A machine learning approach to predict daptomycin exposure from two concentrations based on Monte Carlo simulations.

Daptomycin is a concentration-dependent lipopeptide antibiotic for which exposure/effect relationships have been shown. Machine learning (ML) algorithms, developed to predict the individual exposure to drugs, have shown very good performances in comparison to maximum a posteriori Bayesian estimation (MAP-BE). The aim of this work was to predict the area under the blood concentration curve (AUC) of daptomycin from two samples and a few covariates using XGBoost ML algorithm trained on Monte Carlo simulations. Five thousand one hundred fifty patients were simulated from two literature population pharmacokinetics models. Data from the first model were split into a training set (75%) and a testing set (25%). Four ML algorithms were built to learn AUC based on daptomycin blood concentration samples at pre-dose and 1 h post-dose. The XGBoost model (best ML algorithm) with the lowest root mean square error (RMSE) in a 10-fold cross-validation experiment was evaluated in both the test set and the simulations from the second population pharmacokinetic model (validation). The ML model based on the two concentrations, the differences between these concentrations, and five other covariates (sex, weight, daptomycin dose, creatinine clearance, and body temperature) yielded very good AUC estimation in the test (relative bias/RMSE = 0.43/7.69%) and validation sets (relative bias/RMSE = 4.61/6.63%). The XGBoost ML model developed allowed accurate estimation of daptomycin AUC using C0, C1h, and a few covariates and could be used for exposure estimation and dose adjustment. This ML approach can facilitate the conduct of future therapeutic drug monitoring (TDM) studies.

In Silico Pharmacokinetics Evaluation of Forgiveness for Doravirine and Rilpivirine.

BACKGROUND: This study aimed to evaluate the concentrations of rilpivirine (RLP) and doravirine (DOR) after 3 days-off using simulations from population pharmacokinetics models. METHODS: The authors conducted a series of 500 sets of 10,000 Monte Carlo simulations to examine the steady-state conditions for 2 common dosage levels: 25 mg/d for RLP and 100 mg/d for DOR. These simulations were conducted under 2 scenarios: 1 without drug cessation and another after a 3-day break. The validity of the implementation was established through a comparison of median trough concentrations (C24h) with previously reported data. Subsequently, the proportion of simulated patients with C24h and C72h after 3 days-off (C72h/3do) that exceeded the inhibitory concentration 50 (IC50), 5.2 mcg/L for DOR and 20.5 mcg/L for RLP respectively, was calculated. The inhibitory quotient (IQ) was also computed, which was 6 times IC50 for DOR and 4.5 times IC50 for RLP. Finally, nomograms were constructed to estimate the probability of having C72h/3do > IC50 or > IQ for different ranges of C24h. RESULTS: Simulated C24h median ± SD for RLP were 61.8 ± 0.4 mcg/L and for DOR 397 ± 0 mcg/L. For RLP, 99.3 ± 0.1% exceeded IC50 at C24h, 16.4 ± 0.4% at C72h/3do, and none surpassed the IQ threshold. In contrast, DOR had 100% ± 0% above IC50 at C24h, 93.6 ± 0.2% at C72h/3do, and 58.6 ± 0.5% exceeded the IQ. CONCLUSIONS: These findings suggest that treatment with DOR may offer a more forgiving therapeutic profile than RLP, given the larger proportion of patients achieving effective drug exposure with DOR. However, it is important to acknowledge a significant limitation of this study, namely, the assumption that drug concentration is a perfect surrogate for drug effectiveness.

Application of machine learning to predict tacrolimus exposure in liver and kidney transplant patients given the MeltDose formulation.

PURPOSE: Machine Learning (ML) algorithms represent an interesting alternative to maximum a posteriori Bayesian estimators (MAP-BE) for tacrolimus AUC estimation, but it is not known if training an ML model using a lower number of full pharmacokinetic (PK) profiles (= "true" reference AUC) provides better performances than using a larger dataset of less accurate AUC estimates. The objectives of this study were: to develop and benchmark ML algorithms trained using full PK profiles to estimate MeltDose®-tacrolimus individual AUCs using 2 or 3 blood concentrations; and to compare their performance to MAP-BE. METHODS: Data from liver (n = 113) and kidney (n = 97) transplant recipients involved in MeltDose-tacrolimus PK studies were used for the training and evaluation of ML algorithms. "True" AUC0-24 h was calculated for each patient using the trapezoidal rule on the full PK profile. ML algorithms were trained to estimate tacrolimus true AUC using 2 or 3 blood concentrations. Performances were evaluated in 2 external sets of 16 (renal) and 48 (liver) transplant patients. RESULTS: Best estimation performances were obtained with the MARS algorithm and the following limited sampling strategies (LSS): predose (0), 8, and 12 h post-dose (rMPE = - 1.28%, rRMSE = 7.57%), or 0 and 12 h (rMPE = - 1.9%, rRMSE = 10.06%). In the external dataset, the performances of the final ML algorithms based on two samples in kidney (rMPE = - 3.1%, rRMSE = 11.1%) or liver transplant recipients (rMPE = - 3.4%, rRMSE = 9.86%) were as good as or better than those of MAP-BEs based on three time points. CONCLUSION: The MARS ML models developed using "true" MeltDose®-tacrolimus AUCs yielded accurate individual estimations using only two blood concentrations.

Optimization of Vancomycin Initial Dose in Term and Preterm Neonates by Machine Learning.

INTRODUCTION: Vancomycin is one of the antibiotics most used in neonates. Continuous infusion has many advantages over intermittent infusions, but no consensus has been achieved regarding the optimal initial dose. The objectives of this study were: to develop a Machine learning (ML) algorithm based on pharmacokinetic profiles obtained by Monte Carlo simulations using a population pharmacokinetic model (POPPK) from the literature, in order to derive the best vancomycin initial dose in preterm and term neonates, and to compare ML performances with those of an literature equation (LE) derived from a POPPK previously published. MATERIALS AND METHODS: The parameters of a previously published POPPK model of vancomycin in children and neonates were used in the mrgsolve R package to simulate 1900 PK profiles. ML algorithms were developed from these simulations using Xgboost, GLMNET and MARS in parallel, benchmarked and used to calculate the ML first dose. Performances were evaluated in a second simulation set and in an external set of 82 real patients and compared to those of a LE. RESULTS: The Xgboost algorithm yielded numerically best performances and target attainment rates: 46.9% in the second simulation set of 400-600 AUC/MIC ratio vs. 41.4% for the LE model (p = 0.0018); and 35.3% vs. 28% in real patients (p = 0.401), respectively). The Xgboost model resulted in less AUC/MIC > 600, thus decreasing the risk of nephrotoxicity. CONCLUSION: The Xgboost algorithm developed to estimate the initial dose of vancomycin in term or preterm infants has better performances than a previous validated LE and should be evaluated prospectively.

Distribution of proteinuria- and albuminuria-to-creatinine ratios in preterm newborns.

BACKGROUND: Urine protein assessment is important when glomerular disease or injury is suspected. Normal values of proteinuria already published for preterm newborns suffer from limitation, with small cohorts of patients. This prospective study was conducted to update the urine total protein- and albumin-to-creatinine ratio values. METHODS: Urine samples were collected from 231 preterm newborns within the first 48 h (D0-1) and/or between 72-120 h of life (D3-4). Total protein, albumin, and creatinine were measured, their distribution and upper-limit values determined. RESULTS: At D0-1 and D3-4, respectively, the median for the total protein-to-creatinine ratio were 80 and 107 mg/mmol (upper-limit values 223 and 289 mg/mmol) in the whole studied population, 149 and 214 mg/mmol in children born before 29 weeks of gestational age, 108 and 130 mg/mmol in those born between 29 and 33 weeks, and 61 and 93 mg/mmol in those born after 33 weeks. For the albumin-to-creatinine ratio, the median were 12 and 17 mg/mmol (upper-limit values 65 and 62 mg/mmol) in the whole studied population, 22 and 50 mg/mmol in children born before 29 weeks, 21 mg/mmol in those born between 29 and 33 weeks, and 8 and 12 mg/mmol in those born after 33 weeks. The use of nephrotoxic drugs and mechanical ventilation seems to influence proteinuria and albuminuria values. CONCLUSIONS: We report distribution of proteinuria- and albuminuria-to-creatinine in preterm newborns, including the upper-limit values. These values should be taken into account in the detection and diagnosis of glomerular disease and/or injury in daily clinical practice. Graphical abstract.

Evolution of Muscular Oxygenation during a Walking Test in Preterm Children.

OBJECTIVE: To explore measures of peripheral muscular oxygenation, coupled to gait characteristics, between preterm and full-term children during a 6-minute walking test (6MWT). STUDY DESIGN: Prepubescent children performed a 6MWT. During the test, changes in muscular oxyhemoglobin, deoxyhemoglobin, and total hemoglobin were measured with Near-infrared spectroscopy technology, positioned on subjects' calves. Gait variables were monitored with the OptoGait system. RESULTS: Forty-five children (33 full-term children and 12 preterm children, mean age, 4.9 ± 0.7 and 4.6 ± 0.9 years, respectively) participated in this study. Statistical analysis highlighted a decreased walking performance for preterm children, with significantly lower walking distance (P < .05) than children born full-term (405.1 ± 91.8 m vs 461.0 ± 73.3 m respectively; -9%). A concomitant increase of oxygen extraction (over the time course of Variation of desoxyhemoglobin) was observed from the third minute of the test (P < .05). No statistically significant difference was found for other near-infrared spectroscopy measures. Finally, the analysis of gait variables highlighted a group effect for walking speed (P < .05) and stride length (P < .01). CONCLUSIONS: Premature children showed decreased walking performance and greater change in peripheral muscular oxygen extraction, associated with slower walking speed and stride length. This may point to a muscular maladjustment and reduced functional capacities for children born preterm. These phenomena could be responsible for greater muscular fatigue.

Closure of Isolated Congenital Coronary Artery Fistula: Long-Term Outcomes and Rate of Re-intervention.

Long-term outcome after closure of isolated congenital coronary artery fistula (ICCAF) is poorly documented. To assess late outcome after ICCAF closure, a 1983-2013 retrospective study included all patients who attempted an ICCAF closure and whose follow-up was ≥1 year. ICCAF was diagnosed in 23 patients [median age 6.9 years (0.1-70.5 years), 13 children]. ICCAF was symptomatic in 12 patients (52.2 %). First intervention was either a transcatheter embolization (n = 19 patients, 82.6 %) or a surgical ligation (n = 4 patients, 17.4 %). After a follow-up of 9.0 years (2.8-33.5), neither death nor late ischemic event occurred but one patient was transplanted, because of postoperative myocardial infarction. Late ICCAF recanalization occurred in eight patients, leading to successful embolization of the shunt in all patients after a delay of 9.8 years (5.7-13.8 years) from the first intervention. Re-intervention occurred later in children (p = 0.0027), with a 50 and 37.5 % freedom from re-intervention in adults compared to a 100 and 89.0 % in children, respectively, at 1 and 6 years of follow-up. At last follow-up, coronary artery diameter had decreased from a mean z score of 12.0 ± 7.7 to a mean z score of 6.0 ± 6.0 (p = 0.002). Long-term outcome after ICCAF closure is excellent, with neither death nor late ischemic event, and a significant decrease in coronary artery diameter with time. Late follow-up is of paramount importance, as one-third of patients will require a re-intervention for late shunt recanalization.

Optimization of Ganciclovir and Valganciclovir Starting Dose in Children by Machine Learning.

BACKGROUND AND OBJECTIVES: Ganciclovir (GCV) and valganciclovir (VGCV) show large interindividual pharmacokinetic variability, particularly in children. The objectives of this study were (1) to develop machine learning (ML) algorithms trained on simulated pharmacokinetics profiles obtained by Monte Carlo simulations to estimate the best ganciclovir or valganciclovir starting dose in children and (2) to compare its performances on real-world profiles to previously published equation derived from literature population pharmacokinetic (POPPK) models achieving about 20% of profiles within the target. MATERIALS AND METHODS: The pharmacokinetic parameters of four literature POPPK models in addition to the World Health Organization (WHO) growth curve for children were used in the mrgsolve R package to simulate 10,800 pharmacokinetic profiles. ML algorithms were developed and benchmarked to predict the probability to reach the steady-state, area-under-the-curve target (AUC0-24 within 40-60 mg × h/L) based on demographic characteristics only. The best ML algorithm was then used to calculate the starting dose maximizing the target attainment. Performances were evaluated for ML and literature formula in a test set and in an external set of 32 and 31 actual patients (GCV and VGCV, respectively). RESULTS: A combination of Xgboost, neural network, and random forest algorithms yielded the best performances and highest target attainment in the test set (36.8% for GCV and 35.3% for the VGCV). In actual patients, the best GCV ML starting dose yielded the highest target attainment rate (25.8%) and performed equally for VGCV with the Franck model formula (35.3% for both). CONCLUSION: The ML algorithms exhibit good performances in comparison with previously validated models and should be evaluated prospectively.

Cognitive aid and performance for simulated umbilical venous catheter placement: A randomized trial.

INTRODUCTION: Neonatal resuscitation may require urgent umbilical venous catheter (UVC) placement. Complications can be observed with umbilical venous catheterization, especially in a stressful context. Inspired by the aeronautic environment, medical routine checklists, also called "cognitive aids," secure the equipment and environment for the patients once they are admitted to the operating room. We hypothesized that reading a cognitive aid for UVC placement in the delivery room during neonatal resuscitation simulation scenarios can (a) improve the performance in reducing catheterization duration and (b) can limit complications. METHODS: This was a prospective single-center randomized study. A total of 23 dyads for a simulation scenario were included: 12 in the control group and 11 in the cognitive aid group. In the cognitive aid group, the cognitive aid was read by the same facilitator for every scenario. RESULTS: No significant difference concerning the duration of the procedure was identified between the cognitive aid and control groups: 412 s [342; 420] vs. 374 s [338;402], respectively (p = 0.781). Nevertheless, there were significantly fewer deviations from hygiene guidelines and improved prevention of air embolism in the cognitive aid group compared with the control group. CONCLUSION: The UVC insertion time was similar between the control and cognitive aid groups. Moreover, cognitive aid can limit infectious complications or air embolism by allowing caregivers to follow UVC placement standards.

Dynamics of the digestive acquisition of bacterial carriage and integron presence by French preterm newborns according to maternal colonization: The DAIR3N multicentric study.

Objectives: The study aimed to describe the dynamics and risk factors of Gram-negative bacteria (GNB) acquisition in preterm infants. Methods: This prospective multicenter French study included mothers hospitalized for preterm delivery and their newborns, followed until hospital discharge. Maternal feces and vaginal fluids at delivery, and neonatal feces from birth to discharge were tested for cultivable GNB, potential acquired resistance, and integrons. The primary outcome was the acquisition of GNB and integrons in neonatal feces, and their dynamics, evaluated by survival analysis using the actuarial method. Risk factors were analyzed using Cox models. Results: Two hundred thirty-eight evaluable preterm dyads were included by five different centers over 16 months. GNB were isolated in 32.6% of vaginal samples, with 15.4% of strains producing extended-spectrum beta-lactamase (ESBL) or hyperproducing cephalosporinase (HCase), and in 96.2% of maternal feces, with 7.8% ESBL-GNB or HCase-GNB. Integrons were detected in 40.2% of feces and 10.6% of GNB strains. The mean (SD) length of stay of newborns was 39.5 (15.9) days; 4 died in the hospital. At least one infection episode occurred in 36.1% of newborns. The acquisition of GNB and integrons was progressive from birth to discharge. At discharge, half of newborns had ESBL-GNB or HCase-GNB, independently favored by a premature rupture of membranes (Hazard Ratio (HR), 3.41, 95% confidence interval (CI), 1.71; 6.81), and 25.6% had integrons (protective factor: multiple gestation, HR, 0.367, 95% CI, 0.195; 0.693). Conclusion: In preterm newborns, the acquisitions of GNB, including resistant ones, and integrons are progressive from birth to discharge. A premature rupture of membranes favored the colonization by ESBL-GNB or Hcase-GNB.

Comparative study of dinoprostone and misoprostol for induction of labor in patients with premature rupture of membranes after 35 weeks.

The modalities of induction of labor in the event of premature rupture of membranes are controversial. The main purpose of this study was to compare the modalities of delivery after the use of dinoprostone or misoprostol for labor induction in the preterm rupture of membranes after 35 weeks in women with an unfavorable cervix. We then studied maternal and fetal morbidity for the two drugs. Retrospective, single-center, comparative cohort study in a level 3 maternity unit in France from 2009 to 2018 comparing vaginal administration of misoprostol 50 µg every six hours (maximum 150 µg) and administration of dinoprostone 10 mg, a slow-release vaginal insert, for 24 h (maximum 20 mg), for labor induction in the preterm rupture of membranes after 35 weeks in women with an unfavorable cervix (Bishop score < 6). We included 904 patients, 656 in the misoprostol group and 248 in the dinoprostone group. Vaginal delivery rate was significantly higher in the dinoprostone group (89% vs. 82%, p = 0.016). There were more cesarean sections for abnormal fetal heart rate in the misoprostol group (p = 0.005). The time interval from induction to the beginning of the active phase of labor and the duration of labor were shorter in the misoprostol group than in the dinoprostone group (437 min vs. 719 min, p < 0.001 and 335 min vs. 381 min, p = 0.0023, respectively). Maternal and neonatal outcomes were not significantly different in the two groups. Vaginal dinoprostone used for labor induction in preterm rupture of membranes seems to be more effective for vaginal delivery than vaginal misoprostol (50 µg).

Hemolytic uremic syndrome following complicated appendicitis in a child: what is the missing link?

We herein describe an 18-month-old boy who underwent initially successful surgical and antibiotic treatment of complicated appendicitis with postoperative occurrence of hemolytic uremic syndrome (HUS). This complication was due to Shiga toxin-producing Escherichia coli (STEC) found secondarily in rectal swabs but not in the peritoneal cavity. The literature indicates that a causal link may exist between these two entities, and HUS could be considered an iatrogenic complication of appendicitis management due to a multimodal stress effect in non-symptomatic STEC carriers.