Lipman, Jeffrey FFA (Crit Care), FJFICM; Wallis, Steven C. PhD; Boots, Robert J. FRACP, FJFICM
Cefpirome and cefepime are new fourth-generation cephalosporins (1,2). They both have good Gram-negative cover, with some Gram-positive cover, and can be used for nosocomial sepsis, covering most organisms recovered from critically-ill patients, particularly if used as empirical therapy (1,2). The suggested dosage regimens of the two drugs are the same, and the minimum inhibitory concentrations (MIC) of the two for common organisms are similar (1,2). Which antibiotic should then be chosen for clinical use?
Studies of β-lactam antibiotics on Gram-negative bacilli show a bactericidal activity that is slow, time dependent, and maximal at relatively small concentrations (3). Bacterial killing is primarily related to the time that levels in tissue and plasma exceed a certain threshold. If antibiotic concentrations decrease below this threshold in the in vitro models, bacterial proliferation is immediately resumed (4,5). For the optimal efficacy of β-lactams, it is thus important that the dosing regimen maintain adequate plasma levels for as long as possible through the dosing interval. Preliminary clinical work suggests that low levels may predispose to resistance developing (6). It is not surprising, then, that dosing regimens for β-lactam antibiotics are being reevaluated to keep plasma levels above certain thresholds for longer periods of the dosing interval (7–12), even to the extent of continuous infusions (13–17). An editorial in the intensive care (IC) literature has highlighted the problems of low plasma levels (12).
Dosage recommendations for β-lactam antibiotics, including cephalosporins, incorporate a dosage reduction in the presence of renal dysfunction (1,2). It is common practice to prescribe a standard dose of an antibiotic assuming normal renal function, unless there is significant renal impairment. Previous studies have demonstrated that the standard dosage recommendations in IC patients without renal dysfunction can result in very low levels of β-lactam antibiotics at the end of the dosing intervals (7,9,11,13). This holds true for the fourth-generation cephalosporins cefepime and cefpirome (7,11). Anecdotally it seemed to us that cefpirome levels after 2 g every 12 h were less than cefepime levels with the same dosage (Fig. 1) (7,11).
Questions then arise: What should be considered in choosing which fourth-generation cephalosporin to use? What determinants should be used in choosing the correct dose and dosing interval to maintain better plasma levels for more patients when prescribing these drugs?
In our two previous studies (7,11), we found that some patients had creatinine clearances in excess of the quoted upper range of 80–120 mL/min for our laboratory and the quoted upper norm in the literature (18). This may not be such an unusual occurrence; in a previous study we performed in IC patients, similarly large creatinine clearances were also obtained (19). The phenomenon of large creatinine clearance has also been shown by other authors to occur in some critically-ill patients (20,21). In view of cephalosporin clearance being primarily renal (1,2), we surmised that increased creatinine clearances would result in increased cephalosporin clearance.
The aim of this study was to compare the plasma pharmacokinetics of two fourth-generation cephalosporins, cefpirome and cefepime, in the IC population, specifically addressing the problems of low plasma levels and, hence, time of antibiotic above MICs. We also wanted to examine the relationship of drug clearance to creatinine clearance in patients with normal serum creatinine levels.
Two separate pharmacokinetic studies with very similar protocols, one study for cefepime and the other for cefpirome, were conducted in IC patients (7,11). After IRB approval had been obtained, informed consent was obtained from patients or next of kin for blood sampling of patients receiving clinically indicated antibiotic regimens (either cefepime or cefpirome). Critically-ill adult patients <76 yr old with sepsis (systemic inflammatory response syndrome plus an infected site) and normal renal function receiving clinically indicated cefepime or cefpirome therapy were enrolled.
First-dose pharmacokinetics (onset of cefepime or cefpirome therapy) and a later pharmacokinetic profile (between Day 3 and Day 6 of therapy, inclusive) were measured in enrolled patients. First-dose pharmacokinetics were measured in 13 patients receiving cefepime and in 12 patients receiving cefpirome. Later-dose pharmacokinetics were measured in 12 patients receiving cefepime and in 11 patients receiving cefpirome. A priori–defined dropouts in the cefepime study (3 of 13 patients) and in the cefpirome study (2 of 12 patients) were due to development of renal dysfunction, as defined by abnormal serum creatinine levels. In all these patients, the deterioration in renal function was attributed to sepsis and not to the antibiotic. The second pharmacokinetic profile was performed in 2 of the 3 cefepime dropouts and in 1 of the 2 cefpirome dropouts.
The antibiotic (2 g of cefepime or cefpirome) was infused IV over 3 min. Blood samples were taken before infusion and 5, 10, 20, 30, 60, 90, 120, 240, 360, 480, 600, and 720 min after the infusion was complete. Each 10-mL sample of blood was drawn into a heparinized Vacutainer and stored at 0°C–4°C before centrifugation. Samples were kept at 4°C during centrifuge (2000 g for 10 min), and the plasma was stored at −20°C until it was transferred to a −80°C freezer.
Concentrations of cefepime and cefpirome in plasma were measured by high-performance liquid chromatographic methods modified in-house (7,11). Noncompartmental pharmacokinetic analysis of the plasma concentration versus time data was performed with the WinNonlin software package (Scientific Consulting Inc.).
Serum creatinine levels were routinely measured at least once a day as standard practice in the intensive care unit (ICU). An 8-h urine sample was collected during each sampling period, and the volume and urine creatinine were measured. Creatinine levels were quantified with an enzymatic method free from the cefpirome interference that has been observed with the Jaffe method of creatinine measurement (22). Creatinine clearance was calculated from the serum creatinine measurement that coincided with the sampling interval and from the urine data.
Initial and subsequent dose data for each drug were combined to compare the two treatments and examine the relationship of plasma drug levels to patient creatinine clearance, adjusting for patient weight by using multiple-regression modeling. Regression modeling and descriptive analysis were performed with Stata Version 4 (Stata Corp., College Station, TX). Parametric and nonparametric tests were applied as appropriate. To directly compare the two cephalosporins, we arbitrarily chose a reference level of 4 mg/L.
The mean cefepime and cefpirome concentrations are plotted against time in Figure 1. Patients who developed renal dysfunction and were excluded from data analysis in our original studies (7,11) are included here. One of the three dropouts in the cefepime study and one of the two dropouts in the cefpirome study had only the initial dose and no subsequent-dose pharmacokinetics measured.
The summary pharmacokinetic and concentration data for cefepime and cefpirome (initial and subsequent doses) are shown in Table 1. The mean creatinine clearance and drug clearance values appear slightly larger for the cefpirome study than for the cefepime study. This is apparently reflected in the lower area under the curve (AUC) and by the percentage of the 12-h sampling interval for which the drug concentration was >4 mg/L of the cefpirome population. However, statistical analysis of the combined initial- and subsequent-dose data for each drug revealed no differences in the volume of distribution and time in the dosing interval spent above a representative MIC of 4 mg/L for the two drugs (Table 2). A significant difference in creatinine clearance was seen despite serum creatinine levels being within the reported normal range. Although there was no significant difference in drug clearance, which could be explained by the small numbers in this study, any apparent difference in drug clearance can largely be attributed to differences in creatinine clearance. Figure 2 notes a linear relationship between the creatinine clearance and drug clearance.
Despite clinically normal renal function, on the basis of plasma creatinine concentration, 54% of patients’ plasma concentrations of antibiotic were less than the representative MIC of 4 mg/L for more than 20% of the dosing interval. Thirty-four percent of patients had creatinine clearances more than 144 mL/min (20% more than an expected maximum of 120 mL/min). Although the proportion of patients with a large creatinine clearance was larger in the cefepime group than in the cefpirome group, this was not statistically significant (44% versus 28%; P = 0.36). Only creatinine clearance was an independent predictor of antibiotic clearance, with adjustment for the cephalosporin used and patient weight (r2adj = 0.78; P < 0.0001). Similarly, time of plasma concentration less than MIC was predicted only by creatinine clearance (r2adj = 0.7; P < 0.0001).
In dosing of cephalosporin antibiotics as part of a comparative trial, we found that some septic patients had creatinine clearances that were much more than an expected normal value of 120 mL/min. We and others have documented this phenomenon previously (19–21). A large creatinine clearance directly influences the trough concentration of the drug in plasma and the percentage of time of the dosing interval in which the plasma concentration is less than MIC. This has implications for the efficacy of the treatment of sepsis. Most often, the plasma creatinine concentration is used as a surrogate for creatinine clearance. If it is normal and if there are not the well described phenomena that decrease serum creatinine, such as small size and little muscle bulk, creatinine clearance is assumed to be normal and is generally not more than 120 mL/min. In this septic patient group in the ICU, we have demonstrated that underdosing of these cephalosporins is possible despite presumed normal renal function when, in fact, the patient has supranormal renal function and large drug excretion.
When comparing 2-g 12-hourly IV cefpirome with the same dosage of cefepime, using 25 patients with almost 50 pharmacokinetic profiles, we could demonstrate no statistical difference in common pharmacokinetic variables between the two drugs. The lack of statistical difference may be due to sample size, in that a larger comparative study may show statistical difference (see Fig. 1). In our study, the peak and trough levels of the drugs, volume of distribution, terminal half-life, and clearances of the drugs were all similar. Even time more than MIC of the two drugs was not statistically different. Whatever difference we found could largely be attributed to creatinine clearance differences in the two studies.
Generally, a standard dose of an antibiotic is prescribed assuming a normal renal function, and dosage reductions occur in the setting of renal impairment. We have shown that with the suggested dosing regimens, in some ICU patients, plasma levels of these drugs reach very low levels (7,11,23,24). There are also other confirmatory studies (13,14,16). In this setting, individual differences in the pharmacokinetic profiles of the antibiotics may be important. In patients with normal serum creatinine, the dosages prescribed in the proprietary product information are usually administered. No allowance is made for patients who may have greater creatinine clearance than the serum creatinine would suggest (19–21). It is generally not considered that the creatinine clearance, which is the principle determinant of cephalosporin clearance (1,2), may be more than the assumed normal range in the critically ill as a result of high glomerular filtration. We believe that this is the primary cause of the very low levels of some antibiotics and specifically cephalosporins in some ICU patients (7,11,13,14,16,25–27).
Our data are consistent with findings that younger overweight people have larger creatinine and drug clearance (P < 0.001; data not shown) (18). This is when the serum creatinine is within the normal reported ranges. These patients are at risk of underdosing, resulting in significant periods when the plasma antibiotic concentrations are less than the MIC. It is important to be aware that the dosing of antibiotics in the critically ill adult must take into account not only impaired renal function, but also increased renal clearance. An average creatinine clearance should not be assumed on the basis of a normal creatinine concentration. Standard antibiotic dosing based on a normal creatinine clearance may lead to significant underdosing and time spent with trough antibiotic concentrations less than MIC.
In IC, when patients present with the inflammatory response associated with sepsis, it is common practice to fluid-load these patients, and if their blood pressure is still not corrected, inotropic drugs are often prescribed. It is therefore not surprising that patients with sepsis often have more than normal cardiac outputs (27,28). We assume that the phenomenon of increased cardiac output will produce an increased renal preload, thereby increasing the glomerular filtration rate and creatinine clearance. We therefore attribute the increased creatinine clearance to this increased cardiac output from the above; our patient population was also young and had normal renal function.
With severe sepsis, renal function and, in fact, cardiac function may be diminished, resulting in altered creatinine clearances. In such circumstances, renally excreted antibiotics will accumulate in plasma.
Although we arbitrarily used a level of 4 mg/L for comparison purposes in this study, our findings would have been the same for any other plasma concentration. Our study was not designed to measure outcomes, and therefore we cannot comment on the relevance of the concentrations we have shown to occur with standard dosages (7,11). Time above MIC is the important factor for β-lactam antibiotic efficacy. In relation to cefepime and cefpirome, MIC90 for many ICU-acquired organisms (particularly Pseudomonas aeruginosa) would be in the region of 16 mg/L (29). The individual trials we used for this study both demonstrated low trough levels of the cephalosporins (7,11). In vitro data do suggest the development of resistance with suboptimal β-lactam dosages (30).
To overcome low levels of β-lactam antibiotics, one can either shorten the dosing interval (7,11) or use continuous infusions (7,11,13–17). As well as providing the advantage of less antibiotic requirement each day, continuous infusions may also eliminate the necessity of predicting which patients need shortened dosing intervals. Alternatively, creatinine clearances could be measured more routinely in ICUs. Sladen et al. (31) have suggested that a two-hour creatinine clearance measurement is accurate and useful. Antibiotic doses could then be altered accordingly. This may allow prediction of which patients would have low antibiotic levels at the end of their dosing interval, and hence allow for dosage adjustments.
Unfortunately, unlike Pea et al. (27) and Lugo and Castaneda-Hernandez (25), we did not look at concomitant drug use or interventions, such as fluid loading, in our patients. We do agree with these authors that these interventions are important to be aware of when dosing antibiotics in such patients, particularly in the absence of renal dysfunction. Another limitation of our study is that morbidity and mortality were not end-points. A much larger study would be needed to determine such outcomes. Future studies may, on the one hand, compare antibiotic levels and cardiac output in the presence of normal renal function and, on the other hand, be needed to determine whether low cephalosporin levels do actually result in increased morbidity, mortality, or both (6).
In conclusion, we have shown that the pharmacokinetics of cefepime and cefpirome are similar in IC patients. Therefore, in the presence of similar MICs of the two antibiotics, other issues, such as drug availability, costs, etc., should be used in choosing one drug over the other. One of the main determinants of clearance of both these drugs is creatinine clearance. We have shown that some IC patients have very large creatinine clearances that result in very low levels of these fourth-generation cephalosporins. Either shortening the dosage interval of these drugs or using continuous infusions would prevent low cephalosporin levels and keep levels greater than MIC for longer. Although we used a level of 4 mg/L for comparative reasons in this study, 16 mg/L (MIC90 for many ICU-acquired organisms, particularly P. aeruginosa) is probably a more relevant target plasma level. In view of the lack of bedside measurement of cephalosporin levels, we also suggest that more frequent use of creatinine clearance measurement be made to allow prediction of low cephalosporin levels at the bedside.
1. Barradell LB, Bryson HM. Cefepime: a review of its antibacterial activity, pharmacokinetic properties and therapeutic use. Drugs 1994; 47: 471–505.
2. Wiseman LR, Lamb HM. Cefpirome: a review of its antibacterial activity, pharmacokinetic properties and clinical efficacy in the treatment of severe nosocomial infections and febrile neutropenia. Drugs 1997; 54: 117–40.
3. Vogelman B, Craig WA. Kinetics of antimicrobial activity. J Pediatr 1986; 108: 835–40.
4. Fantin B, Farinotti R, Thabaut A, Carbon C. Conditions for the emergence of resistance to cefpirome and ceftazidime in experimental endocarditis due to Pseudomonas aeruginosa. J Antimicrob Chemother 1994; 33: 563–9.
5. Mouton JW, den Hollander JG. Killing of Pseudomonas aeruginosa during continuous and intermittent infusion of ceftazidime in an in vitro pharmacokinetic model. Antimicrob Agents Chemother 1994; 38: 931–6.
6. Davies J, Wendon J, Wyncoll D, Wade J. Antibiotic infusions reduce the incidence of resistant organisms in liver intensive care [abstract]. Intensive Care Med 2000; 26: s269.
7. Lipman J, Wallis SC, Rickard C. Low plasma cefepime levels in critically ill septic patients: pharmacokinetic modeling indicates improved troughs with revised dosing. Antimicrob Agents Chemother 1999; 43: 2559–61.
8. Turnidge JD. The pharmacodynamics of beta-lactams. Clin Infect Dis 1998; 27: 10–22.
9. Benko AS, Cappelletty DM, Kruse JA, Rybak MJ. Continuous infusion versus intermittent administration of ceftazidime in critically ill patients with suspected gram-negative infections. Antimicrob Agents Chemother 1996; 40: 691–5.
10. Lipman J, Crewe Brown HH, Saunders GL, Gous AG. Subtleties of antibiotic dosages: do doses and intervals make a difference in the critically ill? S Afr J Surg 1996; 34: 160–2.
11. Lipman J, Wallis SC, Rickard CM, Fraenkel D. Low cefpirome levels during twice daily dosing in critically ill septic patients: pharmacokinetic modeling calls for more frequent dosing. Intensive Care Med 2001; 27: 363–70.
12. Crokaert F. Pharmacodynamics, a tool for a better use of antibiotics. Intensive Care Med 2001; 27: 340–3.
13. Hanes SD, Wood GC, Herring V, et al. Intermittent and continuous ceftazidime infusion for critically ill trauma patients. Am J Surg 2000; 179: 436–40.
14. Burgess DS, Summers KK, Hardin TC. Pharmacokinetics and pharmacodynamics of aztreonam administered by continuous intravenous infusion. Clin Ther 1999; 21: 1882–9.
15. Egerer G, Goldschmidt H, Salwender H, et al. Efficacy of continuous infusion of ceftazidime for patients with neutropenic fever after high-dose chemotherapy and peripheral blood stem cell transplantation. Int J Antimicrob Agents 2000; 15: 119–23.
16. Angus BJ, Smith MD, Suputtamongkol Y, et al. Pharmacokinetic- pharmacodynamic evaluation of ceftazidime continuous infusion vs intermittent bolus injection in septicaemic melioidosis. Br J Clin Pharmacol 2000; 49: 445–52.
17. Frame BC, Facca BF, Nicolau DP, Triesenberg SN. Population pharmacokinetics of continuous infusion ceftazidime. Clin Pharmacokinet 1999; 37: 343–50.
18. Siersbaek-Nielsen K, Hansen JM, Kampmann J, Kristensen M. Rapid evaluation of creatinine clearance. Lancet 1971; 1: 1133–4.
19. Wells M, Lipman J. Measurements of glomerular filtration in the intensive care unit are only a rough guide to renal function. S Afr J Surg 1997; 35: 20–3.
20. Snider RD, Kruse JA, Bander JJ, Dunn GH. Accuracy of estimated creatinine clearance in obese patients with stable renal function in the intensive care unit. Pharmacotherapy 1995; 15: 747–53.
21. Brown R, Babcock R, Talbert J, et al. Renal function in critically ill postoperative patients: sequential assessment of creatinine osmolar and free water clearance. Crit Care Med 1980; 8: 68–72.
22. Grotsch H, Hajdu P. Interference by the new antibiotic cefpirome and other cephalosporins in clinical laboratory tests, with special regard to the “Jaffe” reaction. J Clin Chem Clin Biochem 1987; 25: 49–52.
23. Lipman J, Gomersall C, Gin T, et al. Continuous infusion ceftazidime in intensive care: a randomised controlled trial. J Antimicrob Chemother 1999; 43: 309–11.
24. Young RJ, Lipman J, Gin T, et al. Intermittent bolus dosing of ceftazidime in critically ill patients. J Antimicrob Chemother 1997; 40: 269–73.
25. Lugo G, Castaneda-Hernandez G. Relationship between hemodynamic and vital support measures and pharmacokinetic variability of amikacin in critically ill patients with sepsis. Crit Care Med 1997; 25: 806–11.
26. Tang GJ, Tang JJ, Lin BS, et al. Factors affecting gentamicin pharmacokinetics in septic patients. Acta Anaesthesiol Scand 1999; 43: 726–30.
27. Pea F, Porreca L, Baraldo M, Furlanut M. High vancomycin dosage regimens required by intensive care unit patients cotreated with drugs to improve haemodynamics following cardiac surgical procedures. J Antimicrob Chemother 2000; 45: 329–35.
28. Parrillo JE, Parker MM, Natanson C, et al. Septic shock in humans: advances in the understanding of pathogenesis, cardiovascular dysfunction, and therapy. Ann Intern Med 1990; 113: 227–42.
29. Karchmer AW. Cephalosporins. In: Mandell GL, Bennett JE, Dolin R, eds. Mandell, Douglas & Bennett’s principles and practice of infectious diseases. 5th ed. Philadelphia: Churchill Livingstone, 2000: 274–91.
30. Odenholt I, Gustafsson I, Lowdin E, Cars O. Suboptimal antibiotic dosage as a risk factor for selection of penicillin-resistant Streptococcus pneumoniae: in vitro kinetic model. Antimicrob Agents Chemother 2003; 47: 518–23.
31. Sladen RN, Endo E, Harrison T. Two-hour versus 22-hour creatinine clearance in critically ill patients. Anesthesiology 1987; 67: 1013–6.