ABSTRACT
Resistance (intermediate and high) to penicillin amongStreptococcus pneumoniae strains is an emerging problem worldwide. From 1995 to 1997, isolates of S. pneumoniae not susceptible to penicillin were seen with increasing frequency from blood, cerebrospinal fluid, pleural fluid, and middle ear fluid from pediatric patients at the Veterans General Hospital-Kaohsiung. To determine the prevalence of carriage of these penicillin-nonsusceptibleS. pneumoniae isolates, we obtained nasopharyngeal swab specimens from 2,905 children (ages, 2 months to 7 years) attending day-care centers or kindergartens or seen in our outpatient clinic.S. pneumoniae was isolated from 611 children, and 584 strains were available for analysis. The oxacillin disc test was used as a screening test to evaluate penicillin susceptibility. The MICs of 11 antibiotics (penicillin, cefaclor, cefuroxime, ceftriaxone, cefotaxime, imipenem, chloramphenicol, clarithromycin, rifampin, vancomycin, and teicoplanin) were determined by the E-test. Only 169 (29%) of the strains were susceptible to penicillin; 175 (30%) strains were intermediately resistant and 240 (41%) were highly resistant. The isolates also demonstrated high rates of resistance to other β-lactams (46% were resistant to cefaclor, 45% were resistant to cefuroxime, 45% were resistant to ceftriaxone, 31% were resistant to cefotaxime, and 46% were resistant to imipenem). The rate of resistance to macrolide antimicrobial agents was strikingly high; 95% of the isolates were not susceptible to clarithromycin. However, 97% were susceptible to rifampin and 100% were susceptible to the two glycopeptides (vancomycin and teicoplanin). While reports of penicillin-resistant S. pneumoniae increased worldwide through the 1980s, the high prevalence (71%) of resistance reported here is astonishing. Surveillance of nasopharyngeal swab specimen cultures may provide useful information on the prevalence of nonsusceptible strains causing invasive disease. Such information could be used to guide therapy of pneumococcal infections.
Streptococcus pneumoniaeis a leading cause of infectious disease-associated morbidity and mortality worldwide. It is the most common cause of otitis media and sepsis in children under the age of 2 years and is a leading cause of meningitis and pneumonia (28, 35, 59). Until recently,S. pneumoniae has displayed uniform susceptibility to penicillin (PCN). The first case of clinically significant isolate not susceptible to PCN was reported in Australia in 1967 (29). Over the past decade, there has been a dramatic increase in the prevalence of S. pneumoniae with diminished susceptibility to PCN (5, 33, 37, 56). The proportion of PCN-nonsusceptible strains in Spain and South Africa rose from 4.3 and 4.9%, respectively, in 1979 to 40 and 15.4%, respectively, in 1990 (38,42). In France, the prevalence of PCN-nonsusceptible strains increased from 13% in 1984 to 48% in 1990 (27). In the United States, recent nationwide surveys showed that 23.6% of S. pneumoniae strains were not susceptible to PCN (15). In Taiwan, a resistance rate of 12% has been reported (32). However, at the Veterans General Hospital-Kaohsiung, a tertiary-care referral hospital with 1,200 beds, PCN-nonsusceptible S. pneumoniae was cultured with increasing frequency from blood, cerebrospinal fluid, pleural fluid and middle-ear fluid from pediatric patients from 1995 to 1997. The overall resistance rate was 70% among isolates from hospitalized pediatric patients (unpublished data). It is possible that such a high prevalence of resistance resulted in part from the facts that the patient population was from a tertiary-care referral center and that the sample size was relatively small (altogether, 40 clinical specimens). The intention of the investigation described in this report was to further investigate the prevalence of carriage of PCN-nonsusceptible S. pneumoniae strains in the pediatric population.
The nasopharynx as a source of pneumococci has obvious predictive potential for the emergence of resistance in clinically significant isolates and has therefore been used to assess antibiotic resistance among pneumococci from different population groups (36). In Taiwan, no information has been available on S. pneumoniaecolonization or infection in children. The purpose of the study was to obtain data on the susceptibility and rate of prevalence of carriage of PCN-nonsusceptible S. pneumoniae isolates among the pediatric population in southern Taiwan. The approach was a survey of nasopharyngeal carriage among children between 2 months and 7 years of age from day-care centers or kindergartens or from children visiting our outpatient clinic. The susceptibilities of these isolates to other β-lactams, macrolides, chloramphenicol, rifampin, and glycopeptides were also analyzed.
(This paper has been presented as an abstract at the 37th Interscience Conference on Antimicrobial Agents and Chemotherapy, Toronto, Ontario, Canada, 28 September to 1 October 1997.)
MATERIALS AND METHODS
Surveys of nasopharyngeal carriage of S. pneumoniae. Fifteen child day-care centers and kindergartens in Kaohsiung were visited. They were randomly selected and are located in different areas of Kaohsiung. Nasopharyngeal swab specimens were obtained from children attending day-care centers or kindergartens. We also obtained nasopharyngeal swab specimens from children who visited our outpatient clinic. The latter study population comprised children who sought medical care for respiratory or gastrointestinal infection as well as healthy children who came to be examined at a well-baby clinic. Children between 2 months and 7 years of age were enrolled.
Specimen collection.Nasopharyngeal swab specimens for culture were obtained by a single investigator who used a cotton swab placed 1 to 1.5 in. into the nasopharynx. The specimens were immediately plated onto 5% sheep blood (Becton Dickinson Microbiology System, Cockeysville, Md.).
Isolation and identification of S. pneumoniae. All plates were incubated for 24 to 48 h at 37°C in 5% carbon dioxide. S. pneumoniae isolates were identified using typical colonial appearance, α-hemolysis, and Gram staining. Confirmatory tests included optochin sensitivity and bile solubility tests. All strains were kept frozen in −70°C in tryptic soy broth for further analysis. On thawing, the isolates were checked for purity and optochin sensitivity. The isolation and identification procedures and MIC testing were done in the Microbiology Laboratory, Veterans General Hospital-Kaohsiung.
Antibiotic susceptibility tests.All S. pneumoniae isolates were screened for PCN susceptibility by an agar disc diffusion method with a 1-μg oxacillin disc. Isolates that grew within a zone of inhibition of <20 mm were considered resistant. The MICs of PCN, cefaclor, cefuroxime, ceftriaxone, cofotaxime, chloramphenical, clarithomycin, rifampin, vancomycin, and teicoplanin were determined by the E-test (AB Biodisk, Solna, Sweden) following the manufacturer’s instructions. All strains viable upon testing were tested for the MICs. Susceptibility tests were performed from a bacterial inoculum whose turbidity was equivalent to that of a McFarland standard of 0.5. From this suspension, E-tests were performed on blood agar. MICs falling between two marks on the E-test strip were rounded up to the next higher twofold dilution, as recommended in the instructions. Control organisms (Staphylococcus aureus ATCC 29213 and Escherichia coli ATCC 29522) were included in each set of tests. Interpretation of results was performed according to recommendations of the National Committee for Clinical Laboratory Standards (49). Isolates for which the MIC was ≦0.06 μg/ml were considered susceptible; those for which MICs were ≧0.125 μg/ml but ≦1.0 μg/ml were considered intermediately resistant to PCN, whereas those for which MICs were >1 μg/ml were defined as highly resistant. Breakpoints were as follows: for cefuroxime, ceftriaxone, and cefotaxime, susceptible, ≦0.5 μg/ml; nonsusceptible, ≧2 μg/ml; for imipenem, susceptible, ≦0.12 μg/ml; nonsusceptible ≧1 μg/ml; for chloramphenicol, susceptible, ≦4 μg/ml; nonsusceptible ≧8 μg/ml; for clarithromycin, susceptible, ≦0.5 μg/ml; nonsusceptible, ≧2 μg/ml; for rifampin, susceptible, ≦1 μg/ml; nonsusceptible, ≧4 μg/ml; for vancomycin, susceptible, ≦1 μg/ml. MIC results falling between the breakpoints for susceptibility and nonsusceptibility were considered to represent intermediate susceptibility. No reference breakpoints for cefaclor and teicoplanin susceptibility to S. pneumoniae are listed by the National Committee for Clinical Laboratory Standards. For cefaclor, we used the criteria for other cephalosporins; that is, ≦0.5 and ≧2 μg/ml for susceptibility and nonsusceptibility, respectively; for teicoplanin, we used the same standard used for vancomycin.
RESULTS
A total of 2,905 children were enrolled in this study; 611 (21%) of the subjects’ nasopharyngeal swab cultures were positive forS. pneumoniae. Of these 611 isolates, the MICs for 584 were determined. Using MIC as the “gold standard,” we showed that the oxacillin disc screening test has a sensitivity of 96.1% and a specificity of 94.9%. The antibacterial activities of 11 antimicrobial agents against 584 isolates of S. pneumoniae are indicated in Table 1. Table2 illustrates the individual numbers and percentages of susceptibility of these agents stratified by PCN susceptibility.
Susceptibilities of 584 isolates of S. pneumoniae
Susceptibilities to drugs stratified by PCN susceptibility
DISCUSSION
We have detected an extraordinarily high prevalence of resistance to PCN among S. pneumoniae strains isolated from the nasopharynges of children in Kaohsiung, Taiwan. Reports of antimicrobial-resistant S. pneumoniae have increased worldwide during the past two decades, but geographic and temporal patterns vary (3, 19, 21, 26, 30, 34, 40, 43, 44, 45, 50, 54,57). The prevalence of highly resistant strains was higher (41% of all isolates and 58% of nonsusceptible strains) than those published in reports from other geographic areas (3, 16, 18, 28,32, 54, 55). The use of high levels of PCN has been advocated against intermediately resistant strains causing otitis media, pneumonia, and bacteremia (51). Our surveillance study is a warning to pediatric clinicians in Taiwan that caution should be exercised when treating pneumococcal infections, since many strains (41%) are highly resistant to PCN; high doses of PCN may not be adequate as empirical therapy. MICs should be determined for all clinically significant isolates of S. pneumoniae to aid in the selection of appropriate antibiotics for therapy.
Carriage of S. pneumoniae has been correlated with the emergence of clinical disease (28, 31, 39, 48). Thus, the characteristics of carriage isolates could serve as an indicator of the prevalence of resistance strains in the community (12, 18, 46,60). A prevalence of 53% of PCN-nonsusceptible S. pneumoniae among children in a rural Kentucky community has been reported, with 33% of the strains being highly resistant (16). The rate of carriage of PCN-nonsusceptible S. pneumoniae was reported to range from 0 to 92.9% in eastern and central Europe (1). We confirmed that the prevalence of PCN-nonsusceptible S. pneumoniae among nasopharyngeal swab cultures (71%) was correlated with the clinical isolates (70%) in pediatric patients. Our findings of the widespread prevalence of PCN-resistant S. pneumoniae strains in the community documents the value of monitoring nasopharyngeal carriage of S. pneumoniae, since in the present study the day-care centers and kindergartens were located in different areas of Kaohsiung.
The MICs at which 50% of isolates are inhibited (MIC50s) and MIC90s of other β-lactams showed that for the isolates tested there was a trend toward decreased susceptibility which paralleled the PCN susceptibility, as was the case in previous reports (41, 47). Interestingly, a similar trend was not demonstrated for non-β-lactams. Although PCN resistance could be a marker for cephalosporin resistance, the degree of cross-resistance between cephalosporin and PCN among S. pneumoniae is under debate (33). For individual strains the MICs of cephalosporins may vary widely. If susceptibility to individual drugs is stratified according to PCN susceptibility (Table 2), the data show that PCN-susceptible strains may be nonsusceptible to cephalosporins and vice versa. There is a report demonstrating that intermediate PCN resistance in S. pneumoniae is associated with an impaired bacteriologic and clinical responses of acute otitis media to cefaclor and cefuroxime (14). The failure of cefotaxime or ceftriaxone in the treatment of pneumococcal meningitis has been reported previously (7, 8, 10, 20, 22, 55). There is clearly a need to study the correlation of in vitro resistance and the clinical response. Our study showed that the cefaclor MIC for pneumococci is very high, both for those isolates that are PCN susceptible and for those isolates that are PCN nonsusceptible. Although imipenem-resistant strains of S. pneumoniae have been rarely reported (4,11), we documented an unexpectedly high incidence of imipenem-resistant strains. A previous report documented that a progression from colonization with PCN-nonsusceptible S. pneumoniae to invasive disease with resistant strains can occur rapidly and that multiple courses of β-lactams may promote this process (18).
Macrolide resistance in S. pneumoniae has remained at a low level in most countries, although the geographic distribution is variable (6, 13, 19, 37, 44, 56). In Hungary and South Africa, the rate of resistance to erythromycin is reported to be about 50% (37, 43). In the present survey, the nasopharyngeal isolates among children demonstrated a strikingly high incidence of resistance to clarithromycin. Although clarithromycin is a newer macrolide, the susceptibility of S. pneumoniae to that drug was no better than that to erythromycin. The macrolide resistance was thought to have evolved in response to different antibiotic pressures in the community. For example, in Spain, one of the areas with a high prevalence of PCN-nonsusceptible S. pneumoniae, the incidence of erythromycin resistance is low and is probably due to the infrequent use of erythromycin in Spain. However, in Taiwan, macrolides are often prescribed by physicians as first-line antibiotics and are readily available without prescription at drugstores.
Resistance to chloramphenicol in S. pneumoniae was not common in most parts of the world (13, 37, 53, 58). However, our study showed that only 52.4% of nasopharyngeal isolates were susceptible to chloramphenicol and that the MIC90 for the isolates was as high as 24 μg/ml. In the past, chloramphenicol has been suggested as an alternative for the treatment of meningitis caused by PCN-resistant S. pneumoniae (23, 25). Such a recommendation should not be encouraged in Taiwan or in other geographic areas with a relatively high prevalence of strains with chloramphenicol resistance.
Resistance to antibiotics of at least three different groups has been defined as multiple drug resistance (3). A high rate of resistance (between 50 and 70%) was documented in Spain, South Africa, Hungary, Korea, and Pakistan (17, 37, 40, 43). Our study demonstrated that 31% of nasopharyngeal isolates were multiply resistant to multiple antibiotics (penicillin, clarithromycin, and chloramphenicol). Although the rate of multiple drug resistance was not particularly high compared with that in other countries, 98% of our isolates were resistant to more than one of the antimicrobial agents tested. This means that physicians in Taiwan have a limited choice of drugs that can be used against S. pneumoniae. It is difficult to explain such a high incidence of resistance; the injudicious and frequent use of antibiotics has been proposed as a risk factor in other areas (40). We are conducting a study to investigate the risk factors for antibiotic resistance in Taiwan. The availability of over-the-counter antibiotics may be an important risk factor.
In our study, 97% of the isolates were susceptible to rifampin. All isolates were susceptible to vancomycin and teicoplanin as in the previous study (2). Rifampin should not be used alone because of the rapid emergence of resistance (24). It has been shown that the combination of vancomycin and rifampin was effective in sterilizing the cerebrospinal fluid of patients with meningitis caused by PCN-nonsusceptible S. pneumoniae(52), regardless of whether dexamethasone was administered. We agree with the recommendation of the use of the combination of these two drugs.
S. pneumoniae can no longer be considered a pathogen with uniform susceptibility to PCN, cephalosporins, and macrolide antimicrobial agents in Taiwan. The relative frequencies of intermediate and high-level resistance among S. pneumoniaehave important therapeutic implications for the selection of antimicrobial agents to be used for initial empirical treatment of infections frequently caused by S. pneumoniae, e.g., pneumonia and meningitis, particularly in critically ill patients. The Centers for Disease Control and Prevention has recommended that clinicians base their decisions about empirical antibiotic therapy for presumptive pneumococcal infections on local prevalence data (9). Prospective studies of the treatment of invasive infections due to PCN-resistant S. pneumoniae are urgently needed.
ACKNOWLEDGMENTS
We are grateful to Victor L. Yu for critical reading of the manuscript. We also thank Chiang-Ching Shih for helpful suggestions and Chung-Chin Lin and Ya-Fen Tang for technical assistance.
FOOTNOTES
- Received 27 October 1997.
- Returned for modification 23 December 1997.
- Accepted 30 March 1998.
- Copyright © 1998 American Society for Microbiology
