左氧氟沙星群体药代动力学—药效学研究
|
摘要
抗菌药物的体内、体外研究均提示,给药剂量和给药方案均会对细菌的杀灭作用造成影响。优化给药剂量需要综合考虑抗菌药物在体内的药代过程,以及抗菌药物对细菌的药效学特性。药代动力学主要研究药物在体内的吸收、分布、代谢和排泄的过程,着重阐述人体与药物之间的相互作用;抗菌药物的药效学主要研究药物对细菌的杀灭特性,着重阐述抗菌药物和细菌的相互作用。一个好的治疗方案不仅能达到良好的临床和微生物学疗效,而且给药安全。但是临床实际用药过程中,一个固定的给药方案作为治疗目标人群的初始给药方案时,有时会因为剂量不足而达不到有效的治疗浓度;有时则会因为体内的药物暴露水平过高,发生与药物暴露水平相关的不良反应。造成固定给药方案在不同个体中疗效和不良反应差别的原因主要有两方面,首先人群中药代参数受到患者基线状况和不同病理状态的影响,因此药代参数存在个体间差异;其次,药物对某一类病原菌的MICs呈现一定的分布,药物对不同菌株的敏感性不同,即药效学方面的差异。上述两方面的变异给临床用药带来了不确定性。
Sheiner等提出的群体分析方法为这种临床用药的不确定性提供了科学的解决方法。PPK方法能有效分析临床资料中散在的数据点,确定药代参数和药效参数在给药目标人群中的变异,以及造成变异的来源;如果结合抗菌药物的药效学参数,如fAUC/MICs和C_(max)/MICs,提出一系列临床假设,通过临床模拟的方法,能有效制定不同感染人群中相对合理的给药方案。
左氧氟沙星属于新一代喹诺酮类药物,是氧氟沙星的左旋异构体。左氧氟沙星口服生物利用度高,在组织体液中分布广泛,抗菌谱广,对社区获得性肺炎(CAP)和慢性支气管炎急性发作(AECB)的常见病原菌有良好的抗菌作用,甚至对临床中常见的一些耐药菌,如青霉素中介/耐药的肺炎链球菌(PISP/PRSP),产红霉素钝化酶的肺炎链球菌,以及产TEM-1型β-内酰胺酶的流感嗜血杆菌也有良好的杀菌活性,因此是治疗CAP和AECB适宜选用的药物。
迄今为止,中国人群中左氧氟沙星的药代参数主要来自健康志愿者,社区获得性下呼吸道感染患者中的药代资料尚属空白。健康志愿者的基线情况相仿,药代参数的个体间差异小,而感染人群中,年龄、体重和肌酐清除率等病理生理状况复杂,因此健康志愿者的药代参数并不能很好反应感染人群中的实际情况。
本次研究的第一、第二部分将建立左氧氟沙星在社区获得性下呼吸道感染患者中的PPK模型,定量分析感染人群中药代参数的差异,为优化给药方案提供药动学依据。研究的第三部分收集患者的临床和微生物学疗效以及治疗过程中发生的主要不良事件,对药物暴露水平与疗效、药物暴露水平与不良事件进行回归分析,为优化感染人群的给药方案提供药效学依据。研究的第四部分将提出左氧氟沙星对肺炎链球菌的体外药效学模型,从群体的角度探讨如何制定预防或延缓耐药菌生成的给药策略。
第一部分左氧氟沙星群体药代动力学模型的建立和验证
研究目的:建立左氧氟沙星500 mg片在社区获得性下呼吸道感染患者和健康志愿者中的PPK模型。
研究方法:本部分研究共收集164名CAP/AECB患者和18名健康志愿者的血药浓度。对每个感染患者平均采集3～4个散在的血药浓度标本,共分两组采集血样,第一组采样时间为给药后1±0.5 h、4±1 h、12±2 h和停药后24～48 h;第二组采样时间为给药后2±0.5 h、8±1 h、24±2 h和停药后24～48 h。健康志愿者为密集型采血。药物测定方法采用反相高效液相法,检测低限为0.01 mg/L,线性范围为0.01/5 mg/L。同时收集每个受试者的基线情况。采用非线性混合效应模型建立PPK模型(NONMEM VI(?)),并采用1000次非参数Bootstrap法进行模型验证。
结果:左氧氟沙星在感染患者和健康志愿者中符合2房室模型,个体间变异和个体内变异为指数模型。药物表观清除率(CL/F)、中央室分布容积(V_2)、室间清除率(Q)、外周室分布容积(V_3)和口服吸收速率常数(k_a)的群体典型值分别为7.82 L/h、64.8 L、0.286 L/h、5.9 L和1.36/h。肌酐清除率和性别对药物表观清除率有影响,肌酐清除率降低,药物清除率随之减低,女性的药物清除率略低于男性;体重和性别对中央室分布容积有影响,体重升高,中央室分布容积随之升高,女性的分布容积比男性低。CL/F和k_a的个体间变异分别为15.52%和79.94%。Bootstrap验证过程中,最终的PPK模型稳定,模型参数值与1000个Bootstrap验证数据集参数值的均值相比,偏差均小于10%;MAE_(imp)和MSE_(imp)(%)均在5%之内。
结论:PPK模型能很好反应左氧氟沙星在人体内代谢的生理特性,PPK模型能根据临床散在的数据对个体药代参数做出精确的估计,模型的参数与以往研究中左氧氟沙星的药代参数具有可比性。Bootstrap验证提示PPK模型不仅稳定,而且预测功能良好,因此模型可以用于临床模拟。
第二部分左氧氟沙星群体药代动力学模型的应用
研究目的:定量分析不同人群中左氧氟沙星药代参数的差异,为优化给药提供药动学依据。
研究方法:Bayesian反馈法计算164名感染患者和18名健康志愿者的药代参数,统计分析不同人群中药代参数的差异。然后根据PPK模型进行临床模拟,临床模拟在NONMEM ADVAN6 TRANS1中进行,根据感染目标人群的基线状况,即年龄、体重和肌酐清除率的变量-协变量矩阵构建模拟的目标人群。临床模拟可得到目标人群的药代参数分布,然后采用反应曲面分析法(完全二次型模型)定量分析不同人群药代参数的变异,发现可能需要调整给药方案的特殊人群。
结果:肌酐清除率介于50～80 mL/min之间(轻度肾功能减退)的患者,左氧氟沙星500 mg、1天1次连续给药,药物不会在体内蓄积:肌酐清除率介于20～50mL/min之间(中度肾功能减退)的患者,左氧氟沙星在体内的消除半衰期延长至10 h以上,左氧氟沙星500 mg、1天1次连续给药,药物在体内轻度蓄积,如果将给药方案调整为首剂500 mg,随后250 mg、1天1次连续给药,药物不再形成蓄积。肌酐清除率小于20 mL/min(重度肾功能不全)的患者,药物消除半衰期可能延长至27 h以上,左氧氟沙星500 mg、1天1次连续给药,药物在给药第3天起可能会在体内形成明显蓄积,这种情况下需要将给药间隔调整为48 h。
年龄与AUC_(0～∞)呈线性相关,r~2=0.4179,年龄每增加5岁,AUC_(0～∞)平均升高2mg.h/L,70岁以上老年患者的AUC_(0～∞)比人群均值高出20%。老年患者AUC_(0～∞)升高的原因与老年患者肌酐清除率降低有关。女性平均AUC_(0～∞)比男性高20%,但是男性和女性的体内药物浓度均能达到左氧氟沙星的有效治疗浓度。C_(max)在不同人群中分布相对平坦,提高给药剂量可以提高人群的C_(max)。
结论:轻度肾功能减退患者,左氧氟沙星给药方案为500 mg、1天1次给药;中度肾功能减退患者,左氧氟沙星仍可1天1次给药,首剂500 mg,随后可酌情减少至250 mg、1天1次给药;重度肾功能减退患者需要将给药方案调整为首剂500 mg,随后250 mg或500 mg、2天1次给药。不需要单独根据年龄和性别调整左氧氟沙星给药方案。
第三部分左氧氟沙星群体药代动力学-药效学研究(临床研究部分)
研究目的:阐明左氧氟沙星体内药物暴露水平与特定病原菌临床/微生物学疗效,以及药物暴露水平与不良反应之间的关系。
研究方法:CAP/AECB感染患者中,如果给药前合格痰标本经培养分离到病原菌,则这部分患者将进行药物暴露水平与药效的相关性分析,每名患者的药代参数通过Bayesian反馈法计算,PK/PD参数包括每名患者的AUC_(0～∞)/MIC和C_(max)/MIC,药效即临床/微生物学疗效,为二分类变量。164名CAP/AECB患者均有完整的不良事件的数据,因此164名患者都可进行药物暴露水平与不良反应的相关性分析。分析中采用逐步Logistic回归法确定影响疗效和不良反应的各种因素,并确定PPK/PD折点。
随后进行左氧氟沙星对CAP/AECB常见病原菌药效学达标概率的临床模拟,得出左氧氟沙星治疗CAP/AECB的合理给药方案。
结果:164名患者中共78名患者痰培养阳性,这部分患者可进行药物暴露水平与临床/微生物学疗效的量.效关系分析。分析中,患者的临床分离株对左氧氟沙星普遍敏感,仅2名患者评价为微生物学无效,均为AECB患者,分离到的病原菌均为铜绿假单胞菌,这两名患者临床评价为有效,逐步Logistic回归分析没有获得左氧氟沙星对特定病原菌的PPK/PD折点。164名患者均可进行药物暴露水平与不良反应的分析,分析中年龄与中枢神经系统不良反应和可逆的血白细胞降低有关,但是不良反应的程度轻,不影响用药,停药后均能恢复。逐步Logistic回归没有发现与药物暴露水平相关的不良反应。
临床模拟提示,目前临床使用的多种左氧氟沙星给药方案中,仅500 mg、1天1次给药治疗肺炎链球菌感染时,能获得理想的药效学达标概率,AUC_(0～∞)/MICs的药效学达标概率为88.3%(人群中AUC_(0～∞)/MICs超过33.7的概率);但是C_(max)/MICs的达标概率偏低,仅3.8%(人群中C_(max)/MICs超过10的概率)。
结论:左氧氟沙星500 mg、1天1次给药治疗CAP/AECB能获得良好的临床和微生物学疗效,给药安全,研究中没有发现与药物暴露水平相关的不良反应。随后的临床模拟进一步印证了临床常用的几种左氧氟沙星给药方案中,只有500mg每日1次给药有望获得对CAP/AECB常见病原菌理想的药效学达标概率。因此中国人群中,左氧氟沙星对CAP/AECB的标准给药方案应为500 mg、1天1次。
第四部分左氧氟沙星群体药代动力学-药效学研究(体外研究部分)
研究目的:建立描述左氧氟沙星对肺炎链球菌敏感亚群和相对耐药亚群动态杀灭过程的体外药效学模型,评价左氧氟沙星不同给药方案在目标人群中对肺炎链球菌相对耐药亚群的选择压力,探讨在群体中建立预防或延缓耐药菌生成的给药策略。
研究方法:本部分研究首先进行左氧氟沙星对肺炎链球菌8 h和48 h的体外杀菌曲线试验,研究的菌株包括肺炎链球菌ATCC49619和一株临床分离株SPN00205。体外杀菌曲线试验分成2组,第一组初始接种菌量为～10~5 CFU/mL,第二组初始接种菌量为～10~8 CFU/mL,体外杀菌曲线试验中左氧氟沙星药物浓度分别为1/2×MIC、1×MIC、2×MIC、3×MIC、4×MIC、8×MIC和16×MIC,每次试验均设无药生长对照组(空白对照)。空白对照、1/2×MIC、1×MIC、2×MIC、4×MIC和8×MIC浓度组的数据用于计算体外药效学模型的PD参数,3×MIC和16×MIC浓度组的数据用于模型的内部验证。
获得体外药效学模型的PD参数后,根据左氧氟沙星的PPK模型,以及肺炎链球菌野生株的MICs分布,进行2组PPK/PD临床模拟,第一组设定肺炎链球菌的初始菌量为～10~5 CFU/mL,第二组设定肺炎链球菌的初始菌量为～10~8 CFU/mL。模拟中,假设了多种左氧氟沙星给药方案,最后统计分析模拟疗程结束后,肺炎链球菌相对耐药亚群的变化趋势。
结果:左氧氟沙星对肺炎链球菌的体外药效学模型能很好反映左氧氟沙星对肺炎链球菌的杀菌特性,能较好说明体外杀菌曲线的实验数据。临床模拟发现,目前临床常用的几种左氧氟沙星给药方案中,仅500 mg、1天1次给药不会对肺炎链球菌相对耐药亚群造成很明显的选择压力,750 mg、1天1次给药,5天疗程(短期给药方案)对细菌群体中相对耐药亚群的选择压力最低。300 mg、1天1次给药和200 mg、1天2次给药对MIC=1 mg/L的肺炎链球菌中相对耐药亚群的选择压力最大。
结论:500 mg给药方案对肺炎链球菌不仅有较好的药效学达标概率,而且对肺炎链球菌中相对耐药亚群没有明显的选择压力。750 mg短期给药方案对整个菌群中相对耐药亚群的选择压力最低。因此在充分评价该方案临床疗效和安全性后,750 mg、1天1次给药,5天疗程可作为肾功能正常、非老年感染患者的推荐用药。
In vitro and in vivo response-exposure studies have demonstrated that both dose and dosing schedules may influence the effect of an antimicrobial.Dosing optimization actually is the balance between the pharmacokinetics(PK) and the pharmcodynamics (PD).PK is the course of absorption,distribution,metabolism and elimination of the drug,which is considered to be 'how the body does to the drug'.PD is the course of the pharmacological effects against certain pathogen under the fluctuating concentrations in vivo,which can be stated as 'how the drug does to the microbial'.An appropriate regimen should attain a satisfactory PD endpoint and at the same time lowers the drug toxicity.However,it is not unusual that a fixed regimen in the target population can produce clinical failures to some patients because of under-dosage while causing toxicity to the other patients due to excessive dosing.The diversity of drug effects in an infection population comes from the hierarchies of PK variability and the hierarchies of the bactericidal effects of the antimicrobial against different pathogens.So it is always the physician's dream to find out the right dose for their patients.
The population analysis method throws some light on this situation.The intrinsic character of the PPK method has the power to extract the essential information from the sparse clinical data and deduce some useful 'what if' hypothesis which tell the physician how to decide a regimen in different clinical situations.Besides,combining with the PD surrogates,such as fAUC/MICs or C_(max)/MICs,a stochastic simulation can be done according to the PPK model and figures out the probability that a certain dosage attains a certain clinical endpoint.The PPK/PD approaches are not only the promising alternatives to make the rational clinical decisions but also will bring an overwhelming revolution in the process of drug developments.
Levofloxacin,the active(-)-S-isomer of the racemic drug substance of oxfloxacin, has excellent performance against most significant pathogens of the community-acquired lower respiratory infections.It is also highly active against some important pathogens that are resistant to the other type of antibiotics,such as penicillin intermittent Streptococcus pneumoniae(PISP) or penicillin resistant Streptococcus pneumoniae(PRSP),macrolide resistant Streptococcus pneumoniae and TEM-1β-lactamase producing Haemophilus influenzae.Levofloxacin has very good pharmacokinetic characteristics,including its complete oral bioavailability,wide distribution in the body,and high concentration in the infection sites.Its high bactericidal activity and its safety dosing make it one of the first choices of treating the community-acquired pneumoniae(CAP) and the acute extraction of the chronic bronchitis(AECB),which are the two common infection types and the significant morbility and morbidity causes of infectious diseases.
However it is pity that till now we have little PK data of levofloxacin in the CAP and AECB patients.Most of the time,differences of PK data between the volunteer and the patients can not be ignored.The PK parameters of a drug in the patients usually have much wider distributions due to wider distributions of demographical baselines and complex clinical pathological conditions.If we arbitrary embed the PK parameters from the volunteer to the patients,the deviation of estimation might occur and this deviation might form the physician a wrong map of the drug in the patients and thus an improper clinical decision made.
The aim of this thesis is to explore the PPK model of levofloxacin in CAP and AECB patients.The PPK model will then be used to characterize the source of PK variability in the target population quantitatively and form the rational regimens in different clinical situations.The PD data including the clinical and microbial responses as well as safety data are dedicatedly selected.The establishment of a PPK/PD model is expected to describe the relationship between the PK and the PD,which will be used for dosing optimization.The knowledge gained in the PPK/PD model push the study to a more mechanical exploration focusing on finding out a resolution of drug-resistant prevention strategy.
Chapter1.Establishment and validation of the levofloxacin PPK model
Objective:To develop a PPK model of levofloxacin,which will be used to describe the PK characteristics of the drug in different subpopulations.
Methods:A nonlinear mixed effects model was developed with NONMEMⅥ(?) for 164 CAP/AECB patients and 18 healthy volunteers.The average 3～4 sparse drug concentration samples from each patient were collected on the drug absorption, distribution and elimination phases respectively.The data of the volunteer come from a phaseⅠclinical trial of which a density sampling strategy was adopted.The drug concentrations were determined by the reverse high performance liquid chromatography method.The detection limit of the method is 0.01 mg/L with a linear range from 0.01 mg/L to 5 mg/L.The effect of age,weight,sex,creatine clearance (CCR),food,disease status,and drug combinations on the PK parameters were evaluated.The bootstrap method was used to test the stability and predictive function of the model.
Results:A two-compartment linear model with exponential residual errors best described the data collected.Both CCR and sex were found to significantly affect the apparent clearance while weight and sex affected the volume distribution of the central compartment.The typical population values of the apparent clearance(CL/F),volume distribution of the central compartment(V_2),the inter-compartment clearance(Q), volume distribution of the peripheral compartment(V_3) and the absorption rate constant (k_a) were 7.82 L/h,64.8 L,0.286 L/h,5.9 L and 1.36/h respectively.The inter-subject variances of CL/F and ka in the population were 15.52%and 79.94%respectively.In the process of bootstrapping validation,the narrow biases less than 10%of the PK parameters were found between the original data and the mean value of the bootstrap data sets,with an inverse Gaussian distribution of PK parameters in the bootstrap data sets.The MAE_(imp) and MSE_(imp)(%) were both fell within 5%indicated that the model had excellent prediction function.
Conclusion:The PPK model was highly consistent with the physiological characteristics of levofloxacin.The model could robustly estimate the PK parameters by the sparse information.The inter-individual variance of CL/F is relatively narrow indicating its clinical application.The well-performance of the prediction ability of the model was the foundation of the stochastic simulations.
Chapter2.Clinical application of the PPK model
Objective:To develop dosing strategies for levofloxacin in different demographical and pathological situations.
Methods:The definition about the target population of levofloxacin was set according to the indications of the 500 mg once-daily regimen in this study.The variance-covariance matrix of the demographical baselines including age,weight and creatine clearance,the PK parameters and their inter-/intra- variances were embedded into NONMEM ADVAN6 TRANS1.Clinical trial simulation of levofloxacin with 1000 target population was conducted.Response surface analysis(full quadratic model) was then used to quantify the relationship between the covariances and the PK parameters to find out some special clinical situations that the adjustments of the regimen might be required.
Results:The decreased creatine clearance caused a delayed elimination of the drug. A 10-days dosing simulation didn't find out significant accumulation of the drug in mild renal deterioration patients(creatine clearance ranged from 50～80 mL/min). Multiple doses of levofloxacin 500 mg once-daily caused mild accumulation in the moderate renal deterioration patients(creatine clearance ranged from 20～50 mL/min). When the creatine clearance fell below 20 mL/min(the severe renal deterioration),the obvious accumulation might occur on the third day of dosing which diminished if the regimen shifted to once every 48 h.
There was a linear relationship between age and AUC_(0～∞),with r~2 = 0.4179.2mg.h/L of AUC_(0～∞) was escalated by every 5 years in the population,and the old age(70 y above) subgroup had a 20%higher AUC_(0～∞) than the average level of the population. The AUC_(0～∞) in the moderate renal deficient patient was also 20%higher than the average level,which implied that the physician could accommodate the amount of dosage with art for such patient.Gender might have some influences on AUC_(0～∞),that the AUC_(0～∞) in females were 10 mg.h/L higher than that in males.However such differences might have no significant clinical meanings.C_(max) was the only parameter that no surrogate was found to be predicted by.C_(max) was prone to have relatively flat distribution in the population which indicated that the parameter was related to the amount of dosage.Elevated dosing expected a higher C_(max) in the population.
Conclusion:The present regimen of levofloxacin had a highly consistent performance in different subpopulations excepting the moderate and severe renal deficient patients.For the moderate renal deterioration patients,the physician could artistly adjust levofloxacin dosage from 500 mg once-daily to 250 mg once-daily on the second day.However,a strong indication for dosing adjustment was only in patients with severe renal deteriorations,in which the dosing intervals of levofloxacin should be every 48 h.
Chapter3.The PPK/PD study of levofloxacin(clinical study)
Objective:To illuminate the relationship between the drug exposure and the microbial responses as well as toxicity.
Methods:PD study focusing on assessing the microbial responses was conducted on a subset of CAP and AECB patients with available MICs data of causative pathogens. PK parameters of each patient were determined by the Bayesian estimation method. The microbial responses were categorized as eradication,persistent,recurrent, re-infection and replacement.The categories of persistent and recurrent were considered to be microbial failure.The AUC_(0～∞)/MICs and C_(max)/MICs were obtained as the PPKUPD index.Stepwise logistic regression method was adopted to determine the PPK/PD breakpoint that predicting a positive microbial response.Since all infectious patients in the PPK study were safety evaluable,they were all entered into PD study focusing on assessing the relationship between drug exposure and the toxic responses. Stepwise logistic regression method was adopted to estimate the exposure and response (toxic) relationships.
In the process of clinical trial simulation,the drug effects of different regimens of levofloxacin against Streptococcus pneumoniae,which was the most significant pathogen in CAP/AECB were extensively discussed.The probability of different regimens of levofloxacin attaining a positive microbial endpoint in the target population was assessed by stochastic simulations.The MICs distribution data of the wide type strains come from the EUCAST.
Results:The regimen of levofloxacin 500 mg once-daily in this study had so excellent microbial responses that only 2 failure cases in the PPK/PD study which were both AECB patients with positive clinical responses,and the causative pathogens were both Pseudomonas aureginosa,which could be persistent in the bronchitis without any infectious symptoms.The stepwise regression analysis failed to find out a PPK/PD breakpoint for any significant pathogens due to patients' good endpoints in this study. In the safety PPK/PD study,a positive correlationship between the age and the ratio of central nervous system(CNS) adverse effect(AE) was found.However,most of the CNS adverse effect was mild and durable without stopping the use of the drug.The age also related to the transient decrease of white blood cell count.The stochastic simulation indicated that only the regimen of 500 mg once-daily could give a satisfactory target attainment rate against Streptococcus pneumoniae,with AUC_(0～∞) /MICs in 88.3%of the population well above 33.7.However,only 3.8%of the population's C_(max)/MICs were above 10.
Conclusion:The regimen of levofloxacin in this study could attain pretty good clinical and microbial responses.Age was positively related to the CNS adverse effects as well as the transient decrease of white blood cell count.However,the AE was mild and the trade-off between the benefit and risk made the balance obviously to the benefit. No exposure related AE was found.Among the different regimens of levofloxacin,only the regimen of 500 mg once-daily could attain a satisfactory endpoint to Streptococcus pneumoniae.In the process of stochastic simulation,a relatively low C_(max)/MICs ratio in the population stressed necessity to further exploration.
Chapter4.The PPK/PD study of levofloxaein(in-vitro study)
Objective:To establish a mechanism-based mathematical model of levofloxacin's bactericidal effects against both susceptible and relatively resistant subpopulations of Streptococcus pneumoniae.To assess the antibiotic pressure against Streptococcus pneumoniae due to different levofloxacin regimens from a population view.
Methods:Two strains of Streptococcus pneumoniae ATCC49619 and Streptococcus pneumoniae SPN00205,the latter of which was the clinical isolate in the PPK/PD study, were studied.The first step,the MICs and the MBCs of levofloxacin against the two strains were determined by a macrobroth dilution method with initial inoculums of～10~5 CFU/mL.The second step,a 8 h and a 48 h in-vitro time kill curve studies were conducted for each isolate with different constant drug concentrations,including 1/2×MIC,1×MIC,2×MIC,3×MIC,4×MIC,8×MIC and 16×MIC.A drug free growth control group was set for each experiment.The bacteria count quantification of the susceptible group and the relative resistant group in different concentrations was determined on several time points.The data from the blank,1/2×MIC,1×MIC,2×MIC,4×MIC and 8×MIC concentrations was used to estimate the dummy PD parameters that describe the nonlinear growth rate of bacteria and nonlinear kill rate of the drug against the bacteria in a mechanism-based PK/PD model.The data from the 3×MIC and 16×MIC concentrations was used for model internal validation.A stochastic simulation was conducted to estimate the risk of screening the relative resistant Streptococcus pneumoniae by the different levofloxacin regimens from a population view.In the simulation,the PK data was the in vivo data from the PPK study and the PD parameters come from the time kill curve studies.The duration of drug-microbial interaction was set to 10 days which is the standard levofloxacin therapy course for CAP.Two initial microbial loads in the simulation were～10~5 CFU/mL and～10~8 CFU/mL,respectively.The remaining bacteria on the 10~(th) day in the population were summarised.
Results:The PD parameters in the mechanism-based PK/PD model reflected the bactericidal characteristics of levofloxacin against Streptococcus pneumoniae.The model stated that the drug is a concentration-dependent antimicrobial,reaching the maximum bactericidal rate when the concentration attained 8 times of the MICs.The dummy PK parameter of EC_(50) reflected the potency of drug against bacteria with approximately linear relationship to the MICs.In the stochastic simulation,only the regimen of 500 mg once-daily and 750 mg once-daily had limited bacteria selection pressure on the wide type population of Streptococcus pneumoniae.The simulation also indicated that the monitoring on the Streptococcus pneumoniae with the MICs of 2 mg/L should be emphasized.The MICs of 2 mg/L might be the PPK/PD breakpoint of levofloxacin against Streptococcus pneumoniae.
Conclusion:The 500 mg once-daily should be the standard regimen of levofloxacin in clinical due to its excellent clinical/microbial responses and the potential ability to inhibit the development of drug resistance.The 750 mg once-daily short course regimen might perform even better than 500 mg regimen.It is of great value to assess the response and the safety of 750 mg regimen in Chinese population in future researches.
Sheiner等提出的群体分析方法为这种临床用药的不确定性提供了科学的解决方法。PPK方法能有效分析临床资料中散在的数据点,确定药代参数和药效参数在给药目标人群中的变异,以及造成变异的来源;如果结合抗菌药物的药效学参数,如fAUC/MICs和C_(max)/MICs,提出一系列临床假设,通过临床模拟的方法,能有效制定不同感染人群中相对合理的给药方案。
左氧氟沙星属于新一代喹诺酮类药物,是氧氟沙星的左旋异构体。左氧氟沙星口服生物利用度高,在组织体液中分布广泛,抗菌谱广,对社区获得性肺炎(CAP)和慢性支气管炎急性发作(AECB)的常见病原菌有良好的抗菌作用,甚至对临床中常见的一些耐药菌,如青霉素中介/耐药的肺炎链球菌(PISP/PRSP),产红霉素钝化酶的肺炎链球菌,以及产TEM-1型β-内酰胺酶的流感嗜血杆菌也有良好的杀菌活性,因此是治疗CAP和AECB适宜选用的药物。
迄今为止,中国人群中左氧氟沙星的药代参数主要来自健康志愿者,社区获得性下呼吸道感染患者中的药代资料尚属空白。健康志愿者的基线情况相仿,药代参数的个体间差异小,而感染人群中,年龄、体重和肌酐清除率等病理生理状况复杂,因此健康志愿者的药代参数并不能很好反应感染人群中的实际情况。
本次研究的第一、第二部分将建立左氧氟沙星在社区获得性下呼吸道感染患者中的PPK模型,定量分析感染人群中药代参数的差异,为优化给药方案提供药动学依据。研究的第三部分收集患者的临床和微生物学疗效以及治疗过程中发生的主要不良事件,对药物暴露水平与疗效、药物暴露水平与不良事件进行回归分析,为优化感染人群的给药方案提供药效学依据。研究的第四部分将提出左氧氟沙星对肺炎链球菌的体外药效学模型,从群体的角度探讨如何制定预防或延缓耐药菌生成的给药策略。
第一部分左氧氟沙星群体药代动力学模型的建立和验证
研究目的:建立左氧氟沙星500 mg片在社区获得性下呼吸道感染患者和健康志愿者中的PPK模型。
研究方法:本部分研究共收集164名CAP/AECB患者和18名健康志愿者的血药浓度。对每个感染患者平均采集3～4个散在的血药浓度标本,共分两组采集血样,第一组采样时间为给药后1±0.5 h、4±1 h、12±2 h和停药后24～48 h;第二组采样时间为给药后2±0.5 h、8±1 h、24±2 h和停药后24～48 h。健康志愿者为密集型采血。药物测定方法采用反相高效液相法,检测低限为0.01 mg/L,线性范围为0.01/5 mg/L。同时收集每个受试者的基线情况。采用非线性混合效应模型建立PPK模型(NONMEM VI(?)),并采用1000次非参数Bootstrap法进行模型验证。
结果:左氧氟沙星在感染患者和健康志愿者中符合2房室模型,个体间变异和个体内变异为指数模型。药物表观清除率(CL/F)、中央室分布容积(V_2)、室间清除率(Q)、外周室分布容积(V_3)和口服吸收速率常数(k_a)的群体典型值分别为7.82 L/h、64.8 L、0.286 L/h、5.9 L和1.36/h。肌酐清除率和性别对药物表观清除率有影响,肌酐清除率降低,药物清除率随之减低,女性的药物清除率略低于男性;体重和性别对中央室分布容积有影响,体重升高,中央室分布容积随之升高,女性的分布容积比男性低。CL/F和k_a的个体间变异分别为15.52%和79.94%。Bootstrap验证过程中,最终的PPK模型稳定,模型参数值与1000个Bootstrap验证数据集参数值的均值相比,偏差均小于10%;MAE_(imp)和MSE_(imp)(%)均在5%之内。
结论:PPK模型能很好反应左氧氟沙星在人体内代谢的生理特性,PPK模型能根据临床散在的数据对个体药代参数做出精确的估计,模型的参数与以往研究中左氧氟沙星的药代参数具有可比性。Bootstrap验证提示PPK模型不仅稳定,而且预测功能良好,因此模型可以用于临床模拟。
第二部分左氧氟沙星群体药代动力学模型的应用
研究目的:定量分析不同人群中左氧氟沙星药代参数的差异,为优化给药提供药动学依据。
研究方法:Bayesian反馈法计算164名感染患者和18名健康志愿者的药代参数,统计分析不同人群中药代参数的差异。然后根据PPK模型进行临床模拟,临床模拟在NONMEM ADVAN6 TRANS1中进行,根据感染目标人群的基线状况,即年龄、体重和肌酐清除率的变量-协变量矩阵构建模拟的目标人群。临床模拟可得到目标人群的药代参数分布,然后采用反应曲面分析法(完全二次型模型)定量分析不同人群药代参数的变异,发现可能需要调整给药方案的特殊人群。
结果:肌酐清除率介于50～80 mL/min之间(轻度肾功能减退)的患者,左氧氟沙星500 mg、1天1次连续给药,药物不会在体内蓄积:肌酐清除率介于20～50mL/min之间(中度肾功能减退)的患者,左氧氟沙星在体内的消除半衰期延长至10 h以上,左氧氟沙星500 mg、1天1次连续给药,药物在体内轻度蓄积,如果将给药方案调整为首剂500 mg,随后250 mg、1天1次连续给药,药物不再形成蓄积。肌酐清除率小于20 mL/min(重度肾功能不全)的患者,药物消除半衰期可能延长至27 h以上,左氧氟沙星500 mg、1天1次连续给药,药物在给药第3天起可能会在体内形成明显蓄积,这种情况下需要将给药间隔调整为48 h。
年龄与AUC_(0～∞)呈线性相关,r~2=0.4179,年龄每增加5岁,AUC_(0～∞)平均升高2mg.h/L,70岁以上老年患者的AUC_(0～∞)比人群均值高出20%。老年患者AUC_(0～∞)升高的原因与老年患者肌酐清除率降低有关。女性平均AUC_(0～∞)比男性高20%,但是男性和女性的体内药物浓度均能达到左氧氟沙星的有效治疗浓度。C_(max)在不同人群中分布相对平坦,提高给药剂量可以提高人群的C_(max)。
结论:轻度肾功能减退患者,左氧氟沙星给药方案为500 mg、1天1次给药;中度肾功能减退患者,左氧氟沙星仍可1天1次给药,首剂500 mg,随后可酌情减少至250 mg、1天1次给药;重度肾功能减退患者需要将给药方案调整为首剂500 mg,随后250 mg或500 mg、2天1次给药。不需要单独根据年龄和性别调整左氧氟沙星给药方案。
第三部分左氧氟沙星群体药代动力学-药效学研究(临床研究部分)
研究目的:阐明左氧氟沙星体内药物暴露水平与特定病原菌临床/微生物学疗效,以及药物暴露水平与不良反应之间的关系。
研究方法:CAP/AECB感染患者中,如果给药前合格痰标本经培养分离到病原菌,则这部分患者将进行药物暴露水平与药效的相关性分析,每名患者的药代参数通过Bayesian反馈法计算,PK/PD参数包括每名患者的AUC_(0～∞)/MIC和C_(max)/MIC,药效即临床/微生物学疗效,为二分类变量。164名CAP/AECB患者均有完整的不良事件的数据,因此164名患者都可进行药物暴露水平与不良反应的相关性分析。分析中采用逐步Logistic回归法确定影响疗效和不良反应的各种因素,并确定PPK/PD折点。
随后进行左氧氟沙星对CAP/AECB常见病原菌药效学达标概率的临床模拟,得出左氧氟沙星治疗CAP/AECB的合理给药方案。
结果:164名患者中共78名患者痰培养阳性,这部分患者可进行药物暴露水平与临床/微生物学疗效的量.效关系分析。分析中,患者的临床分离株对左氧氟沙星普遍敏感,仅2名患者评价为微生物学无效,均为AECB患者,分离到的病原菌均为铜绿假单胞菌,这两名患者临床评价为有效,逐步Logistic回归分析没有获得左氧氟沙星对特定病原菌的PPK/PD折点。164名患者均可进行药物暴露水平与不良反应的分析,分析中年龄与中枢神经系统不良反应和可逆的血白细胞降低有关,但是不良反应的程度轻,不影响用药,停药后均能恢复。逐步Logistic回归没有发现与药物暴露水平相关的不良反应。
临床模拟提示,目前临床使用的多种左氧氟沙星给药方案中,仅500 mg、1天1次给药治疗肺炎链球菌感染时,能获得理想的药效学达标概率,AUC_(0～∞)/MICs的药效学达标概率为88.3%(人群中AUC_(0～∞)/MICs超过33.7的概率);但是C_(max)/MICs的达标概率偏低,仅3.8%(人群中C_(max)/MICs超过10的概率)。
结论:左氧氟沙星500 mg、1天1次给药治疗CAP/AECB能获得良好的临床和微生物学疗效,给药安全,研究中没有发现与药物暴露水平相关的不良反应。随后的临床模拟进一步印证了临床常用的几种左氧氟沙星给药方案中,只有500mg每日1次给药有望获得对CAP/AECB常见病原菌理想的药效学达标概率。因此中国人群中,左氧氟沙星对CAP/AECB的标准给药方案应为500 mg、1天1次。
第四部分左氧氟沙星群体药代动力学-药效学研究(体外研究部分)
研究目的:建立描述左氧氟沙星对肺炎链球菌敏感亚群和相对耐药亚群动态杀灭过程的体外药效学模型,评价左氧氟沙星不同给药方案在目标人群中对肺炎链球菌相对耐药亚群的选择压力,探讨在群体中建立预防或延缓耐药菌生成的给药策略。
研究方法:本部分研究首先进行左氧氟沙星对肺炎链球菌8 h和48 h的体外杀菌曲线试验,研究的菌株包括肺炎链球菌ATCC49619和一株临床分离株SPN00205。体外杀菌曲线试验分成2组,第一组初始接种菌量为～10~5 CFU/mL,第二组初始接种菌量为～10~8 CFU/mL,体外杀菌曲线试验中左氧氟沙星药物浓度分别为1/2×MIC、1×MIC、2×MIC、3×MIC、4×MIC、8×MIC和16×MIC,每次试验均设无药生长对照组(空白对照)。空白对照、1/2×MIC、1×MIC、2×MIC、4×MIC和8×MIC浓度组的数据用于计算体外药效学模型的PD参数,3×MIC和16×MIC浓度组的数据用于模型的内部验证。
获得体外药效学模型的PD参数后,根据左氧氟沙星的PPK模型,以及肺炎链球菌野生株的MICs分布,进行2组PPK/PD临床模拟,第一组设定肺炎链球菌的初始菌量为～10~5 CFU/mL,第二组设定肺炎链球菌的初始菌量为～10~8 CFU/mL。模拟中,假设了多种左氧氟沙星给药方案,最后统计分析模拟疗程结束后,肺炎链球菌相对耐药亚群的变化趋势。
结果:左氧氟沙星对肺炎链球菌的体外药效学模型能很好反映左氧氟沙星对肺炎链球菌的杀菌特性,能较好说明体外杀菌曲线的实验数据。临床模拟发现,目前临床常用的几种左氧氟沙星给药方案中,仅500 mg、1天1次给药不会对肺炎链球菌相对耐药亚群造成很明显的选择压力,750 mg、1天1次给药,5天疗程(短期给药方案)对细菌群体中相对耐药亚群的选择压力最低。300 mg、1天1次给药和200 mg、1天2次给药对MIC=1 mg/L的肺炎链球菌中相对耐药亚群的选择压力最大。
结论:500 mg给药方案对肺炎链球菌不仅有较好的药效学达标概率,而且对肺炎链球菌中相对耐药亚群没有明显的选择压力。750 mg短期给药方案对整个菌群中相对耐药亚群的选择压力最低。因此在充分评价该方案临床疗效和安全性后,750 mg、1天1次给药,5天疗程可作为肾功能正常、非老年感染患者的推荐用药。
In vitro and in vivo response-exposure studies have demonstrated that both dose and dosing schedules may influence the effect of an antimicrobial.Dosing optimization actually is the balance between the pharmacokinetics(PK) and the pharmcodynamics (PD).PK is the course of absorption,distribution,metabolism and elimination of the drug,which is considered to be 'how the body does to the drug'.PD is the course of the pharmacological effects against certain pathogen under the fluctuating concentrations in vivo,which can be stated as 'how the drug does to the microbial'.An appropriate regimen should attain a satisfactory PD endpoint and at the same time lowers the drug toxicity.However,it is not unusual that a fixed regimen in the target population can produce clinical failures to some patients because of under-dosage while causing toxicity to the other patients due to excessive dosing.The diversity of drug effects in an infection population comes from the hierarchies of PK variability and the hierarchies of the bactericidal effects of the antimicrobial against different pathogens.So it is always the physician's dream to find out the right dose for their patients.
The population analysis method throws some light on this situation.The intrinsic character of the PPK method has the power to extract the essential information from the sparse clinical data and deduce some useful 'what if' hypothesis which tell the physician how to decide a regimen in different clinical situations.Besides,combining with the PD surrogates,such as fAUC/MICs or C_(max)/MICs,a stochastic simulation can be done according to the PPK model and figures out the probability that a certain dosage attains a certain clinical endpoint.The PPK/PD approaches are not only the promising alternatives to make the rational clinical decisions but also will bring an overwhelming revolution in the process of drug developments.
Levofloxacin,the active(-)-S-isomer of the racemic drug substance of oxfloxacin, has excellent performance against most significant pathogens of the community-acquired lower respiratory infections.It is also highly active against some important pathogens that are resistant to the other type of antibiotics,such as penicillin intermittent Streptococcus pneumoniae(PISP) or penicillin resistant Streptococcus pneumoniae(PRSP),macrolide resistant Streptococcus pneumoniae and TEM-1β-lactamase producing Haemophilus influenzae.Levofloxacin has very good pharmacokinetic characteristics,including its complete oral bioavailability,wide distribution in the body,and high concentration in the infection sites.Its high bactericidal activity and its safety dosing make it one of the first choices of treating the community-acquired pneumoniae(CAP) and the acute extraction of the chronic bronchitis(AECB),which are the two common infection types and the significant morbility and morbidity causes of infectious diseases.
However it is pity that till now we have little PK data of levofloxacin in the CAP and AECB patients.Most of the time,differences of PK data between the volunteer and the patients can not be ignored.The PK parameters of a drug in the patients usually have much wider distributions due to wider distributions of demographical baselines and complex clinical pathological conditions.If we arbitrary embed the PK parameters from the volunteer to the patients,the deviation of estimation might occur and this deviation might form the physician a wrong map of the drug in the patients and thus an improper clinical decision made.
The aim of this thesis is to explore the PPK model of levofloxacin in CAP and AECB patients.The PPK model will then be used to characterize the source of PK variability in the target population quantitatively and form the rational regimens in different clinical situations.The PD data including the clinical and microbial responses as well as safety data are dedicatedly selected.The establishment of a PPK/PD model is expected to describe the relationship between the PK and the PD,which will be used for dosing optimization.The knowledge gained in the PPK/PD model push the study to a more mechanical exploration focusing on finding out a resolution of drug-resistant prevention strategy.
Chapter1.Establishment and validation of the levofloxacin PPK model
Objective:To develop a PPK model of levofloxacin,which will be used to describe the PK characteristics of the drug in different subpopulations.
Methods:A nonlinear mixed effects model was developed with NONMEMⅥ(?) for 164 CAP/AECB patients and 18 healthy volunteers.The average 3～4 sparse drug concentration samples from each patient were collected on the drug absorption, distribution and elimination phases respectively.The data of the volunteer come from a phaseⅠclinical trial of which a density sampling strategy was adopted.The drug concentrations were determined by the reverse high performance liquid chromatography method.The detection limit of the method is 0.01 mg/L with a linear range from 0.01 mg/L to 5 mg/L.The effect of age,weight,sex,creatine clearance (CCR),food,disease status,and drug combinations on the PK parameters were evaluated.The bootstrap method was used to test the stability and predictive function of the model.
Results:A two-compartment linear model with exponential residual errors best described the data collected.Both CCR and sex were found to significantly affect the apparent clearance while weight and sex affected the volume distribution of the central compartment.The typical population values of the apparent clearance(CL/F),volume distribution of the central compartment(V_2),the inter-compartment clearance(Q), volume distribution of the peripheral compartment(V_3) and the absorption rate constant (k_a) were 7.82 L/h,64.8 L,0.286 L/h,5.9 L and 1.36/h respectively.The inter-subject variances of CL/F and ka in the population were 15.52%and 79.94%respectively.In the process of bootstrapping validation,the narrow biases less than 10%of the PK parameters were found between the original data and the mean value of the bootstrap data sets,with an inverse Gaussian distribution of PK parameters in the bootstrap data sets.The MAE_(imp) and MSE_(imp)(%) were both fell within 5%indicated that the model had excellent prediction function.
Conclusion:The PPK model was highly consistent with the physiological characteristics of levofloxacin.The model could robustly estimate the PK parameters by the sparse information.The inter-individual variance of CL/F is relatively narrow indicating its clinical application.The well-performance of the prediction ability of the model was the foundation of the stochastic simulations.
Chapter2.Clinical application of the PPK model
Objective:To develop dosing strategies for levofloxacin in different demographical and pathological situations.
Methods:The definition about the target population of levofloxacin was set according to the indications of the 500 mg once-daily regimen in this study.The variance-covariance matrix of the demographical baselines including age,weight and creatine clearance,the PK parameters and their inter-/intra- variances were embedded into NONMEM ADVAN6 TRANS1.Clinical trial simulation of levofloxacin with 1000 target population was conducted.Response surface analysis(full quadratic model) was then used to quantify the relationship between the covariances and the PK parameters to find out some special clinical situations that the adjustments of the regimen might be required.
Results:The decreased creatine clearance caused a delayed elimination of the drug. A 10-days dosing simulation didn't find out significant accumulation of the drug in mild renal deterioration patients(creatine clearance ranged from 50～80 mL/min). Multiple doses of levofloxacin 500 mg once-daily caused mild accumulation in the moderate renal deterioration patients(creatine clearance ranged from 20～50 mL/min). When the creatine clearance fell below 20 mL/min(the severe renal deterioration),the obvious accumulation might occur on the third day of dosing which diminished if the regimen shifted to once every 48 h.
There was a linear relationship between age and AUC_(0～∞),with r~2 = 0.4179.2mg.h/L of AUC_(0～∞) was escalated by every 5 years in the population,and the old age(70 y above) subgroup had a 20%higher AUC_(0～∞) than the average level of the population. The AUC_(0～∞) in the moderate renal deficient patient was also 20%higher than the average level,which implied that the physician could accommodate the amount of dosage with art for such patient.Gender might have some influences on AUC_(0～∞),that the AUC_(0～∞) in females were 10 mg.h/L higher than that in males.However such differences might have no significant clinical meanings.C_(max) was the only parameter that no surrogate was found to be predicted by.C_(max) was prone to have relatively flat distribution in the population which indicated that the parameter was related to the amount of dosage.Elevated dosing expected a higher C_(max) in the population.
Conclusion:The present regimen of levofloxacin had a highly consistent performance in different subpopulations excepting the moderate and severe renal deficient patients.For the moderate renal deterioration patients,the physician could artistly adjust levofloxacin dosage from 500 mg once-daily to 250 mg once-daily on the second day.However,a strong indication for dosing adjustment was only in patients with severe renal deteriorations,in which the dosing intervals of levofloxacin should be every 48 h.
Chapter3.The PPK/PD study of levofloxacin(clinical study)
Objective:To illuminate the relationship between the drug exposure and the microbial responses as well as toxicity.
Methods:PD study focusing on assessing the microbial responses was conducted on a subset of CAP and AECB patients with available MICs data of causative pathogens. PK parameters of each patient were determined by the Bayesian estimation method. The microbial responses were categorized as eradication,persistent,recurrent, re-infection and replacement.The categories of persistent and recurrent were considered to be microbial failure.The AUC_(0～∞)/MICs and C_(max)/MICs were obtained as the PPKUPD index.Stepwise logistic regression method was adopted to determine the PPK/PD breakpoint that predicting a positive microbial response.Since all infectious patients in the PPK study were safety evaluable,they were all entered into PD study focusing on assessing the relationship between drug exposure and the toxic responses. Stepwise logistic regression method was adopted to estimate the exposure and response (toxic) relationships.
In the process of clinical trial simulation,the drug effects of different regimens of levofloxacin against Streptococcus pneumoniae,which was the most significant pathogen in CAP/AECB were extensively discussed.The probability of different regimens of levofloxacin attaining a positive microbial endpoint in the target population was assessed by stochastic simulations.The MICs distribution data of the wide type strains come from the EUCAST.
Results:The regimen of levofloxacin 500 mg once-daily in this study had so excellent microbial responses that only 2 failure cases in the PPK/PD study which were both AECB patients with positive clinical responses,and the causative pathogens were both Pseudomonas aureginosa,which could be persistent in the bronchitis without any infectious symptoms.The stepwise regression analysis failed to find out a PPK/PD breakpoint for any significant pathogens due to patients' good endpoints in this study. In the safety PPK/PD study,a positive correlationship between the age and the ratio of central nervous system(CNS) adverse effect(AE) was found.However,most of the CNS adverse effect was mild and durable without stopping the use of the drug.The age also related to the transient decrease of white blood cell count.The stochastic simulation indicated that only the regimen of 500 mg once-daily could give a satisfactory target attainment rate against Streptococcus pneumoniae,with AUC_(0～∞) /MICs in 88.3%of the population well above 33.7.However,only 3.8%of the population's C_(max)/MICs were above 10.
Conclusion:The regimen of levofloxacin in this study could attain pretty good clinical and microbial responses.Age was positively related to the CNS adverse effects as well as the transient decrease of white blood cell count.However,the AE was mild and the trade-off between the benefit and risk made the balance obviously to the benefit. No exposure related AE was found.Among the different regimens of levofloxacin,only the regimen of 500 mg once-daily could attain a satisfactory endpoint to Streptococcus pneumoniae.In the process of stochastic simulation,a relatively low C_(max)/MICs ratio in the population stressed necessity to further exploration.
Chapter4.The PPK/PD study of levofloxaein(in-vitro study)
Objective:To establish a mechanism-based mathematical model of levofloxacin's bactericidal effects against both susceptible and relatively resistant subpopulations of Streptococcus pneumoniae.To assess the antibiotic pressure against Streptococcus pneumoniae due to different levofloxacin regimens from a population view.
Methods:Two strains of Streptococcus pneumoniae ATCC49619 and Streptococcus pneumoniae SPN00205,the latter of which was the clinical isolate in the PPK/PD study, were studied.The first step,the MICs and the MBCs of levofloxacin against the two strains were determined by a macrobroth dilution method with initial inoculums of～10~5 CFU/mL.The second step,a 8 h and a 48 h in-vitro time kill curve studies were conducted for each isolate with different constant drug concentrations,including 1/2×MIC,1×MIC,2×MIC,3×MIC,4×MIC,8×MIC and 16×MIC.A drug free growth control group was set for each experiment.The bacteria count quantification of the susceptible group and the relative resistant group in different concentrations was determined on several time points.The data from the blank,1/2×MIC,1×MIC,2×MIC,4×MIC and 8×MIC concentrations was used to estimate the dummy PD parameters that describe the nonlinear growth rate of bacteria and nonlinear kill rate of the drug against the bacteria in a mechanism-based PK/PD model.The data from the 3×MIC and 16×MIC concentrations was used for model internal validation.A stochastic simulation was conducted to estimate the risk of screening the relative resistant Streptococcus pneumoniae by the different levofloxacin regimens from a population view.In the simulation,the PK data was the in vivo data from the PPK study and the PD parameters come from the time kill curve studies.The duration of drug-microbial interaction was set to 10 days which is the standard levofloxacin therapy course for CAP.Two initial microbial loads in the simulation were～10~5 CFU/mL and～10~8 CFU/mL,respectively.The remaining bacteria on the 10~(th) day in the population were summarised.
Results:The PD parameters in the mechanism-based PK/PD model reflected the bactericidal characteristics of levofloxacin against Streptococcus pneumoniae.The model stated that the drug is a concentration-dependent antimicrobial,reaching the maximum bactericidal rate when the concentration attained 8 times of the MICs.The dummy PK parameter of EC_(50) reflected the potency of drug against bacteria with approximately linear relationship to the MICs.In the stochastic simulation,only the regimen of 500 mg once-daily and 750 mg once-daily had limited bacteria selection pressure on the wide type population of Streptococcus pneumoniae.The simulation also indicated that the monitoring on the Streptococcus pneumoniae with the MICs of 2 mg/L should be emphasized.The MICs of 2 mg/L might be the PPK/PD breakpoint of levofloxacin against Streptococcus pneumoniae.
Conclusion:The 500 mg once-daily should be the standard regimen of levofloxacin in clinical due to its excellent clinical/microbial responses and the potential ability to inhibit the development of drug resistance.The 750 mg once-daily short course regimen might perform even better than 500 mg regimen.It is of great value to assess the response and the safety of 750 mg regimen in Chinese population in future researches.
引文
1.Davidson R,Cavalcanti R,Brunton JL,et al.Resistance to levofloxacin and failure of treatment of pneumococcus pneumoniae[J].N Engl J Med 2002,346:747-750.
2.Eagle H,Fleishman R,Levy M.'Continuous' vs.'discontinuous' therapy with penicillin[J].N Engl J Med 1953,248:481-488.
3.Craig WA,Ebert SC.Killing and regrowth of bacteria in vitro:A review[J].Scand J Infect Dis 1991,$74:63-70.
4.Craig WA,Andes DR.Correlation of the magnitude of the AUC24/MIC for 6fluoroquinolones against Streptococcus pneumoniae with survival and bactericidal activity in an animal model[R].40th Interscience Conference on Antimicrobial Agents and Chemotherapy,Toronto,Canada 2000.
5.Lacy MK,Lu W,Xu X,et al..Pharmacodynamic comparison of levofloxacin,ciprofloxacin and ampicillin against Streptococcus pneumoniae in an in vitro model of infection[J].Antimicrob Agents Chemother 1999,43:672-677.
6.Drusano GL,Johnson DE,Rosen M,et al.Pharmacodynamics of a fluoroquinolone antimicrobial agent in a neutropenic rat model of Pseudomonas sepsis[J].Antimicrob Agents Chemother 1993,37:483-490.
7.Ambrose PG,Grasela DM,Grasela TH,et al..Pharmacodynamics of fluoroquinolones against Streptococcus pneumoniae in patients with community-acquired respiratory tract infections[J].Antimicrob Agents Chemother 2001,45:2793-2797.
8.Chien SC,Chow AT,Natarajan J,et al..Absence of age and gender effects on pharmacokinetics of a single 500-milligram oral dose of levofloxacin in healthy subjects [J]. Antimicrob Agents Chemother 1997, 41:1562-1565.
9. Preston SL, Drusano GL, Berman AL, et al.. Levofloxacin population pharmacokinetics and creation of a demographic model for prediction of individual drug clearance in patients with serious community-acquired infection [J]. Antimicrob Agents Chemother 1998, 42:1098-1104.
10. Furlanut M, Brollo L, Lugatti E, et al.. Pharmacokinetic aspects of levofloxacin 500 mg once daily during sequential intravenous/oral therapy in patients with lower respiratory tract infections [J]. Journal of Antimicrob Chemother 2003, 51:101-106.
11. Kiser TH, Hoody DW, Obristch MD, et al.. Levofloxacin pharmacokinetics and pharmacodynamics in patients with severe burn injury [J]. Antimicrob Agents Chemother 2006, 50:1937-1945.
12. Traunmuller F, Thalhammer-Scherrer R, Locker GJ, et al.. Single-dose pharmacokinetics of levofloxacin during continuous veno-venous haemofiltration in critically ill patients [J]. Journal of Antimicrob Chemother 2001,47:229-231.
13. Czock D, Husig-Linde C, Langhoff A, et al.. Pharmacokinetics of moxifloxacin and levofloxacin in intensive care unit patients who have acute renal failure and undergo extended daily dialysis [J]. Clin J Am Soc Nephrol 2006,1:1263-1268.
14. Sowinski KM, Lucksiri A, Kays MB, et al.. Levofloxacin pharmacokinetics in ESRD and removal by the cellulose acetate high performance-210 hemodialyzer [J]. Am J Kidney Dis 2003, 42:342-349.
15. Preston SL, Drusano GL, Berman AL, et al.. Pharmacodynamics of levofloxacin, a new paradigm for clinical trial [J]. JAMA 1998, 279:125-129.
16. Lipsitch M, Samore MH. Antimicrobial use and antimicrobial resistance: a population perspective. Emerg Infect Dis 2002, 8:347-354.
17. Allen GP, Kaatz GW, Rybak MJ. Activities of mutant prevention concentration-target moxifloxacin and levofloxacin against Streptococcus pneumoniae in an in vitro pharmacodynamic model [J]. Antimicrob Agents Chemother 2003, 47:2606-2614.
18. David Czock, Frieder Keller. Mechanism-based pharmacokinetic-pharmacodynamic modeling of antimicrobial effects [J]. Journal of pharmacokinetic and pharmacodynamic 2007, 34:727-751.
1. Guidance for Industry population pharmacokinetics [R]. United States Food and Drug Administration, 1999.
2. Challenge and opportunity on the critical path to new medical products [R]. United States Food and Drug Administration, 2004.
3. Sheiner LB, Beal SL. Evaluation of methods for estimating population pharmacokinetics parameters. I. Michaelis-Menten model: routine clinical pharmacokinetic data [J]. J Pharmacokinet Biopharm 1980, 8(6):553-571.
4. Sheiner BL, Beal SL. Evaluation of methods for estimating population pharmacokinetic parameters. II. Biexponential model and experimental pharmacokinetic data [J]. J Pharmacokinet Biopharm 1981, 9(5):635-651.
5. Bies RR, Feng Y, Lotrich FE, et al.. Utility of sparse concentration sampling for citalopram in elderly clinical trial subjects [J]. J Clin Pharmacol 2004, 44(12):1352-1359.
6. Beal SL, Sheiner LB. Estimating population kinetics [J]. Crit Rev Biomed Eng 1982, 8(3): 195-222.
7. Sheiner LB, Rosenberg B, Marathe W. Estimation of population characteristics of pharmacokinetic parameters from routine clinical data [J]. J Pharmacokinet Biopharm 1977, 5(5):445-479.
8. Sheiner LB, Beal SL. Evaluation of methods for estimating population pharmacokinetic parameters. I. Biexponential model and experimental pharmacokinetic data [J]. J Pharmacokinet Biopharm 1981, 9(5):635-651.
9. Sheiner LB. The population approach to pharmacokinetic data analysis: rationale and standard data analysis methods [J]. Drug Metab Rev 1984, 15(2):153-171.
10. Beal SL. Population pharmacokinetic data and parameter estimation based on their first two statistical moments [J]. Drug Metab Rev 1984, 15(2): 173-193.
11. Whiting B, Kelman AW, Grevel J. Population pharmacokinetics: theory and clinical application [J]. Clin Pharmacokinet 1986, 11(5):387-401.
12. King G, Honaker J, Joseph A, Scheve. K. Analyzing Incomplete Political Science Data [J]. American Political Science Review 2001, 95(1):49-69.
13. Rubin D. Multiple Imputation After 18+ Years [J]. Journal of the American Statistical Association 1996, 91: 473-489.
14. Shafer SL. Pharmacokinetic and pharmacodynamic analysis with NONMEM [R]. California, USA, 2006.
15. Lindstrom MJ, Bates DM. Nonlinear mixed effects models for repeated measures data [J]. Biometrics 1990, 46(6):673-687.
16. S-Plus TM [R]. Insightful, Seattle, Washington, 2002.
17. Vonesh EF, Carter RL. Mixed effects nonlinear regression for unbalanced repeated measures [J]. Biometrics 1992, 46(6):673-687.
18. Wolfinger RD. Laplace's approximation for nonlinear mixed models [J]. Biometrika 1993, 80:791-795.
19. Mallet A. A maximum likelihood estimation method for random coefficient regression models [J]. Biometrika 1986, 73:645-656.
20. Beal SL, Sheiner LB. NONMEM Users Guide—Part VII. Conditional estimation methods [R]. University of California, San Francisco, 1992.
21. Phillip G Permutation, parametric, and bootstrap tests of hypotheses: Third edition [M]. Springer, 2004.
22. Schuirmann DJ. A comparison of the two one-sided tests procedure and the power approach for assessing the equivalence of average bioavailability [J]. J Pharmacokinet Biopharm 1987,15(6):657-680.
23. Kaniwa N, Aoyagi N, Ogata H, et al.. Application of the NONMEM method to evaluation of the bioavailability of drug products [J]. J Pharm Sci 1990, 79(12):1116-1120.
24. Maier GA, Lockwood GG, Oppermann JA, et al.. Characterization of the highly variable bioavailability of tiludronate in normal volunteers using population pharmacokinetic methodologies [J]. Eur J Drug Metab Pharmacokinet 1999, 24(3):249-254.
25. Pentikis HS, Henderson JD, Tran JL, et al.. Bioequivalence: individual and population compartmental modeling compared to the noncompartmental approach [J]. Pharm Res 1996,13(7):1116-1121.
26. Reith DM, Hooper WD, Parke J, et al.. Population pharmacokinetic modeling of steady state carbamazepine clearance in children, adolescents, and adults [J]. J Pharmacokinet Biopharm 2001, 28(1):79-92.
27. Csajka C, Verotta D. Pharmacokinetic-pharmacodynamic modelling: history and perspectives [J]. J Pharmacokinet Pharmacodyn 2006, 33(3):1-53.
28. Ette EI, Garg V, Jayaraj A. A rational approach to drug development: the exploratory phase [J]. Clin Res Regul Affairs 2004, 21(2): 155-177.
29. Zhou Y. Choice of designs and doses for early phase trials [J]. Fundam Clin Pharmacol 2004,18(3):373-378.
31. Jennison C, Turnbull BW. Group sequential methods with applications to clinical trials [R]. Chapman & Hall/CRC, Boca Raton, FL, 2000.
32. Krams M, Lees KR, Hacke W, et al.. Acute stroke therapy by inhibition of neutrophils (ASTIN): an adaptive dose-response study of UK279276 in acute ischemic stroke [J]. Stroke 2003, 34(11):2543-2548.
33. Sambol NC, Sheiner LB. Population dose versus response of betaxolol and atenolol: a comparison of potency and variability [J]. Clin Pharmacol Ther 1991,49(1):24-31.
2.Eagle H,Fleishman R,Levy M.'Continuous' vs.'discontinuous' therapy with penicillin[J].N Engl J Med 1953,248:481-488.
3.Craig WA,Ebert SC.Killing and regrowth of bacteria in vitro:A review[J].Scand J Infect Dis 1991,$74:63-70.
4.Craig WA,Andes DR.Correlation of the magnitude of the AUC24/MIC for 6fluoroquinolones against Streptococcus pneumoniae with survival and bactericidal activity in an animal model[R].40th Interscience Conference on Antimicrobial Agents and Chemotherapy,Toronto,Canada 2000.
5.Lacy MK,Lu W,Xu X,et al..Pharmacodynamic comparison of levofloxacin,ciprofloxacin and ampicillin against Streptococcus pneumoniae in an in vitro model of infection[J].Antimicrob Agents Chemother 1999,43:672-677.
6.Drusano GL,Johnson DE,Rosen M,et al.Pharmacodynamics of a fluoroquinolone antimicrobial agent in a neutropenic rat model of Pseudomonas sepsis[J].Antimicrob Agents Chemother 1993,37:483-490.
7.Ambrose PG,Grasela DM,Grasela TH,et al..Pharmacodynamics of fluoroquinolones against Streptococcus pneumoniae in patients with community-acquired respiratory tract infections[J].Antimicrob Agents Chemother 2001,45:2793-2797.
8.Chien SC,Chow AT,Natarajan J,et al..Absence of age and gender effects on pharmacokinetics of a single 500-milligram oral dose of levofloxacin in healthy subjects [J]. Antimicrob Agents Chemother 1997, 41:1562-1565.
9. Preston SL, Drusano GL, Berman AL, et al.. Levofloxacin population pharmacokinetics and creation of a demographic model for prediction of individual drug clearance in patients with serious community-acquired infection [J]. Antimicrob Agents Chemother 1998, 42:1098-1104.
10. Furlanut M, Brollo L, Lugatti E, et al.. Pharmacokinetic aspects of levofloxacin 500 mg once daily during sequential intravenous/oral therapy in patients with lower respiratory tract infections [J]. Journal of Antimicrob Chemother 2003, 51:101-106.
11. Kiser TH, Hoody DW, Obristch MD, et al.. Levofloxacin pharmacokinetics and pharmacodynamics in patients with severe burn injury [J]. Antimicrob Agents Chemother 2006, 50:1937-1945.
12. Traunmuller F, Thalhammer-Scherrer R, Locker GJ, et al.. Single-dose pharmacokinetics of levofloxacin during continuous veno-venous haemofiltration in critically ill patients [J]. Journal of Antimicrob Chemother 2001,47:229-231.
13. Czock D, Husig-Linde C, Langhoff A, et al.. Pharmacokinetics of moxifloxacin and levofloxacin in intensive care unit patients who have acute renal failure and undergo extended daily dialysis [J]. Clin J Am Soc Nephrol 2006,1:1263-1268.
14. Sowinski KM, Lucksiri A, Kays MB, et al.. Levofloxacin pharmacokinetics in ESRD and removal by the cellulose acetate high performance-210 hemodialyzer [J]. Am J Kidney Dis 2003, 42:342-349.
15. Preston SL, Drusano GL, Berman AL, et al.. Pharmacodynamics of levofloxacin, a new paradigm for clinical trial [J]. JAMA 1998, 279:125-129.
16. Lipsitch M, Samore MH. Antimicrobial use and antimicrobial resistance: a population perspective. Emerg Infect Dis 2002, 8:347-354.
17. Allen GP, Kaatz GW, Rybak MJ. Activities of mutant prevention concentration-target moxifloxacin and levofloxacin against Streptococcus pneumoniae in an in vitro pharmacodynamic model [J]. Antimicrob Agents Chemother 2003, 47:2606-2614.
18. David Czock, Frieder Keller. Mechanism-based pharmacokinetic-pharmacodynamic modeling of antimicrobial effects [J]. Journal of pharmacokinetic and pharmacodynamic 2007, 34:727-751.
1. Guidance for Industry population pharmacokinetics [R]. United States Food and Drug Administration, 1999.
2. Challenge and opportunity on the critical path to new medical products [R]. United States Food and Drug Administration, 2004.
3. Sheiner LB, Beal SL. Evaluation of methods for estimating population pharmacokinetics parameters. I. Michaelis-Menten model: routine clinical pharmacokinetic data [J]. J Pharmacokinet Biopharm 1980, 8(6):553-571.
4. Sheiner BL, Beal SL. Evaluation of methods for estimating population pharmacokinetic parameters. II. Biexponential model and experimental pharmacokinetic data [J]. J Pharmacokinet Biopharm 1981, 9(5):635-651.
5. Bies RR, Feng Y, Lotrich FE, et al.. Utility of sparse concentration sampling for citalopram in elderly clinical trial subjects [J]. J Clin Pharmacol 2004, 44(12):1352-1359.
6. Beal SL, Sheiner LB. Estimating population kinetics [J]. Crit Rev Biomed Eng 1982, 8(3): 195-222.
7. Sheiner LB, Rosenberg B, Marathe W. Estimation of population characteristics of pharmacokinetic parameters from routine clinical data [J]. J Pharmacokinet Biopharm 1977, 5(5):445-479.
8. Sheiner LB, Beal SL. Evaluation of methods for estimating population pharmacokinetic parameters. I. Biexponential model and experimental pharmacokinetic data [J]. J Pharmacokinet Biopharm 1981, 9(5):635-651.
9. Sheiner LB. The population approach to pharmacokinetic data analysis: rationale and standard data analysis methods [J]. Drug Metab Rev 1984, 15(2):153-171.
10. Beal SL. Population pharmacokinetic data and parameter estimation based on their first two statistical moments [J]. Drug Metab Rev 1984, 15(2): 173-193.
11. Whiting B, Kelman AW, Grevel J. Population pharmacokinetics: theory and clinical application [J]. Clin Pharmacokinet 1986, 11(5):387-401.
12. King G, Honaker J, Joseph A, Scheve. K. Analyzing Incomplete Political Science Data [J]. American Political Science Review 2001, 95(1):49-69.
13. Rubin D. Multiple Imputation After 18+ Years [J]. Journal of the American Statistical Association 1996, 91: 473-489.
14. Shafer SL. Pharmacokinetic and pharmacodynamic analysis with NONMEM [R]. California, USA, 2006.
15. Lindstrom MJ, Bates DM. Nonlinear mixed effects models for repeated measures data [J]. Biometrics 1990, 46(6):673-687.
16. S-Plus TM [R]. Insightful, Seattle, Washington, 2002.
17. Vonesh EF, Carter RL. Mixed effects nonlinear regression for unbalanced repeated measures [J]. Biometrics 1992, 46(6):673-687.
18. Wolfinger RD. Laplace's approximation for nonlinear mixed models [J]. Biometrika 1993, 80:791-795.
19. Mallet A. A maximum likelihood estimation method for random coefficient regression models [J]. Biometrika 1986, 73:645-656.
20. Beal SL, Sheiner LB. NONMEM Users Guide—Part VII. Conditional estimation methods [R]. University of California, San Francisco, 1992.
21. Phillip G Permutation, parametric, and bootstrap tests of hypotheses: Third edition [M]. Springer, 2004.
22. Schuirmann DJ. A comparison of the two one-sided tests procedure and the power approach for assessing the equivalence of average bioavailability [J]. J Pharmacokinet Biopharm 1987,15(6):657-680.
23. Kaniwa N, Aoyagi N, Ogata H, et al.. Application of the NONMEM method to evaluation of the bioavailability of drug products [J]. J Pharm Sci 1990, 79(12):1116-1120.
24. Maier GA, Lockwood GG, Oppermann JA, et al.. Characterization of the highly variable bioavailability of tiludronate in normal volunteers using population pharmacokinetic methodologies [J]. Eur J Drug Metab Pharmacokinet 1999, 24(3):249-254.
25. Pentikis HS, Henderson JD, Tran JL, et al.. Bioequivalence: individual and population compartmental modeling compared to the noncompartmental approach [J]. Pharm Res 1996,13(7):1116-1121.
26. Reith DM, Hooper WD, Parke J, et al.. Population pharmacokinetic modeling of steady state carbamazepine clearance in children, adolescents, and adults [J]. J Pharmacokinet Biopharm 2001, 28(1):79-92.
27. Csajka C, Verotta D. Pharmacokinetic-pharmacodynamic modelling: history and perspectives [J]. J Pharmacokinet Pharmacodyn 2006, 33(3):1-53.
28. Ette EI, Garg V, Jayaraj A. A rational approach to drug development: the exploratory phase [J]. Clin Res Regul Affairs 2004, 21(2): 155-177.
29. Zhou Y. Choice of designs and doses for early phase trials [J]. Fundam Clin Pharmacol 2004,18(3):373-378.
31. Jennison C, Turnbull BW. Group sequential methods with applications to clinical trials [R]. Chapman & Hall/CRC, Boca Raton, FL, 2000.
32. Krams M, Lees KR, Hacke W, et al.. Acute stroke therapy by inhibition of neutrophils (ASTIN): an adaptive dose-response study of UK279276 in acute ischemic stroke [J]. Stroke 2003, 34(11):2543-2548.
33. Sambol NC, Sheiner LB. Population dose versus response of betaxolol and atenolol: a comparison of potency and variability [J]. Clin Pharmacol Ther 1991,49(1):24-31.