Background
Ciprofloxacin target attainment (TA) is not achieved in 60-80% of adult critically ill patients admitted to Intensive Care (IC; using a minimal inhibitory concentration (MIC) of 0.5mg/L (Pseudomonas aeruginosa). Antibiotic dose optimization can be achieved by therapeutic drug monitoring (TDM) and should ideally only be performed in patients at risk of target non-attainment (TNA).

Aims
Develop ciprofloxacin TNA models in critically ill patients.

Methods
Data from the DOLPHIN, EXPAT studies, and Amsterdam UMC were combined to develop and validate models to predict TNA the day after antibiotic initiation. TA was defined as a plasma concentration ƒ area-under-the-curve 0–24/MIC≥125. We randomly divided 185 patients into a training and validation set (80/20%). Covariates were obtained using the Boruta and elastic net (EN) algorithm. Models were trained using the random forest (RF) and EN. The main metrics used for performance were the area-under-the-receiver-operator-curve (AUROC), sensitivity, specificity, and calibration.

Results
The covariates age, weight, frequency, and Sequential Organ Failure Assessment score (SOFA), albumin, and creatinine were good TNA predictors. In the validation, the RF and EN models showed good calibration and an AUROC of 0.81 and 0.75, a sensitivity of 0.8 and 0.61, and a specificity of 0.70 and 0.77, respectively, without the covariates SOFA score and albumin.

Conclusions
Both ciprofloxacin TNA models showed good performance in the validation set, using 4 readily available predictors in the ICU setting. These -easy to implement- models may improve the allocation of TDM resources in the ICU.

Keywords
PK/PD, TDM, critical care, target attainment, ciprofloxacin, fluoroquinolones

Backgroung : Previous studies have shown that more than a third of intensive care units (ICU) patients do not reach the target concentrations of beta-lactams, despite administration through continuous infusion.

Aims : The objective was to evaluate the impact of a systematic and protocolized Therapeutic Drug Monitoring (TDM) to optimize antibiotic treatment in ICU.

Methods : This prospective, non-randomized, bicentric, comparative interventional study was conducted on 172 ICU patients treated with continuous infusion of five beta-lactams. The pharmacokinetic/pharmacodynamics (PK/PD) target was a free concentration greater than four times the minimum inhibitory concentration of the implicated pathogen. One group of patients received systematic TDM (days 1, 4, and 7), while the control group did not. The primary objective was to evaluate the impact of TDM on PK/PD target attainment.

Results : No difference has been identified regarding the achievement of target concentrations between cohorts with and without TDM. Reaching target concentrations at D1 allows better achievement of targets concentrations at D4 and D7 (OR=8.8 [3.5; 22.3], p<0.001). Several factors were associated with PK/PD target attainment at D1: creatinine serum, age, cefepim treatment, body surface area and sex. No significant impact of TDM was observed on clinical outcomes (ICU free days, SOFA score, neurotoxicity). A toxic concentration was associated with EEG neurotoxicity (p=0.005) and the development of renal failure (p=0.002).

Conclusions : This study highlights the interest of moving towards a better dosage personalization from initiation to the whole treatment of beta-lactams antibiotics in ICU patients.

Keywords: beta-lactams, TDM, pharmacokinetics, ICU, neurotoxicity.

Background: Underexposure of levofloxacin, a key multidrug-resistant tuberculosis (MDR-TB) drug, can lead to treatment failure and acquired drug resistance. Levofloxacin therapeutic drug monitoring (TDM) optimises dosing to ensure efficacy while minimising toxicity. However, plasma-based TDM is invasive and resource-intensive, limiting its use particularly in high-burden, low-resource settings.

Aims: To assess the feasibility of saliva-based TDM for levofloxacin in MDR-TB patients.

Methods: Adults receiving levofloxacin for ≥7 days had blood and saliva samples collected at 0, 2, and 5 hours post-dose at steady state (ACTRN12620000681954). Levofloxacin concentrations were quantified using both HPLC-MS and UV-spectrophotometry. Saliva and plasma concentrations and the two analytical methods were compared via Passing-Bablok regression, Bland-Altman analysis and Pearson's correlation test.

Results: Of the 60 patients, 70% were male, with a mean age of 44 years and weight of 50 kg. Among 360 paired samples, saliva and plasma concentrations on average showed a proportional bias of 1.13, a constant bias of -0.52 and strong correlation (r = 0.86, P < 0.05), with only 3.3% of values outside the 95% limits of agreement. Saliva concentrations were, on average, 0.98 times lower than in plasma. UV-spectrophotometry showed reasonable agreement with HPLC-MS, with an average proportional bias of 1.34, a constant bias of -1.15 and moderate correlation (r = 0.777, P < 0.05).

Conclusion: Levofloxacin concentrations in plasma and saliva were strongly correlated. A simple saliva-based UV-spectrophotometry could be a semiquantitative tool for screening patients at risk of levofloxacin underexposure in resource-limited settings.

Keywords: Levofloxacin, pharmacokinetics, therapeutic drug monitoring, saliva.

Background: Levofloxacin is a cornerstone of oral antibiotic therapy for osteoarticular infections (OAI) caused by Staphylococcus spp. It presents significant pharmacokinetics variability, which may lead to underexposure or unnecessary overexposure. As a result, dosing decisions are challenging for clinicians. To optimize dosing schedule, our center has introduced systematic therapeutic drug monitoring (TDM) using AUC0-24h estimation (target 62.5 to 150 μg.h/mL).

Aims: This monocentric retrospective study describes the dose-exposure relationship in OAI patients and try to determine covariates that could help choosing the right dose.

Methods: AUCs were estimated from three samples (pre-dose, 1h, and 3h post-dose) using popPK modeling. Data from TDM (Sept 2018–Nov 2024) were used to calculate the percentage of target attainment. Thereafter, considering the dose required to reach the target, we classified the patients according to bodyweight (> or < 70kg) and eGFR, or according to sex and eGFR, to determine whether we could determine a priori the right dose for the majority of patients according to these criteria.

Results: 230 patients were included. The median age, weight, and eGFR were 68y, 78kg, and 94.5mL/min/1.73m², respectively. Sixty-five percent of patients reached the target AUC with their initial dose. Our classification according to weight or sex and eGFR helped us to select the right dose for only 35% to 70% of patients, depending on the category.

Conclusion: Estimating the dose according to two parameters remains inaccurate but could be optimised using sex and weight and eGFR. No PK-model currently considers them together in OAI.

Keywords: levofloxacin, osteoarticular infection

Background: Vancomycin dosing in the neonatal intensive care unit (NICU) is challenging. Recently, a population pharmacokinetics (popPK) model and dosing app were developed and implemented by our research group for a Level IV (highest acuity level) NICU. However, external validation is necessary to ensure its generalizability in other NICUs.

Aims: To externally evaluate our derived popPK model and compare to 32 other popPK models with a Level III NICU cohort within the same geographic region.

Methods: This retrospective cohort study included neonates until a postmenstrual age (PMA) of 44 weeks and postnatal age of 4 weeks, who had ≥1 vancomycin concentration. The predictive model performance was compared by evaluating the differences between the predicted and observed vancomycin concentrations. The optimal popPK model was determined based on five criteria: mean error (ME)<0.5mg/L, mean absolute prediction error (MAPE)<15%, median absolute prediction error (mdAPE)<15%, root mean squared error (RMSE)<10, and percentage within 30% of observed value (p30)>45%.

Results: A total of 366 neonates (55% male, PMA: 28.9±3.81weeks, weight 1.06±0.6kg) with 661 vancomycin concentrations were included. Among 33 popPK models, only our popPK model met all five criteria (ME=0.46mg/L, MAPE=12.2%, mdAPE=4.2%, RMSE=8, p30=49.3%). Five other models (Jung 2021, Germovsek 2019, Oudin 2011, Marques-Minana 2010, DeHoog 2000) met at least three criteria. All six were one-compartment models and most incorporated weight, PMA and/or serum creatinine.

Conclusions: Our popPK model had robust predictive performance, outperforming 32 other models. This supports the broader application of this popPK-derived vancomycin dosing app in neonates.

Keywords: vancomycin, neonates, population pharmacokinetics

Background: A study in our institution showed an incidence of invasive fungal disease (IFD) of 38.7% in patients undergoing chemotherapy for AML. Despite Posaconazole prophylaxis, the incidence of IFD in AML and transplant patients in our institution is about 21.3%. Main challenges in posaconazole prophylaxis are lack of fatty meal intake due to Ill health, Intestinal mucositis and Intraday variability of levels.

Aims: To evaluate the role of posaconazole (in tablet or suspension form) in the prevention of IFD in AML patients and bone marrow transplant recipients.
To compare the proportion of patients who get IFD with posaconazole levels.

Methods: A prospective observational cohort study was done over a period of two years. 63 patients were recruited to determine the correlation between PCZ plasma concentration and its effectiveness and adverse effects. Blood samples are collected in 2 visits, and the patient is followed up till end of treatment for IFD.

Results: The overall incidence of IFD among the 63 patients was 28.6%. 90% of patients using oral PCZ suspension did not achieve the recommended concentration (above 700 ng/ml) in visit 2. A higher proportion of patients (p <0.01) administered PCZ tablets achieved recommended prophylactic concentrations. Subgroup analysis showed that patients with PCZ trough of >700 ng/ml in both visits, did not develop IFD.

Conclusions: Median PCZ concentrations were significantly lower in those who developed IFD than in those who did not develop fungal infections. Routine TDM in frequent intervals for PCZ for dose optimization is needed.

Key Words: Posaconazole, IFD

Background: Augmented renal clearance (ARC) enhances vancomycin elimination and potentially compromising therapeutic efficacy.

Aims: This study evaluates the impact of ARC on vancomycin pharmacokinetics (PK) and pharmacodynamics (PD) in paediatric patients and identifies factors associated with ARC.

Methods: A retrospective cohort study analyzed steady-state vancomycin levels in 201 children admitted to a tertiary paediatric hospital between January 2022 and December 2023. PK parameters were estimated using the Sawchuk-Zaske method, while the 24 hours-area under the curve (AUC24H) was calculated using the trapezoidal method. Patients were categorized into ARC and non-ARC groups for comparison.

Results: ARC was observed in 62.7% of patients at the initiation of vancomycin therapy. Children with ARC had significantly lower vancomycin PK/PD indices despite receiving similar initial doses. Compared to non-ARC patients, those with ARC had a significantly lower minimum concentration (Cmin: 6.96mg/L [IQR: 8.31] vs. 11.01mg/L [IQR: 12.90], p<0.001), lower AUC24H (270.35mg·h/L [IQR: 142.41] vs. 392.76mg·h/L [IQR: 289.12], p<0.001), and lower AUC24H/MIC (223.88mg·h/L [IQR: 358.63] vs. 334.36mg·h/L [IQR: 435.22], p=0.031). To achieve target AUC24H, patients with ARC required significantly higher vancomycin doses than those in the non-ARC group (61.20mg/kg/day [IQR: 16.00] vs. 45.00mg/kg/day [IQR: 28.80], p<0.001). ARC was associated with younger age and admission to general wards.

Conclusions: ARC is prevalent in paediatric patients and correlates with lower vancomycin exposure, thus potentially increasing the risk of treatment failure. Higher vancomycin doses are required to achieve optimal PK/PD targets in patients with ARC.

Keywords: vancomycin, pharmacokinetics, pharmacodynamics, augmented renal clearance, paediatrics