Trough levels of teicoplanin in pediatric inpatients 18 years old or younger at Samsung Medical Center, Sungkyunkwan University School of Medicine between September 2014 and April 2018 were collected. Medical records were retrospectively reviewed to collect clinical information. Distribution and characteristics of trough levels in patients with decreased renal function and those with bacteremia were also assessed. Patients who were included in this study were 18 years old or younger who received teicoplanin for at least 4 doses and had available exam data of trough levels. Process of data exclusion is described in Supplementary Fig. 1.

The authors defined decreased renal function as estimated glomerular filtration rate (eGFR) less than 70 mL/min/body surface area (BSA),8 which was calculated with modified Schwartz equation (K × height in centimeter/plasma creatinine [mg/dL], K = 0.413).

Within bacteremia population, patients were regarded to be at defervescence if body temperature lower than 37.5°C was achieved for at least 48 hours.

Teicoplanin was administered intravenously. Study population received teicoplanin at a loading dose of 10 mg/kg/dose at 12 hours' interval, three times in a row. Maintenance dose was commenced at a dose of 10 mg/kg/dose 24 hours after the last loading dose. The maximal dose of administered teicoplanin was 400 mg. Trough levels of teicoplanin were measured with liquid chromatography-tandem mass spectrometry method at the Department of Laboratory Medicine and Genetics of the Samsung Medical Center. Detailed information on standard materials used, preparations of calibrators, quality controls, instrumental condition, and method validation were described in a previous study.9

Steady state sample was defined as the one collected after the fourth or fifth dose of teicoplanin. Drug levels were measured within seven days from the initial dose after reaching steady state in this study. Each of the episodes were considered to be separate if a duration between the last dose of previous teicoplanin and the first dose of restart is 5 days or more.

Descriptive data are expressed as median and ranges. The χ² test and Mann-Whitney U test were used to compare categorical and continuous variables, respectively. Differences in the achievement of target trough level (≥ 10 μg/mL) were evaluated for statistical significance with logistic regression models. Patient age, sex, body weight, serum creatinine, eGFR, interval between initial dose and 1st steady state exam, loading and maintenance doses were included as covariates. Analyses were done with IBM SPSS Statistics, version 25 (SPSS Inc., Chicago, IL, USA). Statistical significance was defined by a two-sided P value < 0.05.

This study was approved by the Institutional Review Board (IRB) of Samsung Medical Center (IRB No. 2018-08-045). Informed consent was waived because this study was retrospective, and measurement of serum teicoplanin drug level was a routine process for patient care.

A total of 187 trough levels were collected from 143 patients (Table 1), including 74 (51.7%) males. The median age of all patients was 6.2 years (range, 0.09–17.7 years). Hematologic and oncologic diseases were the most common underlying diseases (n = 119, 83.2%), followed by gastroenterologic and hepatologic disease (n = 6, 4.2%), primary immunodeficiencies (n = 5, 3.5%), neurologic disease, and cardiologic diseases (n = 3, 2.1% respectively). The most common reason of teicoplanin use was empiric treatment for prolonged neutropenic fever in hematologic and oncologic disease (n = 143, 75.5%).

Among 187 measured trough levels, 180 were first-measured-values. The median value of 1st measured trough levels was 16.2 μg/mL (range, 2.3–100 μg/mL). Interval between initial teicoplanin injection and 1st trough level ranged from 47.6 hours to 179.3 hours and median was 96.5 hours. Data were divided into two groups based on trough levels: trough level < 10 μg/mL (n = 44, 24.4%) and trough level ≥ 10 μg/mL (n = 136, 75.6%). There were 77 data (42.8%) of which trough levels were between ≥ 10 μg/mL to < 20 μg/mL and 55 data (30.6%) for ≥ 20 μg/mL to < 60 μg/mL. Four trough levels (2.2%) were over 60 μg/mL which is generally considered as toxic level. Characteristics of these patients are described in Supplementary Table 1. Patients whose steady state trough level was less than 10 μg/mL tended to be younger, have lower body weight, have longer interval between the 1st injection and trough level measurement, and have better kidney function (Table 2). Patients were also divided into two groups based on age: patients < 6 years old (infants and young children) vs. patients ≥ 6 years old (school-aged children and adolescents). Although trough levels in both groups were within optimal therapeutic range, lower steady state levels were observed in the younger age group (median: 13.5 vs. 18.0 μg/ml, P = 0.038). In a logistic regression model, patients whose trough levels were within target range tended to have higher body weight, shorter interval between initial injection and trough level measurement, lower eGFR, higher maintenance dose (Supplementary Table 2).

Eight patients had decreased renal function (median eGFR: 62.6 mL/min/BSA, range: 43.8–66.5 mL/min/BSA). Of these patients, four received renal replacement therapy including one patient who had received both peritoneal dialysis and continuous renal replacement therapy (Table 3).

Median trough levels at steady-state were significantly higher in this group with decreased renal function than in patients with normal renal function (median trough level: 49.1 vs. 15.6 μg/mL, P < 0.001) (Fig. 1). All trough levels at steady-state in these patients were over 20 μg/mL. One patient showed trough level over 60 μg/mL. Compared to patients with normal renal function, patients with decreased renal function showed no statistical differences in loading dose, maintenance dose, or interval between the initial dose and measurement of trough level.

Eight patients had gram-positive bacteremia in the study population (Table 4). All patients with gram-positive bacteremia had normal renal function (median eGFR: 147.7 mL/min/BSA, range, 98.0–209.0 mL/min/BSA). Median trough level at steady-state was 11.2 μg/mL (range, 8.2–19.3 μg/mL). Three patients failed to reach optimal trough level even at steady-state after five to seven doses of teicoplanin. Staphylococcus aureus was the most common pathogen (n = 4, 50%), followed by Viridans streptococci (n = 2, 25%) and Enterococcus gallinarum (n = 1) or coagulase negative staphylococci (n = 1). These eight isolates were sensitive to teicoplanin (n = 6) or vancomycin (n = 7).

In seven patients (except patient 1), blood culture became negative within three days from the initial blood culture. All the patients reached microbiological cure within 14 days. Defervescence was achieved within a median of 4.5 days (range, 2–13 days). In three patients who did not reach target trough level (patient number: 1, 2, and 3), bacteremia resolved by 14 days, although patient 1 had persistent bacteremia for 3 days.

Four patients had changes in their initial antibiotic agents during treatment. Two patients had change of teicoplanin to vancomycin. One patient (patient 6) had this change of antibiotics because he developed S. aureus bacteremia while he was receiving teicoplanin as an empirical therapy for neutropenic fever. One patient (number 5) started teicoplanin when she developed S. aureus bacteremia. She received teicoplanin for 5 days. After that, she had vancomycin for continuation. Two patients (patient number: 2 and 7) had change of vancomycin to teicoplanin due to adverse reaction to vancomycin (drug rash and drug fever, respectively) and achieved cure for the bacteremia.