C Chauvet (1), S Rochegude(1), N Bleyzac(1,2), P Maire(2), G Putet(3), G Aulagner(1).
(1) Pharmacy department, Debrousse Hospital, Lyon, France; (2) ADCAPT, A.Charial Hospital, Francheville, France; (3) Neonatal intensive care, Debrousse Hospital, Lyon, France
Vancomycin dosage regimens were routinely adapted in new borns by the pharmacist staff of Debrousse hospital with USCPACK PC clinical programs. A two-compartment model was used. Creatinin clearance (CLCr) was estimated using a corrected Schwartz equation:
CLCr (ml/min/1.73m2)=k * height(cm)/ creatinin (mg/dL) where k=0.35
Actually in neonatology, Postconceptional Age (PCA) is the reference covariate of vancomycin elimination. The purpose of this study was to determine which covariate (CLCr or PCA) allowed the best predictions of vancomycin serum levels in neonates.
Methods: Patients criteria selection were : postnatal age less than one month, over two vancomycin serum levels per patient. Serum sample were analyzed using an enzyme immunoassay (EMIT on COBAS Mira Roche). Vancomycin pharmacokinetic parameter values were estimated using a non parametric method (NPEM2 program). Bias and precision of vancomycin serum level predictions obtained with mean or median of pharmacokinetics parameters were tested using PCA or CLCr (Student test).
Results: Thirty-three infants were included in this retrospective study. PCA and CLCr were (mean ± SD) 33.9 ± 5.2 weeks, 19.6 ± 11.3 ml/min/1.73 m2. Ninety six vancomycin serum levels were analyzed. The results for bias and precision are listed in the following table :
| Bias | Precision | |||||
| APC | CLCr | APC | CLCr | |||
| Mean | -1.60 | -1.08 | p<0.001 | 16.88 | 12.10 | NS |
| Median | -0.53 | -0.66 | NS | 15.75 | 12.86 | NS |
| p<0.001 | p<0.05 | NS | NS |
NS: non significant
Precision values obtained using one or the other covariate were not significantly different. But, Bias was significantly lower when CLCr was used as a covariate. Moreover, using median as a summary of pharmacokinetic parameters was always the best to minimize the bias.
Conclusion: If we use median pharmacokinetic parameters values for bayesien forcasting of vancomycin serum levels in neonates, CLCr and PCA are two equivalent covariates of vancomycin elimination. Thus, this study allowed the indirect validation of our formula to estimate CLCr in new borns.
Reference: PAGE 8 () Abstr 141 [www.page-meeting.org/?abstract=141]
Poster: poster