Rapid attainment of pharmacodynamic parameter goals in patients receiving vancomycin or aminoglycosides for serious infections

Introduction:

Pharmacodynamic targets for antibiotic treatment are available. However, there is considerable variation in how clinicians attempt to achieve them when prescribing doses.

Objectives:

Compare AUC₂₄/MIC for initial vancomycin doses or Cmax/MIC for initial gentamicin/tobramycin doses selected by clinicians using estimated population versus actual parameters to those attained after individualization.

Study Design:

Observational design, 152 vancomycin/146 aminoglycoside patients, criterion: clinicians’ self-identified treatment goals: vancomycin AUC₂₄/MIC>400, aminoglycoside Cmax/MIC≥10, culture-documented MRSA or gram-negative infection, MIC, antibiotic plus serum creatinine concentrations.

Methods:

Estimated population AUC₂₄ for vancomycin: AUC₂₄ = D/{[(CrCl_est • 0.79) + 15.4] • 0.06}, where D=dose, CrCl_est=estimated creatinine clearance. Estimated population Cmax for gentamicin/tobramycin: Cmax = [(D/t’)(1-e^-kt’)]/[kV(1-e^-kt)], where D=dose, t’=infusion time, t=dosage interval, k=elimination rate constant (k = 0.00293(CrCl_est)+0.014), V=volume of distribution (0.26 L/kg for nonobese, or ABW for obese [ABW=IBW+[0.4(TBW-IBW)], TBW=total body weight). Estimated population MIC was the institutional average during the previous 6 months. Actual and adjusted AUC₂₄or Cmax were computed using a Bayesian computer program. Initial doses were determined by clinicians, and adjusted doses were prescribed to attain treatment goals.

Results:

While clinicians anticipated all doses initially prescribed would attain the treatment goals, only 51% of the vancomycin regimens and only 76% of the aminoglycoside regimens were expected to achieve goals using population estimates. For initial dosing of vancomycin, only 35% of patients actually achieved an AUC₂₄/MIC>400. For initial dosing of aminoglycosides, only 64% of patients actually achieved a Cmax/MIC≥10. Adjusted dosages achieved the treatment goal in all cases (p<0.01).

Conclusions: Antibiotic doses can be rapidly individualized using Bayesian techniques to attain pharmacokinetic/pharmacodynamic goals. Clinician-prescribed initial doses or doses computed using population estimates will not achieve this goal for all patients, but at the outset a higher percentage of aminoglycoside patients with reach goal pharmacodynamic values compared to vancomycin patients.