Nosocomial Outbreaks of Multiresistant Bacteria: Extended-Spectrum Beta-Lactamases Have Arrived in North America

The history of the antibiotic era shows that widespread use of new antibiotics fosters emergence of resistant clinical isolates, often possessing novel antimicrobial resistance mechanisms. Clinical use of oxyimino--lactams, particularly ceftazidime and cefotaxime, skyrocketed during the last 15 years. In 1983 plasmid-mediated enzymes that confer resistance to oxyimino--lactamsextended-spectrum -lactamases (ESBL)were first observed in Germany. Since then, the number and variety of these -lactamases have increased rapidly and their distribution is now worldwide.

Most of the ESBLs derive from TEM-1 or 2 and SHV-1, the common plasmid-determined -lactamases that confer resistance to penicillins and early-generation cephalosporins. Substitution of a single amino acid in the active site region of the SHV-1 molecule transforms it into the extended-spectrum -lactamase, SHV-2, that can inactivate late-generation cephalosporins. The number of derivatives of the TEM and SHV -lactamases proven to be unique by sequencing or oligotyping has now reached 26 and 5, respectively [1, 2]. Several have been found in many countries (SHV-2, SHV-4, SHV-5, TEM-6), whereas others seem to occur more commonly in one or two countries. For example, TEM-3 seems more prevalent in France and TEM-10 and TEM-12 in the United States and England [1]. Most of the isolates that produce ESBLs have come from hospitalized patients and have frequently caused nosocomial outbreaks, primarily due to Klebsiella pneumoniae strains that often have associated resistance to aminoglycosides. Recent surveys of nosocomial isolates of K. pneumoniae from Europe show that 14% to 16% produce ESBLs [3, 4]. In France the prevalence increased from less than 1% in 1985 to the current level by 1988. Extended-spectrum -lactamases are found much less frequently in other species of Enterobacteriaceae [3, 4].

General statistics regarding the prevalence of ESBLs, however, may be misleading. The prevalence of ESBLs among 12 French hospitals from different provinces varied from zero to 46% [4]. Indeed, three hospitals had a frequency much higher than the total prevalence of 14% and the others had much lower rates. In some hospitals, sporadic nosocomial outbreaks due to strains producing ESBLs seem to lead to an endemic problem. Selection pressure from widespread hospital use of late-generation cephalosporins apparently enhances colonization of the digestive or respiratory tracts of patients, and infection follows [2, 5]. Not surprisingly, outbreaks have often been associated with surgery, prolonged hospital stay, or the presence of urinary or arterial catheters, especially in patients in intensive care units [5]. Patients from nursing homes and other long-term care institutions may recirculate such strains into hospitals [6].

Most of the large outbreaks of nosocomial infection due to ESBL-producing Klebsiella have been reported from Europe, especially from France. In this issue, Meyer and colleagues [7] report the largest outbreak of ceftazidime-resistant Klebsiella infections yet to occur in a general hospital in North America. This report emphasizes that nosocomial infections caused by ESBL-producing bacteria are as serious a threat to hospitals in North America as they have been to hospitals in other countries. The outbreak in the Medical Center of Queens coincided with increasing use of ceftazidime and declined after restricting ceftazidime, implying a causal relationship. However, the simultaneous institution of barrier precautions with ceftazidime restriction undoubtedly contributed greatly to interrupting nosocomial spread. A casecontrol study might have provided more direct evidence of cause and effect. A recent outbreak in a pediatric hospital affords, perhaps, a clearer example of the danger of widespread use of ceftazidime [2]. Approximately 1 year after the hospital adopted a policy of using empiric ceftazidime monotherapy to treat febrile neutropenic patients, clusters of infections due to ceftazidime-resistant K. pneumoniae began to occur. The isolates produced a novel ESBL, TEM-26. After stopping the routine use of empiric ceftazidime monotherapy, the outbreak ended. Similarly, the widespread use of ceftazidime preceded an outbreak of ceftazidime-resistant K. pneumoniae in the Youville hospital, a long-term care and rehabilitation facility in Boston; the incidence declined after restricting ceftazidime use [8]. Interestingly, all three outbreaks involved Klebsiella strains that produced TEM-10 or TEM-26.

Another important warning in the paper by Meyer and colleagues [7] is that the outbreak spread throughout the hospital for a year before it was detected. A major problem is the failure of clinical microbiology laboratories to detect strains producing ESBLs. Most of the ESBLs confer moderately high-level resistance to ceftazidime and aztreonam, but levels of resistance to cefotaxime may be marginal [1]. Laboratories that test only for cefotaxime will often fail to detect such strains. Measuring and recording inhibition-zone diameters would show the lower levels of resistance not detected by existing breakpoints. Using a test based on antibacterial synergy between -lactamase inhibitors and oxyimino--lactams, Sirot and colleagues [4] detected about 50% of the resistant strains misclassified as susceptible by the standard interpretative criteria of disc diffusion. As Meyer and colleagues point out, testing of susceptibility to ceftazidime is more likely to identify ESBL-producing strains. The same is true for aztreonam. Alternate methods of testing or new standards for interpreting existing tests are clearly needed to ensure early detection of multiresistant strains.

Despite much accumulated experience in treating patients infected with ESBL-producing bacteria, few reports detail the specifics of therapeutic outcome with different -lactams [9, 10]. Consequently, little consensus exists on the efficacy of different -lactam antibiotics used alone or combined with -lactamase inhibitors against ESBL-producing bacteria. Many beta-lactams test susceptible with standard inocula but test resistant with larger inocula and show marginal efficacy in animal-model experiments [11, 12]. Using a more stringent criterion, the minimum bactericidal concentration in a macrodilution assay, Meyer and colleagues [7] found that only imipenem had consistent bactericidal activity. Although two patients with nonbacteremic bacteriuria responded to treatment with third-generation cephalosporins, imipenem gave the most favorable clinical results. More studies of the clinical efficacy of different antibiotic regimens in treating infections caused by ESBL-producing bacteria are needed. The cure by cefotaxime, after failure of ceftazidime, of a patient with meningitis due to an ESBL-producing Klebsiella (probably TEM-10) underscores the complexity of this issue [9]. The experience at the Medical Center of Queens illustrates the danger of blanket use of imipenem; namely, the subsequent emergence of an outbreak of infections caused by imipenem-resistant Acinetobacter. Antibiotic roulette is not the answer to limiting the threat of multiresistant microbes. Early detection of ESBL-producing strains and prompt containment through use of barrier precautions is essential to prevent widespread hospital outbreaks, along with judicious use of antimicrobial agents.

Failure to control outbreaks has resulted in the appearance of novel ESBLs in the same institution. In Clermont-Ferrand Hospital, for example, eight different ESBLs appeared after the first surfaced in 1984 [13]. Both plasmid and strain dissemination occurred. The genes that encode at least two ESBLs reside on transposons, mobile genetic elements that can shuttle between plasmids and the chromosome [14, 15]. A TEM-1-bearing plasmid that had been seen 12 years earlier in one hospital recently underwent a single point mutation to produce TEM-12 [14]. Apparently, point mutations and recombination between different -lactamase genes are yielding new ESBLs in the hothouse environments where these genes are endemic [14, 16].

Chromosomal genes that encode even broader spectrum -lactamases have incorporated into plasmids, a long-feared threat. In The Miriam Hospital, nosocomial isolates of K. pneumoniae and Escherichia coli were found to produce the MIR-1 -lactamase, a plasmid-determined chromosomal-type -lactamase derived from Enterobacter cloacae [17]. The strains were highly resistant to all currently licensed -lactams except imipenem. Plasmid-borne genes determining other chromosomal-type -lactamases have also appeared recently in France, Greece, and Pakistan [18-20]. The discovery in Japan of a strain of Pseudomonas aeruginosa producing a plasmid-mediated -lactamase that confers resistance to imipenem and all other -lactams presents an even greater threat [21]. Early detection and prompt containment can limit the spread of these multiresistant pathogens.

• Copyright 2004 by the American College of Physicians