Antimicrobial susceptibility of invasive and lower respiratory tract isolates of Streptococcus pneumoniae , 1998 to 2007

1Alberta Children’s Hospital; 2Calgary Laboratory Services; 3University of Calgary, Calgary, Alberta Correspondence: Dr Otto G Vanderkooi, Division of Infectious Diseases, Alberta Children’s Hospital, 2888 Shaganappi Trail Northwest, Calgary, Alberta T3B 6A8. Telephone 403-955-7407, fax 403-955-3045, e-mail ovanderk@ucalgary.ca Streptococcus pneumoniae remains a leading cause of morbidity and mortality worldwide, both in children and adults. Clinical manifestations of disease associated with this pathogen vary, and include meningitis, bacteremia, pneumonia and otitis media (1-6). Penicillin has been the drug of choice for treatment of pneumococcal infections, but penicillinnonsusceptible (NS; intermediate + resistant) isolates of S pneumoniae (PNSP) have been reported with increasing frequency from many parts of the world (7). Of even more concern is that PNSP are more likely than penicillin-susceptible strains to be NS to other classes of antimicrobials (7). The prevalence of resistance varies from country to country, and also within countries. PNSP rates are reportedly 28% in Europe and Latin America, 31% in Malaysia and 34% in the United States (8,9). Regional variation in pneumococcal antimicrobial susceptibility has been shown to occur in the United States, Europe and Canada (10-14). Data from Canadian sources over the past 30 years show that resistance of S pneumoniae to penicillin and other antimicrobials is increasing in this country. In the late 1970s, 2.4% of S pneumoniae isolates in Alberta and the Northwest Territories were penicillin NS (15). More than 25 years later, penicillin NS rates across Canada are 11.7% to 15%, with 3.3% to 6.5% of pneumococcal isolates being penicillin resistant (14,16,17). Surveillance data of S pneumoniae isolates submitted to participating laboratories in 2002 revealed penicillin originAl Article

Because antimicrobial resistance in S pneumoniae is variable with respect to geographic area and is increasing, continued local surveillance of this problem is required to guide appropriate antimicrobial choices.The present report describes the results of antimicrobial susceptibility rates of S pneumoniae isolates for cases of invasive and lower respiratory tract (LRT) disease in the Calgary Health Region (CHR) from 1998 to 2006.

MethODs study design
Population-based surveillance of invasive and LRT S pneumoniae infections has been conducted in the CHR, one of nine regional health authorities in Alberta, since January 1, 1998, by the Calgary Area Streptococcus pneumoniae Epidemiology Research (CASPER) group.The CHR is an integrated, publicly funded health care system that encompasses the hospitals within the city of Calgary as well as the medical centres serving a large rural area surrounding the city (population 1,111,614 in 2007) (18).
Population-based surveillance is made possible in the CHR by Calgary Laboratory Services (CLS), a centralized regional laboratory service that provides clinical microbiology services to both hospitalized and ambulatory patients.
S pneumoniae isolates from invasive and LRT infections collected at the CLS from January 1, 1998, to December 31, 2007, were included in the study.Patients were deemed to be CHR residents on the basis of town of residence listed in the admission and laboratory records.If this information was unavailable, patients with Alberta health care numbers were considered to be CHR residents if the culture was submitted to a collection site within the CHR boundaries.The study was approved by the Conjoint Health Research Ethics Board of the University of Calgary before commencement.
The date of collection, specimen source, demographic information, serotype and antimicrobial susceptibility are collected for each S pneumoniae case identified by CLS.For all invasive cases, chart reviews were performed to collect clinical information.Chart reviews were conducted prospectively as of August 1, 2003, and retrospectively for earlier cases.Patients and/or their parents or next of kin were also interviewed as part of the prospective component.

Laboratory methods
Isolates were confirmed to be S pneumoniae by standard methodology including Gram stain, colonial morphology on blood agar, bile solubility, susceptibility to optochin and pneumococcal antibody agglutination (Phadebact Pneumococcus, Boule Diagnostic AN, Sweden).Susceptibility testing was performed by broth microdilution utilizing 96-well panels (PML Microbiological Inc, Canada) and results were interpreted according to Clinical and Laboratory Standards Institute guidelines (19)(20)(21).Cefprozil susceptibility testing continued until July 2004.Telithromycin susceptibility testing started in 2004 and is ongoing.
Invasive S pneumoniae infection was defined as the isolation of S pneumoniae from a normally sterile body site.LRT S pneumoniae infection was defined as the isolation of S pneumoniae from bronchoalveolar lavage, bronchial wash, tracheal aspirate, endotracheal tube aspirate and protected brush specimen.Only one isolate was included for analysis per episode (defined as isolates collected within 30 days of each other).In the event that S pneumoniae was isolated from an invasive and LRT specimen collected during the same episode, only the invasive isolate was included in the analysis.NS S pneumoniae isolates are those with susceptibility tests indicating intermediate or resistant susceptibility according to the Clinical and Laboratory Standards Institute guidelines (19)(20)(21).Multidrug resistant (MDR) S pneumoniae isolates were NS to antimicrobials from three or more of the following classes: beta-lactams, sulfonamides, macrolides/ lincosamides, tetracycline, fluoroquinolones, glycopeptides and streptogramins.

Data analysis
Data were analyzed using SPSS 16.0 for Mac OS X (SPSS Inc, USA).Categorical data were summarized as proportions and continuous data were summarized as means and medians with ranges, and standard deviations.Differences between groups were tested by the c 2 test for categorical variables and Student's t test for continuous variables.Logistic regression was used to analyze trends in antimicrobial susceptibility over the study period.Differences were considered statistically significant at a P<0.05.

ResuLts
A total of 2382 isolates were recovered by CLS between January 1, 1998, and December 31, 2007.Analysis was performed on 1666 episodes, excluding 716 isolates for residence outside of the CHR (n=314), not S pneumoniae (n=40), multiple isolates from a single episode (n=308), not saved (n=8), nonviability of the organism (n=13), and isolates that were neither invasive nor LRT in origin (n=33).
One thousand (60.0%) isolates were from male patients, with no sex recorded for one patient.Mean (± SD) patient age was 44.8±26.0 years (range 0.04 to 105.2 years).LRT isolates came from patients who were younger (mean 42.3 years versus 45.9 years, P=0.046) than those patients with invasive isolates, and were also more likely to be male (66.3% versus 57.4%, P=0.001).Two hundred forty (14.4%) were from patients younger than five years of age, 94 (5.6%) from patients five to 17 years of age, 928 (55.7%) from patients 18 to 64 years of age, and 404 (24.2%) from patients 65 years of age and older.
Table 1 demonstrates NS rates for all antimicrobials tested.Isolates that were NS to penicillin and trimethoprimsulfamethoxazole (TMP-SMX) also showed higher NS rates to penicillin, TMP-SMX, parenteral cefuroxime, cefotaxime and erythromycin (Figure 1).Among penicillin and TMP-SMX NS isolates, 40.1% and 14.8%, respectively, were MDR (P<0.001 in both cases).Antimicrobial resistance rates for the study period are shown in Figure 2. Susceptibility results comparing invasive and LRT isolates are demonstrated in Figure 3.
Rates of antimicrobial nonsusceptibility were significantly higher for respiratory isolates compared with invasive isolates.

DIsCussIOn
Rates of antimicrobial nonsusceptibility continue to be relatively low and stable when compared with data from other parts of Canada and other countries (11,12,14,(22)(23)(24).Penicillin nonsusceptibility at 8% overall is one of the lowest reported rates in the modern era, with amoxicillin faring even more favourably with 0.3% intermediate susceptibility.Similarly, resistance rates among the other antimicrobials tested are low in our locale.In the CHR, NS rates to cefuroxime in invasive isolates from 2003 onward demonstrate significant decline (Figure 2).This was coincident with the introduction of the pneumococcal conjugate vaccine for infants and children in the fall of 2002.Similar results were reported following the introduction of 7-valent pneumococcal conjugate vaccine into the United States, including decreases in penicillin and ceftriaxone NS isolates (25)(26)(27)(28).Proposed explanations for the fall in resistance rates among invasive isolates after immunization with the conjugate vaccine include reduction in nasopharyngeal colonization with resistant serotypes, and reduced selection pressure as a consequence of declining rates of pneumococcal disease leading to reduced antimicrobial use (29,30).Extensive review in several of our group's previous studies have revealed a significant shift in nasopharyngeal carriage isolates among pediatric patients (31)(32)(33).Our relatively modest declines are likely due to low prevaccine resistance rates.However, this trend of decreasing resistance postvaccine may not continue as recent data have found some increased PNSP with increase resistance in nonvaccine serotypes (34,35).TMP-SMX rates of NS increased significantly in our adult population from 2005 to 2007.This was due to an outbreak of serotype-specific (serotype 5) pneumococcal disease in the CHR among primarily homeless persons.This particular pneumococcal strain demonstrates intermediate susceptibility or resistance to TMP-SMX.This outbreak is described more fully in another publications, with a full manuscript in preparation (36,37).
Our study found increased NS rates among LRT isolates when compared with isolates from invasive disease.Differences in NS rates as a function of isolate source have been reported by other authors in North America, Europe and Asia (8,11,22,38).The Canadian Bacterial Surveillance Network also reported higher rates of penicillin NS for nonsterile compared with sterile isolates (16.6% versus 12.4%, P=0.005), with LRT specimens constituting 60% of the nonsterile isolates (14).Similarly, in Quebec, noninvasive isolates demonstrated higher rates of resistance to macrolides, penicillin and cephalosporins (39).The Tracking Resistance in the United States Today (TRUST) program in the United States reported higher rates of resistance among LRT isolates versus blood for penicillin, azithromycin, TMP-SMX and levofloxacin (P<0.0001 for all comparisons) (12).Similar results were observed in the United Kingdom and Ireland, with an increased proportion of NS isolates from LRT isolates for penicillin (8.1% versus 6.2%) and tetracycline (7.6% versus 4.0%), similar rates for cefotaxime, clindamycin and amoxicillin, but decreased rates of erythromycin NS isolates (12% versus 14.6%) (40).When respiratory tract isolates are compared, specimens from the head, ear, eye, nose and throat show higher NS rates than those from the LRT (11,(22)(23)(24).
Infections with serotype 19A, especially MDR 19A, have begun to have a significant impact in the epidemiology of pneumococcal infections in many jurisdictions (34,35,(41)(42)(43).Important infections seen with this serotype have included otitis media and invasive disease (35,41).Up to 2007, serotype 19A and especially MDR 19A was an uncommon cause of invasive infection or LRT infection in Calgary (44).There were 12 cases of 19A disease from 1998 to 2006 and eight cases in 2007 (44).
Age was another significant factor determining NS rates in the CHR, with erythromycin and cefotaxime NS and MDR higher in pneumococcal isolates from young children (younger than five years of age).LRT isolates from children younger than five years of age were also more likely to be NS to TMP-SMX, cefotaxime, erythromycin and MDR than isolates from older children and adults.This finding was also reported in Toronto, Ontario (45).Similarly, Powis et al (14) were able to show a significantly higher rate of fluoroquinolone NS among older adults compared with younger adults, but age was not associated with NS rates for the other antimicrobials.Six years of surveillance by the Canadian Respiratory Organism Susceptibility Study (CROSS) of pneumococcal isolates from the respiratory tract showed that age had no impact on penicillin, cefuroxime and clarithromycin NS rates (46).Weiss et al (39) examined the antimicrobial NS rates of pneumococcal isolates in Quebec from 2000 to 2001 showing younger age association with increased NS rates, but only for the macrolides (30% versus 14.8%, younger than 16 years of age versus 16 years of age and older, P<0.01).Other authors have also reported on age as a determinant of resistance (12,(47)(48)(49).The TRUST program showed that rates of resistance to penicillin, azithromycin, TMP-SMX and ceftriaxone were higher for isolates recovered from patients 18 years of age and younger versus patients 18 to 64 years of age (P<0.001)(12).Jones et al showed that isolates recovered from five or fewer years of age had higher rates of NS to penicillin, cefprozil, cefuroxime axetil, macrolides and TMP-SMX (P<0.05)than among the other age groups (47).Hoban et al (48) used data from 25 countries and demonstrated that isolates from infants younger than two years of age were more likely to be penicillin resistant (OR 1.98) and erythromycin resistant (OR 1.89) than isolates from adults.
An important strength of the present study is that the surveillance system identifies cases using a comprehensive centralized laboratory (CLS), which allows for a true population-based assessment of the antimicrobial susceptibility trends.A potential weakness is that respiratory tract samples are obtained in either severe cases or where there is therapeutic failure.The data from these samples are of particular importance because they reflect those clinical circumstances where adequate antimicrobials most indicated.The findings of our study have important implications for the empiric treatment of severe disease suspected to be caused by S pneumoniae in our region.Overall, susceptibility rates to parenteral cefuroxime, cefotaxime and ceftriaxone are greater than 90%.Because only 1.5% of invasive isolates are cefotaxime NS, this antimicrobial remains a good choice for empirical treatment of suspected invasive pneumococcal disease.Given the NS rate of 16% to parenteral cefuroxime among LRT isolates in children younger than five years of age, it would be prudent to use another antimicrobial for empirical treatment of severe pneumonia presumed to be pneumococcal in origin in young children.Physicians in our region also need to be aware of the cefotaxime NS rate of 9% among LRT isolates in children younger than five years of age, which further complicates the issue of treatment.Amoxicillin continues to be an excellent first-line antimicrobial for the treatment of pneumococcal infections that are amenable to oral therapy.
Our study provides more evidence that regional variation in NS rates among pneumococcal isolates occurs, thus emphasizing the need for local surveillance of susceptibility patterns to provide clinically relevant information with regard to treatment options.Both isolates from young children and LRT isolates are more resistant to antimicrobials and thus prudent antimicrobial utilization remains important.

Figure 4 )
Figure 4) Comparison of antimicrobial susceptibility among lower respiratory tract (LRT) isolates and invasive isolates obtained from patients younger than and older than five years (y) of age in the Calgary Health Region, 1998 to 2007.The comparisons between LRT isolates were significant (P<0.05) for trimothoprim-sulfamethoxazole (TMP-SMX) and cefotaxime.The comparisons between invasive isolates were significant (P<0.05) for penicillin, TMP-SMX and multidrug-resistant (MDR) isolates.NS Nonsusceptible