Previous Article | Next Article ![]()
Antimicrobial Agents and Chemotherapy, May 2004, p. 1488-1494, Vol. 48, No. 5
0066-4804/04/$08.00+0 DOI: 10.1128/AAC.48.5.1488-1494.2004
Copyright © 2004, American Society for Microbiology. All Rights Reserved.
Kitasato Institute for Life Sciences and Graduate School of Infection Control Sciences, Kitasato University, 5-9-1 Shirokane, Minato-ku,1 Department of Pediatrics, Tokyo National Medical Center, Higashigaoka, Meguro-ku, Tokyo,3 Pharmaceutical Center, Meiji Seika Kaihsha, Morookacho, Kohoku-ku, Yokohama, Japan2
Received 5 September 2003/ Returned for modification 30 November 2003/ Accepted 31 January 2004
| ABSTRACT |
|---|
|
|
|---|
17 years old) and 73 from adults (
18 years old). Isolates with or without abnormal pbp1a, pbp2x, or pbp2b genes identified by PCR were classified into six genotype patterns and 90% MIC (MIC90) values for PEN: (i) strains with three normal genes (17.2% of isolates; MIC90, 0.031 µg/ml); (ii) strains with abnormal pbp2x (22.1%, 0.063 µg/ml); (iii) strains with abnormal pbp2b (1.0%, 0.125 µg/ml); (iv) strains with abnormal pbp2x and pbp2b (7.4%, 0.25 µg/ml); (v) strains with abnormal pbp1a and pbp2x (12.7%, 0.25 µg/ml); and (vi) strains with three abnormal PBP genes (39.7%, 4 µg/ml), which are termed genotypic PEN-resistant S. pneumoniae (gPRSP). Panipenem, a carbapenem, showed an excellent MIC90 (0.125 µg/ml) against gPRSP, followed by meropenem and vancomycin (0.5 µg/ml), cefotaxime and ceftriaxone (1 µg/ml), and ampicillin (4 µg/ml). Strains of gPRSP were significantly more prevalent in children (45.2%) than in adults (27.4%). The most frequent serotypes were 6B, 19F, 23F, 6A, and 14 in children and 23F, 22, 3, 10, 6B, and 19F in adults. Serotypes 6B, 6A, 19F, 23F, and 14 predominated among gPRSP. In children, 7- and 11-valent pneumococcal conjugate vaccines would cover 76.2 and 81.3% of isolates, respectively, although coverage would be lower in adults (43.9 and 56.0%, respectively). These findings suggest the need for early introduction of pneumococcal conjugate vaccines and continuous bacteriological surveillance for meningitis. | INTRODUCTION |
|---|
|
|
|---|
Surveillance studies of antibiotic susceptibility, serotype distribution of causative strains, and mortality rate in meningitis have been carried out nationwide in many countries (8, 11, 15, 20, 21, 28, 33, 36, 39, 43).
In Japan, prevalence of PEN-resistant S. pneumoniae (PRSP) among clinical isolates from acute otitis media and respiratory tract infections (RTIs) has been increasing rapidly, especially in younger children (40, 41). In parallel with overall increases in incidence of PRSP, meningitis caused by PRSP is being reported increasingly throughout Japan (4). As investigators in that country, we felt that nationwide surveillance had become crucial given increasing isolation of resistant strains that could cause meningitis, including PRSP, as well as ampicillin (AMP)- and cephalosporin-resistant strains of Haemophilus influenzae type b, the most frequent etiologic agent of bacterial meningitis. In addition, accurate up-to-date data are critical to help decision making for introduction of new conjugate antipneumococcal vaccines appropriate to the country and its population in consideration of serotype distributions among people and geographic areas (19, 38).
Based on these considerations, we organized a study group, the Nationwide Surveillance for Bacterial Meningitis (NSBM), in 1999. One of the group's objectives was to characterize S. pneumoniae and H. influenzae isolates responsible for meningitis in terms of serotype and antimicrobial resistance according to gene alteration.
This is the first report in Japan describing susceptibilities of isolates from pneumococcal meningitis to intravenous ß-lactam antibiotics and vancomycin, along with their genotypes of PEN-binding protein (PBP) genes and serotypes.
| MATERIALS AND METHODS |
|---|
|
|
|---|
All clinical isolates were grown on sheep blood agar (Nippon Becton Dickinson, Tokyo, Japan) at 37°C in an atmosphere containing 5% CO2, and a single colony was isolated for storage in 10% skim milk (Difco Laboratories, Detroit, Mich.) at 80°C. Identification of isolates as S. pneumoniae was confirmed by PCR amplification of the autolysin (lytA) gene (18).
Susceptibility testing. MICs of ß-lactam antibiotics and vancomycin against S. pneumoniae were determined by an agar dilution method using Mueller-Hinton agar (MH; Difco Laboratories) supplemented with 5% defibrinized sheep blood (32). Bacterial inocula were prepared according to a previously reported method (40). Antibiotics used in the present study were PEN and AMP (Meiji Seika Kaisha, Tokyo, Japan); cefotaxime (CTX; Nippon Hoechst Marion Roussel, Tokyo, Japan), ceftriaxone (CRO; Nippon Roche, Tokyo, Japan), panipenem (PAM; Sankyo, Tokyo, Japan), meropenem (MEM; Sumitomo Pharmaceuticals, Tokyo, Japan), and vancomycin (VAN; Shionogi, Osaka, Japan). S. pneumoniae ATCC 49619 was used as a quality control strain for susceptibility testing.
Serotyping. Serotypes of all S. pneumoniae strains were determined by the Quellung reaction using antiserum purchased from the Statens Serum Institute (Copenhagen, Denmark).
PCR to identify three PBP genes and the lytA gene. To confirm that isolates were S. pneumoniae, the lytA gene (18) encoding the autolysin enzyme specific to S. pneumoniae was amplified simultaneously with the three PBP genes. Oligonucleotide primers for detection of the three PBP genes were designed to amplify proportions of the normal pbp1a (5, 27, 37), pbp2x (6, 26, 34), and pbp2b (14, 45) genes detected only in susceptible strains. Among the three sets of primers, the primers for the pbp1a gene were newly constructed based on current worldwide data (30): forward, 5'-2037AAACCGCGACTGGGGATCAAC2057-3'; and reverse, 5'-2275GGTTGAGTCCGACCTTGTTT2256-3'. Portions of each gene corresponding to the primers were positioned in blocks of highly divergent sequences within or near conserved amino acid motifs previously identified in the mosaic PBP genes of PEN-nonsusceptible S. pneumoniae. Primer mixture A contained primers for detecting the lytA and pbp1a genes, whereas primer mixture B contained primers for detecting the pbp2x and pbp2b genes.
Bacterial samples received from each institution were suspended into 2 ml of Mueller-Hinton broth, and then the 5 µl of the broth was added in a 0.5-ml microtube containing 30 µl of a lysis solution made up as previously reported (40). A CSF sample was added into a lysis solution after centrifugation at 4°C and 5,000 rpm for 5 min. The tubes were placed in a thermal cycler (Gene Amp PCR System 9600-R; Perkin-Elmer Cetus, Norwalk, Conn.), and bacterial cells were lysed for 20 min at 60°C and for 5 min at 94°C to obtain template DNA.
Next, 2 µl of template DNA solution was added to each of two tubes marked A and B. These tubes contained 30 µl of reaction mixture, consisting of (i) 600 ng of appropriate primer, (ii) 100 µl of a 25 mM deoxynucleoside triphosphate mixture, (iii) 40 U of Tth DNA polymerase (Toyobo, Tokyo, Japan), and (iv) 100 µl of 10x PCR buffer (pH 8.3) per ml of solution. PCR cycling conditions consisted of 35 cycles at 94°C for 15 s, 53°C for 15 s, and 72°C for 15 s. Amplified DNA fragments were analyzed by electrophoresis on a 3% agarose gel. Two DNA fragments of 319 and 239 bp in mixture A corresponded to the products of lytA and pbp1a genes, respectively. Similarly, DNA fragments of 197 and 147 bp in mixture B corresponded to the pbp2x and pbp2b genes, respectively.
When an isolate showed all three DNA fragments corresponding to pbp1a, pbp2x, and pbp2b, the PBP genes were regarded as having essentially the same sequences as in the R6 strain of PSSP. When any of these DNA bands was not detected or was detected in different sizes, the gene in question were regarded possessing different sequences.
| RESULTS |
|---|
|
|
|---|
Based on the PCR results, strains tested were classified into six groups according to genotype as follows: (i) strains with three normal pbp genes (n = 35, 17.2%); (ii) strains with an abnormal pbp2x gene (n = 45, 22.1%); (iii) strains with an abnormal pbp2b gene (n = 2, 1.0%); (iv) strains with abnormal pbp2x and pbp2b genes (n = 15, 7.4%); (v) strains with abnormal pbp1a and pbp2x genes (n = 26, 12.7%); and (vi) strains with three abnormal genespbp1a, pbp2x, and pbp2b (n = 81, 39.7%). In the present study, for convenience, genotypes i and vi were termed PSSP and gPRSP, respectively. The remaining genotypes ii to v were termed genotypic PEN-intermediately resistant S. pneumoniae (gPISP).
Table 1 shows the 50% MIC (MIC50), the MIC90, and the MIC range of seven intravenous antibiotics against the strains classified into the six genotype patterns. Figure 1 shows the MIC distribution for PEN, CTX, MEM, and PAM according to PCR results for comparison of the influence of each PBP alteration on MICs; this influence varied considerably depending on the category of the antibiotic. Two antibiotic types were evident among the six ß-lactam antibiotics. One was the PEN type, for which the strains with an abnormal pbp2x gene did not differ notably from PSSP. AMP, MEM, and PAM belonged to the PEN type. The other ß-lactam category was the CTX type, where strains with an abnormal pbp2x gene differed from PSSP in having an increased MIC 8 to 16 times greater than in PSSP; CTX and CRO belonged to this type. MICs of CTX-type agents were affected most by an abnormal pbp2x gene, while MICs of PEN-type agents were affected most by the pbp2b gene.
|
|
The PSSP strains showing CTX MICs of 0.063 to 0.125 µg/ml had amino acid substitutions located in the area of Lys-Ser-Gly (KSG) or Ser-Ser-Asn (SSN) conserved motifs that could not be detected with our pbp2x primers (data not shown here). In addition, strains with high CTX and CRO MICs of
4 µg/ml had two amino acid substitutions changing a Ser-Thr-Met-Lys (STMK) conserved motif in the pbp2x gene to Ser-Ala-Phe-Lys (SAFK).
The MICs of VAN for all S. pneumoniae strains were distributed from 0.25 to 0.5 µg/ml, and no resistant strain was observed.
Resistance types of isolates and patient ages.
Tables 2 and 3, respectively, show relationships between abnormal PBP genotypes of causative strains and patient age in children (
17 years old) and in adults (
18 years old). The 219 cases consist of 146 children and 73 adults. Among them 15 cases were identified the genotype of the causative agent in the CSF by using direct PCR.
|
|
In adult cases compared to pediatric cases, prevalence of gPRSP was significantly lower at 27.4%, whereas those of gPISP with an abnormal pbp2x gene and PSSP were higher at 27.4 and 27.4%, respectively (
2 = 12.0928, P = 0.0335). The highest adulthood incidence occurred in the sixth and seventh decades.
Year-by-year changes in resistant strains. Table 4 shows frequencies of identified resistance genotypes of isolates in each year. The gPRSP already had increased to 42.9% in 1999, remaining highly prevalent up to the present. The prevalence of gPISP strains such as those with an abnormal pbp2x and with abnormal pbp1a plus pbp2x genes has not changed so much.
|
2 = 66.2876, P < 0.0001).
|
|
Although a few resistant strains were identified as serotype 14, serotypes 6A, 6B, 19F, and 23F showed the greatest prevalence among gPRSP. All serotype 3 isolates were of the mucoid type and were gPISP having an abnormal pbp2x gene. The 7- and 11-valent pneumococcal conjugate vaccines covered serotypes of strains isolated from children in 76.1 and 81.2% of cases, respectively; the corresponding percentages in those from adults were 43.9 and 56.1%.
| DISCUSSION |
|---|
|
|
|---|
Since the early 1990s, isolations of PEN-nonsusceptible strains in RTI and acute otitis media have increased dramatically throughout Japan, particularly in young children (41). Currently, genotypically proven gPISP and gPRSP have been isolated in 2002 at rates of 33.0 and 54.9%, respectively. The prevalence of nonsusceptible strains in Japan now appears to be higher than in other countries (2).
The rapid increase in resistant strains has paralleled clinical introduction of new oral cephalosporins that have been greatly overprescribed for outpatients as a first-choice antibiotic. The prescription of antibiotics for pediatric outpatients in Japan differs fundamentally from patterns in other countries (12, 13, 22). The problem is reflected by the observation that many pneumococcal isolates (20%) possess an abnormal pbp2x gene, which decreases their susceptibility to cephalosporin antibiotics as opposed to PEN (6). In other countries, in which amoxicillin is prescribed more frequently for outpatients, the prevalence of strains possessing an abnormal pbp2x gene is lower (30).
The first case of meningitis caused by PRSP in Japan was reported in 1988 (4). We noted that PEN-nonsusceptible isolates appeared to increase gradually as a causative agent of meningitis in parallel with increases of such isolates in RTI. The prolonged low incidence of pneumococcal meningitis may reflect the Japanese nationwide insurance system, under which every one has access to antibiotic medication without financial barrier.
Before initiating the NSBM study, we established a rapid facsimile and e-mail reply system from our laboratory to local laboratory technicians and physicians, who received reports for their reference data. A predicted MIC for ß-lactam antibiotics could be calculated from PCR results for PBP genes within 2.5 h after we received an isolate. The predicted MIC was derived from a multiple regression analysis between MICs of ß-lactam antibiotics against S. pneumoniae and the presence of abnormal pbp1a, pbp2x, and pbp2b genes (40), with S. pneumoniae strains from RTI collected nationwide in Japan between 1998 and 2000. Subsequently, exact MICs of several intravenous antibiotics against the isolates were redetermined by standard biologic methods.
As described in Results, the antimicrobial activity of PAM (31, 41), a carbapenem, was quite impressive with an MIC90 of 0.125 µg/ml against gPRSP. PAM has also a strong bactericidal activity compared to other intravenous antibiotics. PAM is now establishing a reputation as an antibiotic of first choice for severe pneumococcal infections instead of cephalosporins. Concentrations of PAM in CSF were 6.84 µg/ml at the acute stage and 3.28 µg/ml at the recovering stage after a 1-h drip infusion at a dose of 27.5 mg/kg in pediatric patients with meningitis (17). Notable neurotoxicity or nephrotoxicity have not observed clinically in the pediatric patients thus far. In a rabbit model, neurotoxicity of PAM was approximately half that of imipenem (24). Unfortunately, PAM/betamipron is not available clinically in most countries except in Japan, Korea, and China. In Japan, VAN has not been approved for the treatment of pneumococcal meningitis. This is another reason to use PAM for pneumococcal meningitis to pediatric patients.
Meanwhile, various pneumococcal vaccines have been developed, ranging from a 23-polyvalent polysaccharide vaccine to conjugate vaccines, including 7-, 9-, or 11-polyvalency (7, 16, 47). Although all 90 recognized serotypes of S. pneumoniae appear to be pathogenic, how well a vaccine can cover serotypes representing invasive isolates is a key point. Predominant serotypes vary conspicuously between developing and industrialized countries (38), patient age groups, and disease types (8, 11). Some serotypes are more frequent in children for both carriage and infection, whereas others are rarely found in carriers but are frequent in patients with invasive disease.
As described in Results, isolates belonging to a serogroup associated with carriage were more frequent in meningitis in children no more than 17 years old than they were in adults, probably because younger children have not yet developed antibodies to these common serotypes. Since ß-lactam antibiotic resistance is found mainly in serotypes associated with carriage, the prevalence of resistant strains was significantly higher in the young children in meningitis.
If 7- and 11-valent vaccines were introduced into Japan, up to 76.2 and 81.3% of causative isolates from meningitis in children would be covered by the respective vaccines; for adults, these percentages would be lower (43.9 and 56.0%). The relatively high coverage rate for children in Japan compared to other countries could be explained by the limitation of prevalent resistant strains to serotypes 6B, 19F, 23A, and 14 (9, 10); these are included in the 7-valent vaccine. Since the 11-valent vaccine also provides cross-protection against serotype 6A, a further 10.1% of meningitis cases would be covered, bringing total coverage to 91.4%. However, cross-protection within a given serogroup is not fully assured (42). The added presence of serotypes 1, 3, 5, and 7F in the 11-valent vaccine increases the potential coverage rate in adults considerably.
In conclusion, to prevent severe infections with resistant microorganisms from increasing, vaccination against S. pneumoniae, which is not yet available in Japan, is vitally important (1, 44). In addition, to nationwide surveillance for antibiotic susceptibilities and serotypes of S. pneumoniae (25, 35, 46), postgraduate education for physicians also is necessary to ensure that selection of antibiotics is based on pharmacokinetic, pharmacodynamic, and biologic test data.
| ACKNOWLEDGMENTS |
|---|
This study was partially supported by a grant from the Ministry of Health, Labor, and Welfare of Japan (Research Project for Emerging and Reemerging Infectious Diseases).
| FOOTNOTES |
|---|
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Clin. Vaccine Immunol. | Clin. Microbiol. Rev. |
|---|---|
| J. Clin. Microbiol. | ALL ASM JOURNALS |