Multi-center prospective population pharmacokinetic study and the performance of web-based individual dose optimization application of intravenous vancomycin for adults in Hong Kong: A study protocol.

A recent consensus guideline recommends migrating the therapeutic drug monitoring practice for intravenous vancomycin for the treatment of methicillin-resistant Staphylococcus aureus infection from the traditional trough-based approach to the Bayesian approach based on area under curve to improve clinical outcomes. To support the implementation of the new strategy for hospitals under Hospital Authority, Hong Kong, this study is being proposed to (1) estimate and validate a population pharmacokinetic model of intravenous vancomycin for local adults, (2) develop a web-based individual dose optimization application for clinical use, and (3) evaluate the performance of the application by comparing the treatment outcomes and clinical satisfaction against the traditional approach. 300 adult subjects prescribed with intravenous vancomycin and not on renal replacement therapy will be recruited for population pharmacokinetic model development and validation. Sex, age, body weight, serum creatinine level, intravenous vancomycin dosing records, serum vancomycin concentrations etc. will be collected from several electronic health record systems maintained by Hospital Authority. Parameter estimation will be performed using non-linear mixed-effect modeling techniques. The web-based individual dose optimization application is based on a previously reported application and is built using R and the package shiny. Data from another 50 subjects will be collected during the last three months of the study period and treated as informed by the developed application and compared against historical control for clinical outcomes. Since the study will incur extra blood-taking procedures from patients, informed consent is required. Other than that, recruited subjects should receive medical treatments as usual. Identifiable patient data will be available only to site investigators and clinicians in each hospital. The study protocol and informed consent forms have been approved by the Joint Chinese University of Hong Kong-New Territories East Cluster Clinical Research Ethics Committee (reference number: NTEC-2021-0215) and registered at the Chinese Clinical Trial Registry (registration number: ChiCTR2100048714).

Introduction

Indication of intravenous vancomycin

Intravenous (IV) vancomycin has long been the first-line antimicrobial agent for treating severe methicillin-resistant Staphylococcus aureus (MRSA) infection [1]. It is often prescribed empirically for patients with suspected MRSA infection because of its efficacy in organism eradication. However, a significant drawback of IV vancomycin is the risk of acute kidney injury (AKI), which is associated with large areas under the curve of the concentration-time profile of vancomycin [2]. Given the narrow therapeutic index of IV vancomycin, therapeutic drug monitoring (TDM) of vancomycin is mandatory to balance efficacy against toxicity [3].

Traditional TDM of IV vancomycin

In 2009, the American Society of Health-System Pharmacists, the Infectious Diseases Society of America, and the Society of Infectious Diseases Pharmacists jointly published a consensus report (the 2009 guideline), recommending a steady-state area under the curve over 24 hours (AUC24) to minimum inhibitory concentration (MIC) ratio (AUC24/MIC) of at least 400 mg∙hr/L as the pharmacokinetic (PK) target for successful vancomycin therapy. Nevertheless, given the difficulty in obtaining multiple serum vancomycin concentrations (Cs) for the estimation of AUC24, they recommended the use of steady-state trough Cs as a surrogate marker for the AUC24/MIC target [4].

Updated recommendations for TDM of IV vancomycin

Despite the previous recommendation, over the past decade, there has been minimal to no data on the efficacy and safety profile supporting the use of steady-state trough Cs as a helpful treatment endpoint [5]. In fact, steady-state trough Cs is later found to be poorly correlated with AUC24 [6]. Contrastingly, there has been increasing evidence on the use of AUC24 as the therapeutic target for IV vancomycin [7, 8]. Moreover, on top of the previous target of AUC24/MIC of at least 400 mg∙hr/L to eradicate MRSA, it has also been shown that AUC24 of at least 600 mg∙hr/L is associated with an increased risk of AKI [9]. Of equal importance, Neely et al. conducted a prospective trial demonstrating the superiority of combining the AUC24 target and a Bayesian approach to AUC24 estimation over the traditional trough-based approach [8]. In fact, Bayesian tools are becoming more readily available over the past decade, thanks to rapid advancement in computing efficiency [10-13].

Given the above, the above named societies and the Pediatric Infectious Diseases Society published a revised consensus guideline in 2020 (the 2020 guideline). It recommended (1) against the previous therapeutic target in terms of steady-state trough Cs, and (2) the promotion of the achievement of the therapeutic target of AUC24 between 400–600 (assuming MIC of 1 mg/L) through dose individualization guided by individual AUC24, which should be obtained through Bayesian estimation based on trough Cs (or both trough Cs and peak Cs for better accuracy) [5].

Advantages of Bayesian estimation

Since trough-based TDM is based on observations at the steady state, it is often necessary to wait until the third or fourth infusion, after which the steady state is almost reached, for the sampling of Cs. However, with Bayesian estimation, it becomes possible to extrapolate the pre-steady-state PK profile to predict steady-state behavior, and Cs sampled within the first 24 to 48 hours can be sufficient to inform the optimal dosing regimen [5]. It implies that waiting until the steady state for sampling is no longer necessary. As a result, the Bayesian approach can help reduce the sampling of vancomycin levels and shorten the length of therapy [8].

Performing Bayesian estimation requires a prior distribution, which, as recommended by the 2020 guideline and in the current study, will be represented by a population pharmacokinetic (popPK) model of IV vancomycin among adults in Hong Kong [5]. The popPK model contains the population estimates of the effects of multiple covariates, including but not limited to age, sex, body weight (WT), serum creatinine level (SCr), and distributions. Compared to the traditional trough-based approach, which is based solely on the observed Cs, the currently recommended Bayesian approach also takes the population characteristics into account, thus improving the accuracy of the estimation of AUC24 [14].

Besides, since Bayesian estimation requires numerical approximation and cannot be performed manually, a web-based application that automates the estimation and dose optimization process would be ideal for clinical use. Such an interface will be designed to automatically obtain individual medical data from electronic health record systems, thus reducing the workload of the clinical frontlines and the incidence of medication errors.

Current practice in Hospital Authority and recommendations

The TDM practice of IV vancomycin in medical institutions under Hospital Authority (HA) in Hong Kong has generally been in line with the 2009 guideline that steady-state trough Cs is being monitored for patients put on IV vancomycin. In view of the updated evidence and guideline, we recommend local institutions to advance the current practice to meet the suggestions by the 2020 guideline to improve the treatment outcomes. In support of the implementation of the dose optimization application, the current study is being proposed to (1) estimate and validate a popPK model of IV vancomycin for adults in Hong Kong, (2) develop a web-based dose optimization application for clinical use in HA, and finally (3) evaluate the performance of the web-based dose optimization application by comparing the treatment outcomes and clinical satisfaction against the traditional trough-based TDM approach.

Materials and methods

Study design and population

This is a multi-center prospective study involving hospitals across all seven clusters (which corresponding to different regions of Hong Kong) of HA. All in-patient subjects (1) at least 18 years of age, (2) admitted to one of the following nine HA hospitals: Pamela Youde Nethersole Eastern Hospital, Ruttonjee and Tang Shiu Kin Hospitals, Queen Mary Hospital, Queen Elizabeth Hospital, Kwong Wah Hospital, United Christian Hospital, Princess Margaret Hospital, Prince of Wales Hospital, and Tuen Mun Hospital, and (3) for whom intermittent IV vancomycin is prescribed are eligible for recruitment into the study for data collection. Notwithstanding the above, all subjects prescribed any form of renal replacement therapy that has been started before or is scheduled to start during IV vancomycin treatment shall be excluded.

This study expects to recruit 300 subjects, among which 50 subjects will be set aside for external validation of the popPK model, while the rest belong to the model estimating set. At a later stage of the study, data from another 50 subjects will be collected for the evaluation of the performance of the developed interface. Multi-center involvement is expected to facilitate adequate enrolment.

Expected study timeline

The entire data collection period will last for one year. It is expected that the popPK data collection period (PDCP) will last for about six months. popPK model development will start once sufficient data have been collected for preliminary analyses. Background preparation work for the dose optimization interface will be in progress during the entire popPK study period. After the interface becomes ready-to-use, training to use the interface and evaluation of the performance of the developed interface are expected to start at about 9 months into the study. The last 3 months of the study period is the evaluation data collection period (EDCP). The SPIRIT schedule of enrolment is available in Figs 1 and 2 outlines the study timeline.

Fig 1

The SPIRIT schedule of enrolment.

aInformed consent may be obtained on the next day if it cannot be done on the day the subject is screened. bCollection of dynamic data items will continue until vancomycin treatment ends.

Fig 2

The study timeline.

popPK: population pharmacokinetic. HA: Hospital Authority.

Data items to collect

Data item summary

For each subject, constant data items refer to variables that take only a single value. Meanwhile, all dynamic data items should be collected during each of his/her individual data collection period(s) (IDCP), defined as the period starting from the start of the 1 infusion and until the sampling for the last Cs measurement. Data items required for the study are summarized in Table 1.

Table 1

Data items to be collected.

Item type		Data item
Constant data items (one value per subject)		Date of birth
		Ethnicity
		Hospital and ward/specialty
		Sex
		Baseline WT
		Body height
		Obesity status
		Baseline SCr
		Source of infection
Dynamic data items	Measurements, assessments, or events	Significant changes in WT
		SCr during treatment
		Pathophysiological conditions: AKI, sepsis, severe trauma, severe burns etc.
		Death (if applicable)
		Concomitant drugs (diuretics, aminoglycosides, non-steroidal anti-inflammatory drugs, or other drugs that significantly influence renal functions)
		Renal replacement therapy (intermittent or continuous)
	Vancomycin dosing records	Date and exact timing of start and end of infusion
		Infusion rate
		Total amount infused
	Cs	Date and exact timing of sampling
		Measured concentrations
	Microbial cultures	Date and time of sampling
		Isolated micro-organisms
		MIC of vancomycin against MRSA
	outcomes	Time to achieving therapeutic target
		Time to development of AKI at different stages
		Length of IV vancomycin therapy
		Number of Cs blood samples collected
		Time to recovery in terms of time to afebrile
		Time to normal white blood cell count

AKI: acute kidney injury. Cs: serum vancomycin concentration. IV: intravenous. MIC: minimum inhibitory concentration. MRSA: methicillin-resistant Staphylococcus aureus. SCr: Serum creatinine concentration. WT: body weight. The reference method to measure MIC is broth microdilution but Etest is used in the involved clinics. Previous results have shown comparable measurements between broth microdilution and Etest [15].

For dynamic data items, all records available during the IDCP of a subject will be taken. The most recent Cs measurement will be considered the last one when (1) the subject has recovered and IV vancomycin treatment has been stopped, (2) the subject has been put off IV vancomycin and switched to receive alternative antibiotic(s), (3) the clinician has decided that IV vancomycin is to be halted indefinitely, or (4) the subject has deceased. (Note that either a temporary halt of IV vancomycin due to, for e.g., impaired renal function, high Cs etc., or a change in IV vancomycin dosing regimen will not be considered an interruption of the IDCP.) If, within the PDCP/EDCP, IV vancomycin is started on a subject again after his last IDCP, a new IDCP should be initiated for the same subject. If the PDCP/EDCP ends before an IDCP of a subject, data collection for the subject should be extended until the IDCP ends as defined above. Besides, the dates and times of all dynamic data item records should be noted.

Details for IV vancomycin dosing records

For each infusion, (1) the date and exact timing (accurate to the minute as much as possible) at the start of infusion, (2) the infusion period assuming constant-rate infusion (with timed records of any irregularity in infusion rate and interruption), and (3) the total amount infused will be recorded.

In the case that it comes to the investigators’ attention that the recruited subject either (1) has received IV vancomycin within 72 hours before the start of the 1st infusion or (2) has severe baseline renal impairment and has received IV vancomycin within a week before the start of the 1st infusion, then all available dosing records within these stated periods prior to the 1st infusion should be recorded.

Details for vancomycin concentration records

In this paragraph, dosing interval 1 refers to one of the pre-steady-state dosing intervals and dosing interval 2 refers to one of the steady-state or near-steady-state dosing intervals. Subject to the actual implementation in individual hospital level, the recommended sampling schedule requests six Cs measurements. These include (1) two peak Cs sampled at least one hour and within two hours after the last infusion ends during dosing interval 1 and dosing interval 2, (2) two random levels sampled at least an hour after the last peak Cs and at least an hour before the next trough Cs during dosing interval 1 and dosing interval 2, and (3) two trough Cs sampled within one hour and strictly before the start of the next infusion during dosing interval 1 and dosing interval 2. In the case that not all the six recommended Cs measurements can be performed (e.g. due to failure to obtain informed consent or difficulties in logistics), an alternative sparser schedule should be considered. Details regarding the Cs sampling schedule are available in Fig 3.

Fig 3

Details of the Cs sampling schedule with respect to the dosing schedule.

Cs: serum vancomycin concentration. n: number of Cs measurements in the schedule. Note that the time separation requirements for peak, random, and trough Cs are only shown in the ideal sampling schedule. The same set of requirements applies for all proposed schedules with lower priorities.

In case it is infeasible to sample during a planned dosing interval, sampling should be delayed to the next feasible dosing interval. If vancomycin is put off before the last dose intended within the IDCP, then if vancomycin is restarted later, the schedule should be restarted from dosing interval 1. Note that any change in vancomycin dose and/or any supplementary vancomycin dose administered at once should not interrupt the planned sampling schedule. If there is a change in the administration frequency, the sampling schedule should be updated based on the new frequency. The timings of all samplings of Cs after satisfying the above recommended measurement schedule are not bound by the study protocol but subject entirely to clinical needs as judged by the clinicians.

Data collection method

Clinicians will order extra laboratory assays according to the data items required by the study protocol. Site investigators will collect and organize required data of recruited patient using the Clinical Management System, Medication Administration Record, In-Patient Medication Order Entry System, bedside patient chart, and, when necessary, verbal clarifications. Anonymized patient data will be shared in a confidential manner with the data analysts. The site investigators shall keep a separate, confidential conversion list between anonymized identifiers (available to the data analysts) and in-site patient identifiers, such that clarification will be possible when data validity is in doubt.

Data exclusion and management

All data in any IDCP (1) with any missing constant data item (except body height unless subject is obese, and ethnicity), (2) without baseline SCr, or (3) without at least one peak Cs (or random Cs) plus at least one trough Cs sampled will be removed from subsequent analyses. Besides, in each IDCP, all Cs records that are over 168 hours (7 days) after the last Cs record or the end of the last infusion (whichever later) will be removed from subsequent analyses. All subjects with multiple IDCPs will be treated as the same subject and the vancomycin estimated to be remaining in the subject’s body from the previous IDCP will be carried over for the estimation of Cs in the subsequent IDCP(s).

Each of WT, SCr, and body height will be imputed across each IDCP with linear interpolation and extrapolation (or assumed constant if only one value is present). Preliminary analyses of individual Cs profiles will be performed to identify potential outliers. When necessary, investigations into the possibilities of errors in data collection and unusual patient conditions causing the extreme values will be carried out as soon as possible. Besides, regarding the severity of AKI, KDIGO staging will be determined for each subject from the SCr collected (No AKI, stage 1, 2 or 3).

PopPK model estimation and validation

Structural and parameter models

To characterize the Cs-time profiles of known doses and administration times of IV vancomycin, the one- or two-compartment infusion model with first-order elimination will be tested, where the one demonstrating better goodness-of-fit (demonstrated by smaller objective function value) will be chosen. Between-subject and between-occasion variabilities in PK parameters are assumed to follow log-normal distributions. Residual unexplained variability will be described by a combined proportional-additive error model [16-18].

Covariate model

The effects of WT on PK parameters will be presumably estimated by the power model, where allometric scaling with pre-determined exponents will be tested against estimated exponents [19]. Creatinine clearance will be approximated by the Cockcroft-Gault equation (which is based on sex, age, WT, and SCr) and associated with vancomycin clearance by testing different curve functions [20]. Residual covariate effects are then tested against other potential covariates, including but not limited to ethnicity, concurrent pathophysiological conditions, concomitant drugs, renal replacement therapy, and isolated microbes. Hypothesis testing at α = 0.01 will be conducted to compare the goodness-of-fits between two nested models by assuming that the change in objective function value from the richer model to the sub-model follows the χ-distribution with df equal to the number of constrained parameters [21].

Model evaluation and validation

Predictive plots, residual plots, normalized empirical Bayes estimates plots, prediction-corrected visual predictive check, and normalized prediction distribution errors will be inspected to evaluate the final model and parameter estimates [22]. After the above evaluation, bootstrapping using 1,000 resamples will be done for internal validation [23]. External validation will be done by evaluating the internally validated model against a separate, smaller dataset.

Computer software

NONMEM® 7 will be used to obtain parameter estimates [24]. Below-limit-of-quantification data will be assessed using the M3 method [25]. Perl-speaks-NONMEM will be used to coordinate NONMEM® runs and model evaluation [26]. R and its packages will be used for the generation of model evaluation graphics [27, 28].

Expected outcomes

In predictive plots, observations should scatter around the identity line. Weighted residuals and normalized empirical Bayes estimates should scatter around zero (with 95% of the points lying within -1.96 and 1.96) with no observable trend alone and against time and other variables. Visual predictive checks should demonstrate agreements between corrected observed and predicted Cs in terms of the percentiles. Normalized prediction distribution error plots should resemble the standard normal distribution. Bootstrapping should show reasonable distributions of bootstrap estimates with their medians close to and their 95% confidence intervals containing the model estimates. Apart from bootstrapping, the above applies also to external validation. Overall, all model diagnostics should indicate good predictive performance and stability of the developed popPK model. Successful results in this part will provide the foundation to conduct Bayesian estimation that can improve the TDM practice of IV vancomycin, as agreed by the community [5].

Development of the web-based dose optimization application

The infrastructure and framework of a previously published web-based individual dose adjustment application for high-dose methotrexate in the pediatric population will be replicated in this study. This application was built using R and its packages, including shiny, which are (1) open sources, (2) validated against the proprietary software package, NONMEM®, for the accuracy of individual parameter estimation, and (3) has been shown to be more efficient than relying on NONMEM® in performing individual parameter estimation [29]. The interface will be amended to adapt to the requirements of this study and clinical application.

Statistical analyses for outcomes

The group of subjects treated as informed by the developed application (recruited during the EDCP) is compared against the group treated with the traditional approach (recruited during the PDCP) and historical control data, which will be verified by baseline characteristics comparison and, if necessary, propensity score matching. The probabilities of developing AKI and proportions of AKI stages between approaches will be compared with the chi-square test. The times to achieving therapeutic target, development of AKI at different stages, becoming afebrile, and achieving normal white blood cell count will be compared using Kaplan-Meier analysis and Cox proportional hazard regression. Finally, the length of IV vancomycin therapy and the number of Cs blood samples collected will be compared using the Student’s t-test or the Wilcoxon signed rank test, depending on the distribution of the data.

Ethical considerations

Need of patient data collection

The current trough-based TDM approach of vancomycin is no more recommended due to its poor prediction of vancomycin exposure, and therefore should be phased out. Instead, the currently recommended TDM strategy is to rely on Bayesian estimation of vancomycin AUC24. The application of the strategy requires a prior distribution for the local population that is developed upon rich popPK data, thus patient data collection is indispensable. The result of this study will enable HA institutions to comply with the most updated recommendations and is expected to improve treatment outcome once applied.

Changes to clinical procedures

From the perspective of the patients recruited into this study, the study sampling schedule requires more measurements of Cs than usual and thus will likely incur more frequent blood sampling than usual. Other than that, there is no indication of other procedures and medical treatments in this study. As in routine clinical practice, study subjects will receive medical treatments that are deemed the most appropriate by clinicians. It is well acknowledged that rich data collected in study subjects may alter clinicians’ decisions on vancomycin treatments. However, when compared to the current trough-based TDM approach, the rich sampling scheme will likely enrich the information required for making accurate prediction of vancomycin exposure. This is supported by the 2020 guideline’s recommendation to estimate AUC24 using both steady-state peak Cs and steady-state trough Cs when a Bayesian tool is not yet available [5]. Therefore, it is very unlikely that the extra sampling required in this study will adversely affect the optimality of vancomycin therapy.

From the perspective of the clinicians and site investigators, workload may increase due to more frequent sampling, ordering of assays for extra samples, and collection and validation of detailed patient data. However, TDM has been a routine procedure of vancomycin treatment, so new types of intervention are not introduced in this study. In any case, clinicians and site investigators should prioritize patient care over the consolidation of study data.

Patient privacy

Access to identifiable patient data collected in this study will be available only to site investigators and clinicians in each institution. All data will be anonymized before being sent to the data analysts. Data collected by the data analysts will be handled with encryption and password protection.

Informed consent

The gathering of routinely collected data is not expected to adversely affect patient treatment nor expose patient information. However, since rich sampling will incur extra blood-taking procedures done on patients, a written informed consent must be obtained from the patient or, if the patient is incapable of giving informed consent, his legal representative, who must be a member of his next of kin. Extra blood taking procedures will not be performed for subjects who refuse to participate in the study, and the data available through routine vancomycin treatment based on the updated guideline will continue to be included in this study.

Compliance to ethical standard, registration, and approval

This study will be conducted in compliance with the Declaration of Helsinki and is being submitted for review by multiple Cluster Research Ethics Committees of HA. The study protocol (version 1.1, Apr 21st, 2021) and informed consent forms have been approved by the Joint Chinese University of Hong Kong–New Territories East Cluster Clinical Research Ethics Committee (reference number: 2021.175; approval date: May 21st, 2021). (See S1 File.) Approvals by the Ethics Committees of other clusters are pending. Any protocol modifications will be reviewed by the Ethics Committee. The study has also been registered at the Chinese Clinical Trial Registry (registration number: ChiCTR2100048714) on July 13th, 2021. To ensure homogenous study procedures across study centers, the principal investigator and his team of other investigators at The Chinese University of Hong Kong will explain the study procedures thoroughly to the site coordinators and answer all questions they have before study commencement, as well as closely monitor the recruitment of and study procedures done on the first subject at each center.

Documents approved by the ethics committee.

(DOCX)

Click here for additional data file.

SPIRIT 2013 checklist: Recommended items to address in a clinical trial protocol and related documents *.

(DOCX)

Click here for additional data file.

(PDF)

Click here for additional data file.

(PDF)

Click here for additional data file.

(PDF)

Click here for additional data file.

(PDF)

Click here for additional data file.

5 Jan 2022

Multi-center prospective population pharmacokinetic study and the performance of web-based individual dose optimization application of intravenous vancomycin for adults in Hong Kong: A study protocol

PLOS ONE

Dear Dr. Lam,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

This is an interesting approach that has merit, but the reviewers include several suggestions for improvement.  In particular, please minimize the use of acronyms which often serve only to irritate the reader - we have more than enough 'electron space' to simply spell most of these out save those in the common vernacular (e.g. ICU, KDIGO, etc.).  I look forward to seeing the revision which I will send back to the same referees for decision.

Please submit your revised manuscript by Feb 11 2022 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Steven Eric Wolf, MD

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. In your Data Availability statement, you have not specified where the minimal data set underlying the results described in your manuscript can be found. PLOS defines a study's minimal data set as the underlying data used to reach the conclusions drawn in the manuscript and any additional data required to replicate the reported study findings in their entirety. All PLOS journals require that the minimal data set be made fully available. For more information about our data policy, please see http://journals.plos.org/plosone/s/data-availability.

Upon re-submitting your revised manuscript, please upload your study’s minimal underlying data set as either Supporting Information files or to a stable, public repository and include the relevant URLs, DOIs, or accession numbers within your revised cover letter. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories. Any potentially identifying patient information must be fully anonymized.

Important: If there are ethical or legal restrictions to sharing your data publicly, please explain these restrictions in detail. Please see our guidelines for more information on what we consider unacceptable restrictions to publicly sharing data: http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions. Note that it is not acceptable for the authors to be the sole named individuals responsible for ensuring data access.

We will update your Data Availability statement to reflect the information you provide in your cover letter.

3. Your abstract cannot contain citations. Please only include citations in the body text of the manuscript, and ensure that they remain in ascending numerical order on first mention.

4. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Does the manuscript provide a valid rationale for the proposed study, with clearly identified and justified research questions?

The research question outlined is expected to address a valid academic problem or topic and contribute to the base of knowledge in the field.

Reviewer #1: Yes

Reviewer #2: Partly

Reviewer #3: Yes

**********

2. Is the protocol technically sound and planned in a manner that will lead to a meaningful outcome and allow testing the stated hypotheses?

The manuscript should describe the methods in sufficient detail to prevent undisclosed flexibility in the experimental procedure or analysis pipeline, including sufficient outcome-neutral conditions (e.g. necessary controls, absence of floor or ceiling effects) to test the proposed hypotheses and a statistical power analysis where applicable. As there may be aspects of the methodology and analysis which can only be refined once the work is undertaken, authors should outline potential assumptions and explicitly describe what aspects of the proposed analyses, if any, are exploratory.

Reviewer #1: Partly

Reviewer #2: Partly

Reviewer #3: Yes

**********

3. Is the methodology feasible and described in sufficient detail to allow the work to be replicable?

Descriptions of methods and materials in the protocol should be reported in sufficient detail for another researcher to reproduce all experiments and analyses. The protocol should describe the appropriate controls, sample size calculations, and replication needed to ensure that the data are robust and reproducible.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

4. Have the authors described where all data underlying the findings will be made available when the study is complete?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception, at the time of publication. The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above and, if applicable, provide comments about issues authors must address before this protocol can be accepted for publication. You may also include additional comments for the author, including concerns about research or publication ethics.

You may also provide optional suggestions and comments to authors that they might find helpful in planning their study.

(Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: This is an interesting PK study that primarily describes a carefully planned PK analysis. The statistician is clearly an expert in NONMEM7. Although I am a statistical reviewer, I am more concerned that the paper is primarily written about the statistical procedures rather than the design of the study itself. Some of the omissions that I see are:

--multi-center studies should describe how each center's personnel will be trained on the protocol so that patient procedures are homogeneous

--A lot of data will be collected, and this is well-described, but its relation to the study objectives is less clearly described.

--There is no sample size justification. Why 300? Why 50?

--Bayesian methods seem appropriate. Yet you are using t-tests and standard survival estimators for the comparison.

--More on historical controls. How can we be certain that the historical controls are comparable?

--Far too many acronyms throughout the paper. I know PK people like acronyms, but it is overwhelming for the reader to keep track.

--Details such as in lines 208-212 would be better shown in a figure or flow chart.

--Remove the software used in the abstract. NONMEM is not a statistical procedure, it is a software designed to do certain modeling. It is better to say the statistical modeling technique rather than refer to the software.

--A sweeping conclusion that related all these NONMEM C_s analyses to clinical practice.

Reviewer #2: The introduction needs to clarify the other recommendations and guidelines being used to guide the new study.

Recommend addressing that reference method or “gold standard” for calculating MIC is called broth microdilution (BMD) and how that incorporates to the study.

Target range should be AUC:MIC ratio of 400-600.

Reviewer #3: Multi-center prospective pharmacokinetic study and the performance of web-based individual dose optimization application of intravenous vancomycin for adults in Hong Kong: A study protocol

I commend the authors on this ambitious study to develop an algorithm for area under the curve based individual vancomycin dosing.

Line 111: Please define “HA”

Line 137: explain methods for a priori sample sizes

Table 1

Constant Data:

Consider staging acute kidney injury at baseline and throughout the treatment protocol (ie RIFLE, AKIN, or KDIGO). This can easily be calculated with proposed laboratory values to be collected and can further stratify patients adding greater depth and applicability to the web-based model.

Also, include source of infection

Dynamic Data Items:

Include renal based outcome (ie time to development of AKI, Time to ESRD, time to change in AKI stage, etc.)

Line 127: For patients with a new IDCP, please explain how this new IDCP will be treated during the data analysis. Will this be a unique dataset? Define the timeframe between cessation of vancomycin and a new IDCP.

Line 189: Define dosing interval.

Line 227: 168 hours is between 28 and 42 half-lives of vancomycin. This seems excessive. Consider reducing this time frame or explain why this time frame is chosen.

Line 290: Define TTT, TTA, TTN

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

Reviewer #3: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

15 Mar 2022

Dear Editor,

All authors and I would like to thank the editorial staff and reviewers for handling our manuscript. Please see the following point-to-point responses to the editors’ and reviewers’ comments. Editors’ and reviewers’ comments are quoted in smaller grey font with tight line spacing, while our responses are in blue.

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

Thank you for the information. We have reformatted our manuscript accordingly.

2. In your Data Availability statement, you have not specified where the minimal data set underlying the results described in your manuscript can be found. PLOS defines a study's minimal data set as the underlying data used to reach the conclusions drawn in the manuscript and any additional data required to replicate the reported study findings in their entirety. All PLOS journals require that the minimal data set be made fully available. For more information about our data policy, please see http://journals.plos.org/plosone/s/data-availability.

Upon re-submitting your revised manuscript, please upload your study’s minimal underlying data set as either Supporting Information files or to a stable, public repository and include the relevant URLs, DOIs, or accession numbers within your revised cover letter. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories. Any potentially identifying patient information must be fully anonymized.

Important: If there are ethical or legal restrictions to sharing your data publicly, please explain these restrictions in detail. Please see our guidelines for more information on what we consider unacceptable restrictions to publicly sharing data: http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions. Note that it is not acceptable for the authors to be the sole named individuals responsible for ensuring data access.

We will update your Data Availability statement to reflect the information you provide in your cover letter.

The protocol does not report any result. We have specified so in the Data Availability section.

3. Your abstract cannot contain citations. Please only include citations in the body text of the manuscript, and ensure that they remain in ascending numerical order on first mention.

Our abstract does not contain any citation.

4. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

We have added the caption for our Supporting Information files at the end of our manuscript and updated any in-text citations accordingly.

Reviewer #1: This is an interesting PK study that primarily describes a carefully planned PK analysis. The statistician is clearly an expert in NONMEM7. Although I am a statistical reviewer, I am more concerned that the paper is primarily written about the statistical procedures rather than the design of the study itself.

Thank you for your time reviewing our manuscript and your valuable comments. We understand that our current manuscript is lacking certain details regarding study design. Please see our responses below.

Some of the omissions that I see are:

--multi-center studies should describe how each center's personnel will be trained on the protocol so that patient procedures are homogeneous

We will explain the protocol thoroughly to the site coordinators, answer all questions they have before study commencement, and closely monitor the recruitment of and study procedures done on the first subject at each center. The section “Compliance to ethical standard, registration, and approval” has been supplemented.

--A lot of data will be collected, and this is well-described, but its relation to the study objectives is less clearly described.

Thank you for your reminder. We have supplemented the “Structural and parameter models” and “Covariate model” sections to explain how the data help characterize the concentration-time profiles of IV vancomycin, considering potential covariate effects.

--There is no sample size justification. Why 300? Why 50?

Unlike hypothesis testing, where the minimum sample size required can be approximated based on the expected effect size and the power needed, in a popPK modeling study where the major covariates are also known a priori, there is no established way to calculate the minimum sample size. The sample sizes of previously published popPK models vary significantly, for e.g., 16 and 36 in Hui K. H. et al (2019) DOI: 10.1002/jcph.1349, 326 in Francis J. et al (2019) DOI: 10.1128/AAC.01964-18, and 1,151 subjects in Stringer F. et al (2013) DOI: 10.1177/0091270012447121. We notice that most popPK reports had sample sizes that range from 50 to 300. Therefore, we aim at the higher end of the range by defining a sample size of 300 and hope for more accurate estimates while avoiding further increasing the study burden.

Similarly, there is no established way to calculate the minimum sample size for external validation. Our decision to include 50 subjects for external validation was based on the observation that most popPK studies adopt a sample size for external validation that is about 15-30% of the sample size for model estimation. (300 * 17% = 50)

--Bayesian methods seem appropriate. Yet you are using t-tests and standard survival estimators for the comparison.

Bayesian methods are being applied to estimate the individual PK and thus optimal doses. However, successfully verifying the prior distribution to be used in the Bayesian estimation process does not imply that using the developed interface to guide dosing will improve the overall clinical outcome. Therefore, t-tests and Cox regression are proposed to compare the clinical outcomes of the two treatment approaches: with vs without the developed Bayesian application. These are statistical tests commonly used to test for treatment effects in conventional clinical studies.

--More on historical controls. How can we be certain that the historical controls are comparable?

We will verify the comparability through baseline characteristics comparison and, if necessary, propensity score matching. We have supplemented this under the sub-section “Statistical analyses for outcomes” in the Material and methods section.

--Far too many acronyms throughout the paper. I know PK people like acronyms, but it is overwhelming for the reader to keep track.

We apologize for using too many acronyms. We have removed most acronyms for technical terms (such as Cs,ss,trough) and those used sparingly in the updated manuscript.

--Details such as in lines 208-212 would be better shown in a figure or flow chart.

Thank you for your recommendation. We think it is a great idea to trim the passage to improve readability. We have added a figure to explain the ideal and alternative sampling schedules with respect to the dosing schedule.

--Remove the software used in the abstract. NONMEM is not a statistical procedure, it is a software designed to do certain modeling. It is better to say the statistical modeling technique rather than refer to the software.

We have removed NONMEM from the abstract and replaced it with non-linear mixed-effect modeling techniques.

--A sweeping conclusion that related all these NONMEM C_s analyses to clinical practice.

We have added a conclusive sentence at the end of the sub-section “PopPK model estimation and validation” in the Materials and methods section. Thank you for the recommendation.

Reviewer #2: The introduction needs to clarify the other recommendations and guidelines being used to guide the new study.

Thank you for your time reviewing our manuscript. We have further explained (1) the choice of a popPK model as the prior distribution for Bayesian estimation as a recommendation of the guideline.

Recommend addressing that reference method or “gold standard” for calculating MIC is called broth microdilution (BMD) and how that incorporates to the study.

Target range should be AUC:MIC ratio of 400-600.

We have included an explanation in the footnote of Table 1 about the methods used to measure MIC.

Reviewer #3: Multi-center prospective pharmacokinetic study and the performance of web-based individual dose optimization application of intravenous vancomycin for adults in Hong Kong: A study protocol

I commend the authors on this ambitious study to develop an algorithm for area under the curve based individual vancomycin dosing.

Thank you very much for your support of our work. Please see our responses to your valuable comments below.

Line 111: Please define “HA”

HA refers to Hospital Authority, the statutory body governing public hospitals in Hong Kong. The definition has been supplemented at its first appearance in the main text.

Line 137: explain methods for a priori sample sizes

Unlike hypothesis testing, where the minimum sample size required can be approximated based on the expected effect size and the power needed, in a popPK modeling study where the major covariates are also known a priori, there is no established way to calculate the minimum sample size. The sample sizes of previously published popPK models vary significantly, for e.g., 16 and 36 in Hui K. H. et al (2019) DOI: 10.1002/jcph.1349, 326 in Francis J. et al (2019) DOI: 10.1128/AAC.01964-18, and 1,151 subjects in Stringer F. et al (2013) DOI: 10.1177/0091270012447121. We notice that most popPK reports had sample sizes that range from 50 to 300. Therefore, we aim at the higher end of the range by defining a sample size of 300 and hope for more accurate estimates while avoiding further increasing the study burden.

Similarly, there is no established way to calculate the minimum sample size for external validation. Our decision to include 50 subjects for external validation was based on the observation that most popPK studies adopt a sample size for external validation that is about 15-30% of the sample size for model estimation. (300 * 17% = 50)

Table 1

Constant Data:

Consider staging acute kidney injury at baseline and throughout the treatment protocol (ie RIFLE, AKIN, or KDIGO). This can easily be calculated with proposed laboratory values to be collected and can further stratify patients adding greater depth and applicability to the web-based model.

Also, include source of infection

Dynamic Data Items:

Include renal based outcome (ie time to development of AKI, Time to ESRD, time to change in AKI stage, etc.)

We agree that adding AKI staging enables deeper analyses, such as the dose relationship of AKI. We have added a statement about conversion to AKI stages at the end of the sub-section “Data exclusion and management” and a description of corresponding statistical analyses in the sub-section “Statistical analyses for outcomes” under the Materials and methods section. Besides, we have added the item “source of infection” in Table 1. Similarly, we have included renal outcomes in terms of AKI staging in Table 1 and the sub-section “Statistical analyses for outcomes”.

Line 127: For patients with a new IDCP, please explain how this new IDCP will be treated during the data analysis. Will this be a unique dataset? Define the timeframe between cessation of vancomycin and a new IDCP.

The primary purpose of defining the IDCP is to ensure the necessary rich Cs sampling near the start of vancomycin treatment. The new IDCP will still belong to the same subject in the dataset. There will be no wash-out period between consecutive IDCPs – once a new IDCP is indicated to start due to vancomycin re-initiation, the remaining vancomycin estimated to be in the subject’s body from the doses in the previous IDCP will be carried over for the estimation of Cs in the subsequent IDCP(s). A sentence has been added to the end of the first paragraph of the sub-section ”Data exclusion and management” under the Materials and methods section.

Line 189: Define dosing interval.

Dosing interval has been defined later in the same paragraph. We apologize for putting the definitions after usage, thus causing confusion. We have moved the definition to the start of the paragraph.

Line 227: 168 hours is between 28 and 42 half-lives of vancomycin. This seems excessive. Consider reducing this time frame or explain why this time frame is chosen.

Thank you for pointing this out. The intention of introducing this time frame is to remove any Cs whose prediction might be inaccurate due to missing dosing records. (E.g. it is reasonable to suspect that there are missing dosing records when vancomycin remains detectable in a subject with normal renal function 1 week after the last dose.) It is true that 168 hours is typically 28 to 42 half-lives of vancomycin. However, this ratio is much reduced in subjects with profound renal impairment. Therefore, we would like to take a 7-day period as an approximated average. Besides, we will keep an eye on any significant underestimation for Cs recorded after a suspiciously long period after the last dose in case any missing dosing records are not being picked up by the above 7-day method.

Line 290: Define TTT, TTA, TTN

To reduce the use of acronyms, as suggested by other reviewers, we have removed the acronyms TTT (time to therapeutic target attainment), TTA (time to afebrile), and TTN (time to normal white blood cell count) from the text.

Submitted filename: Responses_to_reviewers_v1.1.docx

Click here for additional data file.

19 Apr 2022

Multi-center prospective population pharmacokinetic study and the performance of web-based individual dose optimization application of intravenous vancomycin for adults in Hong Kong: A study protocol

PONE-D-21-24164R1

Dear Dr. Lam,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Steven Eric Wolf, MD

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Does the manuscript provide a valid rationale for the proposed study, with clearly identified and justified research questions?

The research question outlined is expected to address a valid academic problem or topic and contribute to the base of knowledge in the field.

Reviewer #1: Yes

**********

2. Is the protocol technically sound and planned in a manner that will lead to a meaningful outcome and allow testing the stated hypotheses?

The manuscript should describe the methods in sufficient detail to prevent undisclosed flexibility in the experimental procedure or analysis pipeline, including sufficient outcome-neutral conditions (e.g. necessary controls, absence of floor or ceiling effects) to test the proposed hypotheses and a statistical power analysis where applicable. As there may be aspects of the methodology and analysis which can only be refined once the work is undertaken, authors should outline potential assumptions and explicitly describe what aspects of the proposed analyses, if any, are exploratory.

Reviewer #1: Yes

**********

3. Is the methodology feasible and described in sufficient detail to allow the work to be replicable?

Descriptions of methods and materials in the protocol should be reported in sufficient detail for another researcher to reproduce all experiments and analyses. The protocol should describe the appropriate controls, sample size calculations, and replication needed to ensure that the data are robust and reproducible.

Reviewer #1: Yes

**********

4. Have the authors described where all data underlying the findings will be made available when the study is complete?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception, at the time of publication. The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above and, if applicable, provide comments about issues authors must address before this protocol can be accepted for publication. You may also include additional comments for the author, including concerns about research or publication ethics.

You may also provide optional suggestions and comments to authors that they might find helpful in planning their study.

(Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1:

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

25 Apr 2022

PONE-D-21-24164R1

Multi-center prospective population pharmacokinetic study and the performance of web-based individual dose optimization application of intravenous vancomycin for adults in Hong Kong: A study protocol

Dear Dr. Lam:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Steven Eric Wolf

Academic Editor

PLOS ONE