Evaluation of Several Dosing Regimens of Cefepime, with Various Simulations of Renal Function, against Clinical Isolates of Pseudomonas aeruginosa in a Pharmacodynamic Infection Model

Abstract

The objectives of this study were as follows: (i) to examine the killing activity of 2-g doses of cefepime against two clinical isolates (mucoid and nonmucoid) of Pseudomonas aeruginosa in a pharmacodynamic in vitro infection model, (ii) to compare the percentage of time above the MIC (T > MIC) for each of the regimens against P. aeruginosa, and (iii) to evaluate the area under the bactericidal curve for each regimen. Cefepime was administered at intervals of 8, 12, and 24 h with and without tobramycin, and two different levels of renal function were simulated: normal (creatinine clearance [CL_CR] = 90 ml/min) and decreased (CR_CL = 60 ml/min). Also, the killing activity of cefepime with and without tobramycin was compared to the killing activity of ceftazidime (2 g every 8 h) with and without tobramycin. The T > MIC was 100% in the central chamber except for the regimen in which cefepime was administered every 12 h and the CL_CR was 90 ml/min, which provided concentrations above the MIC for 92% of the dosing interval against the C31 (mucoid; MIC of cefepime, 4 μg/ml) isolate and for 75% of the interval against the C34 (nonmucoid; MIC of cefepime, 8 μg/ml) isolate. All cefepime and ceftazidime monotherapy simulations resulted in 99.9% killing of the nonmucoid isolate within 4 to 8 h and within 4 to 6 h, respectively. Against the mucoid isolate, 99.9% killing was achieved only with combination therapy. The results of this study indicate that cefepime dosed at 2 g every 12 h under conditions of normal renal function and every 24 h with decreased creatinine clearance (60 ml/min) is effective both as monotherapy and in combination therapy against a nonmucoid strain of P. aeruginosa. With cefepime MICs of 4 and 8 μg/ml, the single-agent regimens provided T > MIC values in the central chamber for 92 and ≥75% of the dosing interval against the mucoid and nonmucoid isolates, respectively. Cefepime dosed at 2 g every 12 h, with a creatinine clearance of 90 ml/min, and every 24 h, with a creatinine clearance of 60 ml/min, resulted in killing activity equivalent to that of ceftazidime dosed at 2 g every 8 h. None of the monotherapies provided adequate killing of the mucoid strain of P. aeruginosa despite drug concentrations being above the MIC for ≥92% of all dosing intervals. Finally, combination therapy with tobramycin and either cefepime or ceftazidime enhanced the killing of both the mucoid and nonmucoid P. aeruginosa isolates.

Cefepime is a new cephalosporin with a broad spectrum of activity against gram-positive cocci, enteric gram-negative bacilli, and Pseudomonas spp. It is widely used for the treatment of patients with serious infections, including pneumonia, complicated urinary tract infections, and febrile neutropenia. The pharmacokinetic characteristics of cefepime allow for a dosing regimen of 2 g every 12 h (q12h), with anticipated peak and trough concentrations of approximately 160 and 7 μg/ml, respectively, in patients with normal renal function. Cefepime has good activity even against strains of enteric gram-negative organisms, especially Enterobacter spp., which hyperproduce chromosomally encoded Bush class I β-lactamases (12). Several properties of cefepime (rapid passage through the outer membrane, low affinity for Bush class I β-lactamases, and high affinity for PBP 2 and PBP 3) (5, 6) are considered to be the basis for its excellent activity against organisms which hyperproduce β-lactamases. The third-generation cephalosporins have poor activity against these more resistant gram-negative organisms (12). The MICs of cefepime against some of these resistant isolates are eightfold or more lower than the corresponding MICs of ceftazidime (14). These lower MICs (often <1 μg/ml) are easily exceeded by the trough concentrations obtained by dosing cefepime q12h (11). However, the MIC of cefepime at which 90% of Pseudomonas aeruginosa isolates are inhibited is 4 to 8 μg/ml, and the concentration-time profile for q12h dosing may result in suboptimal drug concentrations and, thus, less killing. Rather than being treated with monotherapy, most infections caused by P. aeruginosa are treated with combination therapy, such as a β-lactam and an aminoglycoside, since this combination is often synergistic. Therefore, with combination therapy, the q12h dosing interval may be very effective against Pseudomonas infections. The dosage administration recommendations in the manufacturer’s prescribing information suggest changing the interval to q24h when the creatinine clearance (CR_CL) equals or falls below 60 ml/min. Again, this may result in lower trough concentrations and less killing activity against P. aeruginosa, although synergism with an aminoglycoside may provide effective killing. The objectives of this study were as follows: (i) to examine the killing activity of 2-g doses of cefepime administered at intervals of 8, 12, and 24 h with and without tobramycin against two clinical isolates of P. aeruginosa in a pharmacodynamic in vitro infection model; (ii) to compare the percentage of time above the MIC (T > MIC) for each of the regimens against P. aeruginosa; and (iii) to evaluate the area under the bactericidal time curve (AUBC) for each regimen. Two different levels of renal function were simulated: normal (CR_CL = 90 ml/min) and decreased (CR_CL = 60 ml/min). Also, the killing activity of cefepime with and without tobramycin was compared to the killing activity of ceftazidime dosed at 2 g q8h with and without tobramycin.

MATERIALS AND METHODS

Organisms.

Two clinical strains of P. aeruginosa were tested: C31, a mucoid isolate recovered from a cystic fibrosis patient, and C34, a nonmucoid isolate.

Susceptibility testing.

The MICs and MBCs of cefepime, ceftazidime, and tobramycin were determined by the microtiter broth dilution method in accordance with National Committee for Clinical Laboratory Standards guidelines (15). Mueller-Hinton broth (Difco) supplemented with calcium (25 mg/liter) and magnesium (12.5 mg/liter) was used for all susceptibility and model experiments.

In vitro model.

The in vitro model, consisting of two chambers, allows simulation of human pharmacokinetic parameters in the presence of bacteria (8). The 190-ml central compartment mimics the bloodstream, and in this chamber the pharmacokinetic parameters are simulated. The previously described infection model was modified to accommodate a SpectraPor 2 dialysis membrane (12,000- to 14,000-Da molecular mass cutoff) as the peripheral or infection compartment. One end of the membrane was tied into a knot, creating a sac which contained 15 ml of the organism suspension, providing a surface area-to-volume ratio of 4.8. The membrane allowed the passage of fresh medium and drug into the chamber but retained the organisms (approximately 10⁶ CFU/ml initially) within it. Drugs were injected through a port at the top of the model to achieve the desired peak concentration of each antimicrobial agent. Fresh medium was pumped into the central compartment via a peristaltic pump, and drug-containing medium was displaced at a rate equivalent to the half-life (t_1/2) of the antibiotic. The entire model apparatus was placed in a 37°C water bath. All experiments were performed over 48 h and tested in duplicate. The drug regimens simulated were as follows: cefepime at 2 g q8h and q12h with a t_1/2 of 2.3 h (3) (CR_CL ≈ 90 ml/min), with or without tobramycin at ∼6 mg/kg of body weight q24h; cefepime at 2 g q12h and q24h with a t_1/2 of 5 h (4) (CR_CL ≈ 60 ml/min), with or without tobramycin at ∼6 mg/kg q24h; and ceftazidime at 2 g q8h with a t_1/2 of 2.0 h (CR_CL ≈ 90 ml/min), with or without tobramycin at 6 mg/kg q24h. Growth controls were also used in the model over 48 h. The central-chamber target peak concentration for both cefepime and ceftazidime was 160 μg/ml, and that for tobramycin was 18 μg/ml.

Pharmacokinetics and pharmacodynamics.

The pharmacodynamic response for each regimen was evaluated by removing 0.1-ml samples from the infection chamber at various times over 48 h. These samples were serially diluted in normal saline, and 20-μl aliquots were plated in triplicate on tryptic soy agar (TSA). The 10-fold serial dilutions were adequate for minimizing antibiotic carryover. The colony counts were graphed versus time, and the AUBC from 0 to 48 h for each regimen was determined by the trapezoidal rule, using the program Lagran (version 2.1) (17).

The pharmacokinetic parameters for the central compartment were assessed by removing 0.5-ml samples over the time course of the experiment to ensure that the targeted peak concentration and t_1/2 of each drug were achieved. The timing of the samples varied with the dosing intervals of the regimens. From the central compartment the area under the curve (AUC) from 0 to 24 h for each drug regimen was determined by the trapezoidal rule, using the program Lagran (version 2.1) (17), and the AUC/MIC ratio for each organism and regimen was calculated. The peak concentration in the central chamber was also used to calculate the peak concentration/MIC ratio for each regimen. Drug concentrations at the site of the infection were also evaluated by removing 0.2-ml samples from the infection chamber over the course of the experiments. Cefepime and ceftazidime concentrations were determined by bioassay with Escherichia coli ATCC 25922 as the test organism. The assays had a limit of detection of 0.25 μg/ml and an r² of 0.995, and the coefficients of variation for drug concentrations of 5 and 0.25 μg/ml were <10%. Tobramycin concentrations were determined by bioassay, using Bacillus subtilis as the test organism. The assay had a limit of detection of 0.25 μg/ml and an r² of 0.992, and the coefficients of variation for drug concentrations of 10 and 0.25 μg/ml were <10%.

Emergence of resistance.

The development of resistance to cefepime or ceftazidime during therapy was assessed by inoculating samples from the 0-, 24-, and 48-h time points onto TSA containing drug at concentrations four and eight times its MIC for the organism. Colony counts from these antibiotic plates were compared to the number obtained on the non-drug-containing TSA plates to determine the fraction of resistant mutants.

Statistical analysis.

The AUBCs, T > MIC values, and colony counts at 24 and 48 h for each of the regimens were compared by analysis of variance and Tukey’s test for multiple comparisons. A P value of ≤0.05 was considered significant.

RESULTS

Susceptibility.

The MICs of cefepime, ceftazidime, and tobramycin against the C31 (mucoid) isolate were 4, 2, and 1 μg/ml, respectively, and their respective MBCs were 4, 8, and 2 μg/ml. Against the C34 (nonmucoid) isolate, their MICs were 8, 1, and 1μg/ml, respectively, and their respective MBCs were 8, 2, and 2 μg/ml.

Pharmacokinetics.

The mean peak concentrations of cefepime and ceftazidime ± the standard deviations (SDs) for all simulated regimens were 153.6 ± 7.3 and 175.4 ± 13.6 μg/ml, respectively. The average cefepime t_1/2s ± the SDs in the central chamber with CR_CL values of 90 and 60 ml/min were 2.4 ± 0.1 and 5.25 ± 0.8, respectively. The ceftazidime t_1/2 ± SD was 2.0 ± 0.1 h for a CL_CR of 90 ml/min. The mean tobramycin peak concentration ± SD for simulations with creatinine clearances of 90 and 60 ml/min was 15.5 ± 0.7 μg/ml, and the mean t_1/2s ± SDs were 2.2 ± 0.0 and 5.0 ± 0.2 h, respectively. The β-lactam drug concentrations obtained in the infection chamber are shown in Table 1. Each of the drugs achieved its peak concentration in the infection chamber within 4 h (the complete concentration profile is not shown). The T > MIC was 100% in the central chamber except for the cefepime q12h regimen with a CR_CL of 90 ml/min, which provided T > MIC for 92 and 75% of the interval against the C31 and C34 isolates, respectively.

TABLE 1.

β-Lactam concentrations in the infection bag

Drug and dose	CRCL (ml/min)	Drug concn ± SD (μg/ml) at h:			T > MIC (%) for strain C34a
		4	24	48
Cefepime
q8h	90	79.4 ± 2.3	50.2 ± 10.6	42.0 ± 22.6	100
q12h	90	40.6 ± 2.1	7.0 ± 0.0	7.8 ± 2.2	≥85
q12h	60	54.8 ± 9.7	32.5 ± 2.9	27.1 ± 4.5	100
q24h	60	49.7 ± 2.6	7.0 ± 0.4	6.3 ± 0.1	≥85
Ceftazidime, q8h	90	47.9 ± 5.2	30.3 ± 0.6	27.4 ± 1.7	100

^aEstimated percentage of the dosing interval during which the drug concentration was above the MIC. For strain C31, all values were 100%. 

Time-kill studies of nonmucoid isolate C34.

The results of the monotherapy regimens are shown in Fig. 1A and B, and those of the combination therapies are shown in Fig. 2. The AUBC for the monotherapy regimens ranged from 112 to 140 log CFU · h/ml, and for the combination regimens it ranged from 104 to 125 log CFU · h/ml. There was no statistically significant difference in the AUBCs or colony counts at 24 and 48 h between any of the monotherapy regimens with cefepime and ceftazidime. Also, there was no difference in the AUBCs or colony counts between any of the monotherapy cephalosporin regimens and the combination therapies. All cefepime and ceftazidime monotherapy simulations resulted in 99.9% killing within 4 to 8 h and within 4 to 6 h, respectively. All combination regimens resulted in 99.9% killing within 2 to 3 h of initiation of therapy. The peak concentration/MIC ratios for all of the cefepime regimens were similar (∼19), since all regimens simulated a 2-g dose. For the ceftazidime regimen, the peak concentration/MIC ratio was 175. All AUC/MIC ratios for the cephalosporin monotherapies were >200 except for that of the regimen consisting of cefepime at 2 g q12h with a CR_CL of 90 ml/min, which was 140. Tobramycin monotherapy resulted initially in rapid killing to the limit of detection, with regrowth at 24 h for all three simulations. There was less killing with the second dose of tobramycin than with the first, and regrowth recurred at 48 h. The peak concentration/MIC ratio for the simulation with a CR_CL of 60 or 90 ml/min was 15. The AUC/MIC ratios for the 60- and 90-ml/min simulations were 110 and 52, respectively. No cefepime or ceftazidime resistance developed during any therapy simulation.

FIG. 1.

Monotherapy cephalosporin (A) and tobramycin (B) regimens against the nonmucoid isolate C34. (A) Symbols: •, cefepime q8h with a CR_CL of 90 ml/min; , cefepime q12h with a CR_CL of 90 ml/min; , cefepime q24h with a CR_CL of 60 ml/min; ▾, cefepime q12h with a CR_CL of 60 ml/min; , ceftazidime q8h with a CR_CL of 90 ml/min; , growth control. (B) Symbols: •, tobramycin q24h with a CR_CL of 90 ml/min; , tobramycin q24h with a CR_CL of 60 ml/min; , growth control.

FIG. 2.

All combination regimens against the nonmucoid isolate C34. Symbols: •, cefepime q8h plus tobramycin q24h with a CR_CL of 90 ml/min; , cefepime q12h plus tobramycin q24h with a CR_CL of 90 ml/min; , cefepime q24h plus tobramycin q24h with a CR_CL of 60 ml/min; ▾, cefepime q12h plus tobramycin q24h with a CR_CL of 60 ml/min; , ceftazidime q8h plus tobramycin q24h with a CR_CL of 90 ml/min; , growth control.

Time-kill study of mucoid isolate C31.

The killing results of all β-lactam monotherapy regimens for the mucoid isolate are shown in Fig. 3. The cephalosporin therapies resulted in only about a 2 log₁₀ CFU/ml decrease in this mucoid isolate. The killing curves for all of the tobramycin monotherapy regimens (data not shown) were similar to those obtained for the C34 isolate. The tobramycin peak concentration/MIC and AUC/MIC ratios were also the same as those for the C34 isolate. The AUBCs for all of the cefepime and ceftazidime monotherapy regimens were equivalent and ranged from 214 to 233 log CFU · h/ml. Also, there was no difference in colony counts at 24 or 48 h between any of the β-lactam regimens. The peak concentration/MIC ratio for each of the cefepime regimens was ∼38, and that for the ceftazidime regimens was 87. The AUC/MIC ratio for each of the monotherapy cefepime and ceftazidime regimens was ≥430 except for that of the q12h regimen with a CR_CL of 90 ml/min, which was 280. The results of the combination regimens are shown in Fig. 4. All combinations tested achieved 99.9% killing within 2 to 3 h. These low counts were essentially maintained over 48 h for most regimens. At 24 h, three regimens demonstrated increases in colony counts of ≤1 log over the 12-h time point, and at 48 h, five regimens had increases in colony counts of <2 logs. These increased colony counts at 24 and 48 h were likely due to detachment of membrane inner surface-adherent bacteria during the sampling process. For each regimen, the number of bacteria adherent to the inner surface of the membrane at 48 h was at least 2 log₁₀ CFU/ml more than the planktonic counts observed at this time. The range of AUBCs for each of the combination therapies was 102 to 125 log CFU · h/ml, and these values were not significantly different from one another. Each of the combination regimens achieved statistical significance over each of the corresponding monotherapy regimens with respect to colony counts and AUBC (P ≤ 0.02). No cefepime or ceftazidime resistance developed during any therapy simulation.

FIG. 3.

Cephalosporin monotherapy against P. aeruginosa mucoid isolate C31. Symbols: •, cefepime q8h with a CR_CL of 90 ml/min; , cefepime q12h with a CR_CL of 90 ml/min; , cefepime q24h with a CR_CL of 60 ml/min; ▾, cefepime q12h with a CR_CL of 60 ml/min; , ceftazidime q8h with a CR_CL of 90 ml/min; , growth control.

FIG. 4.

All combination regimens against P. aeruginosa mucoid isolate C31. Symbols: •, cefepime q8h plus tobramycin q24h with a CR_CL of 90 ml/min; , cefepime q12h plus tobramycin q24h with a CR_CL of 90 ml/min; , cefepime q24h plus tobramycin q24h with a CR_CL of 60 ml/min; ▾, cefepime q12h plus tobramycin q24h with a CR_CL of 60 ml/min; , ceftazidime q8h plus tobramycin q24h with a CR_CL of 90 ml/min; , growth control.

DISCUSSION

The pharmacodynamic parameter most closely associated with efficacy for β-lactam therapy is the T > MIC for the organism. The relationship of T > MIC to antibacterial effect is best described by the E_max model (10). Using this model, Craig found an in vivo bacteriostatic effect against gram-negative bacilli when β-lactam concentrations exceeded the MIC for 40% of the dosing interval, and maximal killing was approached when the MIC was exceeded for 60 to 70% of the interval. In a neutropenic mouse model, Craig and colleagues (9) found that when cephalosporin concentrations were above the MIC for about 40 to 60% of the dosing interval, a static effect was achieved against gram-negative bacilli. As stated earlier, it has been questioned whether cefepime dosed by the manufacturer’s recommendations would provide a sufficient drug concentration over the interval to produce good killing and satisfactory T > MIC against P. aeruginosa. In assessing the T > MIC in the central chamber for the various cefepime regimens against the two clinical isolates of P. aeruginosa used in this study, concentrations were above the MIC for the mucoid isolate 100% of the interval for all regimens except for the q12h regimen with a CR_CL of 90 ml/min, which covered 92% of the dosing interval. The results were the same for the nonmucoid isolate, except that due to the higher MIC of 8 μg/ml, the q12h regimen with a t_1/2 of ∼2.3 h covered only 75% of the interval. The regimen consisting of cefepime dosed at 2 g q12h with simulation of normal renal function also resulted in 99.9% killing of the nonmucoid isolate within the first 24 h, with no resistance developing during the 48-h experiment. When adjusting the cefepime dose to q24h for a CL_CR of 60 ml/min, the T > MIC was 100% of the interval against the nonmucoid isolate, and 99.9% killing was also achieved within 24 h with no development of resistance. Thus, the manufacturer’s suggested dosing appears to provide effective drug concentration profiles for nonmucoid P. aeruginosa based on T > MIC values and achievement of 99.9% killing activity.

In the infection chamber, all cefepime regimens provided concentrations above the MIC against the mucoid isolate, which had a cefepime MIC of 4 μg/ml for 100% of the dosing intervals. Against the nonmucoid isolate, the cefepime q12h regimen with a CR_CL of 90 ml/min and the q24h regimen with a CR_CL of 60 ml/min provided infection site concentrations above the MIC of 8 μg/ml for at least 85% of the interval; with all other regimens, the concentration exceeded the MIC 100% of the time. The exact percentage cannot be determined for these two regimens since samples collected from the infection chamber were insufficient to perform a complete pharmacokinetic analysis. The peak concentrations of cefepime (40 to 50 μg/ml) obtained in the infection chamber with the CR_CL simulations of 60 and 90 ml/min were similar to those obtained in cantharides-induced (noninflammatory) blisters (30 to 35 μg/ml) in healthy volunteers (16); however, they were lower than those obtained in the inflammatory fluid of suction-induced blisters, which were around 80 μg/ml (13). The estimated rate of elimination of cefepime from the infection chamber, or drug t_1/2, varied with each dosing regimen (for example, with the q8h regimen, the t_1/2 was about 6 h, while with the q12h and CR_CL and = 90 regimen, the t_1/2 was about 3.1 h). In general, these values are lower than the elimination rate from tissue in vivo, in part due to the surface area-to-volume ratio of the infection chamber and rates of equilibration between the central and infection chambers (7, 18, 19). However, since the relationship between drug concentrations at the site of the infection to the killing activity of various drugs and drug regimens in vivo has not been fully elucidated, interpretation of the infection chamber concentrations and T > MIC values should be limited at this time.

In contrast to the killing results obtained with the nonmucoid C34 isolate, the association between T > MIC and killing or efficacy of the various regimens was much less predictable for the mucoid P. aeruginosa C31 isolate. Each regimen was above the MIC for ≥92% of the dosing interval; however, the regimens reduced the inoculum by only about 1.5 to 2 log CFU/ml. The reason for the less-efficient killing is likely a decreased penetration of cefepime or ceftazidime through the mucous and/or biofilm produced by and thus surrounding this strain of P. aeruginosa. It is well known that P. aeruginosa can produce a biofilm which serves as a protective barrier for adherent organisms. Monotherapy often results in minimal killing activity against these strains, whereas combination therapy has variable activity. In two studies, Anwar and colleagues (1, 2) investigated young versus old biofilms and the ability of tobramycin plus piperacillin to kill mucoid P. aeruginosa. The results suggested that P. aeruginosa with young biofilm (<2 days old) was effectively killed by the combination regimen whereas the cells associated with the old biofilm were not. Since the C31 isolate continuously expresses the mucoid phenotype, it may readily and rapidly produce biofilm, protecting it from single-antibiotic activity; thus, only a minimal decrease in colony count results. Combination therapy was effective against this isolate; this may reflect good activity against relatively young biofilm as was seen in the Anwar studies. Since biofilm formation was not assessed in this study, one can only postulate that biofilm accounts for the differences in the killing activities observed for the two P. aeruginosa isolates.

The results of this study indicate that cefepime dosed at 2 g q12h under conditions of normal renal function is effective both as monotherapy and in combination therapy with tobramycin against nonmucoid strains of P. aeruginosa. However, combination therapy is often preferred for treating P. aeruginosa infections. With cefepime MICs of 4 and 8 μg/ml, the single-agent 2-g regimens provided a T > MIC in the central chamber for 92 and 75% of the dosing intervals, respectively, which in the in vitro model produced a bactericidal effect against the nonmucoid strain. The dose change to 2 g q24h in patients with a CR_CL of ≤60 ml/min, as recommended by the manufacturer, provided adequate drug concentrations and T > MIC values to effectively treat infections caused by nonmucoid P. aeruginosa with a drug MIC of ≤8 μg/ml. Against a mucoid strain of P. aeruginosa, monotherapy did not provide adequate killing despite drug concentrations being above the MIC for ≥92% of all dosing intervals. Cefepime dosed at 2 g q12h with normal renal function or q24h with decreased renal function was as effective as ceftazidime dosed at 2 g q8h against both strains of P. aeruginosa. Finally, combination therapy with tobramycin plus either cefepime or ceftazidime enhanced the killing of both the mucoid and nonmucoid P. aeruginosa isolates.