Kozlowski KH., Idzik M., Milewska-Bobula B., Dunin-Wasowicz D., Lyson-Wojciechowska G.
The Children's Memorial Health Institute, 04-736 Warsaw, Poland
Introduction: Gancyclovir (GCV) is drug of choice for the treatment of active cytomegalovirus infection in newborns and infants. There are limited research studies concerning GCV dosage strategy and following drug-concentration vs. clinical effects in these special population patients [5]. We confirmed previous clinical studies [4] indicating that GCV in newborns and infants aproximates 1-compartment model kinetics (in adults 2-compartment model). GCV is minimally bound to plasma protein and glomerular filration and tubular secretion are the main elimination pathways. Because of possible cytomegalovirus resistance with inadequate GVC levels and bone-marrow hypertoxicity with silent symptomps after GCV overdosing, population pharmacokinetic service is necessary.
Purpose of the study was to evaluate population pharmacokinetic model able to characterize GCV distribution and elimination parameters as function of anthropometric and physiologic variable as well as individualization dosage strategy in infected newborns and infants.
Methods: Clinical material consisted of 86 patients was qualified into GCV treatment, in which uniformed diagnostic tests was performed including anthropometric mesurements and specific laboratory serological and analytical tests, as well as Gancyclovir determination in 300 samples of 100 ul serum samples using validated HPLC-SF method. Each patient initially was obtained test dose of 5 or 7.5 mg/kg GCV as 1 h i.v. infusion and blood samples were drawn after 0 and 4.5 h after the end of infusion. On the next occasions doses and dosing intervals (6,8, 12 or 24 h) were adjusted using pharmacokinetic approach and optimal target therapeutic levels as Cmin=0.5-1.0 ug/ml and Cmax<8.0 ug/ml. Population pharmacokinetic computations were done using NONMEM program [1] with FOCE INTERACTION option and NMWIN front-end [4]. Final model considered elimination rate constant KE (1/h) as nonlinear function of creatinine clearance CLcr (ml/min.) and volume of distribution VD (L}as nonlinear function of body lenght (cm) and body weight (kg). CLcr values were estimated anthropometrically [2]. Proportional error model for KE, exponential error model for VD and power error model for residuals was selected.
Results are presented in the Table
TABLE: GANCICLOVIR – NONMEM RESULTS – FINAL MODEL, 300 MEASURED POINTS; 134 OCCASIONS; 86 NEWBORNS & INFANTS
FINAL MODEL ESTIMATES – FIXED EFFECTS
| ESTIMATE ±CV% | COVARIATE MODEL | CONDITION |
| TH1=0.4477± 1.3% | TVKE=TH1 | IF CLCR>TH2 |
| TH2=5.1196± 5.0% | TVKE=TH1*EXP(-LN(TH2/CLCR)) | in other cases<TH2< td> |
| TH3=0.0733± 1.5% | TVVD=EXP(TH3*(BW*BL)**0.5) | UNITS |
| TH4=0.6399± 3.7% | CLCR [mL/min] |
FINAL MODEL ESTIMATES – RANDOM EFFECTS
| VARIANCE ±CV% | ERROR MODEL | UNITS | |
| VAR-KE=0.0209±21.0% | KE=TVKE*(1+ETA(1)) | KE [1/h] | DOSE [mg/kg] |
| VAR-VD=0.0548±12.8% | VD=TVVD*EXP(ETA(2)) | VD [L] | BW (kg) |
| VAR-CP=0.0294±15.0% | Y=F+F**TH4*ERR(1) | CP [ug/mL] | BL [cm] |
Conclusions: 1. NONMEM model with clinical covariates was succesfully elaborated for Gancyclovir pharmacokinetics in newborns and infants. 2. Population pharmacokinetic approach is necessary for individualization GCV regimens beacause of nonlinear relation of drug pharmacokinetics and patients clinical state (renal function, body size) as well as ethical considerations (limited sampling strategy).
References: (1) Beal S.L., Sheiner L.B.: NONMEM User’s Guide. Part V. University of California at San Francisco, CA, USA, 1992. (2) Brion L.P. et al.: J. Ped. 109, 698-707, 1986. (3) Trang J.M. et al. : Clin Pharmacol. Ther. 53, 18-21, 1993. (4) Vielhaber J.P., Barrett J.S., Pharm. Res. 11, 709, 1989. (5) Zhou X.-J. et al.: Antimicrob. Agents Chemother. 40, 2202-2205, 1996.
Reference: PAGE 10 () Abstr 220 [www.page-meeting.org/?abstract=220]
Poster: poster