Beta-lactam antimicrobial dosing optimization in obese patients compared to non-obese patients using population pharmacokinetic/pharmacodynamic approach

Eun Kyoung Chung, Purdue University

Abstract

Obesity is a significant global health problem and has been associated with altered pharmacokinetics and pharmacodynamics of many drugs. However, little is known regarding the effect of obesity on the pharmacokinetics and pharmacodynamics of many broad-spectrum, beta-lactam antibiotics such as piperacillin/tazobactam, meropenem, and cefepime. The objective of this study is to evaluate the population pharmacokinetics and pharmacodynamics of piperacillin/tazobactam, meropenem, and cefepime in hospitalized obese patients in order to determine dosing regimens that provide similar exposures between obese and non-obese patients. For piperacillin/tazobactam, a retrospective analysis was conducted using prospectively collected serum concentration-time data from two previous studies (Study 1 and Study 2) published by our research group. Hospitalized, adult patients who required antimicrobial therapy for a suspected or documented bacterial infection were eligible to participate in both studies. In Study 2, only patients with total body weight (TBW) greater than 120 kg were eligible to be enrolled. Patients were classified as either obese [body mass index (BMI) ¡Ý 30 kg/m2] or non-obese (BMI < 30 kg/m2). In Study 1, all patients received piperacillin/tazobactam 4.5 g every 8 hours (q8h), infused over 4 hours. In Study 2, patients received piperacillin/tazobactam either 4.5 g or 6.75 g q8h, infused over 4 hours. After 2 or more days of therapy, serial blood samples were collected from an indwelling IV catheter immediately prior to drug administration, and at 1, 2, 3, 4 (end of infusion), 5, 6, 7 and 8 hours after the start of infusion. Piperacillin and tazobactam serum concentrations were determined by the previously validated high performance liquid chromatography (HPLC) method. Population pharmacokinetic parameters were estimated using NONMEM, and the final pharmacokinetic model was built by evaluating the effects of covariates on the pharmacokinetic parameters of piperacillin and tazobactam using the stepwise forward inclusion followed by the backward elimination process. Tested covariates included: 1) age; 2) sex; 3) body size descriptor, including TBW, ideal body weight (IBW), lean body weight (LBW), and BMI; 4) creatinine clearance (CRCL); and 5) admission to an intensive care unit (ICU; ICU=1, general medical ward=0). In the stepwise forward inclusion process, covariates that reduced the model objective function value (OFV) > 3.84 (p < 0.05; ¦Ö2 distribution; 1 df) were considered significantly associated with the pharmacokinetic parameters in the model. In the backward elimination process, a covariate was removed if its elimination increased the model OFV by < 5.024 (p > 0.025; ¦Ö2 distribution; 1 df). Using the final pharmacokinetic model, Monte Carlo simulations were performed for three 4-hour dosing regimens to calculate probability of target attainment (PTA) using ¡Ý 50%fT>MIC. Overall, a convenience sample of 27 patients (11 non-obese and 16 obese) were studied. TBW ranged from 60 kg to 211 kg, BMI from 19.7 kg/m2 to 72.9 kg/m2, and measured creatinine clearance (CRCL) from 23 mL/min to 260 mL/min. Patient demographics [median (range)] in non-obese vs. obese group are: age, 53 (27-76) vs. 48 (35-69) years; CRCL, 88 (23-148) vs. 111 (28-260) mL/min; height, 175 (163-190) vs. 175 (157-190) cm; TBW, 74 (60-100) vs. 151 (98-211) kg; LBW, 54 (39-72) vs. 78 (50-94) kg; IBW, 71 (55-84) vs. 71 (50-84) kg; BMI, 24.8 (19.7-29.4) vs. 50.1 (32.7-72.9) kg/m2. The number of male patients was seven in non-obese and ten in obese patient groups, and the number of patients admitted to an intensive care unit (ICU) was seven each in non-obese and obese patient groups. Compared to non-obese patients, obese patients had significantly larger TBW, LBW, and BMI (p < 0.05); other demographics were similar between non-obese and obese patients. Observed serum concentration-time profiles of both piperacillin and tazobactam were best described by a one-compartment model with zero-order input and first-order, linear elimination. The final model for piperacillin was: clearance (CL; L/h) = 11.3 + [0.0646*(CRCL-105)] + [0.0579*(BMI-35)]; and volume of distribution (V; L) = 31.3 + [0.132*(TBW-120)]. The final model for tazobactam was: CL (L/h) = 10.1 + [0.0272*(CRCL-105)]; and V (L) = 34.3. For both piperacillin and tazobactam, obese patients had significantly increased CL and V compared to non-obese patients. The pharmacokinetic parameters [median (range)] in non-obese vs. obese patients were: piperacillin CL, 9.0 (4.8-14.2) vs. 13.1 (6.8-20.0) L/h (p=0.026); piperacillin V, 24.6 (17.1-37.8) vs. 32.5 (19.8-69.8) L (p=0.014); tazobactam CL, 6.8 (4.4-15.5) vs. 13.1 (5.6-26.4) L/h (p=0.005); and tazobactam V, 17.1 (9.4-70.3) vs. 45.5 (10.5-116.6) L (p=0.019). Based on the pharmacodynamic analysis using Monte Carlo simulation, at the piperacillin MICs ¡Ü 16 mg/L in the presence of tazobactam, which is the susceptibility breakpoint for Enterobacteriaceae and Pseudomonas aeruginosa, PTA was > 90% for 4-hour infusion dosing regimens ¡Ý 3.375 g q8h in non-obese patients and ¡Ý 4.5 g q8h in obese patients, respectively. For meropenem, a retrospective analysis was conducted using prospectively collected serum concentration-time data from three previous studies (Study 3, Study 4, and Study 5) published by our research group. Hospitalized, adult patients who required antimicrobial therapy for a suspected or documented bacterial infection were eligible to participate in all three studies. Although patients with CRCL less than 50 mL/min were eligible to participate in Study 3, they were excluded in Study 4 and 5 due to different study objectives. In Study 3, only patients with BMI ¡Ý 40 kg/m2 were enrolled, and in Study 4, only patients with BMI ¡Ý 40 kg/m2 or TBW ¡Ý 100 pounds over their IBW were enrolled. Patients were classified as either obese (BMI ¡Ý 30 kg/m2) or non-obese (BMI < 30 kg/m2). In Study 3, patients received the following meropenem dosing regimens: 500 mg q6h if CRCL > 60 mL/min; 500 mg q8h if CRCL was 40 to 60 mL/min; and 500 mg q12h if CRCL was 10 to 39 mL/min. In Study 4, all patients received either 500 mg or 1000 mg q6h. In Study 5, all patients received 1000 mg q8h. In all studies, all dosing regimens were infused over 30 minutes. After 2 or more days of therapy, serial blood samples were collected from an indwelling IV catheter as scheduled in each study: immediately prior to drug administration, 0.5 (end of infusion), 0.75, 1, 1.5, 2, 3, 4, 5, 6, 8 (if receiving q8h or q12h dosing regimens), and 12 hours (if receiving q12h dosing regimens) after the start of infusion in Study 3; prior to drug administration, 0.5 (end of infusion), 1, 2, 3, 4, and 6 hours after the start of infusion in Study 4; and prior to drug administration, 0.5 (end of infusion), 1, 1.5, 2, 3, 4, 5, 6, and 8 hours after the start of infusion in Study 5. Serum meropenem concentrations were determined by previously described analytical methods: HPLC in Study 3 and Study 4; and ultraperformance liquid chromatography in Study 5. Population pharmacokinetic parameters were estimated using NONMEM, and the final pharmacokinetic model was built by evaluating the effects of covariates on the meropenem pharmacokinetic parameters using the stepwise forward inclusion followed by the backward elimination process. Tested covariates included: 1) age; 2) sex; 3) body size descriptor, including TBW, IBW, LBW, and BMI; 4) CRCL; and 5) admission to an ICU (ICU=1, general medical ward=0). In the stepwise forward inclusion process, covariates that reduced the model OFV > 3.84 (p < 0.05; ¦Ö2 distribution; 1 df) were considered significantly associated with the pharmacokinetic parameters in the model. In the backward elimination process, a covariate was removed if its elimination increased the model OFV by < 5.024 (p > 0.025; ¦Ö2 distribution; 1 df). Using the final pharmacokinetic model, Monte Carlo simulations were per...

Degree

Ph.D.

Advisors

Kays, Purdue University.

Subject Area

Pharmacology|Pharmaceutical sciences

COinS