An optimal sampling schedule for neonates, infants & children receiving cefazolin +/- vancomycin for cardiopulmonary bypass

Background: Dosing of prophylactic antibiotics in children during cardiopulmonary bypass (CPB) remains poorly defined. Pharmacokinetic (PK) studies can be improved using optimal design when sampling is limited, or multiple factors influence PK. We aimed to optimize a sampling schedule designed to determine cefazolin and vancomycin PK in children undergoing CPB. Methods: A one compartment distribution model for vancomycin and a three compartment distribution model for cefazolin was used with theory based allometric scaling and maturation to describe first-order elimination clearance. The CPB circuit was represented by an additional compartment. We assumed 60 subjects received cefazolin 50 mg.kg-1, with 50 of these subjects undergoing a procedure with CPB. We assumed 15 subjects also received 15 mg.kg-1 vancomycin. Optimal times for up to 8 samples per patient were estimated, ignoring CPB effects, using WinPOPT (University of Otago, New Zealand). Optimal sampling times for determination of CPB related changes were considered separately. Designs were selected based on relative standard errors (RSEs) for model parameters and comparison of criterions summarizing design efficiency. Results: Sample times were 0.001, 0.001, 0.108, 0.36, 1.05, 1.85 h following the first dose, and 0.36 and 2.5 h after the second dose, for With CPB subjects. Sample times were 0.127, 0.43, 0.43, 1.3, 3.18, 6, 6 h after the first dose and 6 h after the second dose, for Without CPB subjects. Five samples, taken directly from the CPB circuit, were required to adequately capture CPB related changes in CPB V and CL. RSE estimates of cefazolin, vancomycin and CPB circuit parameters for the final design were ≤ 30%, with the exception of one of the cefazolin volumes (V2) for which RSEs were 49%. Conclusion: The sampling schedule may be used in the planning of a clinical study of children receiving cefazolin and vancomycin during CPB.