Cefotaxime

John F. Cade, M.D., Ph.D., F.C.C.P.; Jeffrey Presneill, M.B.; Craig Keighley, M.B., F.C.C.P.; and Vincent Sinickas, M.B., Ph.D.

The optimal antibiotic dosage in serious chest infections is not established and commonly used regimens may well be excessive. We have compared the efficacy of a low dose of cefotaxime (2 g every 12 h) with a more usual dose (2 g every 8 h) in a prospective, randomized study of the treatment of chest infections in the seriously ill. Fifty intensive care unit patients received either regimen for five days. The two groups appeared demographically comparable. Clinical resolution occurred in 86 percent, no change occurred in 4 percent, and deterioration occurred in 10 percent. Microbiologic clearance occurred in 52 percent of those in whom a pathogen was isolated (46 percent of patients). There was no significant difference in clinical or microbiologic response between the two regimens. It is concluded that cefotaxime in a dose of 2 g twice daily is effective in the treatment of serious chest infections.

Chest infection is a major problem in the critically ill patient.[1-7] Such infections are mostly hospital-acquired complications and are thus usually due to Gram-negative microorganisms.[2-4,6,8] They are commonly treated with a third-generation cephalosporin, such as cefotaxmine,[7-9] but the dosage regimen used may vary considerably.[10-14]

The usual dosage recommendation for cefotaxime in serious infections has been 6 g or more daily in three or more divided doses, a common regimen being 2 g thrice daily.[14,15] Although lower doses are used in less severe infections and collation of results in patients given greater and less than 4 g daily has shown similar cure rates, there is a lack of direct comparative information on the efficacy of lower doses in serious infections.[12,14]

We have compared a relatively low dose of cefotaxime (viz 4 g daily, given as 2 g twice daily) with the more traditional dose (6 g daily, given as 2 g thrice daily), in the treatment of chest infections in seriously ill patients, in a prospective randomized controlled single-blind study, using both clinical and microbiologic responses as end points. This study has implications not only for the cost of treatment, but also for the prediction of effective dosing regimens when these are based solely on classic pharmacokinetic considerations.

PATIENTS AND METHODS

Patients

Patients eligible for entry were those seriously ill or injured and in a general intensive care unit (ICU). The specific criteria for entry included the presence of three or more of the following[15]: macroscopically purulent sputum, chest roentgenographic evidence of recent consolidation (<48 h), fever (>38 [degrees] C), and leukocytosis (>12,000 [mm.sup.-3]).

The criteria for exclusion were either of the following[15]: known or suspected need for another antibiotic or antibiotic requirement for extrathoracic infection with Gram-negative microorganisms. The former exclusion applied if the known or suspected microorganism was or was likely to be resistant to cefotaxime. The latter exclusion was because the assessment of the response of the chest infection could have then been confounded.

The concomitant or continuing indication for certain other prophylactic antibiotics was not a criterion for exclusion, provided the response of the chest infection was not thereby obscured. For example, flucloxacillin for prophylaxis after orthopedic or vascular surgery, penicillin after splenectomy, and metronidazole following bowel surgery were permitted, as these antibiotics had no activity against aerobic Gram-negative bacilli. Such additional antibiotic use was recorded separately and the microbiologic responses in these patients was analyzed separately to confirm that the additional antibiotic had no confounding effect on the assessment of cefotaxime (in the event only one patient receiving concomitant antibiotics had a microbiologic cure).

Patient demographic data were recorded, including underlying condition, severity of illness class,[16] therapeutic intervention score (TISS),[17] acute physiology and chronic health evaluation (APACHE II),[18] duration of ICU stay, and outcome.

The tracheobronchial aspirate in ventilated patients and expectorated sputum in nonventilated patients were examined microbiologically by microscopy and culture within 24 h either side of entry and of completion of the study. The culture was based on the Gramstain, using accepted criteria.[19] Sensitivity testing was carried out on all pathogens and was by agar dilution, using standard methods and break points.[20]

Treatment

Patients were randomized to receive cefotaxime (Claforan, Roussel Pharmaceuticals) 2 g intravenously (IV) either every 8 h or every 12 h. A double-blind design was not considered reasonable as it would have required injections every 4 h, every second one being placebo in the 8-h treatment group and two of three being placebo in the 12-h treatment group. Treatment was given for five days, unless either complete cure or therapeutic failure requiring another antibiotic occurred sooner. Randomization was by means of a computer-generated set of random numbers in blocks of ten. The number of patients chosen to study was based on an 80 percent power to show a global clinical score of 62 percent as different from a score of 50 percent (the estimated "benchmark") with a significance of 0.05. The clinical score of 62 percent was chosen because it was the score produced by the less effective regimen in a previous comparative study.[15] [TABULAR DATA OMITTED]

Assessment

The primary response was the clinical response as it could be analyzed on an intention-to-treat basis.[15] The microbiologic results were used to show comparability between groups and to support the clinical assessment with an efficacy analysis.

The clinical response was assessed on the basis of change between entry and completion of the study in the four objective entry criteria (viz, purulent sputum, chest roentgenographic changes, fever, and leukocytosis). Improvement was recorded if there was resolution of the abnormal criterion; deterioration was recorded if there was any worsening of an abnormal criterion or if a previously normal criterion became abnormal (ie, would have met entry requirements). The response was coded separately for each of these criteria, using the following semiquantitative scoring system: improvement was scored as 1, no change as 2, and deterioration as 3. In the event that the specific criterion was initially normal and remained normal, it was omitted as it was unassessable. To avoid bias, no subjective global assessments were made by the investigators.

The scores for the clinical responses were summed, since there was a semiquantitative grading of each parameter ranging from 1 (best response) to 3 (worst response). The summed scores were divided by the maximum possible score in any individual patient to give a final global score, expressed as a percentage. Thus, if all five clinical criteria were assessable in any individual patient, the best possible score would be 5 (33 percent) of 15 and the worst possible score 15 (100 percent) of 15; no change would be 10 (67 percent) of 15, improvement would be any score <67 percent, and deterioration would be score any score >67 percent.

The microbiologic response was classified as cure, failure, or not assessable.

Cure indicated the presence of an initial pathogen and absence of any subsequent pathogen. It was subclassified as to whether other antibiotics were given concomitantly (as indicated above).

Failure was subclassified into persistence of a sensitive initial pathogen, persistence of a resistant initial pathogen, occurrence of resistance during treatment, and superinfection. Failure could occur with more than one code in an individual patient.

Not assessable indicated that both initial and subsequent cultures were pathogen free or that the result of either culture was unvailable.

Hypothesis and Analysis

For the purposes of statistical analyses, the null hypothesis of the study was that a dose of 2 g every 12 h was not worse than 2 g every 8 h (it was considered that the lower dose could not be better). The data were tabulated using a database program on a microcomputer, so that the data could be checked and analyzed blindly, ie, without reference to a specific treatment regimen until the final analyses were complete.

The two groups of patients were compared for demographic matching, clinical response, and microbiologic response. Continuous variables were compared using the unpaired t test and discrete variables using the [X.sup.2] test. For the analyses of matching, the tests of significance were two-tailed as a difference either way was sought. For the response analyses, the tests of significance were one-tailed as the original null hypothesis was that the lower dose was not worse than the higher dose.

The clinical response was analyzed on an intention-to-treat basis, with all patients included. The microbiologic response was analyzed on an efficacy basis, with only patients with respiratory pathogens included. [TABULAR DATA OMITTED]

Ethics

This study was approved by the Hospital's Board of Medical Research, Ethics Committee, and Pharmaceutical Advisory Committee.

RESULTS

There were 50 patients, 25 of whom received either dosage regimen. The chief clinical details and microbiologic findings in all patients are shown in Table 1. All except three patients (Nos. 4, 6, and 17) had hospital-acquired chest infection. Most (78 percent) were mechanically ventilated.

As shown in Table 2, the two groups of patients were comparable on demographic grounds, except that those receiving cefotaxime every 12 h were significantly older, although age did not correlate with any of the outcome indices. The diagnostic criteria for chest infection met the formal Centers for Disease Control (CDC) criteria for pneumonia in 46 patients (92 percent).[21]

Renal function was clearly abnormal in nine patients (serum creatinine greater than 0.20 mmol/L throughout the study), although none required dialysis or hemofiltration at the time. In a further 11 patients, minor abnormalities of renal function were observed during the study (transient elevations of serum creatinine above 0.12 mmol/L but below 0.20 mmol/L). There was no statistically significant difference in renal function between the two treatment groups (Table 2). In addition, the clinical response was similar whether renal function was normal or not (global score 49 percent vs 50 percent, respectively), and the clinical response did not correlate with the serum creatinine.

Thirty-three patients (66 percent) completed the planned five-day course of antibiotic treatment (Table 3). Six patients improved sufficiently before five days for antibiotic therapy to be discontinued. In one patient, clinical failure (ie, worsening entry criteria) led to antibiotic change before five days. In five patients, antibiotics were changed because the initial culture showed pathogens more appropriately treated with another antibiotic--two with Streptococcus pneumoniae, one with methicillin-resistant Staphylococcus aureus, one with Xanthomonas maltophilia, and one with Legionella. In two patients, the occurrence of systemic sepsis from sites other than the chest (abdomen in one and central venous line in one) prompted a change of antibiotic. Three patients died during treatment, none due to infection (one due to cerebral hemorrhage and two due to cardiogenic shock). A further nine patients died 6 to 67 days subsequently in the ICU (five due to adult respiratory distress syndrome and multiorgan failure, two due to systemic sepsis, one due to cardiogenic shock, and one due to cerebral hemorrhage).

Incomplete

The clinical response of all patients is shown in Table 4. Resolution of infection occurred in 86 percent, no change in occurred 4 percent, and deterioration occurred in 10 percent. The global score was similar in both groups, as were the numbers who were improved, unchanged, and worse.

(*1) For global score, values are mean with SD and 95 percent confidence intervals in parentheses.

The microbiologic findings are shown in Table 1, the comparisons of the two treatment groups are shown in Table 5, and the responses are shown in Table 6. Pathogens were isolated from initial cultures in 19 patients (38 percent) and appeared comparable between the two treatment groups (Table 5). Mixed upper respiratory tract flora were identified in a further 12 patients (24 percent) but were not classified as pathogens. Pathogens were isolated from final cultures in 11 patients (22 percent), in four of whom there was no initial pathogen. Twenty-three patients (46 percent) were thus microbiologically assessable. Of these, microbiologic cure was obtained in 12 (52 percent) and failure occurred in 11 (48 percent). There were no instances of failure due to persistence of a sensitive initial pathogen or to development of resistance during treatment. Failure occurred in three patients (13 percent) due to persistence of a resistant pathogen and in eight patients (35 percent) due to superinfection. There were no statistically significant differences in microbiologic response between the two treatment groups. In the microbiologically assessable patients, the demographic details and clinical outcome were not statistically different between treatment groups. In addition, there was no statistically significant difference for these clinical parameters between patients microbiologically assessable and not assessable. [TABULAR DATA OMITTED]

DISCUSSION

The documentation of formal dose-response relationships for antibiotics can present considerable practical and ethical difficulties where the response is the effective treatment of an infection. The recommended dosage regimen is thus usually based on the antibiotic's pharmacokinetic and microbiologic profile and confirmed by clinical trial.[14] Yet the extrapolation of pharmacokinetic and susceptibility data to effective dosage recommendations can be misleading, especially in the seriously ill.[15] Other factors, some known (eg, tissue penetration, postantibiotic effect where applicable, inoculum effect, effects on phagocytes)[14] and perhaps some unknown presumably contribute to efficacy. Moreover, except in a few situations in which dosage may be more clearly defined, most schedules are empiric and, while effective, may well be excessive.[14] The appropriately structured clinical trial must therefore be the yardstick for establishing or confirming optimal dosage, which for reasons of cost and convenience should be the lowest effective dose administered least frequently.

Cure

Failure

(*1) This patient received concomitant flucloxacillin for vascular surgical prophylaxis, while H influenzae was cleared from the respiratory tract.

Based on its pharmacokinetics and microbiology, the usual dosage recommendation for cefotaxime in serious infections has been 6 to 12 g daily in three or more divided doses, with the higher doses generally reserved for infections in less accessible sites (eg, meningitis, endocarditis) or in immunocompromised hosts.[11,14] A lower dose of 3 to 4 g daily in three or fewer divided doses has often been used in infections that were either less severe or due to very susceptible microorganisms. A daily dose of 4 g or less has thus been regarded as a low dose and greater than 4 g has been regarded as a high dose.[14] Yet collation of clinical results from more than 10,000 patients given cefotaxime for infections of various types and of varying severity showed similar cure rates for those given high or low doses.[12] In more than 2,000 patients, microbilogic results were also comparable over a range of dosages.[14] However, these results are of limited value because they were uncontrolled.

Although a twice daily dosage of cefotaxime has been reported previously to be effective in lower respiratory tract infections,[22] the more common regimen has been 2 g thrice daily, particularly in the seriusly ill.[15] In the present controlled study, a low dose of cefotaxime (2 g twice daily) was shown to be directly comparable in clinical and microbiologic efficacy with the more traditional dose (2 g thrice daily). Although a low dose of cefotaxime of 4 g daily was found to be effective, this is not necessarily the optimal dose. It is possible that even lower or less frequent doses could be similarly effective, but this suggestion would require formal clinical evaluation.

All patients were in the ICU and had serious chest infections; most (78 percent) were being mechanically ventilated. This can be a difficult group of patients to study, particularly because of diagnostic and therapeutic constraints. The present study thus had a number of limitations, related mainly to design and analysis. First, the study could not reasonably require the identification of a bacterial pathogen prior to commencing treatment. While this is a practical clinical approach, as would be expected, only about half the patients turned out to be microbiologically assessable. The resultant intention-to-treat analysis based on clinical response will tend to minimize any difference between the two regimens. This, together with the relatively small number of patients studied (the study had a power of 80 percent to detect a 12 percent difference in clinical score with a confidence of 95 percent), should prompt caution in interpreting any lack of difference between treatments. Second, the efficacy analysis based on microbiologic response can be considered only supportive of the main clinical analysis, since the number of patients microbiologically assessable was too small for this analysis to exclude a significant [unkeyable] error.

The results of the present study have both theoretical and practical implications. The theoretical conclusion relates to the disparity between the long dosing interval found to be effective (12 h) and the short half-life of cefotaxime (about 1 h). On the other hand, we showed in a previous study that ceftriaxone, an antibiotic with a similar antimicrobial spectrum to cefotaxime but with a much longer half-life (about 8 h), was poorly effective when the dosing interval was 24 h, as recommended, in serious chest infections.[15] Renal impairment could have been a potential mechanism for extending the effective dosing interval of cefotaxime, but this possibility was excluded in both studies. A more plausible reason for the unexpectedly greater efficacy of cefotaxime and for its seemingly long dosing interval may perhaps lie with its unique active metabolite, desacetylcefotaxime. This metabolite confers synergy with the parent compound, particularly against a number of Gram-negative aerobic bacilli, and has a half-life about twice as long as cefotaxime itself.[23-26]

The practical conclusion relates to the much greater cost-effectiveness of the lower-dose regimen with its lower cost and greater convenience. A 33 percent saving in the drug and delivery costs of treating a major complication is a considerable aid to cost-containment in the expensive environment surrounding the care of the seriously ill.

ACKNOWLEDGMENTS: The authors thank Roussel Pharmaceuticals, Sydney, Australia, for their support.

REFERENCES

1 Bartlett JG. Pneumonia. In: Bone RC, ed. Critical care: a comprehensive approach. Park Ridge, Ill: American College of Chest Physicians, 1984; 434-45

2 Bender BS, Bartlett JG. Nosocomial infections. In: Bone RC, ed. Critical care: a comprehensive approach. Park Ridge, Ill: American College of Chest Physicians, 1984; 446-63

3 Craven DE, Driks MR. Nosocomial pneumonia in the intubated patient. Semin Respir Med 1987; 2:20-33

4 Hessen MT, Kaye D. Nosocomial pneumonia. Crit Care Clin 1988; 4:245-57

5 Hoyt JW. Complications of infection in the critically ill patient. In: Lumb PD, Bryan-Brown CW, eds. Complications in critical care medicine. Chicago: Year Book, 1988; 205-19

6 Marino PL. Nosocomial pneumonia. In: The ICU book. Malvern, Pa: Lea & Febiger, 1991; 588-602.

7 Wolinsky E. Pulmonary infection in critical care medicine. In: Shoemaker WC, Thompson WL, Holbrook PR, eds. Textbook of critical care. Philadelphia: WB Saunders, 1984; 577-79

8 Bartlett JG, O'Keefe P, Tally FP. Bacteriology of hospital-acquired pneumonia. Arch Intern Med 1986; 146:868-74

9 Summer WR, Nelson S. Severe community-acquired pneumonia: how to decide on initial therapy. Pulmonary Perspectives, ACCP 1988; 5:4-8

10 Kunin CC. Dosage schedules of antimicrobial agents. Rev Infect Dis 1981; 3:4-11

11 Neu HC, Carbon C, Pechere J-C (eds). Cefotaxime: proceedings from the 15th international congress of chemotherapy. Drugs 1988; 35(suppl 2):1-231

12 Parker RH. Effect of frequency of administration on therapeutic efficacy of cefotaxime. Clin Ther 1984; 6:488-99

13 Rondanelli R, Dionigi RV, Calvi M, Dell'Antonio M, Corsico G, Mapelli A. Cefotaxime therapy of lower respiratory tract infections in intensive-care patients. Int J Clin Pharmacol Res 1987; 7:73-6

14 Simon A, d'Aubrac CA, Safran C, Carbon C. Cefotaxime optimal dosage in adult patients. Drugs 1988; 35(suppl 2):221-30

15 Reeves JH, Russell GM, Cade JF, McDonald M. Comparison of ceftriaxone with cefotaxime in serious chest infections. Chest 1989; 96:1292-97

16 Cullen DJ. Results and costs of intensive care. Anesthesiology 1977; 47:203-16

17 Cullen DJ, Civetta JM, Briggs BA, Ferrara LC. Therapeutic intervention scoring system: a method for quantitative comparison of patient care. Crit Care Med 1974; 2:57-60

18 Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification system. Crit Care Med 1985; 13:818-29

19 Isenberg HD, Washington JA, Balows A, Sonnenwirth AC. Collection handling and processing of specimens. In: Lennette EH, Balows A, Hausler WJ, Shadomy HJ, eds. Manual of clinical microbiology, 1985; 79-81

20 Washington JA. Susceptibility tests: agar dilution. In: Lennette EH, Balows A, Hausler WJ, Shadomy HJ, eds. Manual of clinical microbiology. 4th ed. Washington, DC: American Society for Microbiology. 1985; 967-71

21 Garner JS, Jarvis WR, Emori TG, et al. CDC definitions for nosocomial infections, 1988. Am J Infect Control 1988; 16:128-40

22 Miki F, Shiota K. Cefotaxime in lower respiratory tract infections compared to cefazolin. J Antimicrob Chemother 1980; 6(suppl A):169-75

23 Jehl FX, Jones RN, Monteil HF, Neu HC. Cefotaxime and desacetylcefotaxime: a synergistic antibacterial interaction. Roussel Uclaf Monograph, 1990; 1-79

24 Jones RN, Neu HC (eds). Symposium: new findings on the antibacterial interaction between cefotaxime and desacetyl-cefotaxime. Diagn Microbiol Infect Dis 1989; 12:23-122

25 Oizumi K, Hayashi I, Aonuma S, Konno K. In vitro activity of desacetylcefotaxime and the interaction with its parent compound, cefotaxime. Drugs 1988; 35(suppl 2):57-61

26 Piedrola G, Galan I, Leyva A, Maroto MC. Comparison of in vitro activity of cefotaxime and desacetylcefotaxime alone and in combination against 320 Gram-negative clinical isolates. Drugs 1988; 35(suppl 2):62-4

COPYRIGHT 1992 American College of Chest Physicians
COPYRIGHT 2004 Gale Group