Pharmacokinetic and Pharmacodynamic Evaluation of Doripenem in Critically Ill Trauma Patients with Sepsis

Abstract

Background:

Doripenem is approved by the Food and Drug Administration for the treatment of patients with complicated intra-abdominal infections and complicated urinary tract infections. While studies have described the pharmacokinetics/pharmacodynamics (PK/PD) of doripenem in the critically ill, no study has described the probability of target attainment profile among trauma patients with sepsis.

Patients and Methods:

This study was a prospective, open-label, pharmacokinetic study in the surgical intensive care unit (SICU) at Grady Health System. Thirty trauma patients with sepsis admitted to the SICU received doripenem 1 g infused over 4 hours every 8 hours for three doses. Blood samples were taken just before and after the third dose. A two-compartment model was fit to the data using non-parametric population PK modeling software. Embedded with the final PK model, a Monte Carlo Simulations (MCS) was performed to determine the PK/PD profile of doripenem 1 g, infused over 4 hours, every 8 hours after administration of the first and fourth doses.

Results:

Overall, the model fit the data well, and mean (standard deviation) clearance and volume of the central compartment were 16.9 (11.4) L/h and 28.5 (16.0) L, respectively. In the MCS analyses, doripenem 1 g, infused over 4 hours, administered every 8 hours, conferred >90% probabilities of achieving 30–50% time greater than the minimum inhibitory concentration (30–50% T>MIC) for MICs ≤2 mg/L after infusion of both the first and fourth doses. The MCS indicated that more intensive doripenem dosing schemes should be considered for organisms with MIC values in excess of 2 mg/L.

Conclusions:

This is the first study to describe the doripenem PK/PD in critically ill patients with trauma. Among these patients, the MCS analyses suggest that current dosing strategies may be ineffective when the MIC value for the infecting pathogen is expected to be above 2 mg/L.

Doripenem is a synthetic, broad-spectrum, anti-pseudomonal carbapenem that is currently approved by the Food and Drug Administration for the treatment of patients with complicated intra-abdominal infections and complicated urinary tract infections [1]. Similar to other beta-lactams, the pharmacokinetic/pharmacodynamic (PK/PD) target for doripenem is the time that free drug concentrations exceed the minimum inhibitory concentration (MIC) during the dosing interval (fT>MIC) [2 –5]. As with any antibiotic, a firm understanding of its PK/PD profile across various patient populations is required to develop appropriate dosing regimens [6].

For critically ill patients with trauma, it is especially important to maximize the PK/PD profile. Critically ill patients are highly vulnerable to infections with gram negative bacteria with high MIC values, and published data suggest that these patients often require more intensive dosing schemes because of their altered physiology [7 –9]. Although critically ill trauma patients tend to be younger and have fewer comorbidities, they are often hypermetabolic and are frequently administered aggressive fluid resuscitation resulting in increased renal clearance of drugs and larger volumes of distribution [7,9]. Thus, it is often difficult to reliably achieve critical PD targets, such as fT>MIC, with standard dosing regimens in critically ill patients with trauma.

Although other studies have evaluated the PK/PD of doripenem in the critically ill, there are no studies to date describing the PK/PD in the critically ill trauma population with sepsis [3 –5]. The primary objective of this study is to evaluate the PK and PD of doripenem in a population of critically ill trauma patients with sepsis. As part of this examination, we examined the probability of target attainment (PTA) profile after the first and fourth doses to determine the need for loading doses in this population.

Patients and Methods

This was a prospective, open-label, steady-state pharmacokinetic study of critically ill trauma patients with sepsis. Patients were recruited from Grady Health System (Atlanta, GA), a level I trauma center. Inclusion criteria were: age ≥18 years, admission to the Emory surgical intensive care unit (SICU) service for greater than 24 hours, and a diagnosis of sepsis that necessitated empiric antimicrobial therapy [10]. Patients were excluded from the study if they had acute or chronic renal dysfunction (urine output less than 0.5 mL/kg/h or calculated creatinine clearance (CL_cr) of less than 50 mL/min by Cockroft-Gault equation), pregnancy, known allergy to beta-lactam antibiotics, non-English–speaking, or if they were enrolled in another clinical trial [11].

The study was approved by the Institutional Review Board (IRB), and written informed consent was obtained from the patient or a legally authorized representative (LAR) if the patient was unable to give informed consent because of his or her medical condition before initiation of any study procedure. Verbal consent was obtained in patients with IRB permission in special circumstances when the LAR was unable to physically come to the hospital to provide written informed consent.

Interventions and measurements

Patients received doripenem 1000 mg every 8 hours over 4 hours for three doses. Blood samples were collected from an indwelling intravenous catheter immediately before initiation of the third dose (time 0), and 5, 6, and 8 hours after the start of the third dose infusion. All blood samples were collected in non-anticoagulant (red-top) tubes. The samples were allowed to clot, and the serum was obtained by centrifugation within 15 minutes of collection. The serum samples were stored at –70°C until assayed by a validated high performance liquid chromatography methodology with a correlation coefficient of ≥0.998, slope of 0.0575 ± 0.0037 (mean ± standard deviation [SD]), and intercept of 0.0013 ± 0.0057 (mean ± SD) [12]. After completion of the 24-hour study period, doripenem was discontinued, and patients were administered the standard SICU empiric antibiotic of choice.

Data collection

The following data points were collected for each patient: demographic information (age, gender, and weight), allergies, serum creatinine (SCr), CL_CR, temperature, heart rate, respiratory rate, white blood cell count, Sequential Organ Failure Assessment (SOFA) score, Injury Severity Score (ISS), hospital length of stay (LOS) before study enrollment, exact time of blood withdrawals in relation to dose administration.

Pharmacokinetic data analysis

All data were analyzed in a population PK model by use of big non-parametric adaptive grid (BigNPAG) [13]. The PK model was parameterized as a two-compartment model with zero-order infusion and first-order inter-compartmental transfer and elimination. The inverse of the estimated assay variance was used as the first estimate for weighting, and the analysis was performed with adaptive gamma. On attaining convergence, Bayesian estimates for each patient were obtained using the “population of one” utility within BigNPAG.

Mean, median, and modal values were used as measurements of the central tendency of the population parameter estimates and were evaluated in the maximum a posteriori probability (MAP) Bayesian analysis. Scatter plots were examined for individual patients and the overall population. Goodness of fit was assessed by regression with an observed versus predicted plot, coefficients of determination, and log-likelihood values. Predictive performance was based on weighted mean bias and bias-adjusted weighted mean precision.

Post hoc exploratory analyses were completed to examine the relations between individual Bayesian-posterior doripenem clearance estimates and estimates of renal function using several different equations. The renal function estimates included: CL_cr equation with ideal body weight, CL_cr equation with lean body weight, CL_cr equation with adjusted body weight, modification of diet and renal disease (MDRD) formula, and Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) formula [14 –16]. As part of these analyses, these relations were stratified by a renal function estimate of >130 versus ≤130 mL/min, the renal function associated with augmented renal function [17 –19].

Monte Carlo Simulation (MCS)

Embedded with the final population PK model, a 5000-subject MCS using ADAPT 5 was performed for a prolonged 4-hour infusion of doripenem 1000 mg administered every 8 hours for four doses [20]. The population simulation without process noise option was used, and log-normal distributions of PK parameters were selected. The fractions of simulated subjects who achieved 30%, 40%, and 50% T>MIC were calculated for the range of doripenem MICs from 0.25–8 mg/L. Because early appropriate therapy is a key determinant of clinical outcomes for patients with serious infections, this analysis focused on the probabilities of achieving 30–50% T>MIC after the first dose (hours 0–8 post-initial dose) and fourth dose (hours 24–32 post-initial dose) [21,22].

The MCS was also used to evaluate the predictive performance of the final population PK model. We simulated the prolonged 4-hour infusion of doripenem 1000 mg administered every 8 hours for four doses using the mean parameter vector values, then dispersed the patients’ data points around the simulated points. The fidelity by which the concentration-time curves mirrored the raw data was assessed by visual inspection.

Results

Thirty patients were enrolled between May 2010 and June 2011, and data were collected and evaluable for 27 patients. Baseline data are provided in Table 1. The study cohort was predominately male (81%), and the mean (SD) age was 43.6 (15) years. Nearly 90% of patients were receiving mechanical ventilation, and the average CL_CR was 135.1 mL/min. The mean (SD) SOFA score was 5.2 (1.8), and the average (SD) LOS in the ICU before enrollment was 35.3 (39.5) days.

Table 1.

Demographics

Characteristic	Value (n = 27)
Mean (SD) age	43.6 ± 15 y
No. (%) of males	22 (81)
Mean (SD) weight	91.6 ± 27.3 kg
Mean (SD) ISS	22.7 ± 16.4
Mean (SD) SOFA score	5.2 ± 1.8
Mean (SD) CL_cr	135.1 ± 66.7 mL/min
No. (%) Mechanically ventilated	24 (89)
Mean (SD) ICU LOS	35.3 ± 39.5 d

SD = standard deviation; No. = number; ISS = Injury Severity Score; SOFA = Sequential Organ Failure Assessment; CL_cr = creatinine clearance; LOS = length of stay.

Population pharmacokinetic analysis

The mean (SD) population parameter estimates identified by BigNPAG are shown in Table 2. The predictive performance of the PK model after the Bayesian step is displayed in Table 3. Using the population mean parameter values as the measure of central tendency, the overall fit of the model to the data was good, and the observed-predicted plots were highly acceptable.

Table 2.

Pharmacokinetic Parameters from Population Pharmacokinetic Analysis

Parameter	Value (SD)
Vc (L)	28.52 ± 16.01
CL (L/h)	16.94 ± 11.4
kcp (h⁻¹)	1.47 ± 2.24
kpc (h⁻¹)	1.99 ± 3.15

SD = standard deviation; Vc = volume of the central compartment; CL = clearance of the central compartment; kcp = microtransfer rate constant from the central to the peripheral compartment; kpc = microtransfer rate constant from the peripheral to the central compartment.

Table 3.

Goodness of Fit and Predictive Performance of the Population Model for Plasma Compartments

Parameter	Value or description
Regression line^a	Observed = −1.00 × predicted + 0.311
R^2b	0.996
Mean weighted error (mg/L)	−0.032
Mean weighted squared error (mg/L)²	−0.970
Log likelihood	−100.711

Best-fit regression for the observed-predicted plot after the Bayesian step.

Coefficient of determination for the best-fit linear regression line for the observed-predicted plot after the Bayesian step.

The simulated concentration-time serum curve from the mean parameter vector values from the final population PK model for doripenem 1000 mg, infused over 4 hours, administered every 8 hours is shown in Figure 1. As shown in Figure 1, the raw data points from study subjects are generally evenly dispersed around the simulated concentration-time curve. Most importantly, as doripenem is a characteristic T>MIC antibiotic, doripenem concentrations at the midpoint and end of the dosing interval were well captured around the central tendency.

FIG. 1.

Concentration-time profile of doripenem 1 g administered every 8 hours as a 4-hour infusion as calculated from the mean parameter vector. Empty circles are observed concentrations from critically ill patients with sepsis.

The relations between individual Bayesian-posterior doripenem clearance estimates versus different renal function equation-based estimates are displayed in Figures 2A–E. No demonstrable relations were noted between the individual Bayesian-posterior doripenem clearance estimates and the different renal function equation-based estimates. In the regression analyses, the R² ranged between 0.07 and 0.21 for all comparisons. As shown in Figures 2A–E, the individual Bayesian-posterior doripenem clearance estimates were also similar for the patients with an estimated renal function estimate >130 mL/min relative to those with an estimated renal function estimate <130 mL/min.

FIG. 2.

Regression of 27 subjects’ mean individualized Bayesian clearance (L/h) to different renal function estimate equations. (A) Creatinine clearance (CL_CR) with ideal body weight (IBW) (R² = 0.0760). (B). Creatinine clearance (CL_CR) with lean body weight (LBW) (R² = 0.1086). (C). Creatinine clearance (CL_CR) with adjusted body weight (AdjBW) (R² = 0.1005). (D). Modification of Diet and Renal Disease (MDRD) formula (R² = 0.0732). (E). Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) formula (R² = 0.2049).

Probability of target attainment analysis

The PTA analyses are displayed in Figures 3A and 3B. With the first dose simulations (Fig. 3A), the probability of achieving 30–50% T>MIC exceeded 95% for MIC values ≤2 mg/L (range of MIC values considered susceptible by Clinical and Laboratory Standards Institute, or CLSI). At an MIC value of 4 mg/L, the probability of achieving 30% T>MIC was 90% while the probabilities of achieving 40% and 50% T>MIC were 86% and 76%, respectively. The probability of achieving 30–50% T>MIC was less than 60% at an MIC value of 8 mg/L.

FIG. 3.

(A) Probability of target attainment profile for doripenem 1000 mg, infused over 4 hours, every 8 hours, after the first dose. (B). Probability of target attainment profile for doripenem 1000 mg, infused over 4 hours, every 8 hours, after the fourth dose. MIC = minimum inhibitory concentration.

A modestly improved PTA profile for doripenem 1 g, infused over 4 hours, administered every 8 hours, was observed after the fourth dose (Fig. 3B). Similar to the first dose simulations, the probability of achieving 30–50% T>MIC exceeded 95% for MIC values ≤2 mg/L. Subtle improvements in the PTA profiles between the first and fourth dose simulations were noted at MIC values of 4 and 8 mg/L. The probabilities of achieving 30–50% T>MIC, however, were still less than 80% at an MIC of 8 mg/L with the fourth dose simulations.

Discussion

The primary objective of this study is to characterize the PK/PD profile of doripenem in a population of critically ill trauma patients with sepsis. These data are timely given the suboptimal outcomes observed with doripenem relative to imipenem among patients with augmented renal function, a population highly consistent with our study cohort, in the recent ventilator-associated pneumonia (VAP) randomized clinical trial [23]. Overall, there were several notable findings in this investigation. Compared with studies conducted in healthy volunteers, higher mean estimates of clearance and central volume of distribution were noted in this study. The elevated clearance among trauma patients is most likely related to their augmented renal function; the higher volume of distribution is likely because of the administration of aggressive fluid resuscitation for sepsis.

Our mean estimates and coefficients of variation for volume and clearance aligned closely with several other studies conducted in medical ICUs among patients with documented infections and augmented CL_cr as shown in Table 4. Although our clearance estimate appears to be considerably higher than the PK study of ICU patients by Bhalodi et al. [4], it is important to recognize that they made overall clearance proportional to CL_CR plus an intercept term (clearance = (0.096 × CL_CR) + 3.2. If one were to substitute the average CL_CR of 135 mL/min from our study in the Bhalodi clearance formula, the overall clearance estimates would be nearly identical (17 L/h vs. 16 L/h, respectively). The higher doripenem clearance observed in our study, however, did not directly correlate with the patients’ estimated renal function as shown in Figures 2A–E. The collective data reflect highly variable drug clearance and expanded volume of distributions observed in critically ill patients with sepsis.

Table 4.

Comparison of Clearance and Central Volume of Distribution Parameter Estimates across Doripenem Pharmacokinetic Studies of Patients in Intensive Care Units ^*

Study	Sample Size	Study population	Creatinine clearance in mL/min	Clearance in L/h	Central volume of distribution in L
Roberts and Lipman [5]	31	NPNA	137 (71)	20.4 (13.8)	24.4 (26)
Bhalodi et al. [4]	13	Predominately NPNA	74 (34.9)	10.1 (10.2)	17.3 (11.2)
Jaruratanasirikul et al. [3]	11	VAP	119.7 (38.4)	24.8 (1.6)	17.2 (1.8)
Nandy et al. [2]	303	Healthy subjects, CUTI, pyelonephritis, NPNA	Median = 98	13.6 (0.2)	11.6 (0.3)

NPNA = nosocomial pneumonia; VAP = ventilator associated pneumonia; CUTI = complicated urinary tract infection.

All data presented as mean (standard deviation) unless otherwise noted.

Although the highly variable estimates of clearance and volume of distribution observed suggest that therapeutic drug monitoring may be warranted in this patient population, the MCS analyses indicate that doripenem 1 g, infused over 4 hours, administered every 8 hours, had a highly favorable PTA profile against the range of MIC values considered susceptible by CLSI. Given recent data suggesting that loading doses are needed for prolonged beta-lactam infusions, we evaluated the PK/PD profile after the first dose (hours 0–8 post-initial dose) and at steady state conditions (hours 24–32 post-initial dose). As shown in Figures 3A and 3B, the probabilities of achieving 30–50% T>MIC exceeded 90% for MIC values ≤2 mg/L. Modest differences in the PTA profiles were noted at an MIC value of 4 mg/L, and both were suboptimal at an MIC value of 8 mg/L.

From a PK/PD perspective, the similar PTA profiles between first dose and steady-state simulations are not surprising. A loading dose serves a two-fold purpose: (1) to rapidly achieve concentrations that mimic concentrations at steady-state conditions and (2) to maximize total exposure (e.g., maximum serum concentration and area under the curve) at the onset of therapy. These are not critical concerns for short half-life drugs such as beta-lactams where the interest is maintaining T>MIC for as long as possible during the dosing interval (e.g., time-dependent activity). Loading doses are only required for antibiotics with long terminal half-lives and whose bactericidal activity is concentration-dependent. Further, even though overall clearance is generally higher among trauma patients with sepsis, the half-life is often prolonged because of their elevated volumes of distribution. As clearance = volume of distribution × Ke and half-life = 0.693/Ke, the net effect of this is similar PK/PD profiles with the first and fourth doses.

Although the MCS data are reassuring, there are several things to consider. First, the favorable PTA profile was only observed for MIC values ≤2 mg/L. In clinical practice, it is not uncommon to encounter an infection from a gram negative organism with a doripenem MIC value in excess of 2 mg/L. Given the rising rates of resistance across many gram negative pathogens, many clinicians are often required to use antibiotics for organisms with MIC values in excess of the susceptibility breakpoints [24 –34]. In these circumstances, it would be prudent to consider the risk versus benefit of empirically using more intensive doripenem dosing schemes than the one evaluated in this study.

Second, these simulation studies focused on concentrations in the blood stream. Blood stream concentrations are not always indicative of concentrations at the site of infection, particularly for infection involving the lower respiratory tract [35]. For doripenem, the median epithelial lining fluid/plasma penetration ratio among patients with VAP is only ∼30% [36]. Because the critical drug-pathogen interaction is the same regardless of the site of infection, the only way to achieve a favorable T>MIC profile in the lower respiratory tract with doripenem is to again consider the risk versus benefit of administering a higher doripenem dose (>1 g) over prolonged periods [37,38].

This PK/PD study focused on PD targets associated with maximal bacterial killing and clinical success [37,38]. The next frontier in PK/PD is identifying antibiotic dosing schemes and drug combinations that minimize the emergence of resistance [39]. The existing data suggest that PD targets that prevent resistance to carbapenems are typically 2–4-fold higher than PD targets for success [40 –43]. Further study is still needed in the area of resistance suppression, but the current data suggest that obtaining the PK/PD target against the range of MICs encountered clinically is not likely with the doripenem dosing scheme examined in this analysis and that more intensive dosing schemes will be required. Another limitation to note is that the CL_cr was calculated in this study; however, a measured CL_cr would have been the gold standard in assessing CL_cr.

Last, this study was completed in a critically ill patient population that required a sparse sampling approach. Four blood samples were acquired that included measurement before initiation of the third dose (time 0), and 5, 6, and 8 hours after the start of the third dose infusion. Although a more robust sampling scheme with collections during infusion would have been preferred, we chose to better characterize the elimination phase of this time-dependent antimicrobial. This was important, because our objective was to determine the probability of time above the MIC. Ultimately, the PK/PD profile of time-dependent beta-lactam antibiotics is driven by the elimination phase because of the short half-life.

Conclusions

The current study demonstrates that, as expected, critically ill patients with trauma have higher clearance rates and volumes of distribution for doripenem compared with healthy volunteers. Moreover, because of such drastic alterations in the PK in this population, 1 g doses infused over 4 hours and administered every 8 hours are required to achieve reliably the PD targets associated with maximal bactericidal killing for gram negative pathogens with MIC values considered susceptible by CLSI. If pathogens with higher MICs (>2 mg/L) to doripenem are encountered, or if the infection involves the lower respiratory tract, our data suggest that the risk versus benefit of using more intensive doripenem dosing schemes should be considered. Because this is a PK/PD study, all findings need to be evaluated prospectively in the clinical arena.

Footnotes

Author Disclosure Statement

This study was funded by Ortho McNeil Janssen Scientific Affairs

References

Doribax^™ package insert. Raritan, NJ: Ortho-McNeil Pharmaceutical, Inc, 2007.

Nandy

, Samtani

, and Lin

. Population pharmacokinetics of doripenem based on data from phase I studies with healthy volunteers and phase 2 and 3 studies with critically ill patients. Antimicrob Agents Chemother, 2010; 54:2354–2359.

Jaruratanasirikul

, Wongpoowarak

, Kositpantawong

, et al: Pharmacodynamics of doripenem in critically ill patients with ventilator-associated Gram-negative bacilli pneumonia. Int J Antimicrob Agents, 2012; 40:434–439.

Bhalodi

, Keel

, Quintilani

, et al. Pharmacokinetics of doripenem in infected patients treated within and outside the intensive care unit. Ann Pharmacother, 2013; 47:617–627.

Roberts

, and Lipman

. Optimal doripenem dosing simulations in critically ill nosocomial pneumonia patients with obesity, augmented renal clearance, and decreased bacterial susceptibility. Crit Care Med, 2013; 41:489–495.

Rodvold

KA.

Pharmacodynamics of anti-infective therapy: Taking what we know to the patient's bedside. Pharmacotherapy, 2001; 21:319S–330S.

Boucher

, Wood

, Swanson

. Pharmacokinetic changes in critical illness. Crit Care Clin, 2006; 22:255–271.

Bush

, Levison

. Antibiotic selection and pharmacokinetics in the critically ill. Crit Care Clin, 1988; 4:299–324.

Krishnan

, Murray

. Pharmacologic issues in the critically ill. Clin Chest Med, 2003; 24:671–688.

10.

Levy

, Fink

, Marshall

, et al. 2001 SCCM/EISCM/ACCP/ATS/SIS International Sepsis Definitions Conference. Crit Care Med, 2003; 31:1250–1256.

11.

Cockcroft

, Gault

. Prediction of creatinine clearance from serum creatinine. Nephron, 1976; 16:31–41.

12.

Sutherland

, Nicolau

. Development of an HPLC method for the determination of doripenem in human and mouse serum. J Chromatogr B Analyt Technol Biomed Life Sci, 2007; 853:123–126.

13.

Leary

, Jelliffe

, Schumitzky

, and van Guilder

. An adaptive grid, non-parametric approach to pharmacokinetic and dynamic (PK/PD) models p. 389–394. In: Proceedings of the Fourteenth IEEE Symposium on Computer-based Medical Systems. IEEE Computer Society. July 26–27, 2001; Bethesda, MD.

14.

Levey

, Bosch

, Lewis

, et al. A more accurate method to estimate glomerular filtration rate from serum creatinine: A new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med, 1999; 130:461–470.

15.

Levey

, Coresh

, Greene

, et al. Using standardized serum creatinine values in the modification of diet in renal disease study equation for estimating glomerular filtration rate. Ann Intern Med, 2006; 145:247–254.

16.

Levey

, Stevens

, Schmid

, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med, 2009; 150:604–612.

17.

Udy

, Varghese

, Altukroni

, et al. Subtherapeutic initial beta lactam concentrations in select critically ill patients: Association between augmented renal clearance and low trough drug concentrations. Chest, 2012; 142:30–39.

18.

Udy

, Baptista

, Lim

, et al. Augmented renal clearance in the ICU: Results of a multicenter observational study of renal function in critically ill patients with normal plasma creatinine concentrations. Crit Care Med, 2014; 42:520–527.

19.

Udy

, Roberts

, Shorr

, et al. Augmented renal clearance in septic and traumatized patients with normal plasma creatinine concentrations: Identifying at-risk patients. Crit Care, 2013; 17:R35.

20.

D'Argenio

, Schumitzky

, Wang

. ADAPT 5 User's Guide: Pharmacokinetic/Pharmacodynamic Systems Analysis Software. Los Angeles, CA: Biomedical Simulations Resource, 2009.

21.

Kollef

, Sherman

, Ward

, Fraser

. Inadequate antimicrobial treatment of infections: A risk factor for hospital mortality among critically ill patients. Chest, 1999; 115:462–474.

22.

Lodise

Jr , Patel

, Kwa

, et al. Predictors of 30-day mortality among patients with Pseudomonas aeruginosa bloodstream infections: Impact of delayed appropriate antibiotic selection. Antimicrob Agents Chemother, 2007; 51:3510–3515.

23.

Kollef

, Chastre

, Clavel

, et al. A randomized trial of 7-day doripenem versus 10-day imipenem-cilastatin for ventilator-associated pneumonia. Crit Care, 2012; 16:R218.

24.

Intensive Care Antimicrobial Resistance Epidemiology (ICARE) Surveillance Report, data summary from January 1996 through December 1997: A report from the National Nosocomial Infections Surveillance (NNIS) System. Am J Infect Control, 1999; 27:279–284.

25.

Friedland

, Gallagher

, King

, Woods

. Antimicrobial susceptibility patterns in Pseudomonas aeruginosa: Data from a multicenter Intensive Care Unit Surveillance Study (ISS) in the United States. J Chemother, 2004; 16:437–441.

26.

Harris

, Torres-Viera

, Venkataraman

, DeGirolami

, Samore

, Carmeli

. Epidemiology and clinical outcomes of patients with multiresistant Pseudomonas aeruginosa. Clin Infect Dis, 1999; 28:1128–1133.

27.

Jones

, Kirby

, Beach

, et al. Geographic variations in activity of broad-spectrum beta-lactams against Pseudomonas aeruginosa: Summary of the worldwide SENTRY Antimicrobial Surveillance Program (1997–2000). Diagn Microbiol Infect Dis, 2002; 43:239–243.

28.

Karlowsky

, Draghi

, Jones

, et al. Surveillance for antimicrobial susceptibility among clinical isolates of Pseudomonas aeruginosa and Acinetobacter baumannii from hospitalized patients in the United States, 1998 to 2001. Antimicrob Agents Chemother, 2003; 47:1681–1688.

29.

Livermore

DM.

Multiple mechanisms of antimicrobial resistance in Pseudomonas aeruginosa: Our worst nightmare?. Clin Infect Dis, 2002; 34:634–640.

30.

Obritsch

, Fish

, MacLaren

, Jung

. National surveillance of antimicrobial resistance in Pseudomonas aeruginosa isolates obtained from intensive care unit patients from 1993 to 2002. Antimicrob Agents Chemother, 2004; 48:4606–4610.

31.

Patel

, Harrington

, Dihmess

, et al. Clinical epidemiology of carbapenem-intermediate or resistant Enterobacteriaceae. J Antimicrob Chemother, 2011; 66:1600–1608.

32.

Satlin

, Jenkins

, Walsh

. The global challenge of carbapenem-resistant Enterobacteriaceae in transplant recipients and patients with hematologic malignancies. Clin Infect Dis, 2014; 58:1274–1283.

33.

Cuzon

, Naas

, Truong

, et al. Worldwide diversity of Klebsiella pneumoniae that produce beta-lactamase blaKPC-2 gene. Emerg Infect Dis, 2010; 16:1349–1356.

34.

Goren

, Navon-Venezia

, Chmelnitsky

, Carmeli

. Carbapenem-resistant KPC-2-producing Escherichia coli in a Tel Aviv Medical Center, 2005 to 2008. Antimicrob Agents Chemother, 2010; 54:2687–2691.

35.

Lodise

, Sorgel

, Melnick

, et al. Penetration of meropenem into epithelial lining fluid of patients with ventilator-associated pneumonia. Antimicrob Agents Chemother, 2011; 55:1606–1610.

36.

Justo

, Gotfried

, Deyo

, et al. In: Doripenem intrapulmonary pharmacokinetics in healthy adult subjects. Proceedings of the 51st Interscience Conference on Antimicrobial Agents and Chemotherapy; September 17–20, 2011; Chicago, IL. Abstract A1–A1748.

37.

Drusano

GL.

Antimicrobial pharmacodynamics: Critical interactions of “bug and drug.”. Nat Rev Microbiol, 2004; 2:289–300.

38.

Lodise

, Drusano

. Pharmacokinetics and pharmacodynamics: Optimal antimicrobial therapy in the intensive care unit. Crit Care Clin, 2011; 27:1–18.

39.

Lodise

, Butterfield

. Use of pharmacodynamic principles to inform beta-lactam dosing: “S” does not always mean success. J Hosp Med, 2011;(Suppl 1):S16–S23.

40.

Louie

, Bied

, Fregeau

, et al. Impact of different carbapenems and regimens of administration on resistance emergence for three isogenic Pseudomonas aeruginosa strains with differing mechanisms of resistance. Antimicrob Agents Chemother, 2010; 54:2638–2645.

41.

Tam

, Ledesma

, Schilling

, et al. In vivo dynamics of carbapenem-resistant Pseudomonas aeruginosa selection after suboptimal dosing. Diagn Microbiol Infect Dis, 2009; 64:427–433.

42.

Tam

, Schilling

, Neshat

, et al. Optimization of meropenem minimum concentration/MIC ratio to suppress in vitro resistance of Pseudomonas aeruginosa. Antimicrob Agents Chemother, 2005; 49:4920–4927.

43.

Tam

, Schilling

, Poole

, Nikolaou

. Mathematical modeling response of Pseudomonas aeruginosa to meropenem. J Antimicrob Chemother, 2007; 60:1302–1309.