Is "Crop Rotation" of Antibiotics the Solution to a "Resistant" Problem in the ICU?

	ARTICLE

TOP Article REFERENCES

The emergence of multiple-drug resistant pathogens in patients who are seriously ill represents a new challenge for the critical care physician. In the therapy of pneumonia and sepsis, initial empiric therapy may be ineffective if the responsible pathogen is not susceptible to available therapy. This is becoming a reality for certain strains of P. aeruginosa, Acinetobacter spp., and Stenotrophomonas maltophilia, among enteric gram-negative bacteria; and an ongoing problem among gram-positive bacteria such as S. aureus and enterococcus. The factors predisposing to resistance are numerous, but previous investigations have shown that both nosocomial pneumonia and bacteremia caused by resistant gram-negative bacteria can be a consequence of prior antibiotic usage (1). The clinical impact of resistant bacteria is not fully known, but the "attributable mortality" of pneumonia caused by these organisms appears to be increased, compared with pneumonia resulting from more easily treated bacteria in the same types of patients (4, 5).

With the rise in antimicrobial resistance have come a variety of strategies designed to prevent this problem. In a recent consensus conference, the Centers for Disease Control suggested that the problem could be better controlled through optimization of prophylactic, empiric, and therapeutic antimicrobial use and that this could be accomplished by education about appropriate antibiotic use and by providing data to physicians about the types of resistant organisms seen in their own intensive care units as part of an ongoing surveillance program (6). If such a program is to be successful, it may be necessary to use predefined guidelines to dictate which antimicrobials should be used for empiric therapy. In one such program, a computer-driven antibiotic guideline was applied for more than 5 yr and was able to reduce antibiotic costs, increase the appropriateness of prophylactic therapy, and minimize adverse drug reactions, without leading to an increase in antibiotic resistance (7).

The advent of guidelines for the therapy of infection is not without concern, and some have suggested that the use of widespread, broad-spectrum empiric therapy will only add to the resistance problem (8). However, it is theoretically possible that if empiric therapy is given in a timely manner, using highly effective agents, it could lead to rapid bacterial killing and thereby minimize the emergence of resistance, an event that often occurs when therapy involves an agent with "borderline" activity against the target organism. In this issue of the journal, Kollef and colleagues (9) have demonstrated that a scheduled change in defining which antibiotic should be routinely used for nosocomial infections in a cardiac surgical ICU led to a reduced incidence of pneumonia and bacteremia because of resistant organisms (9).

The study is an interesting and important one, showing the relationship of infection with resistant organisms to antibiotic policies. The "crop rotation" theory of antibiotic use has suggested that if we routinely vary our "go to" antibiotic in the ICU, we can minimize the emergence of resistance because selection pressure for bacteria to develop resistance to a specific antibiotic would be reduced as organisms become exposed to continually varying antimicrobials. The study by Kollef and colleagues (9) has not really looked at rotation of antibiotics, but it has shown that a scheduled change in antibiotic policy may be useful. In the study, the empiric therapy for suspected gram-negative infection in the cardiac surgical ICU was a third generation cephalosporin, ceftazadime, in the first 6 mo, whereas in the next 6 mo, empiric therapy was done with a quinolone, ciprofloxacin. The program was generally successful in changing prescribing practices, and the use of third generation cephalosporin went from almost 20% of all patients in the first 6 mo to 7% in the next period, whereas the use of ciprofloxacin increased to 19% of all patients after the scheduled change, in contrast to only 3% of all patients in the first 6 mo of the study. With this scheduled change, the incidence of ventilator associated pneumonia (VAP) was reduced from 11.6 to 6.7%, the incidence of microbiologically confirmed VAP was similarly reduced, and the incidence of VAP caused by antibiotic-resistant gram-negative bacteria was reduced from 4 to 0.9%. Although the incidence of bacteremia did not decline in the second 6-mo period, the frequency of bloodstream infections caused by resistant organisms did drop, although not to a statistically significant degree.

The findings in this study are surprising and provocative, raising a number of important questions. It might have been expected that a change in antibiotic prescribing practices would have had no impact on the incidence of pneumonia. Because the development of respiratory infection is presumably a reflection of the host defense status of at-risk patients, a change in antibiotic usage policy might only have been expected to change the identity of the infecting organisms and the patterns of antimicrobial resistance. The observed reduction in the incidence of pneumonia suggests that antibiotics could have two roles in the pathogenesis of VAP: they predispose to infection with resistant organisms and they may directly contribute to the pathogenesis of pneumonia itself. However, even if this explanation is correct, some of the findings in the study are hard to understand. For example, the use of postoperative antibiotics was high in both study periods (44.5 versus 42.5%, before and after), and was not reduced in the second time period, even though there were fewer suspected and proven infections. Thus, although the scheduled change in antibiotic policy led to fewer episodes of VAP, it may have been accompanied by more inappropriate antibiotic usage since the overall prescribing of antibiotics did not decline. In addition, there was no reduction in mortality in the second 6-mo period, although such an outcome may not have been possible unless more patients were studied.

In trying to understand why a scheduled change in empiric antibiotic therapy led to a reduce incidence of infection with resistant organisms, we need to examine how bacteria develop resistance. In general, bacteria use four different mechanisms to become resistant to antibiotics: they decrease the permeability of their cell wall by changes in the porin channels, a mechanism that is especially important for gram-negative bacteria that have a complex cell wall; they inactivate antibiotics through the production of enzymes such as beta-lactamases or cephalosporinases; they alter the target site for antibiotics such as penicillin binding proteins or the DNA gyrase; or they actively extrude the antibiotic once it enters the cell (10, 11). Resistance to cephalosporins generally occurs by a different mechanism than resistance to quinolones, and thus a scheduled change of antibiotics between these classes might have been successful because the alterations that bacteria must make to resist one agent might have no effect on their susceptibility to another agent.

Cephalosporin resistance among gram-negative bacilli can be the result of the induction of chromosomal beta-lactamases in bacteria after repeated exposure to a given antibiotic. The extended spectrum cephalosporins are rendered ineffective when gram-negative organisms such as the Enterobacteriaceae develop a mutation that allows for constitutive production of chromosomal beta-lactamases that are normally inducible enzymes. Resistance to cephalosporins can occur by other mechanisms as well, including inactivation by plasmid-mediated beta-lactamases and through changes in bacterial cell wall permeability and binding sites (12). The induction of chromosomal beta-lactamases has been identified as a consequence of the use of third generation cephalosporins, and resistance rates decline with restricted use (13, 14). Thus, a scheduled change of antibiotic classes away from third generation cephalosporins could restore their activity, allowing these agents to be used again in the future. The replacement drug could be a beta-lactam/beta-lactamase inhibitor combination (such as ticarcillin-clavulanic acid or piperacillin-tazobactam), a carbapenem (imipenem or meropenem), a fluoroquinolone (as in the study by Kollef and colleagues), or an aminoglycoside (11). The new fourth-generation cephalosporin, cefepime, may be another alternative because it is less susceptible to destruction by beta-lactamases than earlier generation cephalosporins, and thus remains active against bacterial mutants that produce such enzymes. In addition, cefepime may require gram-negative bacteria to have more than one mutation to become resistant, and it has a lower affinity than other cephalosporins for binding many of the common bacterial beta-lactamases (15).

Quinolone resistance arises by different mechanisms, and it is chromosomally mediated, primarily by changes in the bacterial DNA gyrase or in topoisomerase IV, the target sites for antibacterial action, although changes in bacterial permeability or efflux may also lead to quinolone resistance (16). Resistance is more likely to develop with unrestricted use of any quinolone antibiotic and is promoted by the presence of necrotic tissue, indwelling foreign bodies, or poorly drained abscesses. When quinolones are used, it is necessary to use the most potent agent available and at an adequate dose. As Acar and Goldstein (16) have pointed out, in vitro studies have shown that selection of a bacterial clone resistant to quinolones is more likely to occur with the use of less effective compounds. Because resistance to one quinolone may raise the minimum inhibitory concentration (MIC) for other quinolones against that organism, it is important to choose a highly active agent, and to dose it so that serum concentrations far exceed the target organism MIC. In one study of infection in the critically ill, using quinolone therapy, outcome was best if the agent used achieved a high enough serum concentration that the area under the concentration-time curve exceeded the target organism MIC by a ratio of 125:1 (17).

The study by Kollef and colleagues (9) does not address these issues and does not specify what dosage of ciprofloxacin was used, but one recent study of severe pneumonia, with nearly 80% of patients being mechanically ventilated, showed efficacy for ciprofloxacin at a dose of 400 mg every 8 h (18). With this dosage, pneumonia caused by multiple pathogens, including S. pneumoniae, was effectively treated, although monotherapy was unable to prevent the emergence of resistance among 33% of the infections caused by P. aeruginosa. Thus, if P. aeruginosa is a likely pathogen, combination therapy is probably needed to minimize the risk of resistance developing during therapy, despite initial testing suggesting susceptibility. The issue of combination empiric therapy is not discussed by Kollef and colleagues, but clearly some patients are going to need more than one empiric antibiotic, and the extent of adherence to this policy, for appropriate patients, will probably have an impact on the development of infection with resistant organisms.

A variety of new quinolone antibiotics are becoming available, and their appropriate use will depend on their proven efficacy in clinical trials, along with their apparent in vitro activity against target organisms, as defined by MICs. Given the relationship of resistance emergence to the use of borderline effective quinolones, we will want to select an agent with the greatest available activity against the likely etiologic pathogens. In the therapy of outpatient respiratory infections, pneumococcus is a predominant organism and a variety of new quinolones have enhanced activity against this pathogen, compared with some of the older quinolones. The MIC₉₀ against S. pneumoniae for levofloxacin, sparfloxacin, grepafloxacin, clinafloxacin, and trovafloxacin, respectively, are (in micrograms/ml): 1.0, 0.25, 0.25, 0.12, and 0.25 (19). However, quinolones as a drug class penetrate extremely well into respiratory secretions, often exceeding serum concentrations, so that their activity can also be defined by the ratio of maximal serum concentrations (after a standard dose) to the MIC of pneumococcus. This ratio for levofloxacin, sparfloxacin, grepafloxacin, clinafloxacin, and trovafloxacin, respectively, is: 6.6, 6.4, 6.0, 12.0, and 12.0 (19). One other feature of these new quinolones is that they are generally as active against penicillin resistant pneumococci as they are against penicillin sensitive organisms.

If the findings from this study are to be more widely applied, we must ask a number of questions. Is it necessary to continually change the identity of our empiric antibiotic regimen and if so how often: every 3 mo, every 6 mo, etc.? If such a change is made, which antibiotic should be used, and when can we return to our original agent? Was the outcome seen in the study by Kollef and colleagues the consequence of changing specifically from a cephalosporin to a quinolone, or was it simply the result of changing from one antibiotic of any class to another antibiotic of a different class? Quinolones and cephalosporins differ in a number of ways, including their mechanisms of antibacterial action and the ways in which they are made ineffective by mutant bacteria. In addition, quinolones as a drug class penetrate into respiratory secretions more extensively (as a percentage of serum concentrations) than do cephalosporins, suggesting that they may have an advantage in eradicating deep seated lung infections. Finally, are the results likely to apply widely, or are they only applicable to a low risk surgical population? Numerous studies have shown that VAP in surgical and trauma patients differs from that seen in medical patients, with both the incidence and the attributable mortality being greater in medical patients (20, 21). Another difference between these populations is the fact that most surgical patients (98% in the study of Kollef and colleagues) receive routine prophylactic antibiotics, whereas this practice is less common in medical patients. In fact, the benefits seen in the study by Kollef and colleagues were confined to late onset VAP, and it is likely that the use of prophylactic therapy prevented early onset infections. In medical patients, early onset pneumonias are common, and they may not be affected by scheduled changes in the identity of routine antibiotics. Thus, an alteration in antibiotic prescribing practices may not impact high-risk medical patients in the same way that they affect a low risk surgical population.

We do not have answers to these questions, but the issues raised by the increasing prevalence of multiple drug-resistant bacteria in the ICU does require all of us to become familiar with the available choices for empiric antibiotic therapy. With careful monitoring of resistance patterns at our own institutions, we can make more logical and effective antibiotic choices. One approach that we have instituted at our hospital is to receive monthly updates about the organisms present in the ICU and the patterns of resistance among these organisms. In doing this, we have found that antibiotic susceptibility patterns can differ in the medical ICU from those seen in other parts of the hospital, and thus therapeutic choices in the critical care unit may need to differ from those used elsewhere in the hospital. If empiric therapy choices are made with such data in mind, we can probably avoid an increase in resistance patterns, as has been suggested by the results of one study in which investigators applied a similar approach (7). We may also need to consider the "crop rotation" theory, and to vary the antibiotics that we choose to use from among the available and effective therapeutic options suggested by sensitivity patterns. The study by Kollef and colleagues is intriguing and should urge us to further explore the issue of changing our empiric antibiotic choices in a scheduled fashion. This may lead us to using new and potentially more expensive (in terms of acquisition costs) antibiotics in the ICU, but if antibiotic rotation among such agents can minimize the development of resistance, then it may become a maneuver that is in the end highly cost effective.

Director, Medical and Respiratory ICU
Winthrop-University Hospital
Mineola, New York
Professor of Medicine
SUNY at Stony Brook
Stony Brook, New York

	Footnotes

Correspondence and requests for reprints should be addressed to 222 Station Plaza N. Suite 400, Mineola, NY 11501.

	References

TOP Article REFERENCES

1. Rello, J., V. Ausina, M. Ricart, J. Castella, and G. Prats. 1993. Impact of previous antimicrobial therapy on the etiology and outcome of ventilator-associated pneumonia. Chest 104: 1230-1235 [Abstract/Free Full Text].

2. Chow, J. W., M. J. Fine, D. M. Shlaes, J. P. Quinn, D. C. Hooper, M. P. Johnson, R. Ramphal, M. M. Wagener, D. K. Miyashira, and V. L. Yu. 1991. Enterobacter bacteremia: Clinical features and emergence of antibiotic resistance during therapy. Ann. Intern. Med. 115: 585-590 .

3. Fagon, J. Y., J. Chastre, Y. Domart, J. L. Troulliet, J. Pierre, C. Darne, and C. Gibert. 1989. Nosocomial pneumonia in patients receiving continuous mechanical ventilation: prospective analysis of 52 episodes with use of a protected specimen brush and quantitative culture techniques. Am. Rev. Respir. Dis. 139: 877-884 [Medline].

4. Brewer, S. C., R. G. Wunderink, C. B. Jones, and K. V. Leeper. 1995. Ventilator-associated pneumonia due to Pseudomonas aeruginosa. Chest 108: 1655-1662 [Abstract/Free Full Text].

5. Kollef, M. H., P. Silver, D. M. Murphy, and E. Trovillion. 1995. The effect of late-onset ventilator-associated pneumonia in determining patient mortality. Chest 108: 1655-1662 .

6. Goldman, D. A., R. A. Weinstein, R. P. Wenzel, O. C. Tablan, R. J. Duma, R. P. Gaynes, J. Schlosser, and W. J. Martone. 1996. Strategies to prevent and control the emergence and spread of antimicrobial- resistant microorganisms in hospitals: a challenge to hospital leadership. J.A.M.A. 275: 234-240 [Abstract].

7. Pestotnik, S. L., D. C. Classen, R. S. Evans, and J. P. Burke. 1996. Implementing antibiotic practice guidelines thorough computer-assisted decision support: clinical and financial outcomes. Ann. Intern. Med. 124: 884-890 [Abstract/Free Full Text].

8. Berk, S. L.. 1995. Justifying the use of blood cultures when diagnosing community-acquired pneumonia. Chest 108: 891-892 [Free Full Text].

9. Kollef, M. H., J. Vlasnik, L. Sharpless, C. Pasque, D. Murphy, and V. Fraser. 1997. Scheduled change of antibiotic classes: a strategy to decrease the incidence of ventilator-associated pneumonia. Am. J. Respir. Crit. Care Med. 156: 1040-1048 [Abstract/Free Full Text].

10. Grayson, M. L., and G. M. Eliopoulos. 1990. Antimicrobial resistance in the intensive care unit. Semin. Respir. Infect. 5: 204-214 [Medline].

11. Gold, H. S., and R. C. Moellering. 1996. Antimicrobial drug resistance. N. Engl. J. Med. 335: 1445-1453 [Free Full Text].

12. Medieros, A. A. 1997. Evolution and dissemination of -lactamases accelerated by generations of -lactam antibiotics. Clin. Infect. Dis. 24(Suppl. 1):S19-S45.

13. Jones, R. N.. 1992. The current and future impact of antimicrobial resistance among nosocomial bacterial pathogens. Diagn. Microbiol. Infect. Dis. 15: 3S-10S [Medline].

14. Meyer, K. S., C. Urban, J. A. Eagan, B. J. Berger, and J. J. Rahal. 1993. Nosocomial outbreak of Klebsiella infection resistant to late-generation cephalosporins. Ann. Intern. Med. 119: 353-358 [Abstract/Free Full Text].

15. Sanders, C. C.. 1993. Cefepime: the next generation? Clin. Infect. Dis. 17: 369-379 [Medline].

16. Acar, J. F., and F. W. Goldstein. 1997. Trends in bacterial resistance to fluoroquinolones. Clin. Infect. Dis. 24(Suppl. 1):S67-S73.

17. Forest, A., D. E. Nix, C. H. Ballow, T. F. Goss, M. C. Birmingham, and J. J. Schentag. 1993. Pharmacodynamics of intravenous ciprofloxacin in seriously ill patients. Antimicrob. Agents Chemother. 37: 1073-1081 [Abstract/Free Full Text].

18. Fink, M. P., D. R. Snydman, M. S. Niederman, K. V. Leeper, R. H. Johnson, S. O. Heard, R. G. Wunderink, J. W. Caldwell, J. J. Schentag, G. A. Siami, R. L. Zameck, D. C. Haverstock, H. H. Reinhart, R. M. Echols, and the Severe Pneumonia Study Group. 1994. Treatment of severe pneumonia in hospitalized patients: results of a multicenter, randomized, double-blind trial comparing intravenous ciprofloxacin with imipenem-cilastatin. Antimicrob. Agents Chemother. 38: 547-557 [Abstract/Free Full Text].

19. Stein, G. E. 1996. Pharmacokinetics and pharmacodynamics of newer fluoroquinolones. Clin. Infect. Dis. 23(Suppl. 1):S19-S24.

20. Baker, A. M., J. Meredith, and E. F. Haponik. 1996. Pneumonia in intubated trauma patients: microbiology and outcome. Am. J. Respir. Crit. Care Med. 153: 343-349 [Abstract].

21. Papazian, L., F. Bregeon, X. Thirion, R. Gregoire, P. Saux, J. P. Denis, G. Perin, J. Charrel, J. F. Dumon, J. P. Affray, and F. Gouin. 1996. Effect of ventilator-associated pneumonia on mortality and morbidity. Am. J. Respir. Crit. Care Med. 154: 91-97 [Abstract].