Controlling Antibiotic Resistance: One Small Victory?

abstracts & commentary

Synopsis: Control of virtually all cephalosporin use at one hospital was associated with a significant reduction in the prevalence of resistant Klebsiella containing extended spectrum beta lactamase. This was accomplished, however, with an increased use of imipenem as well as an increased prevalence of imipenem resistence in Pseudomonas aeruginosa.

Sources: Rahal JJ, et al. Class restriction of cephalosporin use to control total cephalosporin resistance in nosocomial Klebsiella. JAMA 1998;280:1233-1237; Burke JP. Antibiotic resistance—squeezing the balloon? JAMA 1998;280: 1270-1271.

An outbreak of Klebsiella producing an extend-ed spectrum B-lactamase (ESBL) occurred in Rahal and colleagues’ hospital in 1990. Over the next five years, the prevalence gradually increased despite restrictions upon the use of third generation cephalosporins. In 1995, there were a total of 150 isolations of ESBL-producing Klebsiella, representing 19.6% of all Klebsiella isolates. Approximately 40% of the resistant isolates were resistant to cephamycins (cefotetan, cefoxitin) as well. Prior to 1996, use of third generation cephalosporins or imipenem required approval by the infectious disease service. Beginning in 1996, the hospital adopted new antibiotic use guidelines, requiring approval for use of all cephalosporins and cephamycins with few exceptions, such as the use of ceftriaxone for the treatment of meningitis or gonococcal infections. Restrictions of the use of imipenem continued.

Rahal et al measured the effect of the new restriction by comparing the isolation of ceftazidime-resistant Klebsiella in 1996 with that noted in 1995. They also compared the isolation of imipenem-resistant Pseudomonas during the two-year period. Surveillance methods and infection control practices were identical during the two years.

In 1996, there was an 80% decrease in cephalosporin use hospital-wide, from 5558 g/month to 1106 g/month (P < 0.001) compared with 1995. However, imipenem use increased by 141% (197 g/month in 1995 to 474 g/month in 1996). During 1996, there was a 44% reduction in nosocomially acquired ceftazidime-resistant Klebsiella compared with 1995 (150 vs 84 isolates, respectively). The reduction was most apparent in the ICUs. There was a concomitant 69% increase in isolation of imipenem-resistant Pseudomonas.

Comment by robert muder, MD

ESBLs of Klebsiella are typically plasmid mediated, and confer high-level resistance to ceftazidime and aztreonam, and variable, often less marked, resistance to cefotaxime. Many, but not all, of these ESBL-producing strains remain susceptible to cephamycins. The plasmids often contain resistance determinants to other, unrelated antibiotics such as aminoglycosides. ESBL-producing Klebsiella (as well as other members of the Enterobacteriaceae) are widespread in hospitals and long-term care facilities throughout the world.

Efforts to control resistant microorganisms have generally consisted of a two-pronged approach—prevention of transmission by isolation practices and control of antibiotic use. The reported results of control measures have been (to put the best possible face on the situation) decidedly mixed. Rahal et al took a novel approach and hypothesized that restriction of the entire cephalosporin class, including the related cephamycins, would lead to withdrawal of the selective pressure favoring ESBL-producing Klebsiella. They were remarkably successful in reducing cephalosporin use, and, indeed, there was a marked and statistically significant reduction in resistant Klebsiella that was most apparent in the ICUs. Unfortunately, there was an increase in the frequency of isolation of imipenem-resistant Pseudomonas, in response, no doubt, to the increased use of imipenem during the period of intervention. In an accompanying editorial, Burke compares the result to "squeezing a balloon—constraining one end causes the other end to bulge."

It is difficult to judge the overall effect of the intervention by Rahal et al. Rahal et al don’t provide outcome information in terms of infection rates or deaths due to infection. It would also be important to know what effect the change in antibiotic prescribing had on other resistant flora in the hospital, particularly other potentially cephalosporin-resistant agents such as Enterobacter and Serratia. One might surmise that a decrease in cephalosporin use might have led to a decrease in isolation of methicillin-resistant Staphylococcus aureus and resistant Enterococcus, for example. It would also be important to know the changes in the use of alternative agents such as ciprofloxacin and the effect of these changes upon the frequency of resistance to the these agents. It would also be important to know what happened to the total use of antibiotics and any changes in drug expenditures.

Hospitals can be likened to complex ecosystems, in which one change or perturbation is likely to have not only its intended effect but also multiple secondary effects that may not be predictable. It’s not surprising that the "ecologic niche" occupied by resistant Klebsiella would be taken over something else that might be just as objectionable. It is, undoubtedly, overly optimistic to expect that changing the usage pattern of a single class of antibiotics will solve the problem of drug resistance. But, the experience of Rahal et al demonstrates that antibiotic usage patterns can be changed in a rational way over a prolonged period of time, and that at least some of the effects of the intervention can be quantified. Such studies are an important stepping stone if we hope to devise comprehensive control strategies to reduce the threat of antibiotic resistance.