Antibiotic use and resistance confirming the connection

A substantial body of data supports the connection between antibiotic use and resistance. The Meropenem Yearly Susceptibility Test Information Collection (MYSTIC) Program, a global surveillance network initiated to detect changes in carbapenem-resistance rates, reported antibiotic usage and resistance among Gram-negative isolates from 10-15 US institutions from 1999 to 2001. There was a clear population-level connection between drug use and increased resistance to ciprofloxacin for Pseudomonas aeruginosa and Enterobacteriaceae. For instance, ciprofloxacin use doubled to 6.4 defined daily doses (DDD) per 100 patient days by 2001, while Pseudomonal ciprofloxacin resistance doubled to 22.1 percent (Mutnick etal, 2004; Figure 9.6).

Data linking VRE rates - nearly 29 percent in the CDC's latest National Nosocomial Infection Surveillance (NNIS) system - to antibiotic use paint a more nuanced picture than that described above for antibiotic resistance in Gram-negative bacteria (NNIS System, 2004). Results of studies linking antibiotic use with resistant pathogens, such as VRE, depend on factors such as when the study was conducted within the timeline of the pathogen's local emergence, whether the study is prospective or retrospective, patient-level or population-level, and whether it focuses on infection or colonization. Clearly, there has been a temporal relation between vancomycin use, which has increased as much as 20-fold in response to rising MRSA rates, and the emergence of VRE (Ena et al, 1993). For example, as part of the Intensive Care Antimicrobial Resistance Epidemiology (ICARE) project, researchers charted rates of VRE and antibiotic consumption for 120 US ICUs from 1996 to 1999. In their multivariate regression analysis, they found a significant population-level link between vancomycin use and the rates of VRE (Fridkin et al, 2001). Conversely, a meta-analysis of 20 case-control studies failed to find a clear patient-level connection between vancomycin use and VRE colonization or infection. Specifically, after controlling

-■- DDD/100 Patient Days —Ciprofloxacin resistance

-■- DDD/100 Patient Days —Ciprofloxacin resistance

20.4

6.4 22.1

^4.6

11.9 __________

3.1

ta ist si e tr

1999

2000 Time

2001

1999

2000 Time

2001

20 lat o s

ta ist si e tr

Figure 9.6 Correlation between ciprofloxacin use and Pseudomonas resistance from MYSTIC Program, 1999-2001. Adapted from Mutnick et al. (2004).

the data for length of hospital stay, these authors noted that any significant connection between vancomycin use and VRE dissolved (Carmeli et al, 1999). A large case-control study that compared 233 VRE cases with 647 well-matched controls failed to identify antecedent vancomycin use as a risk factor for VRE infection or colonization (Carmeli et al., 2002).

While VRE may owe its initial emergence to vancomycin use, multi-drug resistance ensures that many antibiotic regimens can now promote VRE. The link between third-generation cephalosporin use and cases of VRE has been especially strong. Broad-spectrum third-generation cephalosporins effectively kill Gramnegative colonic micro-flora, leaving the intrinsically-resistant enterococci to multiply. A case-control study of ventilator-associated pneumonia (VAP) patients with VRE (n = 13) versus VAP patients without VRE (n = 25) found a link between VRE and exposure to third-generation cephalosporins, but not vancomycin (Bonten et al., 1996). Also, a meta-analysis of 19 studies found a strong association between previous third-generation cephalosporin exposure and VRE colonization or infection, as well as significant, though weaker, connections between metronidazole and fluoroquinolone exposure and VRE (Harbarth et al., 2002a).

The connection between antibiotic use and resistance also has been confirmed for community-acquired bacterial pathogens. Group A Streptococci (GAS) cause a wide range of serious infections, including pharyngitis, necrotizing soft-t issue infections, and bacteremia, along with immunologic complications such as rheumatic fever and glomerulonephritis. While GAS have remained sensitive to penicillin, macrolide-resistance has been well documented. The frequency of erythromycin-resistant isolates in Finland increased from 5 percent in 1988 to 19 percent in 1993, during a period of increasing erythromycin consumption (Seppala et al, 1992, 1997; Figure 9.7).

Figure 9.7 Consumption of erythromycin by outpatients in Finland from 1976 through 1988, expressed in doses per 1000 people per day. Adapted from Seppala et al. (1997), with permission.

In addition to crude rates of antibiotic consumption, antibiotic use in relation to other demographic parameters may influence rates of resistance. For instance, in examining E. coli and enterococcal isolates from stool samples from healthy volunteers from three cities, researchers found that antibiotic resistance correlated more strongly with drug consumption as a function of population density (expressed as DDD per 1000 inhabitants per day multiplied by the number of inhabitants per km2, or the DDD per km2 per day) than with overall use (DDD per 1000 inhabitants per day) (Bruinsma et al, 2003; Figure 9.8).

Selection pressure from residual antibiotics found in the environment, as a byproduct of either medicinal or agricultural use, also may have an impact on resistance rates. A group in Spain analyzed bacteria from water samples at various sites along the Arga River, downstream of Pamplona. They found that the urban effluent was associated with tetracycline, beta-lactam, and trimethoprim/ sulfamethoxazole resistance among Enterobacteriaceae and Aeromonas isolates, and that this resistance progressively decreased at sampling sites farther from the urban center (Goni-Urriza et al, 2000).

In both the hospital and the community, interactions between individuals -direct person-to-person, or indirect contact via fomites or health-care workers (HCWs) - also drive resistance rates beyond the effects of antibiotic consumption. For example, in hospitals "colonization pressure" - the number of colonized and infected patients - is a major risk factor for acquisition of VRE, MRSA, and Pseudomonas aeruginosa (Bonten et al., 1998, 1999; Merrer et al., 2000).

0 0

Post a comment