁）与73例非老年感染患者（18～＜65岁）相比，前者CL减慢、T_1／2β延长，AUC增大；CL分别为（3.94±1.73）L／h和（5.89±2.08）L／h，T_1／2β分别为（12.07±12.19）h和（6.79±5.18）h：AUC₂₄分别为（490.16±564.36）mg.h／L和（283.92±161.66）mg.h／L。③去甲万古霉素合并用利尿药后使该药的周边室分布容积（V₂）增大，提示两者合用后可加快前者药物廓清速率，血药浓度降低。④患者的肝功能及合并用氨基糖苷类药物对去甲万古霉素PK参数无明显影响。根据研究结果获得了不同患者群体的主要PPK参数，并据此制订了去甲万古霉素在不同群体患者中的给药方案。第三部分去甲万古霉素群体药物动力学和药效学研究感染性疾病与其他疾病不同，它是外来病原微生物侵入人体引起的疾病，因此，治疗感染病不仅需了解抗感染药在人体内过程（临床药物动力学或PPK），尚需掌握该类药物作用于病原微生物的过程及其特点（pharmacodynamics，PD），只有将两者结合方能制定有效而安全的抗感染治疗方案，因此我们采用群体药物动力学模型所获得患者个体药物动力学参数，结合去甲万古霉素对患者自身病原菌的最低抑菌浓度（MIC）测定结果，计算患者PPK／PD参数，预测感染患者应用去甲万古霉素后临床和细菌学疗效，以进一步优化给药方案。本研究采用琼脂双倍稀释法测定去甲万古霉素等对获自71例患者的78株临床分离病原菌的MIC，并根据患者的药时曲线下面积(AUC₂₄)与其自身病原菌MIC进行AUIC值的计算(AUC₂₄／MIC)，同时计算患者血药浓度超过MIC的时间占给药间期的百分比（％T＞MIC）。为排除患者基础情况、合并用药等因素可能对去甲万古霉素疗效的影响，比较了葡萄球菌及肠球菌感染者经去甲万古霉素治疗后治愈组和未治愈组两组的情况，结果显示葡萄球菌感染者两组间仅有给药剂量、年龄、AUC₂₄和AUC₂₄／MIC的差异具有统计学意义（p＜0.05）。肠球菌感染者两组间仅AUC₂₄和AUC₂₄／MIC的差异具有统计学意义（p＜0.05）。多因素回归分析亦显示，仅有AUC₂₄／MIC是影响治愈率的主要因素。当葡萄球菌感染组的AUC₂₄／MIC平均值为579.90和肠球菌感染组的AUC₂₄／MIC平均值为637.67时，去甲万古霉素对患者的治愈率可达到95％。此提示AUIC值是影响去甲万古霉素疗效的主要PPK／PD参数，因此可以认为该参数可作为最佳治疗方案制定的重要参考指标，依据PPK／PD上述研究结果制定了适用于不同感染患者群体的最佳的给药方案。我们建立了去甲万古霉素快速血药浓度测定方法，为该药的TDM开展提供了技术保证。以NONMEM法建立了感染患者中去甲万古霉素的群体药物动力学模型，制定了去甲万古霉素在肾功能不同情况下的给药方案，并以PPK／PD理念对方案进行了优化，设计了安全有效的适用于不同感染患者群体的给药方案，应用于临床可使去甲万古霉素在感染患者体内最大地发挥杀菌作用，并最大程度降低毒性反应，以提高该药临床应用安全性和有效性。更多还原

【Abstract】 A significant challenge to treat infections with antibiotic-resistant bacteria, like highly resistant methicillin-resistant Staphylococcus aureus （MRSA） and other gram-positive bacteria, remains in the field of the current diagnosis and treatment of infectious diseases. Norvancomycin, vancomycin and other glycopeptide antibiotics are vital drugs to treat MRSA infections. However, the class of glycopeptide antibiotics has lower therapeutic index, a narrow therapeutic window associated with their nephrotoxicity and ototoxicity. Patients with special physiologic and pathogenic alternations, such as elderly patients and patients with renal function impairment, are extremely vulnerable to the adverse effects of the drugs. A study desire to investigate the variability of pharmacokinetics of norvancomycin in different populations of patients with the special physiology, pathology, and underlying disease that may lead variable kinetic behaves of this drug, so called population pharmacokinetics （PPK）. PPK based on clinical pharmacokinetics to determine a rational regimen for drugs with the low therapeutic index has been widely conducted in developed countries. Our institute has established a population pharmacokintic model of norvancomycin, based on study of our previous clinical pharmacokinetic practices of norvancomycin and more than ten other drugs. This study was a project of scheduled applications of new techniques from foreign counties. The model can be applied to estimate or predict an effective and safe regimen of norvancomycin for different subgroups of patients. The target population of this pharmacokinetic research was the patient with the infection caused by gram-position bacteria resistant to antibiotics and the patients treated with norvancomycin. An assay to determine serum concentrations of norvancomycin was requested to establish for PPK studies. Since the study carried out vast samples which were collected from patients who may have different underlying diseases and severe infections as well as receive multiple drugs. A high rapid, accurate, and specific assay was demanded, while a small volume of samples was essential for the assay. Thus, we set up a rapid assay to measure serum concentrations of norvancomycin: fluorescence polarization immunoassay （FPIA）. Moreover, the assay was strictly tested andvalidated. After obtaining PPK parameters, combined with the study of the pharmacodynamics （PD） of norvancomycin and the norvancomycin MIC of bacteria which caused this patient’s infection, PPK was linked to PD in order to design an optimized regimen for different subgroups of patients with infections caused by various Gram-positive bacteria resistant to antibiotics. This approach finally assured the efficacy and safety of anti-infection therapy. The thesis is composed of three parts.First part: Establishment and validation of a rapid assay to determine serum concentrations of norvancomycinAlthough norvancomycin has clinically been used over decades, a rapid assay to measure serum concentrations of norvancomycin is not available to monitor therapeutic drug concentrations in patients with infections. Even if bioassay for norvancomycin is sensitive and accurate, it is time-consuming and non-specific. Results of the assay may be influenced by other drugs which are co-administered to patients. Thus the bioassay is not appropriate for use of the therapeutic drug monitoring （TDM） practices. A FPLA assay commercially used for vancomycin was chosen to establish an assay for norvancomycin, based on the very similarity of the chemical structures of norvancomycin and vancomycin, and availability for the kit and instrument （TDx）. The assay was tested and validated in 20 healthy volunteers in comparison with a traditional bioassay and high pressure liquid chromatography （HPLC） to determine its accuracy, stability, sensitivity, and specificity. The relevance of this assay with other two assays was also evaluated. The establishment procedure of the FPIA method for norvancomycin was to assay drug concentrations of 239 different time-point serum samples from 10 young and 10 elderly volunteers; following an intravenous infusion of norvancomycin at a dose of 800 mg. simultaneously these samples were measured via the bioassay. The results of bioassay were used as the measurement standard to establish a lineal regression of two assays: Y= 0.7534X-0.5948, where X was the value from the FPIA method and Y was the value from the bioassay method, R² = 0.9703. This approach indirectly determined serum concentrations of norvancomycin, resulted from values derived from FPIA and the regression equation. Intraday and interday precision （RSD） of the FPLA method for norvancomycin was ≤ 6.08 % and ≤ 4.75, respectively. The range of the recoveries of this assay was from 87.74% to 114.34%. This assay was not influenced by serum concentrations of beta-lactam and aminoglycoside antibiotics when they were co-administered to patients. The assay was compared with HPLC via simultaneously determining 371 serum samples from patients in TDM practices. The comparison showed serumconcentrations determined by the FPIA method plus use of regression equation was very similar to those by the HPLC method. Y= 1.016X+0.0041, where X was the value from the FPIA method plus use of regression equation and Y was the value from the HPLC method, R² = 0.9782. Therefore, the FPIA method was able to be a major approach for TDM procedures. Eventually, it was utilized to measure serum concentrations of norvancomycin from patients in the PPK study and from healthy volunteers in clinical studies.Part two: Population pharmacokinetics （PPK） of norvancomycinThe objective of the PPK study is to completely illustrate the variability of pharmacokinetic of therapeutic drugs in a diverse population, and thus it provides a reliable approach to design a rational regimen in different groups of patients. Nonlinear mixed effect model （NONMEM） is utilized to PPK analysis. The advantages of this method requires sparse samples from each patient and suitable for the PPK modeling, although it demands a large number of patients to enroll a study. Moreover, the POSTHOC program of the NONMEM software allows estimating individual PK parameters and thus can be used to establish an individualized regimen. Therefore, NONMEM approach was chosen to establish a PPK model for patients given norvancomycin in this investigation. Data came from 146 patients diagnosed and suspected with the infection caused by staphylococcus aureus and other gram-positive bacteria resistant to antibiotics. Norvancomycin was administered to all of the patients. The data included TDM, demography, and clinical information associated with these patients. Simultaneously, the study also used different concentration-time-points data derived from volunteers of 10 young and 10 elderly men, following a single intravenous infusion of norvancomycin at 800 mg. All of a data set consisted of 166 cases. Among them, 110 cases were randomly selected to establish the model. The gender, age, body weight, liver function, and renal function （values of Ccr） and the administration of multiple drugs （like aminoglycoside and diuretics） were initially evaluated to search fixed factors that could alter pharmacokinetic parameters and cause interindividual and intraindividual variability in pharmacokinetic parameters. Fixed factors obtained from the screening were applied to establish PPK model. The model was validated via other 56 cases. After confirming the stabilities and effectiveness of the model, all of 166 subjects were pooled to obtain final PPK parameters. The results demonstrated a two compartmental model fit well the PPK model of norvancomycin via the NONMEM approach. The final PPK parameters were found: CL (Ccr ≤ 85 ml/min, CL （Ccr > 85 ml/min）. V₁, Q, and V₂ of norvancomycin was 2.54 L/h, 5.66 L/h, 23.89L. 8.50 L/h. and 21.70L, respectively. Interindividual variability in CL , V₁, Q and V₂ was35.92 %, 11.40%, 0 and 79.75%, respectively. Residual variability was 3.05 mg/L.The findings of the PPK study were followings as: ① the significant impact of renal function impairment on PK parameters of norvancomycin. Among 146 patients: the mean of the drug clearance （CL） and eliminate half-life (T_1/2β was （0.23 ± 0.09） L/h and （154.26 ± 74.28） h in 14 patients with severe renal impairment, （2.17 ± 0.95） L/h and （22.86 ± 20.73）h in 16 patients with moderate renal impairment, and （4.01 ± 1.27） L/h and （9.57 ± 4.34） h in 45 patients with mild renal impairment, respectively. The results of the study strongly suggested that the elimination of norvancomycin decreased in subjects with mild, moderate, and severe renal impairment. ② the important impact of age on PK parameters of norvancomycin. A comparison of 59 elderly patients （≥ 65 years） versus non-elderly patients （18～< 65 years） showed （3.94 ± 1.73 L/h ） versus （5.89 ± 2.08） L/h for CL, （12.07 ± 12.19 ）h versus （6.79 ± 5.18） h for T_1/2β, and （490.16 ± 564.36） mg.h/L versus （283.92 ±161.66） mg.h/L for values of AUC₂₄ Thus, the reduced drug clearance, prolonged T_1/2β, and increased values of AUC₂₄ were found in the elderly patients. ③ the increased volume of a peripheral distribution as coadministration of norvancomycin with diuretics. It suggested that the coadministration may accelerate the elimination of the former drug and thus decrease its serum concentrations. ④ non-influence of hepatic impairment and coadministration with aminoglycoside antibiotics on PK parameters of norvancomycin.Diverse regimens of norvancomycin were finally established to different patients on the basis of important PPK parameters generated from different groups of patients in this study.Part three: PPK/PD Study of norvancomycinInfectious diseases may differ from other diseases, since they are caused by invasion of microorganisms. Therefore, clinicians who prescribe antibiotics should have knowledge of the kinetics of drugs in body （clinical pharmacokinetics or PPK） and the effects and characteristics of antimicrobial activity against pathogens （Pharmacodynamics, PD）. The design of an effective and safe anti-infection regimen should integrate both principles. The last objective of the present study was to accurately calculate individual PPK/PD parameters in order to predict clinical and microbiologic outcomes in patients after the treatment of norvancomycin, and further to optimize the regimen, based on individual pharmacokinetic parameters generated from the PPK model and norvancomycin MIC values of the pathogens in this patient. The norvancomycin MICs of 78 clinically isolated strains from 71 patients were determined by an agar dilution method. The ratios (AUC₂₄/MIC) were calculated in 68 patients, where data of the area under time-concentration curve (AUC₂₄) and the value of MIC were derived from the individual patient. At the same time, calculations were conducted for the percentage of the time of serum concentrations above MIC during the interval of administration. The possibilities of the influence of underlying conditions and coadministration on efficacy of norvancomycin were excluded. The present study was performed in patients with infections caused by Staphylococcus aureus and Enterococcus species and treated with norvancomycin. The group of patients who had the outcomes of cure after the treatment was compared with the group of patients who had the outcomes of non-cure. The statistical study revealed that there were significant differences in the age, daily norvancomycin dose, AUC₂₄ and AUC₂₄/MIC between two group for patients with Staphylococcus aureus （p<0.05）.There was a significant difference in the AUC₂₄ and AUC₂₄/MIC between two group for patients with Enterococcus infections （p<0.05）. The logistic stepwise analyses for multiple factors on clinical outcome revealed only the ratio of AUC₂₄/MIC was the major factor which could significantly determine the clinical outcomes and bacterial eradication for patients with Staphylococcus aureus and enterococcal infections. As the ratio of AUC₂₄/MIC yielded >579.90 for Staphylococcus aureus infections or > 637.67 for enterococcal infections, would predict approximately 95% of patients with Staphylococcus aureus and enterococcal infections to achieve improved clinical outcomes and bacterial eradication, respectively. The study demonstrated that the value of AUIC was a major PPK/PD parameter to evaluate the efficacy of norvancomycin. Therefore, this value can be applied to design an optimal regimen as an important reference. Finally, an individualized regimen can be established for different subgroups of patients on the basis of the results of the PPK/PD study.In summary, the rapid assay to determine serum concentrations of norvancomycin was established. It provided a robust technique approach to conduct TDM of this drug. The population pharmacokinetic model of norvancomycin in patients with infections was successfully generated and validated, using the NONMEM method. The model was effectively applied to design the regimen for patients with variable renal function. Using the novel PPK/PD theory that was developed in this study, the regimen was further optimized in order to assure the efficacy safe regimen for different subgroups of patients with the infection. In the future, it offers promises of a rational use of norvancomycin to reduce the toxicity and enhance the greatest bactericidal effect and to assure efficacy and safety use of norvancomycin in clinical practices.更多还原