Gauri G. Rao (1), Neang S. Ly (1), Brian T. Tsuji (1), Alan Forrest (1)
(1) School of Pharmacy &Pharmaceutical Sciences, University at Buffalo, SUNY, Buffalo, NY
Objectives:L has a complicated &variable population PK model with parallel saturable (non-renal) & first order (renal) pathways of elimination. The traditional dose of 600mg q12hr provides ~30-fold variability in area under the curve (AUC); only ~19% of this variability can be explained by patient covariates. Efficacy is well related to AUC/MIC with values of 100 to 200 needed to treat most sites of infection. The current MIC breakpoint for susceptibility is 4. TD (decreased platelet (PLT)) related to AUC & duration of L treatment. Our objective is to use population PK/PD/TD models & Monte Carlo simulations (MCS) to evaluate the ability to safely &effectively treat infections with an MIC>1.
Methods: We developed an integrated PK/PD/TD model from the literature1-4. MCS (ADAPT 5)was used to randomly select 5000 adult patients (with a full range of weights, creatinine clearance &baseline PLT) who were treated with 3 different regimens (600mg q12h (R1), 600mg q8h (R2) & a regimen individualized by AFC (R3)). For each subject ®imen resulting AUCs were used to compute, for MICs 2 &4, the probability for success for a lower respiratory tract infection (PLRTI2 & PLRTI4) & bacteremia (PBac2 & PBac4) & predicted platelet nadirs on days 7 (NPlt7) & 14 (NPlt14). For the AFC algorithm, a 1200mg loading dose was followed by plasma samples at 0.5, 1, 4, 12 & 24h, with random observation errors, (times according to optimal sampling theory) &600mg q8h. These samples were fit with a MAP-Bayesian estimator & doses adjusted within 24h.
Results: The randomly selected patients had a median (range) weight (kg) 73.4(39-132), creatinine (mL/min) 67.8 (22.6-178) & baseline PLT 305 (135-595). Results for (R1/R2) for %Bac2 (91.8/95.9), %Bac4 (66.2/93.6), %LRTI2(75.5/77.9), %LRTI4(59.6/76.5), %NPLT7<100,000 (5.96/41.5) & %NPLT14<100,000 (14.7/73.3). R1 appears reasonable up to an MIC of 2. A fixed regimen (like R2) sufficient to treat an MIC of 4 would be associated with unacceptable thrombocytopenia. With R3 (addition of AFC algorithm to R2) the resulting variability in AUCs were reduced from greater than 20 fold to less than 2 fold allowing adequate AUC to safely & effectively treat an MIC of 4.
Conclusions: R1 the MIC cutoff should be 2 & not 4. Given the substantial variability in PK poorly described by covariates, AFC of L appears the best approach to extend this to an MIC of 4.
References:
[1] Meagher, A. K., A. Forrest, et al. (2003). “Population pharmacokinetics of linezolid in patients treated in a compassionate-use program.” Antimicrob Agents Chemother 47(2): 548-553.
[2] Rayner, C. R., A. Forrest, et al. (2003). “Clinical pharmacodynamics of linezolid in seriously ill patients treated in a compassionate use programme.” Clin Pharmacokinet 42(15): 1411-1423.
[3] Forrest A, Rayner CR, Meagher AK, Birmingham MC, Schentag JJ. 40th ICAAC. Sept 2000 Toronto, Ontario, Canada
[4] Sasaki, T., H. Takane, et al. “Population pharmacokinetic and pharmacodynamic analysis of linezolid and a hematologic side effect, thrombocytopenia, in Japanese patients.” Antimicrob Agents Chemother 2011 May;55(5):1867-73. doi(2011 Feb 28): 10.1128/AAC.01185-01110.
[5]D’Argenio, D.Z., A. Schumitzky and X. Wang. ADAPT 5 User’s Guide: Pharmacokinetic/ Pharmacodynamic Systems Analysis Software. Biomedical Simulations Resource, Los Angeles, 2009.
Reference: PAGE 22 (2013) Abstr 2957 [www.page-meeting.org/?abstract=2957]
Poster: Infection