Introduction

Community-acquired pneumonia (CAP) presents a significant public health concern both in China and globally.1 Epidemiological surveys in Chinese adults have identified Streptococcus pneumoniae and Mycoplasma pneumoniae as predominant pathogens, with others such as Staphylococcus aureus, Chlamydia pneumoniae, and Haemophilus influenzae also commonly implicated.2–6 Treatment duration typically ranges from 5 to 7 days for mild or moderate CAP, and 10 to 14 days for severe CAP or cases with extrapulmonary complications. However, infections involving Staphylococcus aureus may necessitate extended treatment periods of 14 to 21 days.7–9 Recently, a few cases of community-acquired methicillin-resistant S. aureus (CA-MRSA) pneumonia have emerged in China,10,11 posing a significant challenge due to its high mortality rate of 41.1%.12 Therefore, it is crucial to enhance treatment efficacy and reduce the duration of therapy for severe pneumonia.

Currently, vancomycin (VCM), linezolid, and norvancomycin (NVCM) are the primary empiric antibiotic therapies for gram-positive cocci pulmonary infections.1 NVCM, a glycopeptide antimicrobial agent developed independently by the North China Pharmaceutical Group Corporation, exhibits a molecular structure distinct from that of vancomycin due to the absence of one methyl group.13 In China, NVCM has become a standard treatment for infections associated with antibiotic-resistant gram-positive bacteria such as MRSA, owing to its similar antibiotic activity and spectrum to VCM.13,14 24-h area under the concentration–time curve to minimum inhibitory concentration ratio (AUC24h/MIC) of 361 was considered as pharmacokinetics/pharmacodynamics (PK/PD) efficacy index of NVCM for staphylococcal infection.13 Current guidelines for the diagnosis and treatment of CAP recommend a standard dose of 800 mg every 12 h (q12h).1

Optimal antimicrobial dosing should consider the pharmacokinetics within the target population, the relationship between drug exposure and outcomes and the vulnerability of the pathogens involved.15 In typical adults, standard dose evaluations of antimicrobials often rely on a “normal” dosage, overlooking the significant impact of underlying diseases on drug exposure in clinical scenarios. Recent data highlights a concerning decline in the efficacy of NVCM and other similar antibiotics against MRSA infections due to increasing minimum inhibitory concentration (MIC) levels.16–18 Conventional dosage of NVCM in clinical practice contributes to prolonged pneumonia treatments and the emergence of drug-resistant bacterial strains.1,19 These findings emphasize the need for population pharmacokinetic (PPK) models and personalized dosage regimens, since there are limited PPK/PD studies of NVCM in pneumonia patients, and the optimal dosing regimen of NVCM in severe pneumonia has not been determined. Therefore, optimizing the NVCM dosage regimens for adult patients with mild (or moderate) and severe pneumonia is essential, based on a defined PK/PD breakpoint.

This prospective study aimed to construct a PPK model, comprehensively assess the pharmacokinetic characteristics and variations of NVCM, and determine optimized dosage regimens of NVCM in Chinese adult patients with mild (or moderate) and severe pneumonia.

Materials and Methods Study Design

This prospective, single-center, open-label observational study was conducted at the Second Hospital of Hebei Medical University between November 2020 and March 2022 (Shijiazhuang, China). The experimental protocols were carried out, and written informed consent was obtained in accordance with the ethical standards established by the 1964 Declaration of Helsinki. The study was registered with the Chinese Clinical Trial Registry (registration number: ChiCTR2000039794, the date of registration: 2020–11-23) and approved by the Ethics Committee of the Second Hospital of Hebei Medical University (2020EC07-05-2).

Patient Cohort and Data Acquisition Patients

This research included hospitalized patients aged ≥18 years who were diagnosed with CAP. All enrolled patients presented suspected or confirmed pulmonary infections attributable to gram-positive bacteria. NVCM therapy was administered via intravenous infusion, with concurrent monitoring of serum NVCM concentrations. Pregnant individuals and those undergoing renal replacement therapy were excluded from the analysis. Demographic features, along with laboratory and clinical examination data, were acquired for all enrolled patients. In addition, information related to NVCM treatment was collected from each enrolled patient, including the dosing information (eg, dosage regimen, date, time, and duration), serum concentrations, sampling time, and concomitant medications.

Dosage Regimen and Sampling

To treat pulmonary infections, NVCM (400 mg/bottle; North China Pharmaceutical Group Corporation, Shijiazhuang, China) was administered intravenously at a dose of 800 mg every 12 h for 1 h via a syringe pump. Serum samples (3 mL) were collected at the baseline (0.5 h before administration) and 0.5–1.5 h after the first dose of NVCM was finished. We repeated the serum sampling at the second, third, fourth, and the last day of NVCM treatment.

Assay of Serum NVCM and Creatinine

NVCM concentrations were determined by a high performance liquid chromatography (HPLC) method using the 2489 ultraviolet detector system (e2695, Waters Corp., MA, USA). A Diamonsil C18 column (150 mm × 4.6 mm, 5 µm) was used, and the detection wavelength was 236 nm. The calibration curve ranged from 0.25 to 100 μg/mL. The limit of detection for the analysis of NVCM was 0.05 μg/mL. Serum samples were processed by protein precipitation, and vancomycin was used as internal standard.20 Serum creatinine (Scr) concentrations were measured by an enzymatic method using a Creatinine Test Kit (Creatinine plus ver.2 (CREP2), Roche Diagnostics GmbH, Germany).

PPK Modeling Basic Model

PPK analysis was conducted using the nonlinear mixed effects modeling software NONMEM (version 7.5.0, ICON Development Solutions, MD, USA). Employing first-order conditional estimation method (FOCE-I) with an interaction option to estimate the pharmacokinetic parameters.

The compartmental model of NVCM was preliminarily established by an exploratory analysis of blood sample data. One- or two-compartment models were determined based on the principle of minimizing the objective function value (OFV) and visual inspection of the routine diagnostic plots of the model.

Interindividual variability in PK parameters was assessed by the following exponential model:

Pi = TV(P) x eηi

where Pi denotes the parameter value of the ith subject; TV(P) represents the typical parameter value in the population; and ηi signifies the variability between subjects, assuming to adhere to a normal distribution with the mean and variance of 0 and ω2, respectively.

The residual variability model selection was considered by OFV and visual inspection of the routine diagnostic plots. And the models are expressed as follows:

Additive error model: Cobs = Cpred+ ε1

Proportional error model: Cobs = Cpredx(1+ ε1)

Exponential error model: Cobs = Cpredxeε1

Combined error model: Cobs = Cpredx (1+ ε1)+ε2

where Cobs is the observed concentration of subjects, Cpred is the simulated concentration of subjects, ε1 and ε2 are proportional and additive residual variabilities, respectively, which are all assumed to follow a normal distribution with a mean of 0 and variance of σ2.

Evaluation of Covariate

Demographic information, including sex, age, weight, and physiological characteristics (hepatic and renal function) of patients were evaluated as possible covariates. Stepwise covariate modeling and likelihood ratio tests were used to assess the effect of each variable.

Covariate selection was conducted through forward inclusion and backward elimination procedure. During forward inclusion, a covariate was deemed significant if a reduction in the OFV surpassed 3.84 with P < 0.05 (chi-squared distribution). All significant covariates were integrated into the whole model. Subsequently, backward elimination was performed, removing a covariate from the whole model if its removal resulted in an increase in the OFV of less than 6.64 (P < 0.01, chi-squared distribution). This process led to the derivation of the final PPK model.

Model Evaluation and Verification

The stability of the final model was examined through the bootstrap approach, implemented using the Perl-speaks-NONMEM (PsN, v.5.2.6; http://uupharmacometrics.github.io/PsN/) module.21 The acquired dataset was replicated via iterative sampling. A 1000-times resampling bootstrap procedure was employed to assess the reliability of the final model. The values (median, 95% confidence interval (CI)) obtained from the bootstrap procedure were compared with the estimated parameters of the final model.

The model was verified based on goodness-of-fit (GOF) plots and statistical criteria. Initially, GOF plots, comprising observed (DV) versus individual prediction (IPRED), DV versus population prediction (PRED), conditional weighted residuals (CWRES) versus time, and CWRES versus PRED, were employed for diagnostic purposes.22 Line of identity was incorporated into the DV-IPRED and the DV-PRED plots. The proximity of the data points to the unit line indicated the quality of the fit of the final model.

Additionally, normalized prediction distribution errors (NPDE) and visual predictive check (VPC) were used to verify the predictive performance of the final model.23 For the VPC, the 5/50/95th percentiles of the simulated data were processed using the RStudio environment (v.4.3.3, http://www.rstudio.com/). For the NPDE, we used an NPDE R package for model validation, and the results were presented graphically by using density histogram and quantile–quantile plot (Q-Q plot). The analysis was based on 1000 simulations of the dataset.

Dosing Optimization Analysis of the Initial Dosage Regimen

The established PPK model was used to simulate individual PK parameters (AUC0-24h, AUC24-48h, AUC48-72h, and AUCss,24h) by Bayesian feedback method. Simulated prediction data were generated based on a dose of 800 mg administered every 12 h, an infusion duration of 1 h, and a dense sampling interval of 0.1 h. PK parameters were calculated as follows:

AUCss,24hi = initial daily dose/CLi

The target AUC was defined based on the PK-PD breakpoint of NVCM in Staphylococcus aureus, with an AUC24h/MIC ratio of 361.13 According to previous literature,16Staphylococcus aureus, including MRSA, typically exhibits MIC values within the range of 0.25–1 mg/L. In our study, MIC values of 0.5 and 1 mg/L were assumed, representing the more frequently observed MIC values for NVCM. The probability of target attainment (PTA) of the initial dosage regimen was calculated and defined as the number of qualified patients divided by the number of total patients.

Dosing Simulations

Monte Carlo simulations were carried out for different subgroups according to the covariates obtained from the final model. According to the instructions, the total daily dose of NVCM ranging from 800 to 2400 mg, divided into 2–3 intravenous doses, with each dose not exceeding 1000 mg, aimed at minimizing the risk of nephrotoxicity. Randomly generated sample simulation data includes dosing frequencies of q8h, q12h, and q24h, and preset single doses of 200, 400, 600, and 800 mg. A total of 5000 simulations were performed using the RStudio mrgsolve package, and the PK profiles were estimated for each simulated patient.

The first set of simulations was performed, and the optimal dosage was defined as the one that maximized the proportion of simulated patients with AUCss,24h/MIC ≥361. According to the 2020 Infectious Diseases Society of America (IDSA) guidelines and others,24–28 achieving targeted exposure within the first 24–48 h of therapy is crucial. Considering that an alternative goal of initial therapy may be to maximize the effect early in some patients with severe CAP, a second set of simulations was performed to identify the dosage that maximized the proportion of patients achieve AUC24-48h/MIC over 361. To achieve the target AUC/MIC in approximately 90% of the patients, the PTA ≥ 90% was considered as acceptable.

Results Patient Characteristics

Initially, 46 patients were assessed for eligibility. Subsequently, four patients withdrew from the study, while 2 excluded due to a treatment duration of NVCM less than 72 h. Additionally, 6 patients were excluded due to missing serum samples. Consequently, 34 patients were included in the final analysis, comprising 23 from the Department of Respiratory and Critical Care Medicine, 6 from the Department of Neurosurgery, and 5 from the Department of Hematology. Serum NVCM concentration monitoring was conducted in all patients, including 17 males, with an average age of 54.8 years. Each patient underwent the collection of 3–8 blood samples for analysis and a total of 231 NVCM concentrations were available. A total of 115 (49.8%) NVCM samples were measured as peak concentrations and 116 (50.2%) were trough concentrations. Patient characteristics are summarized in Table 1.

Table 1 Baseline Characteristics of Patients (n = 34)

Finally, 10 bacterial strains were isolated from the alveolar lavage fluid or sputum culture of 34 enrolled patients. These strains were identified as methicillin-sensitive S. aureus (4), S. epidermidis (2) and MRSA (4). The results from susceptibility testing with five anti-bacterial agents, including norvancomycin, showed that all of the 10 Staphylococcus isolates were susceptible to both norvancomycin and vancomycin. And the MICs of the norvancomycin against the 10 strains of MRSA or Staphylococcus were 1 mg/L or lower. The supplement table we provide also details the information on bacterial culture results and blood sampling.

PPK Modeling

In total, 231 NVCM concentrations were available for PPK modeling. The data conformed better to a two-compartment model (OFV = 513.504) with first-order elimination, as opposed to a one-compartment model (OFV = 585.283). Interindividual variability in PK parameters was best described by an exponential model. And the residual variability was fitted to a combined error model.

Following the establishment of the two-compartment model, the impacts of various covariates on PK parameters were explored. The covariates examined were age, sex, weight (WT), height, C-reactive protein (CRP), alkaline phosphatase (ALP), alanine aminotransferase (ALT), aspartate aminotransferase (AST), total bilirubin (TBIL), Scr, direct bilirubin (DBIL), and creatinine clearance (CLcr).

During the forward inclusion, height, age, and Scr were added to build the initial model, with a significant decrease in OFV exceeding 3.84 (P < 0.05). Furthermore, inappropriate covariates were excluded using backward elimination. Age and Scr on CL (ΔOFV = −23.95) were retained as influential covariates in the final model. Parameter estimates for the final model and bootstrap validation are summarized in Table 2. The bootstrap statistical analysis (n = 1000) showed that the parameters derived from the final PPK model closely resembled the mean bootstrap estimates. The respective parameter values fell within the 95% CI, underscoring a robust fit. The final model and typical PK parameters are shown in the following equations:

Table 2 Population Pharmacokinetic Parameter Estimates for the Final Model and Bootstrap Validation

CL (L/h) = 3.15x (Age/57.5)−0.425xe((Scr−59) x (−0.00886)) xeηCL

V1 (L) = 12.3xeηV1

V2 (L) = 115xeηV2

Q (L/h) = 5.21xeηQ

The routine diagnostic GOF plots are shown in Figure 1. Both the PRED and IPRED concentrations aligned well with observed concentrations, distributed evenly on either side of the line of identity. The CWRES values ranged from −2 to 2, indicating adherence to the acceptable GOF criteria for the final model.

Figure 1 Goodness-of-fit plots for final model (A-D).

Notes: (A) population prediction (PRED) versus observed (DV). (B) individual prediction (IPRED) versus DV. (C) conditional weighted residuals (CWRES) versus PRED. (D) CWRES versus time. Points are individual data. Dashed lines in the middle represent the line of identity (A and B) or the line at zero (C and D). Red solid lines represent the linear regression line.

The VPC outcomes for the final model showed that 93.1% (215/231) of the observed data points fell within the 90% CI of the corresponding percentiles of the predicted data. This suggests that the final model accurately predicted the observed concentrations, thus reflecting its reliability. Visual inspection and statistical tests of the NPDE results based on 1000 simulations show that the final model has good predictive performance. The model is shown to accurately describe typical trends as well as the variability in the populations. The density histogram and scatter plot versus time are shown in Figure 2.

Figure 2 Normalized prediction distribution errors (NPDE).

Notes: (a) histogram of NPDE with the density of the standard normal distribution; (b) quantile–quantile plot of NPDE. (c) scatterplot of NPDE versus time; (d) scatterplot of NPDE versus predicted concentration.

Dosing Optimization

The PTA at different MIC levels for the initial dosage regimen is shown in Figure 3. When the MIC was set to 1 mg/L, only 29.41% of the patients could achieve the target concentration on the second day of administration (AUC24-48h: 341.87 ± 109.10 mg·h/L, Mean ± SD), and 76.47% patients could achieve the target value after the plasma concentration reached a steady state (AUCss, 24h: 505.63 ± 162.11 mg·h/L, Mean ± SD). According to previous literature,16 a MIC of 1 mg/L was more frequently observed, indicating that the majority of patients were at risk of underdosing. When the MIC was set to 0.5 mg/L, the target AUC/MIC was achieved at the initial dose (AUC24-48h: 683.74 ± 218.20 mg·h/L, AUCss, 24h: 1011.26 ± 324.22 mg·h/L, Mean ± SD), indicating a risk of overdose.

Figure 3 Probability of target attainment (PTA) in initial dosage regimen (800 mg every 12 h).

Notes: The 24-h area under the concentration–time curve to minimum inhibitory concentration ratio (AUC24h/MIC) target is 361. AUC0-24h, AUC24-48h and AUC48-72h correspond to the AUC estimated during the first, second, and third days of therapy. AUCss,24h correspond to AUC of steady-state.

Based on the covariate selected in the final population pharmacokinetic model, age and Scr were the most important covariates implemented in CL. Thus, various subgroups (Age: 18–64, 65–80 years; Scr: <60, 60 – <133, 133 – <178 µmol/L) were divided. Dosage regimens with different doses (200, 400, 600, and 800 mg) and frequencies (q8h, q12h, q24h) were simulated according to different subgroups. Fraction of patients achieving the AUC24-48h and AUCss,24h target for different dosage regimens in each subgroup are shown in Figure 4.

Figure 4 Fraction of patients achieving the AUC/MIC target in each subgroup.

Notes: (a) target exposure was set as AUC24-48h/MIC = 361, MIC = 0.5mg/L. (b) target exposure was set as AUC24-48h/MIC = 361, MIC = 1mg/L. (c) target exposure was set as AUCss,24h/MIC = 361, MIC = 0.5mg/L. (d) target exposure was set as AUCss,24h/MIC = 361, MIC = 1mg/L.

The final tailored doses for patients are summarized in Table 3. For mild or moderate pneumonia caused by gram-positive cocci, the recommended dosage regimen was 400–800 mg q12h when the target PK/PD breakpoint was set as AUCss,24h/MIC ≥361 and MIC was assumed as 1 mg/L. For severe pneumonia-like CA-MRSA pneumonia, NVCM-targeted exposure should be achieved early during the course of therapy, preferably within the first 24–48 h, and the recommended dosage regimen was 600–800 mg q8h when the target PK/PD breakpoint was set as AUC24-48h/MIC. Furthermore, when the MIC was assumed as 0.5 mg/L, the recommended dose of administration was significantly reduced compared with the dose of 800 mg q12h in previous guidelines, as shown in Table 3.

Table 3 Tailored Dose of NVCM in Adults with Pneumonia

Discussion

Overall, 85.9% of NVCM is excreted through the kidneys, making renal function a crucial factor in its pharmacokinetics. Renal parameters, such as Scr, urea, and CLcr, are used to evaluate the renal function. Numerous studies on population pharmacokinetic models for NVCM and VCM in adults have demonstrated that CLcr is a key covariate affecting pharmacokinetics.13,29 Our study identified age and Scr as the most important covariates regulating CL parameter in patients with pulmonary infections. In the final population model of NVCM developed by other researchers,13 the typical value of CL in the population was 2.54 L/h, which is close to our finding of 3.15 L/h. Their study included CLcr as a covariate for CL, but the majority of the dataset (60%) consisted of patients with various gram-positive bacterial infections, including skin and bloodstream infections, with the remaining 40% of the modeled dataset coming from healthy volunteers. Thus, extrapolation to patients with CAP was limited, as the pharmacokinetics within the target population may change. The typical value of the volume of peripheral distribution (V2) in their study was 21 L, which is lower to our finding of 115 L, indicating increased V2 was observed in patients with pneumonia. Further studies are needed to confirm our results.

According to the 2020 VCM guidelines developed by IDSA and others,28,30 achieving targeted VCM exposure within the first 24–48 h of therapy is crucial (Grade A-II). Early treatment is also appropriate for NVCM, owing to its similar molecular structure, antibiotic activity, and spectrum to VCM. The current recommended dose of NVCM is 800 mg every 12 h, which poses a high risk of underdosing in the target population. Assuming an MIC of 1 mg/L, only 29.41% of the patients in our study achieved the target value within the first 24–48 h after administration. Therefore, to ensure early and appropriate therapy, the dose of NVCM should be increased in patients with severe pulmonary infections. Bayesian-derived AUC monitoring is suitable in such cases as it allows early assessment of AUC target attainment without requiring steady-state serum drug concentrations.

According to the final tailored dose for patients, higher doses (600–800 mg every 8 h) are required if the PK/PD target was AUC24-48h/MIC ≥361. Assuming an MIC of 1 mg/L, the recommended dosage regimen was 400–800 mg q12 h when the target PK/PD breakpoint was set as AUCss,24h ≥ 361 mg·h/L for mild or moderate CAP. With increasing age and Scr levels, the dose was gradually reduced from 800 mg to 400 mg every 12 h. Patients with abnormal renal function were not included in the model, so extrapolation to those with renal insufficiency or failure is limited. Prospective studies with larger sample sizes are warranted to verify the safety and efficacy of the recommended dose.

With an increase in the MIC of NVCM against MRSA infections, a higher AUC is required to obtain similar efficacy. Consequently, increasing the daily NVCM dosage is necessary to mitigate the risk of therapeutic failure. However, the PK parameters and safety profile of these higher NVCM dosages remain undetermined. Dosing simulations have been premised on the assumption of linear pharmacokinetics. Nonetheless, extrapolating the dosage regimen beyond the observed range may lead to nonlinear pharmacokinetics, which warrants further investigation. Additionally, when administering higher NVCM doses, it is crucial to be aware of infusion-related adverse events, such as “red man syndrome”. To mitigate these risks, a slow infusion rate of at least 2 hours is recommended. Therefore, it is prudent to use regimens similar to those employed in routine clinical practice and adjust doses empirically. TDM should be conducted, with dose adjustments made as necessary.

This research has several strengths. First, the prospective collection of blood specimens enhances the accuracy of model outcomes compared to retrospective studies. Second, the PK/PD targets, namely, AUC24-48h/MIC and AUCss,24h/MIC ≥361, provide guidance for early and appropriate therapy. Third, recommending dosage regimens tailored to various subgroups categorized by age and Scr levels improves the clinical relevance of the findings. Clinicians could consider these covariates when selecting the optimal initial NVCM dose regimen. We also employed a graphical approach to recommend doses for achieving the desired probability of meeting the PK/PD efficacy target. Additionally, the established population pharmacokinetic model allows for precise adjustment of NVCM dosage when combined with TDM results, thereby facilitating individualized drug administration in future work.

Nevertheless, the research also has some limitations. The sample size was relatively small, external validation was not conducted, limiting the generalizability of the conclusions to similar clinical contexts. The number of patients with mild to moderate renal insufficiency was so small, which restricts extrapolation to individuals with renal insufficiency or failure. And the proposed method of adjusting administration dose and interval according to Scr value should be carefully considered. We included patients with confirmed or suspected Gram-positive cocci pneumonia, but not all bacteriological evidence was available, and pharmacokinetics of antibiotics in different pathogens may vary.

Conclusion

In summary, we successfully established a two-compartment PPK model of NVCM in Chinese patients with CAP. Notably, age and Scr emerged as covariates influencing NVCM pharmacokinetic parameters, and we proposed dosage regimens tailored for different subgroups based on these covariate levels. Our findings suggest that the conventional dosage regimen may not be optimal for adult patients. Higher doses may be necessary to achieve targeted exposure within the first 24–48 h of treatment initiation. To further elucidate the clinical safety and efficacy of this optimized dosage regimen, prospective studies with larger samples are warranted.

Data Sharing Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Ethics Approval and Consent to Participate

This study was carried out in the Second Hospital of Hebei Medical University. The experimental protocols were carried out, and written informed consent was obtained in accordance with the ethical standards established by the 1964 Declaration of Helsinki. The study was registered with the Chinese Clinical Trial Registry (registration number: ChiCTR2000039794, the date of registration: 2020-11-23) and approved by the Ethics Committee of the Second Hospital of Hebei Medical University (2020EC07-05-2).

Acknowledgments

The authors express their gratitude to the patients and colleagues who participated in this study. We are grateful to the listed institutes for funding support. This research was funded by the Hebei Provincial Department of Science and Technology, grant number: 192777127D and the Health Commission of Hebei Province, grant number: 20220109.

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising, or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Funding

This research was funded by the Health Commission of Hebei Province, grant number: 20220109.

Disclosure

The authors report no conflicts of interest in this work.

References

1. Cao B, Huang Y, She DY, et al. Diagnosis and treatment of community-acquired pneumonia in adults: 2016 clinical practice guidelines by the Chinese thoracic society, Chinese medical association. Clin Respir J. 2017;12:1320–1360. doi:10.1111/crj.12674

2. Tao LL, Hu BJ, He LX. Etiology and antimicrobial resistance of community-acquired pneumonia in adult patients in China. Chin Med J. 2012;125:2967–2972. doi:10.3760/cma.j.issn.0366-6999.2012.17.002

3. Liu YF, Gao Y, Chen MF, Cao B, Yang X-H, Wei L. Etiological analysis and predictive diagnostic model building of community acquired pneumonia in adult outpatients in Beijing, China. BMC Infect Dis. 2013;13:309. doi:10.1186/1471-2334-13-309

4. Cao B, Ren LL, Zhao F. Viral and Mycoplasma pneumoniae community-acquired pneumonia and novel clinical outcome evaluation in ambulatory adult patients in China. Eur J Clin Microbiol Infect Dis. 2010;29:1443–1448. doi:10.1007/s10096-010-1003-2

5. Bao Z, Yuan X, Wang L, Sun Y, Dong X. The incidence and etiology of community-acquired pneumonia in fever outpatients. Exp Biol Med. 2012;237:1256–1261. doi:10.1258/ebm.2012.012014

6. Liu Y, Chen M, Zhao T. A multicentre study on the pathogenic agents in 665 adult patients with community-acquired pneumonia in cities of China. Chin J Tuberc Respir Dis. 2006;29:3–8.

7. Pinzone MR, Cacopardo B, Abbo L, Nunnari G. Duration of antimicrobial therapy in community acquired pneumonia: less is more. Sci World J. 2014;2014:759138. doi:10.1155/2014/759138

8. El Moussaoui R, De Borgie CA, Van Den Broek P. Effectiveness of discontinuing antibiotic treatment after three days versus eight days in mild to moderate-severe community acquired pneumonia: randomised, double blind study. BMJ. 2006;332:1355. doi:10.1136/bmj.332.7554.1355

9. Eccles S, Pincus C, Higgins B, Woodhead M. Diagnosis and management of community and hospital acquired pneumonia in adults: summary of NICE guidance. BMJ. 2014;349:g6722. doi:10.1136/bmj.g6722

10. Chen J, Luo Y, Zhang S. Community-acquired necrotizing pneumonia caused by methicillin-resistant Staphylococcus aureus producing Panton-Valentine leukocidin in a Chinese teenager: case report and literature review. Int J Infect Dis. 2014;26:17–21. doi:10.1016/j.ijid.2014.02.025

11. Qu T, Feng Y, Jiang Y. Whole genome analysis of a community-associated methicillin-resistant Staphylococcus aureus ST59 isolate from a case of human sepsis and severe pneumonia in China. PLoS One. 2014;9:e89235. doi:10.1371/journal.pone.0089235

12. Li H, Zhang T, Huang J. Analysis of risk factors related to mortality of patients with community-acquired pneumonia due to methicillin-resistant Staphylococcus aureus. Chin Crit Care Med. 2010;22(8):459–464. doi:10.3760/cma.j.issn.1003-0603.2010.08.004

13. Zhang J, Zhang Y, Shi Y, et al. Population pharmacokinetic and pharmacodynamic modeling of norvancomycin. Eur J Clin Microbiol. 2008;27:275–284. doi:10.1007/s10096-007-0435-9

14. Wu Y, Kang J, Wang Q. Drug concentrations in the serum and cerebrospinal fluid of patients treated with norvancomycin after craniotomy. Eur J Clin Microbiol. 2017;36:305–311. doi:10.1007/s10096-016-2803-9

15. Marsot A, Boulamery A, Bruguerolle B, et al. Vancomycin: a review of population pharmacokinetic analyses. Clin Pharmacokinet. 2012;51:1–13. doi:10.2165/11596390-000000000-00000

16. Yang QW, Li X, Jia PY, et al. Determination of norvancomycin epidemiological cut-off values (ECOFFs) for Staphylococcus aureus, Staphylococcus epidermidis, Staphylococcus haemolyticus and Staphylococcus hominis. J Antimicrob Chemother. 2021;76:152–159. doi:10.1093/jac/dkaa414

17. Ho PL, Lo PY, Chow KH, et al. Vancomycin MIC creep in MRSA isolates from 1997 to 2008 in a healthcare region in Hong Kong. J Infect. 2010;60:140–145. doi:10.1016/j.jinf.2009.11.011

18. Steinkraus G, White R, Fredrich L. Vancomycin MIC creep in non-vancomycin-intermediate Staphylococcus aureus (VISA), vancomycin-susceptible clinical methicillin-resistant S. aureus (MRSA) blood isolates from 2001–2005. J Antimicrob Chemother. 2007;60:788–794. doi:10.1093/jac/dkm258

19. Li XB, Qi ZZ, Gan Q, et al. Clinical research on efficacy of teicoplanin and norvancomycin in treatment of hospital-acquired MRSA pneumonia in senile patients. Chin J Nosocomiol. 2011;21:3023–3025. doi:10.1631/jzus.B1000135

20. Dong WC, Guo JL, He WJ, et al. Determination of vancomycin and norvancomycin in human cerebrospinal fluid by HPLC. West China J Pharm Sci. 2016;31:312–314. doi:10.13375/j.cnki.wcjps.2016.03.031

21. Lindbom L, Ribbing J, Jonsson EN. Perl-speaks-NONMEM (PsN)-a Perl module for NONMEM related programming. Comput Methods Programs Biomed. 2004;75:85–94. doi:10.1016/j.cmpb.2003.11.003

22. Hooker AC, Staatz CE, Karlsson MO. Conditional weighted residuals (CWRES): a model diagnostic for the FOCE method. Pharma Res. 2007;24:2187–2197. doi:10.1007/s11095-007-9361-x

23. Brendel K, Comets E, Laffont C, et al. Metrics for external model evaluation with an application to the population pharmacokinetics of gliclazide. Pharma Res. 2006;23:2036–2049. doi:10.1007/s11095-006-9067-5

24. Giuliano C, Haase KK, Hall R. Use of vancomycin pharmacokinetic–pharmacodynamic properties in the treatment of MRSA infections. Expert Rev Anti-Infect. 2014;8:95–106. doi:10.1038/nrmicro862

25. Craig WA. Basic pharmacodynamics of antibacterials with clinical applications to the use of beta-lactams, glycopeptides, and linezolid. Infect Dis Clin North Am. 2003;17:479–501. doi:10.1016/s0891-5520(03)00065-5

26. Drusano GL. Antimicrobial pharmacodynamics: critical interactions of ‘bug and drug’. Nat Rev Microbiol. 2004;2:289–300. doi:10.1038/nrmicro862

27. Rybak MJ. Pharmacodynamics: relation to antimicrobial resistance. Am J Infect Control. 2006;34:S37–S45. doi:10.1016/j.ajic.2006.05.227

28. Rybak MJ, Le J, Lodise T, et al. Executive summary: therapeutic monitoring of vancomycin for serious methicillin-resistant Staphylococcus aureus infections: a revised consensus guideline and review of the American Society of Health-System Pharmacists, the Infectious Diseases Society of America, the Pediatric Infectious Diseases Society, and the Society of Infectious Diseases Pharmacists. J Pediatric Infect Dis Soc. 2020;9:281–284. doi:10.1093/jpids/piaa057

29. Zhou Y, Gao F, Chen C. Development of a population pharmacokinetic model of vancomycin and its application in Chinese geriatric patients with pulmonary infections. Eur J Drug Metab Pharmacoki. 2018;44:361–370. doi:10.1007/s13318-018-0534-2

30. He N, Su S, Ye Z, et al. Evidence-based guideline for therapeutic drug monitoring of vancomycin: 2020 update by the division of therapeutic drug monitoring, Chinese pharmacological society. Clin Infect Dis. 2020;71:S363–S371. doi:10.1093/cid/ciaa1536