SUPPLEMENT ARTICLE. among clinical isolates of S. pneumoniae in the United

Transcription

1 SUPPLEMENT ARTICLE Regional Trends in Antimicrobial Resistance among Clinical Isolates of Streptococcus pneumoniae, Haemophilus influenzae, and Moraxella catarrhalis in the United States: Results from the TRUST Surveillance Program, Clyde Thornsberry, 1 Daniel F. Sahm, 2 Laurie J. Kelly, 2 Ian A. Critchley, 2 Mark E. Jones, 3 Alan T. Evangelista, 4 and James A. Karlowsky 2 Focus Technologies Inc., 1 Nashville, Tennessee, 2 Herndon, Virginia, and 3 Hilversum, The Netherlands; 4 Ortho-McNeil Pharmaceutical Inc., Raritan, New Jersey The ongoing TRUST (Tracking Resistance in the United States Today) study, which began monitoring antimicrobial resistance among respiratory pathogens in 1996, routinely tracks resistance at national and regional levels. The TRUST study analyzed 9499 Streptococcus pneumoniae, 1934 Haemophilus influenzae, and 1108 Moraxella catarrhalis isolates that were prospectively collected from 239 laboratories across the 9 US Bureau of the Census regions. Penicillin-resistant S. pneumoniae varied significantly by region, from 8.3% to 24.8% ( P!.001). In each region, penicillin resistance closely predicted resistance to other b-lactams, macrolides, and trimethoprim-sulfamethoxazole. Levofloxacin resistance was 0.5% nationally (regional range, 0.1% 1.0%). Multidrug resistance also varied significantly ( P!.001 ) by region. b-lactamase production among H. influenzae varied significantly (regional range, 24.0% 34.6%) and M. catarrhalis (86.2% 96.8%) also varied by region. Notable variation in regional antimicrobial resistance rates (S. pneumoniae) and b-lactamase production (H. influenzae, M. catarrhalis) exists throughout the United States. Streptococcus pneumoniae, Haemophilus influenzae, and Moraxella catarrhalis are among the most common bacterial pathogens isolated from respiratory tract infection specimens. Studies published during the last decade have described increasing b-lactam, macrolide, and trimethoprim-sulfamethoxazole (TMP-SMX) resistance among clinical isolates of S. pneumoniae in the United Financial support: Ortho-McNeil Pharmaceutical Inc. (Raritan, New Jersey). Reprints or correspondence: Dr. James A. Karlowsky, Focus Technologies, Dulles Technology Dr., Suite 200, Herndon, VA (jkarlowsky@focus answers.com). Clinical Infectious Diseases 2002; 34(Suppl 1):S by the Infectious Diseases Society of America. All rights reserved /2002/3405S1-0002$03.00 States [1 13]. The most recent US surveillance studies have shown that resistance continues to increase in S. pneumoniae [2, 3, 5 7]. Current evidence also suggests that conducting surveillance studies of S. pneumoniae on at least an annual basis is important because resistance can change significantly in a single year [7]. Other studies have shown that previously increasing rates of b-lactamase production among H. influenzae and M. catarrhalis now appear to be leveling off in the United States, at least temporarily [5, 8, 9, 14 16]. United States national surveillance initiatives, such as those cited above, are often not regionally comprehensive or longitudinal because of the logistical commitment and expense required for such a study. Point prevalence studies make it difficult to establish regional and national S4 CID 2002:34 (Suppl 1) Thornsberry et al.

2 resistance trends over time. Nonetheless, the limited number of previous studies describing regional trends in the United States have suggested that regional variations in penicillin resistance may exist among pneumococci [5, 6, 8]. The ongoing TRUST (Tracking Resistance in the United States Today) study, which began monitoring antimicrobial resistance among key respiratory pathogens in the United States in 1996, is one of the few studies that can provide a comprehensive regional assessment of antimicrobial activity among key respiratory tract pathogens in the United States. Multidrug resistance (defined as resistance to 3 or more antimicrobials of different classes) among pneumococci has also become a concern in the United States after first being reported in 1992 [1, 10, 13, 17]. A publication describing the Active Bacterial Core Surveillance program of the Centers for Disease Control and Prevention exemplifies this concern: the authors saw the proportion of multidrug-resistant (MDR) isolates increase from 9% to 14% among patients with invasive pneumococcal disease [10]. Similarly, in a comparison of TRUST data from the and respiratory seasons, multidrug resistance among pneumococci, predominantly respiratory tract and blood isolates, increased from 5.9% to 11% [7]. However, to date, a comprehensive US regional analysis of multidrug resistance in S. pneumoniae has not been published. To address the paucity of published US regional surveillance data, the current study presents a synopsis of the TRUST study, in which 9499 S. pneumoniae, 1934 H. influenzae, and 1108 M. catarrhalis were prospectively collected from 239 laboratories distributed throughout the 9 US Bureau of the Census regions and centrally tested for antimicrobial susceptibilities. The prevalence and composition of multidrug resistance by region was determined, and susceptibility data from were compared with that from the preceding respiratory season. MATERIALS AND METHODS Bacterial isolates. Clinical isolates of S. pneumoniae (n p 9499), H. influenzae ( n p 1934), and M. catarrhalis ( n p 1108) were prospectively collected from 239 microbiology laboratories distributed across the United States from October 1999 to April Overall antimicrobial susceptibility data from were compared with data from the respiratory season. The data were collected from September 1998 to March 1999 from 99 laboratories, resulting in a total of 4296 S. pneumoniae, 1148 H. influenzae, and 573 M. catarrhalis, as previously reported [7]. Overall, 258 laboratories contributed isolates to at least 1 of these 2 study years, and 79 laboratories (31%) participated in both study years. In both years, isolates were submitted to and verified by Focus Technologies central laboratory (Herndon, VA) as described previously [7, 8]. Participating laboratories served institutions of varying size (!100 to 1999 beds), patient demographics, and specialty, and were representatively distributed by population density throughout the United States. Isolates were accepted without regard to patient age, sex, inpatient or outpatient status, or specimen source. Each laboratory was asked to submit S. pneumoniae, 7 12 H. influenzae, and 4 5 M. catarrhalis samples in each respiratory season. To analyze data regionally, laboratories were allocated to their appropriate US Bureau of the Census region (table 1), as follows: Pacific (Washington, Oregon, California, Alaska, Hawaii); Mountain (Idaho, Montana, Wyoming, Nevada, Utah, Colorado, Arizona, New Mexico); West North (North Dakota, South Dakota, Minnesota, Nebraska, Iowa, Kansas, Missouri); West South (Oklahoma, Arkansas, Texas, Louisiana); East North (Wisconsin, Michigan, Illinois, Indiana, Ohio); East South (Kentucky, Tennessee, Alabama, Mississippi); New England (Maine, New Hampshire, Vermont, Massachusetts, Rhode Island, Connecticut); Mid-Atlantic (New York, Pennsylvania, New Jersey); and South Atlantic (Maryland, Delaware, West Virginia, Virginia, North Carolina, South Carolina, Georgia, Florida). One laboratory in Washington, DC, also participated and was included in the South Atlantic region. The numbers of participating laboratories in each of the 9 regions in the respiratory seasons were as follows: 21, Pacific; 14, Mountain; 25, West North ; 20, West South ; 55, East North ; 13, East South ; 14, New England; 37, Mid- Atlantic; and 40, South Atlantic. Antimicrobial susceptibility testing. The following antimicrobial agents were tested in both study years: penicillin, amoxicillin-clavulanate, cefuroxime, ceftriaxone, azithromycin, clarithromycin, TMP-SMX, vancomycin, and levofloxacin. In , erythromycin and clindamycin were added to the MIC testing panel. These additions resulted in slightly different antimicrobial concentration ranges being tested for other agents in each respiratory season. Susceptibility testing was performed by broth microdilution according to the National Committee for Clinical Laboratory Standards (NCCLS) guidelines (M7- A4) with panels manufactured by TREK Diagnostics (Westlake, OH) [18]. Panels were incubated at 35 C for 20 to 24 h in ambient air before MIC determination. S. pneumoniae American Type Culture Collection (ATCC) and H. influenzae ATCC were used as daily panel quality-control strains. For S. pneumoniae and H. influenzae, MICs were interpreted on the basis of NCCLS recommended breakpoints; we used those in place at the start of the respiratory season ( , M100-S9 [19]; , M100-S8 [20]). MDR S. pneumoniae. Regional prevalence and pheno- Trends in Antimicrobial Resistance CID 2002:34 (Suppl 1) S5

3 Table 1. Geographic distribution of 239 participating laboratories across the 9 US Bureau of the Census regions, No. of Region (total no. of laboratories), state a laboratories Pacific (n p 21) Washington 4 Oregon 2 California 15 Mountain (n p 14) Idaho 1 Nevada 1 Utah 3 Colorado 5 Arizona 2 New Mexico 2 West North (np 25) North Dakota 1 South Dakota 3 Nebraska 4 Kansas 3 Missouri 8 Iowa 4 Minnesota 2 West South (np 20) Texas 12 Oklahoma 2 Arkansas 4 Louisiana 2 East North (np 55) Wisconsin 5 Michigan 19 Illinois 12 Indiana 6 Ohio 13 East South (np 13) Kentucky 4 Tennessee 5 Alabama 4 New England (n p 14) Maine 1 Vermont 1 New Hampshire 1 Connecticut 3 Massachusetts 8 Mid-Atlantic (np 37) New York 15 Pennsylvania 16 New Jersey 6 South Atlantic (n p 40) West Virginia 3 Virginia 4 Maryland 6 Delaware 1 Washington, DC 1 North Carolina 7 South Carolina 4 Georgia 6 Florida 8 a The following states had no participating laboratories: Alaska, Hawaii, Mississippi, Montana, Rhode Island, and Wyoming. types of MDR S. pneumoniae were also studied in isolates from the respiratory season. Multidrug resistance was defined as resistance to 3 or more of the following antimicrobials: penicillin, ceftriaxone, azithromycin, TMP-SMX, and levofloxacin. Patient sex, age, and inpatient or outpatient status were also analyzed to determine the impact of these demographic parameters on the isolates resistant to individual antimicrobial agents and for isolates demonstrating an MDR phenotype. Statistical analyses. Statistical analyses were performed by x 2 testing with Epi Info Statcalc, version 6.0 (Centers for Disease Control and Prevention). Uncorrected P values of!.05 were reported as statistically significant; P!.001 was considered highly statistically significant. RESULTS S. pneumoniae. Results obtained for the antimicrobial agents tested against isolates of S. pneumoniae in (TRUST III) and (TRUST IV) are displayed in table 2. For 7 of the 9 agents tested in both respiratory seasons, there was an increase in the percentage of resistant isolates between years. The increase was statistically significant for amoxicillin-clavulanate ( P!.001), cefuroxime ( P!.05), clarithromycin ( P!.001), and TMP-SMX ( P!.05). Penicillin resistance increased from 14.7% in to 16.0% in and approached statistical significance ( P p.05). Vancomycin-resistant isolates were not identified in either year, and the percentage of isolates resistant to levofloxacin decreased from 0.6% to 0.5%. Azithromycin and clarithromycin susceptibilities remained identical over the 2 respiratory seasons, whereas percentages of azithromycin-intermediate and clarithromycin-resistant isolates increased significantly over the same time. Modal MICs and MIC 90 s remained identical between the 2 respiratory seasons for 6 of the 9 agents tested in both respiratory seasons. Changes in modal MIC or MIC 90 for azithromycin, clarithromycin, and vancomycin were universally one-doubling-dilution decreases or increases in MIC from to (table 2). Although amoxicillin-clavulanate resistance increased by 3.7% from to , penicillin resistance was still higher in both respiratory seasons (table 2). During the respiratory season, a change occurred in pneumococcal MIC interpretative breakpoints (introduced 1 January 2000) for cefuroxime and amoxicillin-clavulanate [21]. MIC data presented from the current study were interpreted according to the NCCLS breakpoints in clinical use at the start of the respiratory season [19]. However, when 2000 NCCLS breakpoints [21] were applied for categorical interpretation of amoxicillin-clavulanate MICs, resistance decreased from 10.5% to 2.6%, reclassifying amoxicillin-clavulanate as 1 of 3 agents with the lowest levels of resistance. The application of new cefuroxime breakpoints, specifically for cefuroxime axetil S6 CID 2002:34 (Suppl 1) Thornsberry et al.

4 Table 2. Antimicrobial susceptibility results for Streptococcus pneumoniae isolated in the United States during the (n p 4296) and (n p 9499) respiratory seasons. Antimicrobial Respiratory season MIC, mg/ml Interpretation, % Range Mode MIC 90 S I R Penicillin a 0.03 to Amoxicillinclavulanate b 0.03 to to to Cefuroxime to to Ceftriaxone to to Erythromycin c to Azithromycin to to Clarithromycin to to Clindamycin c to 10.5 NA NA Levofloxacin to to TMP-SMX to to Vancomycin NOTE. I, intermediate; NA, not available; R, resistant; S, susceptible; TMP-SMX, trimethoprimsulfamethoxazole. a National Committee for Clinical Laboratory Standards (NCCLS) M100-S8 susceptible, intermediate, and resistant breakpoints were used to interpret MIC data [19]. b NCCLS M100-S9 susceptible, intermediate, and resistant breakpoints were used to interpret MIC data [18]. c Erythromycin and clindamycin were tested in only. Clindamycin susceptibility was determined by the presence or absence of growth in a single well containing the agent at a concentration of 0.5 mg/ml. (orally administered) interpretations, to the results of the current data set was less dramatic than for amoxicillin-clavulanate, with 72.6% of isolates interpreted as susceptible, 4.6% as intermediate, and 22.8% as resistant (data not shown) [21]. However, as in any surveillance study, the MIC distribution is a better indicator than the MIC 90 or categorical interpretation for tracking subtle shifts in MIC profiles, as shown in table 3. This is especially true for tracking resistance over successive respiratory seasons in instances of NCCLS breakpoint changes. Clindamycin and erythromycin resistance phenotypes suggested that 684 (7.2%) of 9499 S. pneumoniae isolates likely possess resistance arising from ribosomal methylation (erythromycin resistant, clindamycin resistant), whereas 3 times as many (1847 of 9499; 19.4%) macrolide-resistant isolates appear to be resistant as a result of efflux (erythromycin resistant, clindamycin susceptible) (data not shown) [22, 23]. Among penicillin-resistant isolates, those with MICs of 2 mg/ ml increased from 10.9% in to 11.9% in Isolates with MICs of 4 and 4 mg/ml remained relatively stable at 3.5% and 0.3% ( ) and 3.7% and 0.4% ( ), respectively (data not shown). MIC distributions for each agent tested against pneumococci when studied by the penicillin-susceptible, -intermediate, and -resistant status of isolates demonstrated that MIC 90 s and modal MICs were 5- doubling-dilutions higher among penicillin-resistant isolates than among penicillin-susceptible isolates for all b-lactams and macrolides studied (tables 2 and 3). In contrast, the MIC 90 s and modal MICs for vancomycin and levofloxacin were unaffected by pneumococcal penicillin susceptibility. The prevalence of penicillin-resistant S. pneumoniae varied significantly among the 9 US Bureau of the Census regions ( P!.001) (table 4). It was highest in the South Atlantic region (24.8%) and lowest in New England (8.3%). The prevalence of penicillin-intermediate pneumococci showed less variation Trends in Antimicrobial Resistance CID 2002:34 (Suppl 1) S7

5 Table 3. Antimicrobial MIC distributions according to penicillin-susceptibility for 9499 S. pneumoniae isolates (6262 penicillin-susceptible isolates, 1720 penicillin-intermediate isolates, and 1517 penicillin-resistant isolates) recovered during the respiratory season. Antimicrobial Percentage of isolates per MIC, mg/ml Penicillin Penicillin-S Penicillin-I Penicillin-R Amoxicillinclavulanate Penicillin-S Penicillin-I Penicillin-R Cefuroxime Penicillin-S Penicillin-I Penicillin-R Ceftriaxone Penicillin-S Penicillin-I Penicillin-R Erythromycin Penicillin-S Penicillin-I Penicillin-R Azithromycin Penicillin-S Penicillin-I Penicillin-R Clarithromycin Penicillin-S Penicillin-I Penicillin-R Levofloxacin Penicillin-S Penicillin-I Penicillin-R TMP-SMX Penicillin-S Penicillin-I Penicillin-R Vancomycin Penicillin-S Penicillin-I Penicillin-R NOTE. The MIC 90 is presented in boldface type. I, intermediate; R, resistant; S, susceptible; TMP-SMX, trimethoprim-sulfamethoxazole. S8

6 Table 4. Antimicrobial susceptibility results for Streptococcus pneumoniae isolated in each of the 9 US Bureau of the Census regions during the respiratory season. Antimicrobial, interpretation Pacific (n p 945) Mountain (n p 585) Percentage of isolates per US Bureau of the Census region West North (n p 970) West South (n p 812) East North (n p 2218) East South (n p 679) New England (n p 556) Mid-Atlantic (n p 1245) South Atlantic (n p 1489) Penicillin S I R Amoxicillinclavulanate, S I R Cefuroxime S I R Ceftriaxone S I R Vancomycin S NS Erythromycin S I R Azithromycin S I R Clarithromycin S I R Clindamycin S NS Levofloxacin S I R TMP-SMX S I R NOTE. I, intermediate; NS, nonsusceptible; R, resistant; S, susceptible, TMP-SMX, trimethoprim-sulfamethoxazole. S9

7 Table 5. Multidrug-resistant phenotypes and their prevalences among clinical isolates of Streptococcus pneumoniae collected from across the United States during the respiratory season. MDR phenotype No. of isolates MDR isolates, % Total isolates, % (n p 9499) Resistance to 3 antimicrobials Penicillin, azithromycin, TMP-SMX Penicillin, ceftriaxone, TMP-SMX Ceftriaxone, azithromycin, TMP-SMX Penicillin, ceftriaxone, azithromycin 8 0.7!0.1 Azithromycin, TMP-SMX, levofloxacin 8 0.7!0.1 Penicillin, TMP-SMX, levofloxacin a 3 0.3!0.1 Penicillin, azithromycin, levofloxacin 1 0.1!0.1 Ceftriaxone, TMP-SMX, levofloxacin b 1 0.1!0.1 Total Resistance to 4 antimicrobials Penicillin, ceftriaxone, azithromycin, TMP-SMX Penicillin, azithromycin, TMP-SMX, levofloxacin 7 0.6!0.1 Ceftriaxone, azithromycin, TMP-SMX, levofloxacin b 1 0.1!0.1 Total Resistance to 5 antimicrobials Penicillin, ceftriaxone, azithromycin, TMP-SMX, levofloxacin 4 0.4!0.1 Total 4 0.4!0.1 Overall total NOTE. MDR, multidrug-resistant; TMP-SMX, trimethoprim-sulfamethoxazole. a Isolates are ceftriaxone intermediate. b Isolates are penicillin intermediate. by region, ranging from 21.2% in the East South region to 15.8% in New England. Although nonsusceptibility to penicillin (i.e., combined penicillin-intermediate and -resistant isolates) varied by region, all regions had rates 124%. Geographic trends for penicillin resistance were predictably duplicated for the other b-lactams and macrolides tested. The prevalence of erythromycin resistance ranged 17.1% 36.1% and was 125% in 5 of the 9 regions. Clindamycin resistance ranged 4.5% 9.7% across the 9 regions. Resistance to TMP-SMX exceeded 25% in 7 regions. Other than vancomycin, for which resistance has not been reported, levofloxacin demonstrated the lowest prevalence of resistance, with regional rates ranging 0.1% 1.0%. MDR isolates accounted for 12.2% ( n p 1155) of the 9499 S. pneumoniae isolates in (table 5), an increase from 11.0% (472 of 4296) in (data not shown). The majority of MDR isolates (83.1%) were resistant to 3 antimicrobials and accounted for 10.1% of all isolates. Isolates were also identified that were resistant to 4 agents (16.5% of MDR isolates; 2.0% of all isolates) and all 5 agents used to define multidrug resistance (0.3% of MDR isolates; 0.04% of all isolates). Isolates resistant to all 5 agents were not identified in the respiratory season study or in any previous TRUST study from their inception in 1996 (data not shown). Concurrent resistance to penicillin, azithromycin (representative macrolide), and TMP- SMX was the most common MDR phenotype and accounted for 71.2% of MDR isolates. Resistance to penicillin was a component of 1133 (98.1%) of 1155 of MDR isolates. Similarly, 99.2% ( n p 1146 ), 90.6% ( n p 1046), 36.5% ( n p 422), and 2.2% ( n p 25) of MDR isolates were resistant to TMP-SMX, azithromycin, ceftriaxone, and levofloxacin, respectively. Antimicrobial resistance rates for penicillin, ceftriaxone, azithromycin, TMP-SMX, and levofloxacin, as well as the percentage of isolates demonstrating an MDR phenotype, were stratified by patient demographic characteristics (table 6). Patient sex did not significantly ( P 1.05) affect individual antimicrobial agent susceptibilities or multidrug resistance. Penicillin, ceftriaxone, azithromycin, and TMP-SMX resistance was significantly ( P!.05) more common among patients!18 years of age than among older patients. Levofloxacin-resistant isolates were not identified among patients!18 years and were more common ( P p.04) among patients aged 165 years than among patients aged years; however, the prevalence of levofloxacin resistance in both groups was low at!1%. Multidrug resistance was significantly ( P!.001) more common among S10 CID 2002:34 (Suppl 1) Thornsberry et al.

8 Table 6. Demographic characteristics of patients from whom Streptococcus pneumoniae was isolated during the respiratory season. Patient category a Total isolates, n Percentage of isolates resistant to Penicillin Ceftriaxone Azithromycin TMP-SMX Levofloxacin MDR isolates, % (n) Sex Female (517) Male (608) Age, years! (342) (435) (338) Location Inpatient (650) Outpatient (464) NOTE. MDR, multidrug-resistant; TMP-SMX, trimethoprim-sulfamethoxazole. a Demographic information was provided by the participating laboratories for the majority of isolates tested; however, sex, age, and location were not provided for 235, 294, and 278 patients, respectively. patients aged!18 years than among older patients; the prevalence of MDR phenotypes was lowest among patients aged years. Resistance to penicillin, azithromycin, and TMP- SMX, as well as multidrug resistance, were significantly (P!.05) more common among outpatients than among inpatients. Inpatient or outpatient status did not affect resistance to levofloxacin or ceftriaxone. The most common MDR phenotype overall (table 5) resistance to penicillin, azithromycin (representative macrolide), and TMP-SMX was also individually the most prevalent among male and female patients; among patients!18, 18 65, and 165 years of age; and among inpatients and outpatients (data not shown). Table 7 depicts the prevalence of single-drug resistance and multidrug resistance and the most prevalent single-drug-resistant and MDR phenotypes by region in the United States. Singledrug resistance did not vary significantly by region and demonstrated a narrow range, 10.5% 16.3%. MDR rates demonstrated significant variation (P!.001), from 4.7% in New England to 20.8% in the South Atlantic region. In all 9 geographic regions, resistance to TMP-SMX was the most prevalent singledrug-resistant phenotype, and resistance to penicillin, azithromycin, and TMP-SMX was the most prevalent MDR phenotype. H. influenzae. Table 8 presents a summary of H. influenzae MIC susceptibilities for data collected during the 2 consecutive respiratory seasons. b-lactamase production among H. influenzae decreased from 34.5% (397 of 1152 isolates) in Table 7. Prevalence of single-drug resistance and multidrug resistance among Streptococcus pneumoniae in each of the 9 US Bureau of the Census regions. Region n SDR, % a Most prevalent SDR phenotype (% of SDR) MDR, % b Most prevalent MDR phenotype (% of MDR) Pacific TMP-SMX (61) 7.3 Penicillin, azithromycin, TMP-SMX (75) Mountain TMP-SMX (66) 9.6 Penicillin, azithromycin, TMP-SMX (75) West North TMP-SMX (52) 12.5 Penicillin, azithromycin, TMP-SMX (69) West South TMP-SMX (58) 14.0 Penicillin, azithromycin, TMP-SMX (77) East North TMP-SMX (51) 11.2 Penicillin, azithromycin, TMP-SMX (75) East South TMP-SMX (51) 18.4 Penicillin, azithromycin, TMP-SMX (70) New England TMP-SMX (61) 4.7 Penicillin, azithromycin, TMP-SMX (62) Mid-Atlantic TMP-SMX (51) 6.9 Penicillin, azithromycin, TMP-SMX (63) South Atlantic TMP-SMX (68) 20.8 Penicillin, azithromycin, TMP-SMX (68) All regions TMP-SMX (57) 12.2 Penicillin, azithromycin, TMP-SMX (71) NOTE. MDR, multidrug-resistant; SDR, single-drug resistance; TMP-SMX, trimethoprim-sulfamethoxazole. a Resistance to only one of the following: penicillin, ceftriaxone, azithromycin, TMP-SMX, or levofloxacin. b Resistance to 3 or more of the following: penicillin, ceftriaxone, azithromycin, TMP-SMX, or levofloxacin. Trends in Antimicrobial Resistance CID 2002:34 (Suppl 1) S11

9 Table 8. Antimicrobial susceptibility results for Haemophilus influenzae and Moraxella catarrhalis isolated in the United States during the (1152 isolates of H. influenzae and 573 isolates of M. catarrhalis) and (1934 isolates of H. influenzae and 1108 isolates of M. catarrhalis) respiratory seasons. Antimicrobial Respiratory season H. influenzae M. catarrhalis, MIC, mg/ml Interpretation, % MIC, mg/ml Range Mode MIC 90 S I R Range Mode MIC 90 Ampicillin a 0.12 to to b 0.12 to to Amoxicillinclavulanate to Cefuroxime Ceftriaxone Azithromycin to Clarithromycin to to Levofloxacin TMP-SMX to to NOTE. I, intermediate; R, resistant; S, susceptible; TMP-SMX, trimethoprim-sulfamethoxazole. a National Committee for Clinical Laboratory Standards (NCCLS) M100-S8 susceptible, intermediate, and resistant breakpoints were used to interpret MIC data [19]. b NCCLS M100-S9 susceptible, intermediate, and resistant breakpoints were used to interpret MIC data [18]. to 31.3% (606 of 1934 isolates) in (data not shown). Concomitantly, ampicillin resistance decreased from 33.8% to 30.7%. In the respiratory season, b-lactamase production varied significantly ( P!.001 ), from 24.0% (30 of 125 isolates) in the Mountain region to 37.3% (57 of 153 isolates) in New England (data not shown). As expected, the prevalence of ampicillin-resistant H. influenzae also varied significantly ( P!.001) by region (table 9). It was highest in New England (34.6%) and lowest in the Mountain region (24.0%), correlating closely with b-lactam production. b-lactamase negative, ampicillin-resistant isolates occurred at a rate of 0.5% (4 isolates) in and 0.3% (4 isolates) in among b-lactamase negative isolates. b-lactamase positive, amoxicillin-clavulanate-resistant isolates were rare in both respiratory seasons, with 1 isolate in and 2 isolates in All b-lactamase negative, ampicillin-resistant isolates and b-lactamase positive, amoxicillin-clavulanate-resistant isolates were confirmed by repeat MIC and b-lactamase testing. As expected, MIC distributions demonstrated that the MIC 90 for ampicillin increased from 0.5 mg/ml in b-lactamase negative isolates to 18.0 mg/ml in the b-lactamase positive population (data not shown). The impact of b- lactamase production on the activities of the other agents tested was minimal, with MIC distributions for b-lactamase positive and b-lactamase negative isolates being virtually identical (data not shown). Significant changes ( P!.05) in susceptibility between and were not identified for any of the 8 agents tested against H. influenzae. Resistance to amoxicillinclavulanate, cefuroxime, azithromycin, and clarithromycin was!1% in both years. Ceftriaxone- or levofloxacin-resistant isolates were not identified in either year. TMP-SMX resistance increased the most, from 11.7% in to 14.0% in The prevalence of TMP-SMX resistance in ranged from 11.7% in the East South region to 20.9% in the Pacific region and was 15% in 4 of the 9 regions (table 9). One isolate of H. influenzae with a levofloxacin MIC of 2 mg/ml was identified; it was the only isolate with a levofloxacin MIC 0.25 mg/ml. M. catarrhalis. In , 1035 (93.4%) of 1108 of M. catarrhalis isolates were b-lactamase positive, a prevalence similar to that found in , 94.9% (544 of 573 isolates). In the respiratory season, b-lactamase production varied from 86.2% (50 of 58 isolates) in New England to 96.8% (122 of 126 isolates) in the West North region (data not shown). Because M. catarrhalis does not have NCCLS in- S12 CID 2002:34 (Suppl 1) Thornsberry et al.

10 Table 9. Antimicrobial susceptibility results for Haemophilus influenzae isolated in each of the 9 US Bureau of the Census regions during the respiratory season. Antimicrobial, interpretation Pacific (n p 139) Mountain (n p 125) Percentage of isolates per US Bureau of the Census region West North (n p 200) West South (n p 161) East North (n p 440) East South (n p 137) New England (n p 153) Mid-Atlantic (n p 258) South Atlantic (n p 321) Ampicillin S I R Amoxicillinclavulanate S R Cefuroxime S I R Ceftriaxone S NS Azithromycin S NS Clarithromycin S I R Levofloxacin S NS TMP-SMX S I R NOTE. I, intermediate; NS, nonsusceptible; R, resistant; S, susceptible; TMP-SMX, trimethoprim-sulfamethoxazole. terpretive criteria, the assessment of antimicrobial activity in this species is best accomplished by comparisons of modal MICs and MIC 90 s, and shifts in MIC distributions. Although b-lactamase production elevates ampicillin MICs, 90.2% of all isolates had MICs of 4 mg/ml (data not shown). As expected, the effect of b-lactamase production on MIC distributions for M. catarrhalis was most apparent in the b-lactams and cephalosporins ( 2-doubling-dilution increase in MIC 90 ; data not shown). Modal MICs and MIC 90 s remained identical in and for 3 of the 8 agents tested in both respiratory seasons; changes in modal MIC or MIC 90 for ampicillin, ceftriaxone, clarithromycin, levofloxacin, and TMP- SMX were universally one-doubling-dilution decreases or increases in MIC from to (table 8). MIC distributions for all 8 agents tested against M. catarrhalis were virtually identical for the and respiratory seasons (data not shown). DISCUSSION In the current study, we observed that antimicrobial resistance among pneumococci in the United States varied by region, as did the production of b-lactamase by clinical isolates of H. influenzae and M. catarrhalis. In addition, penicillin resistance (MIC 2 mg/ml) increased from 14.7% in to 16.0% in , accompanied by statistically significant increases Trends in Antimicrobial Resistance CID 2002:34 (Suppl 1) S13

11 in amoxicillin-clavulanate, cefuroxime, clarithromycin, and TMP-SMX resistance during the 2 years (table 2). Our data confirm a continuing trend toward increasing resistance to penicillin, other b-lactams, macrolides, and TMP-SMX among S. pneumoniae in the United States; this has also been identified by testing isolates from previous years [1 13]. The high prevalences of penicillin resistance in other countries (e.g., 36.5% in Spain [24] and 19.6% in Hong Kong [25]) suggest that potential exists for continuing increases in penicillin resistance among S. pneumoniae in the United States. However, as described by other investigators, the broadening use of conjugate pneumococcal vaccines in the United States may influence the evolution of antimicrobial resistance [10]. The prevalence of penicillin resistance among S. pneumoniae isolates in the current study varied significantly by geographic region in the United States ( P!.001). Among the 9 US Bureau of the Census regions, penicillin resistance was highest in the South Atlantic region (24.8%) and lowest in New England (8.3%). The results of the current study were similar to those reported by a respiratory season surveillance study in which penicillin resistance was highest in the East South (21.4%) and South Atlantic (19.5%) regions and lowest in the Mid-Atlantic region (7.7%) and New England (10.4%) [8]. Another study, which defined regions differently, reported that during 1997, the South (46.2%) and South East (40.1%) regions had the highest rates of penicillin resistance among outpatient S. pneumoniae isolates [5]. Another published study, which also defined regions differently than the current study, reported that penicillin resistance varied substantially by region in the United States, ranging from 5.8% in the North East region and 7.7% in Minnesota to 20.1% in Florida [6]. Analysis by MIC interpretative category predictably showed that vancomycin (0%) and levofloxacin (0.5%) were the agents with the lowest levels of in vitro resistance against S. pneumoniae (table 2). In the respiratory season, 51 levofloxacinresistant isolates were found in a study population of 9499 (0.5%). In , 25 levofloxacin-resistant pneumococcal isolates were identified out of 4296 tested (0.6%) [7], suggesting that levofloxacin resistance remains rare. All isolates in both years were confirmed by repeat susceptibility testing. In the respiratory season, 199% of all pneumococcal isolates were levofloxacin susceptible; resistance to all other agents except vancomycin (0%), ceftriaxone (3.8%), and clindamycin (7.2%) exceeded 10%. In the case of macrolides, cefuroxime, and TMP-SMX, resistance was 120%. These findings concur with those of a report authored by investigators at the Centers for Disease Control and Prevention that identified only 7 levofloxacin-resistant isolates among 3475 clinical pneumococcal isolates tested (!1% resistance) [10]. Although it is difficult to estimate the extent to which fluoroquinolone resistance will increase, the potential of pneumococci to acquire resistance to fluoroquinolones dictates that surveillance be ongoing. Currently, fluoroquinolone resistance in the United States does not appear to be due to clonal distribution of strains [7] and was not associated with the most prevalent MDR phenotypes (table 5). In our study, penicillin-resistant isolates of S. pneumoniae were predictably more highly resistant to the other b-lactams tested than penicillin-susceptible and -intermediate isolates (table 3). It has been suggested that this correlation may be used in the clinical microbiology laboratory to report susceptibilities to amoxicillin, amoxicillin-clavulanate, cefuroxime, ceftriaxone, cefpodoxime, cefixime, cefprozil, cefaclor, and loracarbef [26]. However, in our study, we found that some penicillinresistant isolates were susceptible or intermediate to other b- lactams specifically, amoxicillin-clavulanate (susceptible, 1%; intermediate, 21%) and ceftriaxone (susceptible, 12%; intermediate, 67%). Penicillin resistance, however, was found to be a reliable predictor of cefuroxime resistance because 99.9% of penicillin-resistant isolates were also cefuroxime-resistant. Tracking the prevalence of MDR pneumococci is important; these phenotypes, which continue to emerge, currently account for 110% of clinical isolates in the United States and limit therapeutic treatment options for doctors [7, 10]. Although multidrug resistance is a growing concern, it has been addressed in only a few previous studies [1, 2, 4, 7, 10]. In the most recent publication of this type, we reported a near doubling in the percentage of MDR isolates in the 2 previous respiratory seasons of the TRUST longitudinal study: from 5.9% in to 11% in Often, previous studies are difficult to compare and interpret because of differences in methods. For example, in a study by Butler et al. [1], the definition of multidrug resistance included isolates at or above the intermediate breakpoint, whereas our criteria included only isolates resistant by NCCLS standards [19]. Other studies that used different definitions of multidrug resistance [2, 4] have been published; some included both penicillin-intermediate and penicillin-resistant isolates or considered isolates resistant to only 2 agents (e.g., penicillin, and TMP-SMX) as MDR [2]. From the current study, it is important to note that 89.8% of MDR phenotypes (table 5) included resistance to a b-lactam and a macrolide, the antimicrobial classes most commonly used to treat pneumococcal infections. This finding is consistent with data gathered in the 2 previous respiratory seasons of the TRUST surveillance initiative [7]. Monitoring MDR phenotypes is a key component of pneumococcal surveillance initiatives and should now be considered mandatory. Overall, the current prevalence of b-lactamase production among H. influenzae in the United States has been reported to be 30% 40% [8, 12, 14, 16]. In the current study, as has been noted in other US studies, b-lactamase production among H. influenzae and M. catarrhalis appears to have plateaued na- S14 CID 2002:34 (Suppl 1) Thornsberry et al.

12 tionally [5, 8, 9, 14 16]. Previously, b-lactamase production among H. influenzae has been on the increase in the United States: 16.5% in 1990 [27], 26.3% in 1992 [12, 28], 28.2% in 1993 [12, 28], 30.1% in 1994 [12, 28], 36.4% in [14], 33.4% to 34.2% in 1997 [8, 9, 16], and 34.5% in (table 8). In , 31.3% of 1934 H. influenzae were identified as b-lactamase positive (table 8). It appears generally that b-lactamase production among H. influenzae leveled off from 1998 through b-lactamase negative, ampicillin-resistant strains and b-lactamase positive, amoxicillinclavulanate-resistant strains were rare in this study (0.3% and 0.4%, respectively) but have been previously described in the literature [14]. A single previous study has described antimicrobial resistance among H. influenzae by region in the United States [5]. In that study, b-lactamase production was found to be highest in the Northeastern (50.3%) and North (48.3%) regions and lowest in the Northwestern (36.7%) and Southwestern (35.4%) regions. North American and European studies have reported rates of b-lactamase production in clinical isolates of M. catarrhalis of 190% [8, 12]. The current study found an overall rate of 93.4% and regional rates that varied from 86.2% (50 of 58 isolates) in New England to 96.8% (122 of 126 isolates) in the West North region. Previous reports of regional rates of b-lactamase production in the United States are not available for comparison; however, it appears that until the late 1970s, isolates of M. catarrhalis had not yet acquired b-lactamases [29]. In contrast, currently, nearly all isolates harbor 1 of 2 BRO b-lactamases [30]. In conclusion, substantial regional variations in antimicrobial susceptibility were demonstrated for S. pneumoniae, H. influenzae, and M. catarrhalis in the United States. Ongoing surveillance of antimicrobial activity is essential because initial therapy for respiratory tract infections is initiated largely empirically, and surveillance is the only mechanism by which changing resistance trends can be identified at the local, regional, and national levels. Acknowledgments We thank David Diakun of Focus Technologies Information Systems for providing technical assistance in preparation of this manuscript. We are especially grateful to all of the clinical testing institutions that participated in both TRUST surveillance studies by contributing valuable data to this study. References 1. Butler JC, Hofmann J, Cetron MS, Elliott JA, Facklam RR, Breiman RF, and the Pneumococcal Sentinel Surveillance Working Group. The continued emergence of drug-resistant Streptococcus pneumoniae in the United States: an update from the Centers for Disease Control and Prevention s Pneumococcal Sentinel Surveillance System. J Infect Dis 1996; 174: Doern GV, Brueggemann AB, Huynh H, Wingert E, Rhomberg P. Antimicrobial resistance with Streptococcus pneumoniae in the United States, Emerg Infect Dis 1999; 5: Doern GV, Pfaller MA, Kugler K, Freeman J, Jones RN. Prevalence of antimicrobial resistance among respiratory tract isolates of Streptococcus pneumoniae in North America: 1997 results from the Sentry Antimicrobial Surveillance Program. Clin Infect Dis 1998; 27: Doern GV, Brueggemann AB, Holley HP, Rauch A. Antimicrobial resistance of Streptococcus pneumoniae recovered from outpatients in the United States during the winter months of 1994 to 1995: results of a 30-center national surveillance study. Antimicrob Agents Chemother 1996; 40: Jacobs MR, Bajaksouzian S, Zilles A, Lin G, Pankuch GA, Appelbaum PC. Susceptibilities of Streptococcus pneumoniae and Haemophilus influenzae to 10 oral antimicrobial agents based on pharmacodynamic parameters: 1997 US surveillance study. Antimicrob Agents Chemother 1999; 43: Mason EO Jr, Lamberth LB, Kershaw NL, Prosser BL, Zoe A, Ambrose PG. Streptococcus pneumoniae in the USA: in vitro susceptibility and pharmacodynamic analysis. J Antimicrob Chemother 2000; 45: Sahm DF, Karlowsky JA, Kelly LJ, et al. Need for annual surveillance of antimicrobial resistance in Streptococcus pneumoniae in the United States: 2-year longitudinal analysis. Antimicrob Agents Chemother 2001; 45: Thornsberry C, Jones ME, Hickey ML, Mauriz Y, Kahn J, Sahm DF. Resistance surveillance of Streptococcus pneumoniae, Haemophilus influenzae and Moraxella catarrhalis isolated in the United States, J Antimicrob Chemother 1999; 44: Thornsberry C, Oglivie PT, Kahn J, Mauriz Y, and the Laboratory Investigator Group. Surveillance of antimicrobial resistance in Streptococcus pneumoniae, Haemophilus influenzae, and Moraxella catarrhalis in the United States in the respiratory season. Diagn Microbiol Infect Dis 1997; 29: Whitney CG, Farley MM, Hadler J, et al. Increasing prevalence of multidrug-resistant Streptococcus pneumoniae in the United States. N Engl J Med 2000; 343: Felmingham D, Grüneberg RN, and the Alexander Project Group. The Alexander Project : latest susceptibility data from this international study of bacterial pathogens from community-acquired lower respiratory tract infections. J Antimicrob Chemother 2000; 45: Felmingham D, Washington J, and the Alexander Project Group. Trends in the antimicrobial susceptibility of bacterial respiratory tract pathogens findings of the Alexander Project J Chemother 1999; 11: McDougal LK, Facklam R, Reeves M, et al. Analysis of multiply antimicrobial-resistant isolates of Streptococcus pneumoniae from the United States. Antimicrob Agents Chemother 1992; 36: Jacobs MR, Koornhof HJ, Robins-Browne RM, et al. Emergence of multiply resistant pneumococci. N Engl J Med 1978; 299: Doern GV, Brueggemann AB, Pierce G, Holley HP, Rauch A. Antibiotic resistance among clinical isolates of Haemophilus influenzae in the United States in 1994 and 1995 and detection of b-lactamase positive strains resistant to amoxicillin-clavulanate: results of a national multicenter surveillance study. Antimicrob Agents Chemother 1997; 41: Doern GV, Brueggemann AB, Pierce G, Hogan T, et al. Prevalence of antimicrobial resistance among 723 outpatient clinical isolates of Moraxella catarrhalis in the United States in 1994 and 1995: results of a 30-center national surveillance study. Antimicrob Agents Chemother 1996; 40: Doern GV, Jones RN, Pfaller MA, Kugler K, and the Sentry Participants Group. Haemophilus influenzae and Moraxella catarrhalis from patients with community-acquired respiratory tract infections: antimicrobial susceptibility patterns from SENTRY antimicrobial surveillance pro- Trends in Antimicrobial Resistance CID 2002:34 (Suppl 1) S15

13 gram (United States and Canada, 1997). Antimicrob Agents Chemother 1999; 43: National Committee for Clinical Laboratory Standards. Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically. Approved standard M7-A4. Wayne, PA: National Committee for Clinical Laboratory Standards, National Committee for Clinical Laboratory Standards. Performance standards for antimicrobial susceptibility testing. 9th informational supplement, M100-S9. Wayne, PA: National Committee for Clinical Laboratory Standards, National Committee for Clinical Laboratory Standards. Performance standards for antimicrobial susceptibility testing. 8th informational supplement, M100-S8. Wayne, PA: National Committee for Clinical Laboratory Standards, National Committee for Clinical Laboratory Standards. Performance standards for antimicrobial susceptibility testing. 10th informational supplement, M100-S10. Wayne, PA: National Committee for Clinical Laboratory Standards, Shortridge VD, Doern GV, Brueggemann AB, Beyer JM, Flamm RK. Prevalence of macrolide resistance mechanisms in Streptococcus pneumoniae isolates from a multicenter antibiotic resistance surveillance study conducted in the United States in Clin Infect Dis 1999; 29: Johnston NJ, de Azavedo JC, Kellner JD, Low DE. Prevalence and characterization of the mechanism of macrolide, lincosamide, and streptogramin resistance in isolates of Streptococcus pneumoniae. Antimicrob Agents Chemother 1998; 42: Baquero F, Garcia-Rodriguez JA, Garcia de Lomas J, Aguilar L, and the Spanish Surveillance Group for Respiratory Pathogens. Antimicrobial resistance of 1,113 Streptococcus pneumoniae isolates from patients with respiratory tract infections in Spain: results of a 1-year ( ) multicenter surveillance study. Antimicrob Agents Chemother 1999; 43: Kam KM, Luey KY, Fung SM, Yiu PP, Harden TJ, Cheung MM. Emergence of multiple-antibiotic-resistant Streptococcus pneumoniae in Hong Kong. Antimicrob Agents Chemother 1995; 39: Brueggemann AB, Pfaller MA, Doern GV. Use of penicillin MICs to predict in vitro activity of other b-lactam antimicrobial agents against Streptococcus pneumoniae. J Clin Microbiol 2001; 39: Jorgensen JH, Doern GV, Maher LA, Howell AW, Redding JS. Antimicrobial resistance among respiratory isolates of Haemophilus influenzae, Moraxella catarrhalis, and Streptococcus pneumoniae in the United States. Antimicrob Agents Chemother 1990; 34: Grüneberg RN, Felmingham D. Results of the Alexander Project: a continuing, multicenter study of the antimicrobial susceptibility of community-acquired lower respiratory tract bacterial pathogens. Diagn Microbiol Infect Dis 1996; 25: Livermore DM. Beta-lactamases in laboratory and clinical practice. Clin Microbiol Rev 1995; 8: Bootsma HJ, van Dijk H, Verhoef J, Fleer A, Mooi FR. Molecular characterization of the BRO b-lactamase of Moraxella (Branhamella) catarrhalis. Antimicrob Agents Chemother 1996; 40: S16 CID 2002:34 (Suppl 1) Thornsberry et al.