BACKGROUND:Meropenem is widely used to treat hospital-acquired pneumonia in critically ill patients, with efficacy dependent on the time that free concentrations exceed the minimum inhibitory concentration (fT > MIC). In practice, single-sample therapeutic drug monitoring (TDM) may not ensure target attainment, particularly in patients requiring continuous renal replacement (CRRT). Model-informed precision dosing (MIPD) enables individualized, real-time adjustments, but implementation is limited. Herein, we developed a Shiny application for Bayesian meropenem dose optimization. METHODS:A previously published nonparametric model was translated into a parametric Bayesian framework with maximum a posteriori (MAP) updating implemented through a Shiny interface. Plasma pharmacokinetic (PK) data from critically ill patients served for external validation. Posterior predictions were benchmarked against population and individual (MAP) outputs from Pmetrics 3.09 and population median profiles from 1000 Monte Carlo simulations per patient. Predictive performance was assessed using relative median prediction error (rMPE) and relative median absolute prediction error (rMAPE), for the combined cohort overall and stratified by CRRT status. Sparse sampling scenarios (one, two, and three observations per patient) were also evaluated. RESULTS:Eighteen patients were evaluated (non-CRRT n = 13; CRRT n = 5). In Pmetrics, rMPE ranged from -7.7% to 9.4% and rMAPE from 18.5% to 25.5% across renal replacement strata and prediction types. Simulation-based population predictions yielded greater error, with rMPE/rMAPE of -25%/34% in CRRT patients and -21%/30% in non-CRRT patients. Across Bayesian Dosing Simulator (BDS) sampling scenarios, prediction error remained within prespecified limits (rMPE ±20%, rMAPE ≤ 30%), with F20 ranging from 80% to 94% and F30 from 87% to 94%. CONCLUSION:A validated nonparametric meropenem model was successfully implemented within an open-source Bayesian framework yielding predictive accuracy comparable to the reference model with robust performance under sparse sampling, supporting its feasibility for individualized meropenem dosing. Prospective evaluation of its clinical safety and effectiveness is needed.

OBJECTIVES:Small-scale studies often lack sufficient statistical power to identify clinically relevant covariates affecting drug pharmacokinetics (PK) and dosing. To overcome these limitations, we pooled data from clinical trials and performed a population PK analysis using individual participant data meta-analysis (IPDMA) approach for polymyxin B. The aim was to uncover covariate-parameter relationships that may remain unidentifiable in isolated analyses of data from smaller studies. METHODS:Data from five polymyxin B clinical studies, comprising both single- and multiple-dose regimens in healthy and critically ill subjects, were pooled. The dataset included dosing information, concentration-time profiles, and demographic and clinical characteristics. A stepwise modelling approach was used to develop a population PK model and identify significant covariate-parameter relationships. Monte Carlo simulations were conducted to evaluate the probability of target attainment (PTA) and the safety and efficacy of various dosing strategies. RESULTS:Polymyxin B PK was best described using a two-compartment model, based on 1482 plasma concentrations from 315 subjects. A statistically significant association was observed between creatinine clearance (CrCL) and sex with drug clearance, with CrCL reducing interindividual variability (IIV) by 2.2% and sex contributing an additional 1.2% reduction. Although age and sex associated with the volume of distribution of the central compartment, with age explaining 14.3% of its IIV, inclusion of sex further reduced IIV by 2.3%. PTA analysis indicated that a 100 mg loading dose followed by 75 mg every 12 hours achieved PK/pharmacodynamic targets for nonpulmonary infections within the first 24 hours and maintained an optimal safety-efficacy balance at steady state. CONCLUSIONS:PTA significantly declined for MICs >1 mg/L, highlighting the need to reassess current MIC breakpoints. Although IPDMA enhances data integration across studies, its capacity to identify additional covariates is limited by data heterogeneity and underrepresentation of specific patient subgroups.

Tuberculosis (TB) is a major cause of morbidity and mortality in children globally. This study developed models to describe population pharmacokinetics (PK) of pyrazinamide (PZA) and ethambutol (EMB) in children with TB with or without human immunodeficiency virus (HIV) coinfection. Ghanaian children with TB with or without HIV coinfection receiving first-line antituberculosis therapy for at least 4 weeks had blood samples collected at time 0 (pre-dose), 1-, 2-, 4-, 8-, and 12-h post-dose. PZA and EMB concentrations were quantified using liquid chromatography tandem mass spectrometry. Nonlinear mixed-effects models were applied to describe the population PK using Monolix2024R1. Maximum concentrations (Cmax) and 24-h area under the time concentration curve (AUC0-24) were compared to published values in adults. A total of 85 children (41 TB, 44 TB/HIV) were included. The median (range) age was 5 years (0.3-14.5), and 61.2% were male. Median (range) doses for PZA and EMB were 31.6 (21.4-49.7) and 21.4 mg/kg (14.3-34.2), respectively. PZA was best described using a one-compartment model and EMB by a two-compartment model. Allometric scaling improved both model fits. Children with TB/HIV coinfection had approximately 18.5% faster PZA clearance and 25% faster EMB clearance. Optimized dosing to achieve adult-equivalent exposures required higher-than-currently recommended doses, particularly among children in the lowest weight bands and those with HIV. The population PK of PZA and EMB was well described by the final models, but the higher-than-currently recommended doses needed to achieve adult-equivalent exposures raise concerns regarding risks for drug-associated toxicities and will require further evaluation.

Ceftriaxone (CRO) PKs are not well studied in children with Multiple Organ Dysfunction Syndrome (MODS). We evaluated CRO PK against current CRO breakpoints (PD). This is a multi-center prospective study (R01HD103755) of antibiotic PK in critically ill children (< 18 years) with MODS (≥ 2 organ failures) prescribed CRO; those on extracorporeal support were excluded. Up to 15 PK samples were collected over 3 days using volumetric absorptive microsampling (VAMS; 20 µL/sample). Covariate data were collected. CRO was quantified in whole blood VAMS using a validated LC-MS/MS assay. Population PK modeling was performed using Monolix 2024R1. One- and two-compartment models were tested. Fixed allometric scaling of clearance (CL) and central volume of distribution (V1) defined the base models. Other covariates (e.g., Chronic Kidney Disease in Children Under 25 eGFR or U25) were evaluated for inclusion based on 1) the corrected Bayesian Information Criterion 2) reductions in between subject variability (CV%) and 3) and physiological relevance. First-24 hr patient-level PK exposures were calculated using Empirical Bayes Estimates from exact dosing and covariate histories. A literature-based free fraction of 10% was applied. The percentage of time wherein free drug concentrations exceeded the minimum inhibitory concentration (MIC) or fT>MIC was calculated for MICs of 1 (susceptible), 2 (intermediate), and 4-8 mg/L resistant pathogens. 44 patients (2 months to 17 years old) with estimated GFR ranging from 25 to 225 mL/min/1.73 m2 (median 100) were enrolled. An allometrically scaled two-compartment model with clearance adjusted based on eGFR sufficiently described the data (Figure 1). Between subject variation (CV%) for V1 and CL was 43.1% and 45.5%, respectively (Table 1). fT>MIC was >50% at CRO MICs ≤1 mg/L, variable for CRO MICs of 2-4 mg/L, and < 50% at 8 mg/L (Figure 2). CRO concentrations in critically ill children with MODS displayed significant variability, even after allometric scaling and accounting for renal function estimates. Variable and low fT>MIC (values less than 50%) were seen in MICs >1 mg/L. These data indicate that therapeutic drug monitoring approaches should be developed and are needed for all MICs ≥2 mg/L. Mark Hall, MD FCCM, Abbvie: Advisor/Consultant|Kiadis: Licensing income unrelated to this submission|Partner Therapeutics: Partner Therapeutics provides study drug for a clinical trial for which I am PI. This trial is unrelated to the submitted abstract|Sobi: Sobi provides study drug for a clinical trial for which I am PI. This trial is unrelated to the submitted abstract Nathaniel J. Rhodes, PharmD MS, Apothecademy, LLC: Advisor/Consultant Kevin J. Downes, MD, Paratek Pharmaceuticals, Inc.: Grant/Research Support|Veloxis Pharmaceuticals, Inc.: Grant/Research Support Marc H. Scheetz, PharmD, MSc, Doseme: Advisor/Consultant|other: Additional not relevant to this abstract. If more information is needed about unrelated relationships, I can provide it.

Up to half of critically ill children experience multiple organ dysfunction syndrome (MODS), which worsens outcomes and increases mortality risk. Cefepime is a broad-spectrum antibiotic commonly used in pediatric sepsis, but its pharmacokinetics during MODS has not been evaluated. We performed a prospective, observational study of critically ill children with MODS who were administered cefepime and collected up to 15 whole blood samples over 3 days using volumetric absorptive microsampling (VAMS). We performed nonlinear mixed-effects modeling to develop a population PK model for cefepime in children with MODS. We then reran our final population PK model using estimated plasma concentrations based on a 0.58:1 VAMS:plasma ratio derived ex vivo. A two-compartment model with allometric scaling best described the data. Estimated glomerular filtration rate and age were significant covariates on total body clearance and central volume, respectively. Final model parameter estimates for clearance and central volume in VAMS were 5.14 L/h and 20.5 L, respectively, for a 36-kg child. Estimates for clearance and central volume in plasma were 2.97 L/h and 11.51 L, respectively. We provide the first population PK model of cefepime in critically ill children with MODS based on VAMS, along with estimated plasma PK derived from VAMS data. Our study demonstrates the feasibility of using a novel microsampling approach to evaluate antimicrobial PK in critically ill children and provides the groundwork to better understand how this approach can optimize treatment and monitoring in this population.

BACKGROUND:Population pharmacokinetic (PK) models can be combined with Bayesian estimation to optimize dosing regimens. The impact of sample collection time on the accuracy and precision of Bayesian predictions was evaluated. METHODS:Data from adult and pediatric patients were used to develop a cefepime population PK model for Bayesian prior use. Holdout data were used for model evaluation. Clinical dosing regimens in the latter cohort were used to conduct optimal sample-time analysis. The accuracy and precision of the Bayesian predictions were assessed as a function of infusion duration and the differences between the observed and optimal sampling times. Analyses were conducted using Pmetrics for R. RESULTS:An allometrically scaled 2-compartment model was fitted (n = 71 patients, 685 observations). In the holdout group (n = 116 patients, 203 observations), the posterior Bayesian fit was acceptable (R2 = 0.923; relative bias -3%; median absolute error, 11.2%; F20, 72%; and F30, 86%). Mid-interval sampling was the optimal 1-sample design for 11/16 regimens. In the 2-sample design, a peak (8/16 regimens) and trough (9/16 regimens) approach was frequently optimal. The 2-sample design yields a lower Bayesian risk of misclassification. For 0.5-hour infusions, Bayesian predictions were similarly accurate but significantly more imprecise when samples were collected >2 hour away from the optimal time versus within ±1 hour of the optimal time (ΔRMSE: 8.98 mg/L, 95% CI: 3.61-15.7 mg/L). For 3 hours infusions, no significant differences in the accuracy or imprecision of the Bayesian predictions were noted. CONCLUSIONS:The nonparametric cefepime population PK model fit as a Bayesian prior in the holdout group. The optimal timing of PK sample collection varied according to regimen type and infusion duration. The precision of Bayesian estimates was lower for 0.5-hour infusions when samples were collected further from the model-predicted regimen-specific optimal collection times.

Population pharmacokinetic (PK) models may not generalize across critically ill populations. We evaluated the transportability of published parametric and nonparametric meropenem models as Bayesian priors. Seven published meropenem models were implemented in Monolix 2024R1. Nonparametric models were converted to log-normal priors from reported medians and CV%. An independent cohort was used for evaluation. A priori and a posteriori predictions were summarized across eligible observations and separately for observations obtained during continuous renal replacement therapy (CRRT) and in the absence of CRRT. Predictive performance was assessed using relative mean prediction error (rMPE) and relative median absolute prediction error (rMdAPE). An rMPE within ±20% and an rMdAPE ≤30% were defined as acceptable. Seventeen patients, including six receiving CRRT, contributed 74 plasma concentrations. In the pooled cohort, only the Rohani and Shekar models met predefined performance thresholds at both population and individual levels (a priori rMdAPE 24.6%-26.4%; a posteriori rMdAPE 6.7%-11.1%). In CRRT patients, no model achieved acceptable a priori performance; after Bayesian updating, all CRRT-specified models met accuracy and precision criteria, with the nonparametric models (Reed and Rohani) demonstrating low imprecision. In non-CRRT patients, several models showed population-level bias, but all met predefined accuracy and precision thresholds after Bayesian updating. Among evaluated models, Rohani, Reed, and Shekar demonstrated the most consistent predictive performance across the CRRT and non-CRRT subgroups. Nonparametric model summaries can be translated into parametric priors, but model selection remains critical when applying Bayesian dosing in critically ill patients.

Vancomycin (VAN) is a first line antibiotic for severe gram-positive bacterial infections in pediatrics but is associated with exposure-dependent kidney injury. In children with multiple organ dysfunction syndrome (MODS), most initial VAN dosing is guided via weight-based approaches. We developed a population pharmacokinetic (popPK) model for children with MODS to evaluate the relationship between acute kidney injury (AKI) and VAN exposures (area under the concentration-time curve [AUC]).Table 1.Population PK parameter fixed effects estimates and random effects (i.e. between subject variability) for central volume of distribution(V1), clearance (CL), beta coefficients on weight adjusted (individual weight/28.4kg) for V1 and CL, peripheral volume of distribution (V2), and intercompartmental bi-directional flow (Q).Figure 1.Observed vs. (A) population and (B) individual predictions for the model. We conducted a multi-center prospective observational PK study (AMPLE) embedded in a larger study of critically ill children with MODS (PARADIGM). Up to 15 PK samples were collected via volumetric absorptive microsampling over 3 days. Parametric popPK modeling was performed using Monolix 2024R1. Covariate inclusion was based on objective function and physiologic relevance. VAN AUC0-24 and AUC24-48 for each child were calculated using Empiric Bayes Estimates in Simulx 2024R1. AKI was considered a 0.3 mg/dL or 50% increase from ICU baseline. Sensitivity and specificity for classifying AKI at each AUC was calculated. A stepwise multivariate analysis determined factors (including AUC) associated with AKI.Figure 2.Sensitivity and specificity rates of specific AUC values.Table 2.Logistic regression analysis for stage 1 AKI. 66 subjects were included (median age of 10 y [range: 1 m – 17 y], median weight of 30 kg [range: 3 – 214]). A two-compartment model with clearance adjusted for allometric scaling (weight0.75) and glomerular function (calculated using CKiD Under 25 [U25] equation) best described the data (Table 1, Fig. 1). Median AUC0-24 and AUC24-48 were 454 mg*hr/L (range: 194-1569) and 505 (range: 8-1994), respectively. 7 total subjects met criteria for ICU-emergent AKI. An AUC24-48 of 465.8 mg*hr/L maintained the best sensitivity (100%) whereas at AUCs ≥ 545.5 mg*hr/L sensitivity dropped to ≤ 20% (Fig. 2). In the multivariate analysis only the Proulx score for classifying MODS was significant for AKI (6.68 OR, 95% CI 1.1 – 40.8, Table 2). Children with MODS had widely varying AUCs. AUCs ≥ 465.8 best identified AKI. In children with MODS, ICU-emergent AKI occurred within the current targeted therapeutic range. AUCs should be maintained as low as feasible. Nathaniel J. Rhodes, PharmD MS, Apothecademy, LLC: Advisor/Consultant Mark Hall, MD FCCM, Abbvie: Advisor/Consultant|Kiadis: Licensing income unrelated to this submission|Partner Therapeutics: Partner Therapeutics provides study drug for a clinical trial for which I am PI. This trial is unrelated to the submitted abstract|Sobi: Sobi provides study drug for a clinical trial for which I am PI. This trial is unrelated to the submitted abstract Kevin J. Downes, MD, Paratek Pharmaceuticals, Inc.: Grant/Research Support|Veloxis Pharmaceuticals, Inc.: Grant/Research Support Marc H. Scheetz, PharmD, MSc, Doseme: Advisor/Consultant|other: Additional not relevant to this abstract. If more information is needed about unrelated relationships, I can provide it.

In 2023, international consensus recommendations suggested the use of prolonged-infusion beta-lactam antibiotics over short-infusion dosing in severely ill adult patients to improve mortality or clinical cure, although the certainty of evidence was very low. With the availability of several new randomized controlled trials (RCTs), this focused update re-evaluates the efficacy of prolonged versus short infusion among severely ill adult patients. The original Population, Intervention, Comparator, and Outcome (PICO) question VII was separated into two questions addressing mortality (VIIa) and clinical cure (VIIb). Methods used for study identification, screening, and evidence grading were consistent with the original guideline, with the addition of Bayesian meta-analyses. A total of 28 RCTs evaluating mortality in severely ill adults were analyzed, with pooled estimates favoring prolonged infusion (RR 0.91; 95% confidence interval 0.85 to 0.97), supported by Bayesian analyses demonstrating a 98.76% posterior probability of benefit. Across 21 RCTs evaluating clinical cure, prolonged infusion improved outcomes compared with short infusion (RR 1.11; 95% confidence interval 1.06 to 1.17), supported by Bayesian analyses demonstrating a 99.97% posterior probability of benefit. Certainty of evidence was strengthened to moderate for mortality and low for clinical cure. Consistent with the original recommendation (PICO X), subgroup analyses evaluating loading doses suggested improved outcomes when a loading dose is used before continuous infusion. Overall, the cumulative evidence reinforces and strengthens the recommendation for use of prolonged-infusion beta-lactam therapy in severely ill adult patients to improve survival and clinical cure. Continued use of loading doses when initiating continuous-infusion therapy is suggested. Future research should identify subpopulations most likely to benefit, clarify optimal prolonged-infusion strategies, and define the role of therapeutic drug monitoring.

Cefepime is associated with neurotoxicity, particularly in the setting of renal impairment where drug exposure increases. This study evaluated cefepime concentrations in the brain tissue during neurotoxicity in a rat model of acute kidney injury (AKI). Male Sprague-Dawley rats (n = 18) received daily intravenous cefepime at 1,250 or 1,593 mg/kg for 5 days. Acute kidney injury was induced using folic acid (250 mg/kg on day 1, then 100 mg/kg/day prior to cefepime). Seizure activity was assessed using a modified Racine scale. Plasma samples were collected three to four times per day; the brain tissues (cerebral cortex and hippocampus) were collected at euthanasia. Cefepime concentrations were measured via liquid chromatography-tandem mass spectrometry (LC-MS/MS). Pharmacokinetic/pharmacodynamic (PK/PD) modeling was performed using Monolix 2024R1. Rats with seizure scores >1 had significantly higher median cefepime levels in the cerebral cortex (64.2 vs 14.1 μg/g, P = 0.0014) and hippocampus (66.2 vs 15.0 μg/g, P < 0.0001). Median (IQR) estimated exposures in the cortex were AUC0-24 = 565 (161.5-1,346) mg·h/L and Cmax = 36.0 (16.3-75.6) mg/L, and in the hippocampus were AUC0-24 = 694.8 (151.5-1,152) mg·h/L and Cmax = 41.9 (13.7-62.8) mg/L. Neurotoxicity in the rat correlated with plasma AUC0-24 > 30,000 mg·h/L and brain tissue concentration of approximately 40 mg/L. Neurotoxicity was achieved via intravenous dosing of cefepime in this rat model of acute kidney injury. Cefepime concentrations were higher in the cerebral cortex and hippocampal brain tissues in animals that had seizure stages >1.

Accurate drug concentration measurement is essential for precision pharmacotherapy, but conventional therapeutic drug monitoring (TDM) requires venous sampling, increasing patient burden, and potentially limiting participation in TDM and model-informed precision dosing (MIPD). Volumetric microsampling enables self-collection of small, fixed volume capillary samples for centralized analysis and may reduce hematocrit-related volumetric bias associated with conventional dried blood spots. This scoping review, conducted according to the PRISMA Extension for Scoping Reviews, synthesized 67 patient-cohort studies published from 2014 through 2026. Studies spanned at least 18 countries and multiple clinical specialties. Liquid chromatography-tandem mass spectrometry was used in 61 studies (91%), and most reported validation using or referencing EMA, FDA, IATDMCT, or CLSI criteria. Recovery was minimally affected across evaluated hematocrit ranges for most analytes, and many analytes remained stable in dried samples, although stability varied by analyte and storage conditions. Where assessed, patient preference was high (80%-100%). However, interpretation against established plasma therapeutic ranges often required whole-blood-to-plasma or capillary-to-venous conversion, with inconsistent derivation and validation methods. Fixed conversion factors performed well for lacosamide and levetiracetam but were less reliable for lamotrigine at high concentrations. Several analyte-matrix combinations, including cefepime, vancomycin, meropenem, paclitaxel, abiraterone, and mitotane, did not consistently meet agreement criteria. Volumetric microsampling can support patient-centered drug monitoring when validated for the specific analyte, matrix, device, and clinical application. Future research should prioritize drug-specific optimization, multicenter validation, standardized conversion reporting, and integration with MIPD platforms.

Accurate assessment of renal function is fundamental to dosing renally eliminated drugs. For decades, serum creatinine-based estimating equations have been used to approximate glomerular filtration rate (GFR) for this purpose. While these equations perform reasonably well at the population level, they exhibit substantial individual-level imprecision and inaccuracy. Simultaneously, recent efforts to remove race from estimating equations addressed important ethical concerns but did not resolve the core limitation of population-based estimation for individual drug dosing. In this perspective, we argue that personalized dosing requires a shift away from demographic surrogates toward direct measurement of renal clearance and clarify when incremental therapeutic complexity is warranted. We review the limitations of endogenous biomarkers, discuss the role of cystatin C, and propose exogenously measured GFR using iohexol integrated within a Bayesian pharmacometric framework as a clinically actionable approach to individualized drug dosing. Practically, this framework can be implemented by administering a small iohexol dose, collecting sparse timed samples, estimating absolute mGFR and its uncertainty with Bayesian software, and using that posterior estimate to refine renal dose selection.

Vancomycin continuous infusion (CI) has been associated with lower nephrotoxicity. This study compared a 3-minute bolus/standard infusion comparator (SI) with 24-hour CI in a validated rat model. Male Sprague-Dawley rats (n=14) received a target total dose of 150 mg/kg as a 3-minute intravenous bolus followed by a 24-hour saline infusion (SI, n=7) or as an initial loading administration followed by continuous infusion to 24 hours (CI, n=7). Urine and serial plasma samples were collected, and terminal whole-kidney homogenates were assayed for vancomycin. Day 1 urine output increased from baseline in SI animals (19.4 vs 8.9 mL/24 h; p<0.001) but not in CI animals (12.4 vs 10.1 mL/24 h; p=0.446). Urinary KIM-1 and clusterin excretion increased on Day 1 in SI compared with baseline and CI (both p<0.001). Median AUC0-24h was 354.1 mg·h/L (IQR, 277.1-427.0) with SI and 379.9 mg·h/L (IQR, 337.7-500.9) with CI (exact Wilcoxon W=28; p=0.710). Whole-kidney vancomycin concentrations were numerically higher with SI (88.02 ± 85.12 vs 27.97 ± 18.74 µg/g; p=0.11). Urinary KIM-1 showed a four-parameter sigmoidal relationship with whole-kidney vancomycin concentration (R²=0.62). Continuous infusion was associated with lower biomarker-defined renal tubular stress and lower whole-kidney drug accumulation without a detectable difference in total systemic AUC0-24h.

Cefepime (FEP) is used for hospital- and ventilator-associated pneumonia when Pseudomonas aeruginosa is involved. However, its pharmacokinetics (PK) in severe pneumonia necessitating extracorporeal membrane oxygenation (ECMO) remain unclear. This single-center, prospective study enrolled 70 mechanically ventilated patients with suspected pneumonia (n = 9 on ECMO), excluding those on renal replacement therapy. Dosing followed institutional renal function-based protocols. Plasma concentrations were quantified by liquid chromatography-tandem mass spectrometry, and a two-compartment PK model was developed using Pmetrics for R, with volume of distribution (Vd) scaled to body weight and ECMO status, and clearance (CL) scaled to renal function. Target attainment was calculated from Bayesian posterior predictions, and Monte Carlo simulations evaluated the cumulative fraction of response (CFR) for regimens of 2 g IV every 8 h, administered as either 0.5 h intermittent or 4 h extended infusion with or without a 2 or 3 g loading dose (LD) (0.5 h). Success was defined as achieving 100% fT >1xMIC within 24 h for 80% of isolates. Seventy patients (60% male, n = 9 ECMO) contributed 114 plasma samples (1-14 per patient). The model fit the data well. ECMO was associated with a 2.8-fold increase in Vd without altering CL. Monte Carlo simulations demonstrated that standard dosing without an LD failed to achieve CFR ≥ 80% in ECMO patients. Incorporating a 3 g but not 2 g LD restored CFR to ≥80% in ECMO. ECMO significantly increased FEP Vd in intensive care unit patients, suggesting sub-optimal target attainment at higher minimum inhibitory concentrations. A 3 g LD appears essential for target attainment, underscoring the need for revised dosing strategies in ECMO.

OBJECTIVES:Estimated glomerular filtration rate is more accurate with combined creatinine and cystatin C equations (eGFR cr-cys ) than creatinine alone. This study created and evaluated a cefepime dosing nomogram based on eGFR cr-cys for initial dosing in the critically ill. DESIGN:Pharmacokinetic modeling and simulation study. SETTING:Academic medical center. PATIENTS:Critically ill adults treated with cefepime. INTERVENTIONS:None. MEASUREMENTS AND MAIN RESULTS:Data from 120 patients with baseline cystatin C and follow-up cefepime levels were used to develop a nomogram based on eGFR cr-cys and weight for initial cefepime dosing. The predicted proportion of patients who achieved a free cefepime concentration above the minimum inhibitory concentration of the organism for 100% of the dosing interval in the first 24 hours (100% ƒT > MIC at 24 hr) was compared between administered doses and those predicted by the nomogram doses. Overall drug exposure was estimated with the free area under the concentration time curve from 0 to 24 hours (ƒAUC 0-24 ) and compared between administered and nomogram doses. Achievement of 100% ƒT > MIC at 24 hours was predicted to be significantly better with the nomogram compared with the administered dose (76% vs. 38%; p < 0.001). The median ƒAUC 0-24 as predicted by the nomogram (666 mg·hr/L) was slightly higher than the actual ƒAUC 0-24 with administered doses (612 mg·hr/L; p = 0.01), but the nomogram led to fewer ƒAUC 0-24 values which were either too high (> 900) or too low (< 300) (7% vs. 20%; p = 0.004). CONCLUSIONS:Use of a cystatin C-inclusive dosing nomogram for cefepime could improve target attainment without increasing the risk of potentially toxic levels in the critically ill.

Optimizing antibiotic exposures in hospital-acquired pneumonia (HAP) is crucial; however, dosing decisions often overlook target levels for the beta-lactamase inhibitor tazobactam (TAZ), and the impact of renal dysfunction and renal replacement therapy on TAZ pharmacokinetics (PK) remains poorly described. We developed a population PK model of TAZ using 162 plasma samples from 35 ICU patients with HAP, including those receiving continuous renal replacement therapy (CRRT). TAZ concentrations were quantified using validated LC-MS/MS methods. A one-compartment model was fit using Monolix, with clearance modeled as a function of creatinine clearance, CRRT effluent flow rate, and intermittent hemodialysis. Monte Carlo simulations assessed the probability of unbound TAZ concentrations exceeding 1, 2, or 4 mg/L for 100% of the interval between 24 and 48 h across intermittent, extended, and continuous infusion (CI) regimens, stratified by renal function and CRRT. TAZ clearance was driven by renal function and CRRT flow rate. All regimens achieved ≥90% probability of target attainment (PTA) at thresholds of 1-2 mg/L. At 4 mg/L, PTA fell below 90% in patients with creatinine clearance ≥150 mL/min for all regimens, with low-dose CI performing the worst. High-dose CI improved PTA at this higher threshold but may increase the risk of piperacillin overexposure. TAZ exposures sufficient for enzymatic beta-lactamase inhibition are generally achieved with standard dosing for lower thresholds but not for more aggressive targets (e.g., high extended-spectrum β-lactamase expression), particularly in patients with preserved or augmented renal function, who may require therapeutic drug monitoring or alternative therapy.