Kazimierz H. Kozlowski
Laboratory of Pharmacokinetics, The Childrens Memorial Health Institute, 04-736 Warsaw, Poland.
Introduction: Infusable preparations of human immunoglobulin-G (IgG) have been considered for administration to premature very low birth weight (VLBW) infants in the prevention of infectious diseases. Knowledge about IgG pharmacokinetics and dosage optimisation are scanty in VLBW infants, children and adults. Some difficulties in the handling of IgG pharmacokinetic data from VLBW infants have been discussed (1). IgG is an endogenous drug and immature newborn infants are not able to produce any of the antibodies in the first period of their lives. The drug has a distribution phase followed by a very long elimination half-life of 3-4 weeks. However, fitting A*EXP(-alpha*t) + B*EXP(-beta*t) or even A*EXP(-alpha*t) + B*EXP(-beta*t)+BASE to concentration-time data has not been considered adequate (1). In the present study IgG was considered as an endogenous drug and the predose level CPZERO was included as a pharmacokinetic parameter. The data were analysed using the NONMEM program (2) with NMWIN (3) installed on a PC-Pentium 200MMX. Concentration-time data taken from the literature (3,4) were used to verify the model.
Methods: ADVAN6 parameterised for CLE, CLD, BETA, V1 and CPZERO, was applied. Doses of 0.50 or 0.75 g/kg BW were modelled using the $DES procedure as a four-step increasing rate constant during the period 0-125.5 min (low dose) or 0-175.5 min (high dose). The variability of the first 4 pharmacokinetic parameters was described by 4 ETA variables according to a multiplicative model. However, no interindividual variability was allowed on the CPZERO pharmacokinetic parameter. A multiplicative model was assumed for the output function Y = F + CPZERO. Measured IgG level data (in mg/kg) sampled before and from 15 min until 28 days following the infusion were taken from 27 VLBW newborns (194 cp-data records).
Results: A successful fit could not be obtained when the predose CPCZERO parameter was considered as a population variable because of its high inter-subject variability. However, when all CPZEROs were considered as intraindividual subject-matched parameters, successful fitting was obtained with 13 subjects of the low dose subgroup. All 27 CPZEROs were then fixed to their respective CP(0) measured values for the full set of data from 27 subjects. The mean population pharmacokinetic parameters THETA() ± SE as well as parameters controlling the populational variances ETA() ± SE are presented in the table:
Table 1. Final parameter estimates for fixed and random effects:
| Parameter | Mean ± SE | ETA Mean ± SE |
| CLE (ml/d/kg) | 4.25 ± 0.25 | 0.0902 ± 0.0256 |
| CLD (ml/d/kg) | 82.6 ± 13.3 | 0.582 ± 0.559 |
| BETA (1/d) | 0.0315 ± 0.0014 | 0.0455 ± 0.0143 |
| VI (ml/kg) | 56.7 ± 2.8 | 0.0397 ± 0.0158 |
| ERROR (1) | 0.0039 ± 0.0008 |
The final estimates of the mean population parameters appeared to be realistic, especially BETA which corresponded to a tl/2 of 20 days for IgG, and the initial distribution volume of 5.7% of BW, which corresponds well with the volume of blood plasma in newborns.
References:
(1). Nolan B.N. et al.: J. Ped. 112,325,1988.
(2) Beal S.L., Sheiner L.B.: NONMEM User’s Guide. Part V. University of California at San Francisco, CA,USA,1992.
(3) Vielhaber J.P., Barrett J.S.: Pharm. Res. 11,709,1994.
(4). Noya F.J.D. et al.: J. Ped. 112,278,1988.
(5). Kohl S. et al.: J.Ped. 115,135,1989.
Reference: PAGE 7 (1998) Abstr 675 [www.page-meeting.org/?abstract=675]
Poster: oral presentation