Pharmacokinetics of once-daily arnikacin in pediatric patients

ORIGINAL ARTICLE Pharmacokinetics of once-daily arnikacin in pediatric patients Laurence Belfayol', Philippe Talon2, atthieu Eveillardl, Patrice Alet' and rancise auvelle' 'Laboratoire de Pharmacie Clinique and 2Service de Pkdiatrie, ontfermeil, rance Objective: To study the pharmacokinetic parameters of a once-daily regimen of amikacin ( mg/kg) in association with other antimicrobial agents in 3 children with severe Gram-negative infections. ethods: A Bayesian approach was developed to optimize the amikacin regimen. The predictive performance was assessed by computing bias and precision. Each patient was evaluated for toxicity after days of treatment. Results: Peak amikacin concentrations on days 2 and of therapy averaged 3.3 *.0 mg/l and 32.4 * 7.4 mg/l, respectively. To achieve peak serum concentrations between 30 and 40 mg/l, individualized dosage was necessary in of 3 children. The pharmacokinetic parameters showed large interindividual variations, with a mean half-life of 2 h and a mean volume of distribution of 0.36 Ukg. No nephrotoxicity was observed in any of the children. After individualization of dosage on the basis of one measurement of peak concentration, no significant differences were observed between predicted and subsequently measured amikacin concentrations. Conclusions: Once-daily dosage of arnikacin ( mg/kg) is well tolerated in pediatric patients; however, a loading dose of mg/kg is recommended to achieve a therapeutic peak value between 30 and 40 mg/l. Initial serum monitoring is essential in a population such as children, with wide interpatient variability. Using the Bayesian approach, the amikacin regimen in children can then be predicted with minimal bias and good precision. Keywords: Amikacin, pediatric patients, pharmacokinetics, drug monitoring INTRODUCTION Aminoglycoside antibiotics are very important, in association with other antimicrobial agents, in the treatment of Gram-negative infections. A daily regimen administered in a single dose has been effective in several studies. Once-daily regimens of amikacin [l], netilmicin [2] or gentamicin [3] are less toxic than, and equally efficient as, more frequent administration. The volume of distribution is inversely correlated with age [4] and is greater in children than in adults []. This interpatient variation in volume of distribution has a direct effect on the therapeutic peak levels achieved [6]. Because of this wide variation in pharmacokinetic parameters and the low therapeutic index, aminoglycosides are drugs that require an individualized dosage. A Bayesian forecasting program has been described to adjust the aminoglycoside dosage [7-lo]. Corresponding author and reprint requests: L. Belfayol, CHI ontfermeil, Service Pharmacie, rue du General Leclerc, 3370 ontfermeil, rance Tel: (33) 4 70 82 26 ax: (33) 4 70 82 82 Accepted 22 April 6 The Bayesian method is a good approach in pehatric patients, because few serum samples per patient are required to estimate the pharmacokinetic parameters for optimizing dosage regimens. The pharmacokinetics of amikacin have been documented in critically [ll] and neutropenic children [2]. While the serum concentrations and the clinical efficacy of amikacin have been studied in children with severe Gramnegative infections [3,4], no information has been reported on amikacin pharmacokinetics in these patients. Therefore, the aim of this study was to investigate the pharmacokinetics (in terms of dosage, concentrations achieved and pharmacokmetic parameters) and to evaluate the toxicity of a once-daily amikacin regimen ( mg/kg per day) in pediatric patients, as well as the Bayesian forecasting method for predicting serum amikacin concentrations. ATERIALS AND ETHODS Patient selection Thirty-five children, ranging in age from to years, who were receiving amikacin for serious infection or 86

Belfayol et al: Pharmacokinetics of amikacin in pediatric patients 87 presumed infection, were included in the study. Severe infections were defined as follows: pyelonephritis (n= 2), pulmonary infection (n=4), meningitis (n=), febrile syndrome in AIDS (n = 2), bacteraemia (n = l), arthritis (n = 3), other locahzed infections (n = 8). Twenty-two bacterial isolates were identified as responsible: Escherichia coli(n = lo), Aeudomonus aeruginosa (n =2), Proteus mirabilis (n = l), Staphylococcus spp. (n = 4), Neisseria meningitidis (n = 3), Haemophibs inzuenzae (n = 2). Drug administration and dosage Each patient received mg/kg amikacin therapy at doses ranging &om 0 to 70 mg as a 30-min intravenous infusion in 0.% sodium chloride. Amikacin was given to all patients in once-daily doses combined with one or more other antimicrobial agents. The other drugs co-administered were ceftriaxone (n = 22), cefotaxime (n = 7), ciprofloxacin (n = l), amoxycillin (n = 3), oxacillin (n = l), teicoplanin (n = ) and fosfomycin (n = 3). or inclusion, all patients had to receive amikacin for a minimum of days. Venous blood samples were collected from an indwelling catheter in the non-infused forearm just before infusion (trough) and 30 min after the end of infusion (peak). Blood samples for follow-up peak and trough amikacin concentrations were obtained at days 2 and. On day 2, the amikacin dosage was adjusted to achieve a peak level of 0 mg/l and a trough level < mg/l. On day, peak and trough samples were collected, to ensure that the actual levels in the serum were in fact close to the calculated levels. Two pairs of peak and trough values were required for each patient for inclusion in the study. Serum concentrations of amikacin were determined by a fluorescence polarization immunoassay (TDx, Abbott Diagnostic Division, Rungis, rance). The sensitivity of the assay was 0.8 mg/l. Replicate measurements (n = ) of representative samples with concentrations in the region of,, and 3 mg/l were made in order to determine the mean and the standard deviation of each sample. The error pattern of the TDx assay for amikacin is shown in igure. The resulting polynomial equation found for this assay is: SD (mg/l) = 0.00078C2-0.0334C + 0.0700 where SD represents the standard deviation of the measurement and C represents the measured serum concentration. 0 0 2 0 2 3 0 3 Serum amikacin (rngh) igure Assay error pattern of the TDx assay for amikacin. SD = standard deviation (mg/l). Bayesian forecasting of dosage The pharmacokinetics of amikacin were described by a one-compartment open model. Individual pharmacokinetic parameters were calculated according to the maximum Bayesian analysis a posteriori (Abbott PKS, Abbott Diagnostic Division, Rungis, rance). The data required included age, gender, height, weight, serum creatinine, amikacin dose, dosage interval and measured serum concentrations. The initial population parameters in this model were as follows: volume of distribution (Vd) = 0.3 L/kg; non-renal clearance (CINR) = 0.047 ml/min per kg; slope of the line between amikacin clearance (C~T) and creatinine clearance (CIcreat) = 0.8. Total clearance is calculated by the following equation: C~T = C~NR + (slopex Clcre,,). The coefficients of variation were set at 30% for Vd at 2% for C~NR and at 40% for the slope. Creatinine clearance was estimated by the method of Schwartz et al. []. Predictive analysis The predictive performance was evaluated by the methods described by Sheiner and Beal [6]. The predicted amikacin serum concentrations were compared with those measured. Bias and precision were assessed, respectively, by mean prediction error (E) with % confidence interval and root mean squared error (RSE) with the corresponding equations: Evaluation of nephrotoxicity The criterion of toxicity was an increase in serum creatinine of % or more between days 2 and of the amikacin therapy.

88 Clinical icrobiology and Infection. Volume 2 Number 3, December 6 where C,,, and C,,l,,, represent the predicted and measured concentrations. The bias was considered significantly different if the % confidence interval dd not include zero. A paired t-test between predicted and measured concentrations was used to evaluate their differences. A p-value < 0.0 was considered statistically significant. RESULTS The demographic parameters of indvidual children, and the amikacin pharmacohnetic parameters are listed in Table. The half-life of amikacin ranged from. to 4.4 h. The distribution volume of amikacin varied widely from patient to patient. The mean peak and trough amikacin serum concentrations on days 2 and are shown in Table 2. The measured serum concentrations showed large interindividual variations. On day 2, 0% of the chddren had trough levels < mg/l, 4% of which were below the limit of detection of the assay. On day, two children had a trough level > mg/l (6. and 8. mg/l). To achieve therapeutic concentrations between 30 and 40 mg/l, individual dosage was necessary in of 3 children. The dose required was higher than the recommended dose of mg/kg, with a mean of 2 mg/kg. easured and predicted serum amikacin concentrations at both peak (n = 3) and trough (n = 3) were compared in all patients &om whom samples were available on day. The bias and precision of Bayesian predictive performance are shown Table Characteristics of the chddren and pharmacokinetic data 2 3 4 6 7 8 2 3 4 6 7 8 2 22 24 2 26 27 28 2 30 3 32 33 34 3. 2. 8 2 4 4 4. 3 3 4.8 7 4 8 2 s 6 4. 2 3 33 3. 26 2.2 4 22 6.8 60 38. 7 8 3 8. 4 6 34 40 8 4 40 34 4 26 37 8 73 30 2 27 76 88 2 4 3 4 30 73 40 3 72. 40 0 40 8 0 3 0 60 40 0.24 0.32 0.24 0.3 0.2 0.44 0.42 0.3 0.48 0.30 0.22 0.4 0.3 0. 0.26 0.37 0.3 0.48 0.4 0.33 0.38 0.42 0.46 0.33 0.32 0.26 0.42 0.30 0.43 0.34 0.43 0.34 0.3 0.27 0.44..8.3. 2.2 2.2.6. 3. 0..7..6. 2..3 2.8.3 2.8.3 4..8 4.4.8.. 3...3.4.6.4 3..2 4.0 8.3 62.7 27.3 30.3 7.2 47.3 8. 8.7 3.3 68.3 36. 48.7 4.2 7.2 4.2 64.3 46.8 34.3 2. 37. 48.0.2 3.8 38.0 4.3 28. 0.0 48.8 0. 43.2 44.3 3.3 30. 2. ean ISD 7.0 k4.3. f2.4 4 I2 0.36 ko.08 2.0 kl.o 6.3 k33.8 T,2 = half-life; KI = volume of distribution; C/T = total clearance

Belfayol et al: Pharmacokinetics of amikacin in pediatric patients 8 Table 2 ean SD trough and peak amikacin concentrations (mg/l) Day 2 Day Trough (n = 3).? 0..4 I.6 (0.8-2.4) (0.8-8.) Peak (n = 3) 3.3.0 32.4 7.4 (7.-4.6) (7.3-0.0) in Table 3. or both peak and trough levels, the bias is not significant ( p > 0.0). Among the mean measured serum concentrations, the precision is greater for the peak (+7.06 mg/l for 32.4 2 7.4 mg/l) than for the trough (+.70 mg/l for.4 &.6 mg/l). No significant differences were observed between measured and predicted trough (respectively,.4 4.6 mg/l and.2 4 0.8 mg/l) and peak (respectively, 32.4 -t 7.4 mg/l and 33.3 &.4 mg/l). These data indicate that serum amikacin concentrations in children can be predicted with minimal bias and reliable precision. The mean length of therapy was 6 days. Nephrotoxicity could be evaluated in all children. No statistically significant difference was found in serum creatinine between days 2 and, with, respectively, 6.2 4 30.2 pmol/l and 4.3 -t 7.2 pmol/l. No nephrotoxicity was observed in any of the patients. DISCUSSION The once-daily dosage regimen resulted in a higher peak and a lower trough level. Of the patients 46% and 4% had peaks of > 30 mg/l and > 40 mg/l, respectively, and 0% had a trough < 2. mg/l on day 2. No significant difference was noted between the mean peak and trough amikacin serum concentrations observed on days 2 and. There was no tendency for amikacin accumulation during treatment. The mean and individual peak amikacin levels in our study were lower than those previously reported in adult patients with severe infections [7,8] or in healthy young Table 3 Predictive performance of serum amikacin concentrations Ed RSEh Trough (n = 3) -0.3 +.70 Peak (n = 3) (-0.88; +0.3) +.37 +7.06 Peak + Trough (n = 70) (-0.; +3.73) +0.33 +. (-2.02; +2.68) "ean prediction error (mg/l). hroot mean squared error (mg/l). adult volunteers ( to 3 years) [] receiving a similar dose of amikacin. These results could be explained by a greater volume of distribution and a more rapid total clearance in pediatric patients. These results, however, are comparable with those observed in critically ill patients over the age of year (33.7 k 4.8 mg/l [ll] and in pediatric patients receiving mg/kg who were suffering from severe Gramnegative infections (36. 4. mg/l [3]). However, the once-daily dosage allows the maintenance of higher aminoglycoside peak levels than conventional regimens. oore et al. [] demonstrated in patients with severe Gram-negative infections that a high peak concentration relative to the IC is a major determinant of the clinical response to aminoglycoside therapy. High peak levels are effective, because the bactericidal activity of aminoglycosides is concentration dependent, but the ideal therapeutic peak has not been defined. Beaucaire et al. [2] reported in critically ill adults a significantly higher mortality rate from infection when the first amikacin peak was below 40 mg/l. To achieve a peak value of 40 mg/l on the first day of treatment, a loading dose is required, especially in pediatric patients with large distribution volume. This loading dose could correspond to 2% of the daily dose [22]. Indeed, in neutropenic children [2,] and critically ill children [ll], a dose of mg/kg per day has been proposed. We noted large interindividual variations in the concentrations and pharmacokinetic parameters of amikacin in pediatric patients. Therapeutic drug monitoring early in treatment is thus necessary to optimize dosage regimens. Predictive performance of the Bayesian program can be enhanced when initial population parameters reflect the patient population being monitored. In clinical routine, monitoring of aminoglycosides has been based on both the peak and trough levels. The frequency of these measurements, however, could be reduced to a single sampling. Although trough concentrations are measured to avoid aminoglycoside toxicity, it may be difficult to adjust dosage on the basis of these concentrations, which are currently below the limit of detection of the assay. In neutropenic children undergoing bone marrow transplantation receiving mg/kg per day of amikacin, serum concentrations declined rapidly and became undetectable within 6 h [24]. Therefore, the serum peak level seems the most appropriate sample for monitoring. Sampling at other time intervals has been proposed. Blaser et al. [2] reported a significant correlation between nephrotoxicity and 8-h serum netilmicin concentrations. In comparison with serum gentamicin concentrations h and 8 h after the initial administration, Bayesian analysis using the 4-h

0 Clinical icrobiology and Infection, Volume 2 Number 3, December 6 concentration was the least biased and the most precise for predicting peak and trough concentrations [26]. However, these stuhes found no relationship between the serum concentrations and therapeutic efficacy. Our results regarding toxicity should be interpreted cautiously. Nephrotoxicity was defined as a rise of 2 % in serum creatinine, and no difference was noted between days 2 and. The higher trough concentrations observed on day (6. and 8. mg/l) were not accompanied by an increase in serum creatinine, and probably reflect wide intrapatient variability in amikacin handling. The fiequency of nephro- and ototoxicity seems to be lower in children than in adults [24]. Other investigations to evaluate ototoxicity must be performed. In conclusion, once-daily dosage of amikacin ( mg/kg per day), combined with another antimicrobial agent, is well tolerated in pediatric patients. Serum concentration monitoring early in treatment is essential in children who show wide interpatient variability in pharmacokinetic parameters. oreover, as it seems that the initial peak value is an important factor in clinical outcome, a loading dose of mg/kg per day is recommended, in order to obtain satisfactory concentrations in most children. Drug monitoring will then indicate whether or not dosage has to be reduced. This approach provides security in achieving the best results in children. Bayesian analysis is suitable for individualization of amikacin dosage for pediatric patients. Indeed, a single non-steady-state concentration is all that is required to adjust dosage early in treatment, preferably in the first 24 h. Acknowledgments This study was supported by the Bristol-yers-Squibb Company, Paris, rance. References aller R, Ahrne H, Eilard T, Eriksson I, Lausen I, the Scandinavian amikacin once ddy study group. Efficacy and safety of amikacin in systemic infections when given as a single daily dose or in two divided doses. J Antimicrob Chemother ; 27(suppl. C): 2-8. Ter Braak EW, de Vries PJ, Bouter Kp, et al. Once-daily dosing regimen for aminoglycoside plus R-lactam combination therapy of serious bacterial infections: comparative trial with netilmicin plus ceftriaxone. Am J ed 0; 8: 8-66. Prins J, Bdler HR, Kuijper EJ, Tange RA, Speelman l? Once versus thrice ddy gentamicin in patients with serious infections. Lancet 3; 34: 33-. Siege JD, ccracken GH, Thomas L, Threkeld N. Pharmacokinetic properties of netilmicin in newborn infants. Antimicrob Agents Chemother 7; : 246-3.. Schevuk Y, Taylor D. Aminoglycosides, volume of distribution in pematrics patients. Ann Pharmacother 0; 24: 273-6. 6. Zaske DE, Cipolle RJ, Rotschafer JC, Kohls PR, Strate RG. Individualizing amikacin regimens: accurate method to achieve therapeutic concentrations. Therapeutic Drug onitoring ; 3: 02-6. 7. Carlsted BC, Uaamnuichi, Day U3, Bowman L, Brater DC. Ammoglycoside dosing in pediatric patients. Ther Drug onit 8; : 38-43. 8. Garraffo R, Iliads A, Can0 J!?, Dellamonica P, Lapalus l? Application of Bayesian estimation for the prehction of an appropriate dosage regimen of amikacin. J Pharmacol Sci 8; 78: 73-7.. Rodvold KA, Pryka RD, Kuehl PG, Blum RA, Donahue l? Bayesian forecasting of serum gentamicin concentrations in intensive care patients. Clin Pharmacokinet 0; 8: 40-8.. Kosirog JL, Rospond R, Destache C, Hall P. Aminoglycoside forecasting in neutropenic patients with cancer. Clin Pharmacokinet 3; 24; 7-87.. arik PE, Havlik I, onteagudo SE, Lipman J. The pharmacokinetics of amikacin in critically ill adult and paediatric patients: comparison of once- versus twice-dady dosing regimens. J Antimicrob Chemother ; 27(suppl. C): 73-. 2. Bouffet E, uhrmann C, rappaz D, et al. Etude clinique, bactkriologique et pharmacocinktique de l amikacine en dose unique journalkre en oncologie pkhatrique. kd al Infect 3; : -. 3. Kafetzis DA, Sianidou L, Vlachos E, et al. Clinical and pharmacokinetic study of a single daily dose of amikacin in paediatric patients with severe Gram-negative infections. J Antimicrob Chemother ; 27(suppl. C): -2. 4. Trujdlo H, Robledo J, Robledo C, et al. Single daily dose amikacin in paedlatric patients with severe Gram-negative infections. J Antimicrob Chemother ; 27(suppl. C): 4-7.. Schwartz GJ, Haycode GR, Edelman G, Spitzer A. A simple estimate of glomerular filtration rate in chddren from body length and plasma creatinine. Pediatrics 76; 8: 2-63. 6. Sheiner LB, Beal SL. Some suggestions for measuring predictive performance. J Pharmacokinet Biopharmacol 8; : 03-2. 7. Tulkens P. Pharmacokmetic and toxicological evaluation of a once-daily regimen versus conventional schedules of netilmicin and amikacin. J Antimicrob Chemother ; 27(suppl. C): 4-6. 8. Giamarellou H, Yiallouros K, Petrikkos G, et al. Comparative kmetics and efficacy of adacin administered once or twice daily in the treatment of systemic Gramnegative infections. J Antimicrob Chemother ; 27 (suppl. C): 73-.. Garraffo R, Drugeon HB, Dellamonica P, Bernard E, Lapalus P. Determination of optimal dosage regimen for amikacin in healthy volunteers by study ofpharmacokinetics and bactericidal activity. Antimicrob Agents Chemother 0; 34: 64-2.

Belfayol et at: Pharmacokinetics of amikacin in pediatric patients. oore RD, Lietman PS, Smith CR. Clinical response to aminoglycoside therapy: importance of the ratio of peak concentration to minimal inhibitory concentration. J Infect Dis 87; : 3-. 2. Beaucaire G, Leroy 0, Beuscart C, et al. Clinical and bacteriological efficacy, and practical aspects of amikacin given once daily for severe infections. J Antimicrob Chemother, 27 (Suppl. C): 2-8. 22. Petitjean 0, Prevot, Lortholary 0, Tod, Nicolas P Pharmacocinitique comparie des aminosides uulists en dose unique journalkre. id al Infect 3; (Special): 22-36.. Schoepfer C, De Champs C, Pouille 0, Demeocq E Sirot J. Empirical treatment of fever in neutropenic cancer patients with amikacin single daily dose and aztreonam [abstract 6/P]. In: Program and abstracts of the 3e RPunion Interdisciplinaire de ChimiothPrapie Anti-Infectieuse. Paris, 3: 36. 24. Viscoli C, Dudley, errea G, et al. Seruni concentrations and safety of single-ddy dosing of amikacin in children undergoing bone marrow transplantation. J Antimicrob Chemother ; 27(suppl. C): 3-. 2. Blaser J, Konig C, Simmen HP, Thurnheer U. onitoring serum concentrations for once-daily netilmicin dosing regimens. J Antimicrob Chemother 4; 33: 34-8. 26. Chrystyn H. Validation of the use of Bayesian analysis in the optimization ofgentamicin therapy &om the commencement of dosing. Drug Intell Clin Pharmacol 88; 22: 4-3.