Pharmacokinetics and pharmacodynamics of linezolid treatment of bacterial infections in children of different age groups: what do we know today?

Background

Child mortality has been and remains one of the most important problems of global healthcare [1, 2], being a significant characteristic of the demographic situation in a country, demonstrating the features of the national healthcare system [3].

The causes of mortality in children of different age groups vary; nevertheless, there is a common feature in the mortality structure—infectious diseases are always present, although their share may vary. In particular, in the structure of infant mortality, endogenous causes—i.e., conditions arising in the perinatal period (81.8%) and congenital anomalies—come first. The leading components of these conditions are hemorrhagic disorders in the newborn (19.8%), respiratory distress syndrome (16.1%) and bacterial sepsis (12.8%). In older children, exogenous causes—injuries, poisonings and external causes—are in first place, but infectious and parasitic diseases, as well as respiratory diseases, play an important role [3].

It is also necessary to note a special group of concern for neonatologists and pediatricians worldwide—premature newborns. Infectious diseases and complications play an important role in the mortality structure of this group of patients—in particular, according to WHO data in 2010, 64% of deaths of children under 5 years of age worldwide were caused by infectious diseases [4, 5, 6]. In Russia, respiratory diseases, especially pneumonia, ranked second among causes of death in children under 1 year of age (82.7 per 100,000 live births), second only to congenital anomalies [7]. Bacterial sepsis in newborns also plays an important role—39.4‰ [8, 9].

Currently, the proportion of the premature newborn population is gradually increasing, due to the overall increase in the availability of medical care, as well as the improvement and development of modern medical technologies for nursing newborns. In addition, the concept of "late childbirth" leads to an increase in the proportion of women with chronic pathology [10, 11].

In the structure of primary outpatient morbidity of children aged 0–14 years, respiratory diseases occupy leading positions, with pneumonia (74.3%) playing the leading role in the structure of these diseases [12]. In the pathogen structure, Streptococcus pneumoniae (41.2–94%) plays a leading role [12–14], and over the years there has been an increase in the resistance of pneumococcal strains to penicillins, macrolides and other antibacterial drugs, which requires the use of reserve drugs.

Sepsis is the main nosocomial cause of mortality in children of all ages. The highest incidence is characteristic of intensive care units, since it is in these departments that the most invasive procedures (mechanical ventilation, catheterization) are performed, which are always associated with a high risk of nosocomial flora acquisition and the development of bacterial infections [15–20]. Depending on the localization, the etiological structure of sepsis may vary; nevertheless, Gram‑positive flora predominates in the development of sepsis due to nosocomial pneumonia, skin and soft tissue infections, and bloodstream infections (Staphylococcus spp. and Streptococcus spp.) [21–23].

Also, Gram‑positive microflora occupies an important place in the structure of infections detected in premature newborns. Gram‑positive pathogens are the main causative agents of sepsis, both with early onset (EOS) and late onset (LOS) (according to various sources, from 33.2% in EOS to 71.6% in LOS). The most common are:

• Staphylococcus spp. (6.3–77.0%) [20, 24–27], especially methicillin‑resistant Staphylococcus aureus (MRSA) (6.3–20.9%) [20], and coagulase‑negative staphylococci (CoNS) (up to 29.3%) [20, 24–27]. They are most often the causative agents of pneumonia—26–30% [28, 29], as well as meningitis—26.0–28.5% [30, 31].

• Streptococcus spp. (8.4–18.8%), in particular group B streptococci (8.4–18.8%) [25, 26]. They are most often the etiological factor in the development of meningitis (up to 50%) [32, 33], as well as pneumonia, but to a lesser extent—9–13% [28, 29].

• Enterococcus spp. (5.3–9.5%) [20, 24–26]. They play an important role in the development of pneumonia (19–33%) [28, 34], as well as urinary tract infections (10–19.6%) [35–38].

For the eradication of the identified pathogens, it is extremely important to choose the optimal antibacterial drug and an adequate dosing regimen. For successful elimination of the pathogen, it is necessary to take into account the peculiarities of pharmacokinetics (PK) and pharmacodynamics (PD) in children, and this is even more important to consider when treating newborns and premature newborns. Unsatisfactory outcomes of infectious diseases can often be explained by incorrect selection of the dosing regimen of an antibacterial drug (ABP), which does not allow achieving the plasma and infection site concentrations necessary for eradication of the pathogen [39, 40], or more precisely, the target values of PK/PD indices (parameters) reflecting the pharmacological activity of the ABP.

Due to the growing resistance of pathogens to traditional antibiotics such as ampicillin, aminoglycosides and cefotaxime, adequate therapy often requires glycopeptides, for example vancomycin, but its use is limited by high nephrotoxicity [41–47].

One of the most widely used narrow‑spectrum ABPs with high activity against the main representatives of Gram‑positive microflora, in particular against staphylococci, is linezolid. According to generally accepted recommendations, the dosing regimen for children under 12 years of age is 10 mg/kg every 8 hours, for adolescents 12–17 years old — 600 mg twice daily, as in adult patients. In premature newborns (gestational age (GA) <34 weeks and postnatal age (PNA) <8 days), due to pharmacokinetic features, longer intervals between administrations may be recommended than in older children—up to 12 hours [48].

However, the literature contains studies in which, in adult patients receiving standard dosing regimens, according to therapeutic drug monitoring (TDM) data, almost half of the patients had linezolid concentrations outside the recommended range, while more than 30% of patients had a risk of concentration‑dependent hematological adverse events due to excess exposure, and this risk was especially pronounced in patients with reduced renal function [49, 50]. In special patient populations, including children and newborns, this percentage was even higher, leading to an increased risk of both underdosing and overdosing. All this indicates the importance of optimizing linezolid therapy taking into account pharmacokinetics and pharmacodynamics.

Linezolid therapy may be associated with adverse events: gastrointestinal symptoms (nausea, diarrhea, vomiting), hematotoxicity (myelosuppression), and lactic acidosis [51–56]. For linezolid, a strong association between exposure and efficacy, as well as some types of toxicity, is known, which also supports the optimization of linezolid dosing based on population PK/PD modeling.

During the period of linezolid use in clinical practice, population pharmacokinetics of linezolid in children has been studied in several studies, mainly in mixed age populations [57, 58]. At the same time, an extremely limited number of PK/PD studies of linezolid have been conducted in the newborn population, and even fewer in the premature newborn population.

Objective

The aim of our study was to search, review, analyze and summarize the results of conducted PK/PD studies of linezolid in children of different age groups.

Materials and methods

Search methodology: systematic review, carried out in accordance with the requirements of the PRISMA guideline for systematic reviews and meta‑analyses. After selection of studies, the titles of articles were exported into a Microsoft Office (USA) 2019 spreadsheet: Microsoft Excel. After data analysis, tables were created in Microsoft Office (USA) 2019: Microsoft Word. Quantitative data were analyzed using arithmetic means, medians, minimum and maximum values (min‑max), and standard deviation. Qualitative data were described using absolute and relative values (%).

Databases searched: scientific electronic library eLIBRARY.ru, PubMed database.

Search period: from 01.01.2000 to 31.12.2025.

Keyword search: information was searched in the scientific electronic library eLIBRARY.ru and the PubMed database using combinations of keywords: «дети или новорождённые» и «линезолид» и «фармакокинетика или фармакокинетический (ая) или ФК/ФД или популяционный (ая) или модельный (ая) или терапевтический лекарственный мониторинг или ТЛМ», as well as «(infant OR newborn OR neonate OR pediatric) AND (linezolid) AND (pharmacokinetic OR PK/PD model (ling) OR population model (ling) OR therapeutic drug monitoring OR TDM».

Inclusion criteria:

1. Study population: pediatric (premature and term newborns, as well as children under 18 years of age);

2. Investigational drug: linezolid;

3. Studies involving the use of linezolid in patients with various bacterial infections;

4. Studies containing original PK studies of linezolid;

5. The concentration sampling scheme, dosing regimens, demographic and clinical characteristics of patients had to be described in detail;

6. The methods used for interpreting TDM data, pharmacokinetic, pharmacodynamic and/or population PK/PD modeling also had to be described in detail;

7. Years of publication: 2000–2025.

Studies were excluded if:

1. They were single case reports or review articles;

2. Studies describing the same population model;

3. Studies lacking data on linezolid PK parameters;

4. Articles including tuberculosis therapy;

5. The work did not present demographic and anthropometric data of the studied patient population;

6. Works included data from adult patient populations;

7. Studies were written in a language other than Russian or English.

Ethical approval. For this systematic review, ethical approval is not required.

Study identification. The PRISMA flow diagram is presented in Figure 1.

Figure 1. PRISMA Flow Diagram

Discussion

PK/PD studies in mixed populations of term and preterm neonates

Information on the main studies in this category is presented in Table 1. In the earliest PK study, 42 newborns were included and divided into 4 groups, depending on postnatal age (PNA, <8 days vs 8 days – 12 weeks) and gestational age (GA, <34 weeks vs ≥34 weeks). Sixteen patients had GA <34 weeks. The study revealed that in all patients with different GA and PNA ≤3 months, after a single intravenous dose of linezolid 10 mg/kg, pharmacokinetics varied significantly during the first week of life depending on PNA [59]. It was shown that linezolid clearance in newborns with PNA <7 days was on average lower than in newborns with longer PNA. At the same time, in premature newborns (GA <34 weeks) during the first week of life, clearance was lower than in the other patients. Clearance increased significantly during the first week of life, regardless of GA, so that in newborns with PNA >7 days it was on average 3 times higher than in adults [59, 60].

In a retrospective study by Sicard et al., PK data from 16 premature newborns (24 linezolid concentrations) were studied, with a mean PNA at the start of therapy of 3 weeks; 12 patients had GA less than 28 weeks. Patients received linezolid intravenously as a prolonged intravenous infusion and orally: a dose of 10 mg/kg every 8 hours for oral administration or 30 mg/kg for intravenous infusion. This work can be considered the first PK/PD study of linezolid in premature newborns based on TDM data, although without calculation of a population model and PK parameters. The main limitations of this study were the small sample size and intravenous/oral dosing of linezolid, as well as the absence of quantitative estimates of PK parameters and their variability [61].

In the PK/PD study by Thibault et al. in premature newborns, data from 26 patients (78 linezolid concentrations: median 3 (1‑7) per patient) were retrospectively studied, and a one‑compartment PK model was developed [55]. The median (min‑max) values of PNA, GA and body weight in the patients included in the study were 24 days (8‑88), 27 weeks (25‑36) and 1423 g (810‑3256), respectively. Twenty‑three patients (88%) had GA <34. The median value of the linezolid doses received was 10 mg/kg (8.1‑15) every 8 hours in 18 newborns (69%) or every 12 hours in 1 patient (4%). Seven patients (27%) received linezolid both every 8 h and every 12 h due to dose interval adjustment during TDM. The median duration of therapy was 13 days (2‑45). Samples for linezolid concentration measurement during the TDM procedure were taken at least after 48 hours of therapy at three time points: before the next dose, 0.5 h and 2‑3 h after the end of infusion. PK data were analyzed based on population modeling (a one‑compartment PK model adequately described the data). The final population PK model of linezolid included PNA and body weight as significant covariates for clearance, and only body weight for volume of distribution. The effect of PNA likely reflects the maturation of hepatic and renal function, which is fundamentally important for linezolid pharmacokinetics. Estimates of population values of linezolid clearance (0.17±0.17 L/h/kg and 0.16±0.03 L/h/kg in newborns with GA <34 weeks and ≥34 weeks, respectively) [59]. It should be borne in mind that there were only 3 patients with GA ≥34 weeks in the study, so the results should be interpreted with caution. Kearns et al. did not include clinically unstable patients with organ dysfunction in their study [59], whereas such patients were included in the Thibault C. study, so the study may have included more severely ill patients. Hepatic dysfunction in severe illness may probably explain the lower clearance values, although no association was found between linezolid clearance and AST and ALT values, possibly due to the limited number of measurements of these laboratory parameters available for analysis. Another explanation for the difference in average parameter values may be that Kearns et al. studied linezolid PK after a single dose, whereas in the Thibault C. study linezolid pharmacokinetics was at steady state. The fact that linezolid clearance at steady state is somewhat lower than after a single dose in healthy volunteers has been previously described [55]. A hypothesis was put forward that linezolid may inhibit its own metabolism by inhibiting the mitochondrial respiratory chain, leading to a decrease in clearance over time [62]. The authors Thibault C. et al. suggested that the absence of concentration measurements in their work before reaching steady state may have prevented observation of this phenomenon. The volume of distribution was larger in patients with GA <34 weeks, as expected, due to differences in total body water and body weight in premature and extremely premature infants. Estimates of volume of distribution were at the upper limit of previously published values (0.99±0.17 L/kg and 0.72±0.03 L/kg versus 0.87±0.13 L/kg and 0.65±0.21 L/kg in patients with GA <34 weeks and 34 weeks, respectively) [59].

In a prospective PK/PD study of linezolid [63] in the premature newborn population, 54 patients (PNA = median 24 days (8‑88 days)) receiving the standard dosing regimen were included. Eighty‑four linezolid concentrations at steady state after 4 maintenance doses were measured: trough concentrations and concentrations within the dosing interval. Each patient had 1‑2 concentrations measured, at least one of which was a trough. The study used a one‑compartment PK model with first‑order linear elimination. Among covariates, body surface area (BSA) was selected as the most significant for clearance (CL) and volume of distribution (V). It was shown that in premature newborns, BSA may be considered the main covariate affecting the total clearance of linezolid, possibly due to the development of organs and systems, especially the kidneys [86].

Table 1

Demographic data and estimated PK parameters in studies of term and preterm neonates

Note: ¹ AUC — area under the plasma concentration‑time pharmacokinetic curve.

PK/PD studies in populations of term newborns and older children

The main studies in this category are presented in Table 2. A population PK/PD model of linezolid including children from 0 to 12 years (body weight from 2.1 to 46 kg) was presented in the work of Li et al. [64]. Of the 91 patients included in the study, 91 had normal renal function (GFR ≥90 mL/min1.73 m²), 17 patients had reduced renal function (GFR 60 — <90 mL/min1.73 m²), and 4 patients had moderate renal impairment (GFR 30 — <60 mL/min*1.73 m²). Patients received linezolid therapy for at least three days, the dosing regimen was a 1‑hour intravenous infusion of 10 mg/kg every 8 hours. A total of 135 linezolid concentrations were included in the PK analysis: 47 were trough concentrations, 5 were peak concentrations at 0.5 hours after the end of infusion, and the remaining 83 samples were taken at time points between Cmax and Ctrough. In this study, body weight and GFR were found to be significant covariates for linezolid clearance. The mean value of linezolid clearance in the study was estimated as 0.13 L/h/kg, which is consistent with estimates from another PK/PD study in children, which reported a mean clearance of 0.152 L/h/kg estimated in 14 patients aged 2 to 11 years [65].

In another PK/PD study [66], 15 patients aged 0‑13 years (3 patients with PNA ≤3 months) were included, receiving linezolid orally and/or via intravenous infusion, usually the prescribed dosing regimen was 10 mg/kg three times daily. All further dose adjustments were made according to the attending physician's decision. The analysis was based on 92 measurements of total linezolid concentration and 70 measurements of its free concentration. As in previous studies of linezolid pharmacokinetics in children, clearance depended on patient age, and PNA was a significant covariate. This is consistent with the results of Thibault et al. [55]. The authors [66] included PNA in the PK model in the form of a maturation function (Hill model). The estimated parameters of this model indicate that clearance significantly increases with increasing PNA, and the observed marked variability in CL values should be taken into account when dosing linezolid. Body weight was used in the model for allometric scaling of CL and V. The mean values for linezolid half‑life in the study in a patient with body weight 9.9 kg and age 24 months (median values in the study) were estimated as 3 hours. This value is not inconsistent with estimates from other pediatric PK studies of linezolid (1.8–2.9 h) [60, 64, 67]. The risk of linezolid‑associated thrombocytopenia (observed in three patients in the study) was estimated as 21.4%. In these patients, the mean trough concentration was 19.8 mg/L, significantly higher than in patients without: 6.84 mg/L. Therefore, dosing adjustments based on linezolid concentration measurements may be recommended to improve therapy safety. Although dose adjustments based on patient covariates—body weight and maturation (based on PNA)—were performed in the study, and cases of thrombocytopenia were observed and occurred at high trough linezolid concentrations, it can be concluded that dosing adjustment based on TDM may be recommended for optimizing linezolid therapy in children.

Only the free fraction of the drug produces clinical response and adverse effects, while it is extremely rare in PK studies of linezolid in children that free drug concentration in blood was measured. The final estimate of protein binding in the study [66] was estimated as 18.9%, and no significant interindividual variability in this parameter was observed in the study. The results obtained are consistent with the results of studies in adult patients [68, 69], and the authors conclude that there is no need to measure the free fraction of linezolid in clinical practice for routine TDM. Although this point needs further study, especially in patients with potential clinically significant changes in binding.

Another PK/PD model in children (63 patients) with severe infections in critical condition, aged 0.10‑15.30 years and body weight 4.2‑70.0 kg (of which 14 patients under 1 year of age) was presented in the article [70]. Linezolid was administered intravenously over 1‑2 hours at a dose of 10 mg/kg three times daily for patients <12 years and 600 mg twice daily for patients 12‑18 years. Four blood samples after reaching steady state were taken from each patient: trough and peak concentrations, as well as two samples at different intermediate time points. A two‑compartment model with first‑order linear elimination was used to describe the PK data. Body weight and AST were identified as significant covariates for clearance. In the study, patients with AST >40 U/L had much higher AUC values than patients with AST ≤40 U/L (205.45 vs 159.96, p <0.05). Interindividual variability for linezolid clearance was estimated as 52‑53%.

In the PK/PD study by Tian et al. [71], 80 patients were included, median age = 3.3 years (5‑95th percentile = 0.1‑12.6 years), and 157 linezolid concentrations after steady state (median = 2 measurements per patient) were used for PK/PD modeling. Linezolid therapy was administered via intravenous infusion at the standard dose of 10 mg/kg every 8 hours for children under 12 years and 600 mg every 12 hours for children 12 years and older. Based on a linear one‑compartment model, body weight and glomerular filtration rate (GFR) were considered statistically significant covariates for linezolid clearance. In this study, the median estimated GFR was 190.1 mL/min/1.73 m², i.e., most patients included in the study had augmented renal clearance (ARC, GFR >130 mL/min/1.73 m²). Total linezolid clearance values ranged from 0.13 to 0.171 L/h/kg, and apparent volume of distribution from 0.42 to 0.92 L/kg. Mean clearance and volume of distribution values of 0.15±0.06 L/h/kg and 0.77±0.50 L/kg, respectively, are consistent with the results of other PK/PD studies of linezolid in the pediatric population [64, 67].

In most PK/PD studies of linezolid in the pediatric patient population, the traditional linear one‑compartment model was used as the structural model. The choice of the simplest traditional PK model may be due to the fact that most of these studies used relatively sparse TDM measurements. The observed differences in estimates of population PK parameter values may largely be explained by differences in the studied patient populations, including differences in age, severity of illness, renal and hepatic function, as well as differences in the design of the studies themselves, including the concentration sampling scheme.

If we try to summarize the results presented in the literature, it can be said that the pharmacokinetics of linezolid in the pediatric population demonstrates age dependence. Thus, in children under 12 years of age, linezolid clearance is on average faster and half‑life is on average shorter compared to adult patients. Although in newborns in the first days after birth, linezolid clearance values are on average similar to those in adult patients, during the first week of life clearance increases fairly rapidly, and increases approximately 2‑3 times by the end of the first week. Subsequently, linezolid clearance decreases, reaching levels characteristic of adult patients in adolescents [72]. The authors of [66] showed that postnatal age is associated with linezolid clearance. This is consistent with the results of previous studies showing significant variability in linezolid pharmacokinetics during the first week of life. On average, linezolid clearance observed in 7‑day‑old infants was lower than in older children. Clearance continued to increase during the first 2‑3 months of life [60]. Consideration of the development of organs and systems in the pediatric patient population is very important for optimal dosing [72].

However, in the study on which the recommended dosing regimen for newborns was based [59], only 9 patients were premature newborns with GA <34 weeks and PNA <7 days. The linezolid dosing recommended in the package insert for newborns with PNA >7 days does not consider gestational age at all. Therefore, it can be said that the recommended dosing regimen does not take into account the peculiarities of linezolid pharmacokinetics and its age‑related changes in newborns depending on GA and PNA in patients older than 7 days. In children of different ages, physiological changes occur that can significantly affect pharmacokinetic processes in the body. This makes it practically impossible to extrapolate dosing regimens estimated for adult patients or for older children to younger children, and especially to newborn patients.

The authors of the study [73] for premature newborns, based on modeling, proposed the following dosing scheme based on patient BSA—a significant covariate for clearance and volume of distribution of linezolid in the study: 6 mg/kg every 8 hours for BSA = 0.11 m², 7 mg/kg every 8 hours for BSA = 0.13 m² and 9 mg/kg every 8 hours for BSA = 0.15 m² at MIC ≤1 mg/L; 7 mg/kg every 8 hours for BSA = 0.11 m², 8 mg/kg every 8 hours for BSA = 0.13 m² and 10 mg/kg every 8 hours for BSA = 0.15 m² at MIC = 2 mg/L.

In the works [73, 74], the estimated glomerular filtration rate (GFR) was considered a significant covariate for linezolid clearance. Linezolid clearance was higher in patients with augmented renal clearance (ARC) (GFR >130 mL/min/1.73m²) compared to patients with normal renal function (90 ≤ GFR ≤ 130 mL/min/1.73m²) in [74], which is similar to conclusions made for the adult population. In [73], it was shown that AST can be considered a covariate for total linezolid clearance, reflecting the role of hepatic metabolism in linezolid elimination. The authors proposed to consider AST level when choosing the linezolid dosing regimen: dose of 10 mg/kg every 12 hours for children under 12 years and 600 mg every 48 hours for children 12‑18 years with AST >200 U/L at MIC = 1 mg/L.

The contradictory results observed in population PK/PD studies of linezolid, including with regard to optimized dosing regimens, are likely related to the high variability of linezolid clearance in pediatric populations [57]. In addition, most of these studies are based on very small patient samples. This may also be the reason that in different studies, different covariates for the main PK parameters of linezolid turn out to be statistically significant and then participate in the development of optimal dosing regimens. The distributions of adverse events during linezolid therapy presented in the studies also differ, which may also be related to the small and heterogeneous patient samples in the conducted studies.

Table 2

Demographic data and estimated PK parameters in studies involving neonates and/or older children

Note: ¹ Cmin — minimum (trough) concentration of the drug in blood plasma.

Target ranges of PK/PD indices for linezolid therapy in the pediatric population

The antibacterial activity of linezolid can best be characterized by two PK/PD parameters: the area under the pharmacokinetic curve over 24 hours at steady state divided by the MIC (AUC0‑24/MIC), usually for optimizing efficacy in adult patients in the range ≥80‑120, and the percentage of time during the day that the drug concentration exceeds the MIC (%T>MIC), usually 85‑100% [75‑78].

Target ranges of PK/PD parameters for efficacy and safety of therapy in the pediatric population and the newborn population have not yet been strictly justified. In the work of Rao et al. [79], it was suggested that AUC0‑24/MIC values in the lower range for adult patients of 80‑100 may be used as a suitable target for linezolid therapy in children, since in children linezolid clearance is on average significantly faster than in adults. In [55], AUC0‑24/MIC >80 and AUC0‑24 <300 were chosen as target ranges for efficacy and safety in premature newborns, respectively. A large number of studies have reported a linear relationship between trough linezolid concentrations and AUC0‑24 [79]. Given the clinical and ethical difficulties of obtaining multiple blood samples for reliable estimation of AUC0‑24/MIC or %T>MIC, measurement of only trough concentration and interpretation taking into account MIC may be a practical indicator of therapeutic efficacy and safety of linezolid therapy. Measurement of trough concentrations is a fairly simple and frequently used method for monitoring linezolid toxicity. Various studies have proposed the following target ranges for trough concentrations: 2‑8 mg/L, 3.6‑8.2 mg/L or 2‑7 mg/L [79]. Experts recommended maintaining trough concentrations in the range of 2‑8 mg/L in both children and adult patients [63]. In the study [73], the following target ranges were used for modeling: 2‑8 mg/L for trough linezolid concentrations and 80‑300 mg*h/L for AUC0‑24.

In the PK/PD study by Tian et al. [71], the AUC0‑24/MIC ≥80 ratio was used as the main PK/PD index for efficacy, and the safety threshold was exceeding the trough concentration Cmin = 7 mg/L.

Target ranges of PK/PD parameters can be used within the framework of population modeling to evaluate the performance of ABP dosing regimens—the probability of target attainment (PTA) in patients of the studied population. Assuming a MIC level, the PTA can be estimated from data on patients receiving a particular dosing regimen or from simulation results (in the latter case, as the proportion of virtual "patients" for whom the target values of PK/PD indices were "attained" according to modeling results). At PTA ≥90%, the ABP dosing regimen can be considered acceptable.

Assessment of the ability of linezolid dosing regimens to achieve recommended therapeutic targets in the pediatric patient population

In most of the reviewed PK/PD studies, patients received the recommended intravenous linezolid dosing regimen for the treatment of bacterial infections: a dose of 10 mg/kg three times daily for children under 12 years and 600 mg twice daily for children 12 years and older. In accordance with the recommendations, in the PK/PD study by Duan et al. [73] in the premature newborn population, patients received the dosing regimen: 10 mg/kg twice daily at GA <34 weeks and PNA <7 days, and 10 mg/kg three times daily at GA ≥34 weeks and at GA <34 weeks with PNA >7 days.

In the work of Li et al., for children from 0 to 12 years, based on the obtained PK parameter estimates, the risk of underdosing was assessed when receiving the linezolid dosing regimen of 10 mg/kg every 8 hours for the treatment of infections with MIC ≥2 mg/L. According to the study, when receiving this dosing regimen, only 81.70% and 35.81% of patients (PTA) achieved the recommended level of PK/PD parameter AUC0‑24/MIC ≥80 against pathogens with MIC = 2 mg/L and 4 mg/L, respectively, indicating the need for dose increase. The authors proposed to increase doses to 15‑20 mg/kg every 8 hours in this patient population at MIC ≥2 mg/L. The conclusion about the risk of suboptimal dosing based on the recommended linezolid dosing regimen is consistent with the results of another PK/PD study of linezolid in children [65]. In this study, a retrospective PK analysis of TDM data was performed, and it was shown that the recommended dosing was suboptimal (Cthrough below the recommended range) in half of patients aged 2 to 11 years, and that higher doses may be required in the presence of bacteria with MIC above 1 mg/L. Increasing the dose to 15 mg/kg every 8 hours may be required for MIC = 2 mg/L. For the recommended dosing regimen, PTA ≥90% for AUC/MIC >100 is achieved only at MIC ≤1 mg/L. None of the studied linezolid dosing regimens in the study guarantee an acceptable PTA at MIC = 4 mg/L.

The authors of [65] conclude that the use of TDM may help optimize linezolid dosing, especially for the treatment of infections caused by pathogens with intermediate susceptibility to linezolid and/or for patients receiving concomitant therapy with drugs such as proton pump inhibitors, dexamethasone, phenobarbital and amiodarone, which significantly affect linezolid plasma concentrations. TDM may also help avoid toxic linezolid concentration levels, which is especially important during long‑term therapy, and thus reduce the frequency of the most well‑known dose‑dependent hematological adverse event—thrombocytopenia.

The authors of [55] demonstrated that linezolid doses up to 12 mg/kg every 8 hours may be required in the premature newborn population to achieve PTA ≥90% at MIC = 2 mg/L.

In the study [70] in the pediatric population of critically ill patients, based on the PK parameter estimates obtained from PK/PD modeling, higher doses of up to 15 mg/kg every 6 h were calculated as optimal for pathogens with MIC = 2 mg/L to achieve the target AUC/MIC >80. No dosing regimen was proposed by the authors for MIC = 4 mg/L.

According to the results of [74], the standard dosing regimen in children does not require adjustment even for MIC = 2 mg/L. For children under 12 years, in the case of pathogens with MIC = 4 mg/L, only the dosing regimen of 600 mg every 12 hours can achieve an acceptable PTA of 90.5%. However, such a dose increase leads to an increased risk of exceeding the safety threshold (up to 48.6%). Therefore, in this case, to optimize linezolid therapy, an attempt should be made to achieve a balance between efficacy and safety. For children aged 12 years and older, in the case of MIC = 4 mg/L, for the standard dosing regimen of 600 mg every 12 hours, PTA = 0.0%. Increasing the dose to 900 mg every 12 hours or 600 mg every 8 hours allows achieving PTA = 83.3% at MIC = 4 with a low probability of exceeding the safety threshold (trough concentration = 7 mg/L).

Thus, it can be concluded that the dosing regimens considered optimal in different studies in the pediatric population differ somewhat due to differences between PK/PD studies, and external validation is required for each such regimen to recommend its use in real clinical practice, taking into account population characteristics.

Conclusion

The reviewed PK/PD studies of linezolid demonstrate that the pharmacokinetics of linezolid in all age groups of the pediatric population is characterized by significant interindividual variability; therefore, the standard dosing regimens developed based on these studies are not optimal for all patients [49, 80‑86]. In the presented population studies of linezolid in children, interindividual variability of PK parameters reached 53% for total clearance and 82% for apparent volume of distribution. Residual unexplained PK variability by the final population model was also considerable, reaching 60%.

Such pronounced variability creates a risk of both underdosing and overdosing when using standard regimens, even adjusted for covariates, in pediatric patients. The dosing regimens of linezolid proposed in different studies themselves differ due to differences between PK/PD studies and extremely limited sample sizes.

All this supports the personalization of linezolid dosing based on TDM data in the pediatric patient population. This is especially important for special patient groups, in particular newborns and premature newborns, as well as patients with impaired renal function, augmented renal clearance, hepatic dysfunction, critically ill patients, and in the case of suspected or estimated pathogen MIC values of 2 mg/L and above.

The use of software for the Bayesian approach and population modeling will allow individualization of therapy based on sparse TDM measurements from real clinical practice, start monitoring concentrations and adjusting therapy without waiting for steady state, as well as calculate PK/PD indices to assess target attainment based on TDM data. It should be noted that in domestic hospitals, TDM has not yet become a routine procedure, due to the difficulties of staff training, the high cost of equipment, and the lack of clear clinical guidelines. Nevertheless, large centers are gradually introducing TDM procedures into practice, which allows more effective and safe prescribing of therapy to patients in special age groups.