Clonal Dissemination of Macrolide-Resistant and Penicillin-Susceptible Serotype 3 and Penicillin-Resistant Taiwan 19F-14 and 23F-15 Streptococcus pneumoniae Isolates in Japan: a Pilot Surveillance Study

Abstract

Large-scale surveillance studies using molecular techniques such as pulsed-field gel electrophoresis (PFGE) have revealed that the spread of antibiotic-resistant pneumococci is due to clonal spread. However, in Japan, surveillance studies using such molecular techniques have never been done. Therefore, we conducted a pilot surveillance study to elucidate the present situation in Japan. Among the 145 isolates examined, the most prevalent serotype was type 19F (20%), for which most isolates were not susceptible to penicillin (86.2%) but were positive for the mef(A)/mef(E) gene (89.7%). The secondmost prevalent was serotype 3 (16.6%), for which most isolates were susceptible to penicillin (87.5%) and positive for the erm(B) gene (91.7%). PFGE analysis showed that both serotypes consisted mainly of clonally identical or related isolates and, in particular, 38% of the type 19F isolates were indistinguishable from or closely related to the Taiwan 19F-14 clone. In addition, some of the Japanese type 23F isolates with the erm(B) gene were indistinguishable from or related to the Taiwan 23F-15 clone as analyzed by PFGE. Based on the results of our pilot study performed in a single institution, it is likely that international antibiotic-resistant clones have already spread in Japan; therefore, a nationwide surveillance study should be urgently conducted.

Streptococcus pneumoniae is a leading etiologic agent of bacteremia, bacterial meningitis, pneumonia, bronchitis, sinusitis, and otitis media (3, 4, 10, 14, 26, 28, 30). S. pneumoniae isolates that are resistant to multiple classes of antibiotics were initially reported in South Africa in 1977 (15) and subsequently spread to geographically different areas through today's rapid transport and movement of people throughout the world (1). In Japan, the incidence of penicillin-nonsusceptible strains has increased from <1.0% during 1974 through 1982 to 65.3% during 1996 through 1997 (16). In the meantime, macrolide-resistant strains are also increasing in number (9, 13, 14, 18, 22, 24, 32) and fluoroquinolone-resistant strains have been found (5, 14, 16, 33) both in Japan and worldwide. This spread of drug-resistant pneumococci in a relatively short period of time has made accurate epidemiological surveillance of this common pathogen of the utmost importance. Large-scale international surveillance studies such as PROTEKT, SENTRY, Alexander, and the Asian Network for Surveillance of Resistant Pathogens have been ongoing and have revealed that the spread of such multidrug-resistant strains is due mostly to clonal spread; however, it has not yet been ascertained whether this is the case in Japan. So far, no surveillance studies using pulsed-field gel electrophoresis (PFGE) analysis have been conducted in Japan, though some PFGE studies of the Japanese strains of S. pneumoniae are available (12, 20, 34, 35). We therefore carried out a 3-year pilot surveillance study in order to assess how antibiotic-resistant pneumococcal isolates disseminate in Japan. In this report, we determined the MICs of several antibiotics and the macrolide-resistant genes. The serotype prevalence that is very important for vaccination strategy was also described in relation to antibiotic susceptibility, since only a few data for Japanese pneumococcal isolates have been reported in the English literature.

MATERIALS AND METHODS

Bacterial strains.

The present study was conducted at Nara Medical University, a tertiary care general hospital with 850 beds in the Nara prefecture located in the midst of Japan. S. pneumoniae strains were consecutively collected from clinical specimens from both inpatients and outpatients submitted to the laboratory of clinical microbiology at our university hospital over a 3-year period (2001 to 2003). In order to avoid duplication of samples, strains consecutively isolated from the same individual within 6 months were excluded. Blood agar medium containing 5% sheep' blood (Kyokuto Pharmaceutical Co., Tokyo, Japan) was used for the cultivation of these isolates at 37°C in a humidified atmosphere supplemented with 5% CO₂. S. pneumoniae cells were identified with standard techniques, including Gram stain, colonial morphology, optochin susceptibility, and bile solubility.

Serotyping.

Determination of capsular serotype and/or serogroup was accomplished by the slide agglutination method using a panel of commercially obtained antisera (Seiken Co., Tokyo, Japan). This included polyclonal antibodies for 39 frequently isolated serotypes and/or groups of S. pneumoniae. All serogroups were further serotyped by capsular swelling reaction with antiserum (the Statens Serum Institut, Copenhagen, Denmark).

Antibiotic susceptibility testing.

Commercially available MIC panels (MICroFAST 3J; DADE Behring, Tokyo, Japan) were used to determine the MICs for the isolates. The following antibiotics were tested: penicillin, cefditoren, cefotaxime, cefepime, chloramphenicol, levofloxacin, meropenem, rifampin, tetracycline, trimethoprim-sulfamethoxazole, and vancomycin. In order to evaluate higher MICs of erythromycin, clarithromycin, and clindamycin, customized MIC panels of antimicrobial agents were purchased (Eiken Co., Tokyo, Japan). Cultures for preparing the inoculum were grown on sheep blood agar in a humidified atmosphere supplemented with 5% CO₂, and a cell suspension at a 0.5 McFarland turbidity standard was prepared in sterilized phosphate-buffered saline (10 mM, pH 7.2). Twenty-five microliters of the suspension was transferred to Mueller-Hinton broth (Eiken Co.) containing 5% lysed horse blood, and then 100 μl of this suspension was inoculated into each well of the MIC panels. The MIC panels were incubated for 20 to 24 h at 35°C and read with a mirror reader (Dynex, Chantilly, Va.). The control strain used was S. pneumoniae ATCC 49619. The MIC breakpoints for susceptibility or resistance of all drugs were determined according to the recommendations by NCCLS (21).

PCR of mef(A)/mef(E) and erm(B) genes.

Because PCR using the primers below amplified a segment of the mef(A) gene but did not discriminate between the mef(A) and mef(E) genes, we refer to this as the mef(A)/mef(E) gene in this study. Detection of erm(B) and mef(A)/mef(E) genes was conducted as previously described by using a test kit commercially available in Japan (Wakunaga Pharmaceutical Co., Hiroshima, Japan) with primers modified as reported by Ubukata et al. (32) as follows: ermB1, 5′-₇₂₁CGTACCTTGGATATTCACCG₇₄₀-3′, and ermB2, 5′-₉₄₄GTAAACAGTTGACGATATTCTCG₉₂₂-3′, for the erm(B) gene and mefA1, 5′-₂₈₈CTGTATGGAGCTACCTGTCTGG₃₀₉-3′, and mefA2, 5′-₅₈₁CCCAGCTTAGGTATACGTAC₅₆₂-3′, for the mef(A)/mef(E) gene. Briefly, a single colony on the blood agar medium was suspended in a microtube containing 30 μl of a lysis solution. The tube was placed in a thermal cycler (Gene Amp PCR System 9600-R; Perkin-Elmer, Norwalk, Conn.), and bacterial cells were lysed under reactive conditions of 60°C for 10 min followed by 94°C for 5 min. Next, 2 μl of the lysate was placed in a PCR tube containing 25 μl of reaction mixture. One milliliter of the reaction mixture consisted of 60 ng of a primer for each of erm(B) and mef(A)/mef(E), 80 μl of 10 mM dinucleoside triphosphate mixture, 40 U of Tth DNA polymerase, and 100 μl of 10× PCR buffer. The PCR conditions were 94°C for 20 s, 52°C for 20 s, and 72°C for 15 s for 30 cycles total.

PFGE study.

PFGE was performed using the GenePath strain typing system (Bio-Rad, Hercules, Calif.) according to the manufacturer's protocol. Briefly, SmaI restriction fragments of chromosome DNA were separated on agarose gels and stained with ethidium bromide, and subsequently, PFGE patterns were analyzed visually and compared. Strains were considered identical if they shared every band, closely related if they differed by one to three bands, possibly related if they differed by four to six bands, or unrelated if they differed by seven or more bands (31).

RESULTS

A total of 187 isolates were collected over the study period; however, after exclusion of duplicate samples from the same patients, 145 were subject to the final analysis. They originated from sputum (n = 84 [57.9%]), nasopharyngeal swabs (n = 35 [24.1%]), ear swabs (n = 9 [6.2%]), bronchial secretions (n = 8 [5.5%]), eye swabs (n = 4 [2.8%]), and other sources, including blood, aspiration fluid from empyema, and female genital specimens (n = 5 [3.4%]). Strains were isolated from adults >15 years of age (n = 98 [67.6%]), from children 2 to 15 years of age (n = 28 [19.3%]), and from infants and young children <2 years of age (n = 19 [13.1%]).

Table 1 shows the serotype distribution of the 145 isolates of S. pneumoniae. In the present study, 139 isolates (95.9%) belonged to 17 different serotypes and 6 (4.1%) of the isolates were nontypeable. Serotypes 19F, 3, 6B, 6A, and 23A accounted for 68.3% of all isolates. Overall, 71.0% of the isolates were included in the serotypes covered by the 23-valent pneumococcal vaccine.

TABLE 1.

Distribution of serotype inclusion in the 23-valent vaccine or in the 7-valent conjugate vaccine and the number of penicillin- and erythromycin-susceptible and-resistant pneumococcal isolates in each serotype

Category and serotype	No. of isolates (%)	Antibiotic sensitivity toa:
		Penicillin			Erythromycin
		S	I	R	S	I	R
Included in the 7- and 23-valent vaccines
19F	29 (20.0)	4	12	13		1	28
6B	20 (13.8)	5	6	9	3		17
23F	10 (6.9)	1	5	4	1	1	7
14	6 (4.1)	3	3		1		5
18C	2 (1.4)	1		1	1		1
4	1 (0.7)	1			1
9V	1 (0.7)	1			1
Included in the 23-valent vaccines only
3	24 (16.6)	21	3		2		22
11A	3 (2.1)	2		1	2		1
15B	6 (4.1)	5	1				6
19A	1 (0.7)	1					1
Others	0
Not included in vaccines
6A	15 (10.3)	9	5	1	1		14
23A	11 (7.6)		10	1	1		10
NT	6 (4.1)	4	2		2		4
35B	7 (4.8)		5	2	2		5
15A/C/F	1 (0.7)		1				1
16	1 (0.7)		1				1
29	1 (0.7)	1			1

^aS, susceptible; I, intermediately susceptible; R, resistant.

Table 2 gives the MIC range, the MICs at which 50% of the isolates were inhibited (MIC₅₀s), and the MIC₉₀s of the 14 antimicrobial agents for the 145 isolates of S. pneumoniae. Sixty-one percent of the isolates were not susceptible to penicillin, exhibiting either intermediate resistance (37.9%) or high-level resistance (23%). Among the cephalosporins tested, cefditoren was the most active, with a MIC₅₀ of 0.25 μg/ml and a MIC₉₀ of 0.5 μg/ml. Extremely large numbers of the isolates were resistant (including intermediate and high-level resistance) to erythromycin (84.8%), clarithromycin (80.0%), and tetracycline (87.6%). All isolates were susceptible to levofloxacin, rifampin, and vancomycin.

TABLE 2.

Susceptibilities of 145 S. pneumoniae isolates to 17 antimicrobial agents

Agent	MIC			% of isolates susceptible or resistanta
	Range	50%	90%	S	I	R
Penicillin	0.03-4	0.25	2	39.3	37.9	23.0
Cefditoren	0.06-1	0.25	0.5	—c	—	—
Cefotaxime	0.06-2	0.25	1	82.1	13.8	4.1
Cefepime	0.5-2	<0.5	2	54.5	37.2	8.3
Chloramphenicol	4-16	<4	8	68.3	31.7	—
Erythromycin	0.06-64	8	>64	15.2	4.1	80.7
Clarithromycin	0.06-64	4	>64	20.0	7.6	72.4
Clindamycin	0.06-64	<0.06	>64	50.3	0.7	49.0
Levofloxacin	0.25-8	1	1	100	—	—
Meropenem	0.12-2	<0.12	0.5	82.1	17.9	—
Rifampin	1-4	<1	—	100	—	—
Tetracycline	0.5-4	>4	>4	12.4	1.4	86.2
Trim/Stb	0.5/9.5-4/76	<0.5/9.5	1/19	71.7	24.8	3.5
Vancomycin	0.12-1	0.5	0.5	100	—	—

^aS, susceptible; I, intermediate; R, resistant.

^bTrim/St, trimethoprim and sulfamethoxazole. Ratio of concentrations of trimethoprim and sulfamethoxazole, 1:19.

^c—, no NCCLS breakpoints were available.

The relationship between serotypes and penicillin or macrolide susceptibility is summarized in Table 1. The most frequently isolated serotype, type 19F, was not susceptible to penicillin (25 of 29, 86.2%), whereas the secondmost frequent serotype, type 3, was mostly susceptible to penicillin (21 of 24, 87.5%). There was no apparent association between serotype and macrolide susceptibility, whereas in regard to macrolide-resistant genes (Table 3), the macrolide resistance of type 19F was associated with the mef(A)/mef(E) gene and that of type 3 was associated with the erm(B) gene. In order to clarify this characteristic phenotype and genotype, PFGE was performed for type 19F and type 3 strains.

TABLE 3.

Relationship between macrolide-resistant genes and serotypes

Resistant gene	No. of pneumococcal isolates tested of serotypes:
	19F	3	6B	6A	23A	23F	35	15	NT	14	11	18	4	9	16	19A	29
DNa		2	2	1			2		2		2	1	1	1			1
erm(B)	3	22	17	14	5	6		1	1	3	1	1				1
mefb	25		1		4	4	5	6	3	3					1
DPc	1				2

^aDN, neither erm(B) nor mef(A)/mef(E) was positive.

^bmef, mef(A)/mef(E).

^cDP, both macrolide-resistant genes [mef(A)/mef(E) and erm(B)] were positive.

Figure 1 shows the representative PFGE patterns of serotype 3 with the number of isolates included in each representative group. There were only two PFGE patterns, and 22 out of 24 isolates of serotype 3 pneumococcal isolates [all possessing the erm(B) gene] belonged to pattern A (lane A) and 2 isolates [all not possessing the erm(B) gene] belonged to pattern B (lane B). There was only one band difference between patterns A and B, suggesting that genetically all of the 24 isolates are closely related. Figure 2 shows the representative PFGE patterns of serotype 19F with the number of isolates included in each group. The PFGE pattern of Taiwan 19F-14 is shown in lane 1. The PFGE pattern of the most commonly isolated clone in serotype 19F isolates is shown in lane 2, which is indistinguishable from that of the Taiwan 19F-14 clone. The PFGE pattern of serotype 19F was divided into 12 patterns, indistinguishable for 8 isolates (lane 2), closely related for 3 isolates (lanes 3 through 5), possibly related for 12 isolates (lanes 6 through 10), and different for 3 isolates (lanes 11 through 13), suggesting that most of them are variants of the Taiwan 19F-14 clone. The mef(A)/mef(E)⁺ erm(B)⁺ isolate exhibited the closely related pattern (lane 5). A serotype 19F isolate, harboring the erm(B) gene, exhibited PFGE patterns (Fig. 3) different from that of the Taiwan 19F-14 clone (Fig. 2, lane 1).

FIG. 1.

PFGE patterns of SmaI-digested DNAs of the 24 serotype 3 isolates. The first and fourth columns of the gel (λ) denote the ladders. A and B represent patterns A and B, respectively (see Results for details). The numbers of isolates belonging to each representative PFGE pattern are indicated at the bottom of the figure.

FIG. 2.

PFGE patterns of SmaI-digested DNAs of the 26 serotype 19F pneumococcal isolates with the mef(A)/mef(E) gene. The first and last columns of the gel (λ) denote the ladders. The numbered columns are as explained in Results. The numbers of isolates belonging to each representative PFGE pattern are indicated at the bottom of the figure.

FIG. 3.

PFGE patterns of SmaI-digested DNAs of the three serotype 19F isolates with only the erm(B) gene. The first and last columns of the gel (λ) denote the ladders. The numbers of isolates belonging to each representative PFGE pattern are indicated at the bottom of the figure.

Following the above results, we looked into the possibility of domestic dissemination of another Taiwan penicillin-resistant S. pneumoniae (PRSP) clone, 23F-15; we compared the PFGE patterns of penicillin-resistant Japanese type 23F isolates with those of Taiwan 23F-15. As shown in Fig. 4, they were grouped into eight patterns: indistinguishable for two isolates (lane 3), closely related for two isolates (lanes 2 and 9), possibly related for three isolates (lanes 4 and 7), and different for three isolates (lanes 5, 6, and 8). Similar to the results for type 19F, most of the serotype 23F isolates were considered variants of the Taiwan 23F-15 clone.

FIG. 4.

PFGE patterns of SmaI-digested DNAs of the nine serotype 23F isolates with reduced susceptibilities to penicillin G. The first and last columns of the gel (λ) denote the ladders. Lane numbers for the columns are as explained in Results. The numbers of isolates belonging to each representative PFGE pattern are indicated at the bottom of the figure.

DISCUSSION

In Japan, the first nationwide surveillance of serotype distribution was conducted in 1984 by Fukumi et al. (7). The 590 identified isolates belonged to 43 Danish types, and the 5 most common serotypes were types 3 (12.7%), 19F (9.3%), 23F (6.8%), 6B (5.9%), and 14 (4.9%), in decreasing order. They reported that 72.9% belonged to 1 of the 23 pneumococcal types to be included in the commercial 23-valent pneumococcal polysaccharide vaccine. From 1994 to 1996, a nationwide surveillance study analyzing 4,255 strains of S. pneumoniae in Japan was performed (16). This revealed that the four most common serotypes and/or groups in Japan were types 19 (26%), 6 (20%), 23 (16%), and 3 (12%). The frequency of isolation of serotype 3 pneumococci remained unchanged; however, the isolation frequency of serogroups 19, 6, and 23 had substantially increased. The MIC analysis clearly showed that serogroups 19, 6, and 23 were mainly not susceptible to penicillin, whereas no PRSP was found in serotype 3 pneumococci (16). Thus, antibiotic selection pressure might have made serogroups 19, 6, and 23 much more prevalent. Similar to these findings, our results showed that the most prevalent serogroups are 19, 3, 6, and 23, in decreasing order. It is well known that the isolation frequency of serotypes differs according to the source and age of patients: serotype 3 is more often recovered from middle-ear fluid and from adults but less commonly found from other sources and infants. Serotypes 1 and 14 are more often isolated from blood (10). Our results showed that serotype 3 was much more commonly recovered than serogroups 6 and 23; most of the former were derived from adults (data not shown). In other countries, however, the overall frequency of isolation of serotype 3 does not reach as high as 10% (10, 17); therefore, the relatively high isolation frequency of serotype 3 might be characteristic of Japan. Our findings that 71.0% of the isolates were covered by the 23-valent vaccine suggest that the vaccine may be effective to a certain degree. Very recently, however, the emergence of type 35B multidrug-resistant invasive S. pneumoniae has been reported (2). Furthermore, isolation of serotype 35B was relatively common in our study (4.8%): they were all multidrug resistant, and one of seven isolates was recovered from blood culture (data not shown). In the age of vaccination, serotypes that are not included in the current vaccine could replace previously dominant serotypes in the future, which may be of growing concern. It is indeed important to continue monitoring the serotype distribution carefully.

The first clinical isolate with reduced susceptibility (MIC of 0.6 μg/ml) to penicillin G was reported in 1967 (9). By 1977, resistant strains had been described from multiple geographical areas (15). In Japan, the first penicillin-nonsusceptible pneumococcus was identified in 1981 (23). Afterward, the frequency of penicillin-intermediate S. pneumoniae and PRSP gradually increased. In the recently conducted Alexander study, analysis of 404 pneumococcal isolates from Japan revealed that the frequencies of penicillin-sensitive S. pneumoniae, penicillin-intermediate S. pneumoniae, and PRSP in Japan were 49.3, 22.3, and 28.5%, respectively (14); this finding is similar to our results. Resistance to β-lactam antibiotics is also common in Japan (16); actually, the most effective β-lactam antibiotic in the present study was cefditoren. There are many reports regarding the high rate of resistance to macrolides in S. pneumoniae in Asian countries including Japan (13, 16, 18, 22, 24, 30, 32). The reported incidence of macrolide-resistant S. pneumoniae in Japan is as high as 70 to 80%, which is consistent with our results. Recently, fluoroquinolone-resistant S. pneumoniae was also reported. However, such clones represent far less than 1% of the isolates (33) and none were found in our study, thus indicating that fluoroquinolone resistance is not yet a pressing problem.

There have been many epidemiological studies regarding association between serotypes and penicillin resistance (16, 17, 27). These studies have suggested that antibiotic selection pressure may have selected dominant serotypes associated with penicillin resistance: such examples were serogroups 6, 14, 19, and 23. However, our results and previous reports (16, 32) have shown that no PRSP was found among serotype 3 isolates in Japan. This situation for serotype 3 isolates in Japan is presumably explained by macrolide selection pressure, since this serotype exerts the high frequency and high levels of macrolide resistance with the erm(B) gene, as shown in this study.

In contrast, there seems to be no relationship between serotype and macrolide susceptibility, presumably because the frequency of macrolide resistance is extremely high among all serotypes. However, we revealed that there was a clear association between serotypes and macrolide-resistant genes. The distribution of these genes may vary with epidemiological settings (8, 11, 13, 18, 22, 24, 32). Japanese investigators (22, 32) showed that all of the macrolide-resistant serotype 3 possessed the erm(B) gene, whereas mef(A) was prevalent (75%) in serogroup 19. These findings are consistent with our present results; serotype 19F is associated with the mef(A)/mef(E) gene, and serotype 3 is strongly associated with the erm(B) gene. Although we did not look into the mef(E) gene, there was a report on the association of several pneumococcal serotypes (14, 19F, and 23F) in Atlanta with the mef(E) gene (11). In our study, we found one isolate that has both the mef(A)/mef(E) and erm(B) genes and is closely related to Taiwan19F-14. Farrell et al. have reported that such an isolate carries the mobile genetic elements Tn1545 and mega and suggested that the acquisition of such elements seems to be very effective in global dissemination (6). We therefore have to pay close attention to the increase of such isolates by continuous surveillance studies in Japan.

Molecular typing such as PFGE or multilocus sequence typing of pneumococcal isolates has been extensively performed throughout the world (19, 27, 29). However, information using such molecular techniques is very limited in Japan (12, 20, 30, 34, 35). International clonal dissemination of antibiotic-resistant pneumococcal isolates is well documented elsewhere (19, 27, 30), and the Taiwan 19F-14 clone is one of the representative penicillin-resistant strains in the United States (27). The Asian Network for Surveillance of Resistant Pathogens Study revealed the spread of the Spanish 23F clone in Asia (30), while only one isolate was related to this Spanish clone in Japan. Furthermore, serotypes of 39 Japanese isolates analyzed in the study were not specified, including the Taiwan 19F-14 and Taiwan 23F-15 clones. The SmaI-restricted PFGE patterns of the mef(A)-positive type 19F PRSP Japanese isolates were reported by Hoshino et al. (12). Most of the PFGE patterns seemed to be indistinguishable or closely related to the PFGE pattern of Taiwan 19F-14. Interestingly, those isolates were recovered in the Nagasaki prefecture, located 600 km west of our province, or in the Chiba prefecture, located 700 km east of our province; thus, this fact suggests that the Taiwan 19F-14 clone has disseminated throughout Japan. As shown in the present study, the Taiwan 23F-15 clone and its variants are also spreading in Japan.

The PFGE results of type 3 pneumococcal isolates clearly showed their isogenecity. Reports from other countries have shown that epidemiologically unrelated serotype 3 clones are genetically related to each other (17, 25), suggesting that serotype 3 has evolved only recently or has remained unchanged over long periods. Though all of our serotype 3 pneumococci were isolated at a single institution, genetically related serotype 3 clones may have disseminated throughout Japan, based on the study by Nagata et al. (20) that was conducted in a hospital with 350 beds in the Kumamoto prefecture, located 550 km west of our province. They reported three cases of hospital-acquired infection with type 3 S. pneumoniae. They showed identical PFGE patterns of type 3 S. pneumoniae isolates from three patients in the same ward; moreover. This PFGE pattern was identical to that of a type 3 pneumococcal isolate from an outpatient attending the same hospital. Importantly, the PFGE pattern by SmaI digestion of those four isolates was apparently identical to that of the type 3 isolates in our present study. Taken together, it might be speculated that Japanese serotype 3 isolates have been genetically unchanged over a long period of time except for the acquisition of macrolide resistance by horizontal transfer. In Japan, most isolates of serotype 3 pneumococci have been reported to be susceptible to penicillin (16), and this fact may be accounted for by the clonal spread of serotype 3 in Japan.

Finally, our results show that there is already some diversity among Taiwan 19F-14- and Taiwan 23F-15- related clones, respectively. This suggests that these related clones may easily acquire mutations for resistance to other antibiotics. After comparing the high homogeneity of penicillin-sensitive S. pneumoniae type 3 pneumococcal isolates, it can be speculated that among pneumococcal isolates, there are some differences in the ability of acquiring genetic elements, especially for antibiotic resistance.

In conclusion, the present results may reflect the current epidemiological situation of S. pneumoniae in Japan, although this was a pilot study conducted in a single local hospital. To validate our present conclusion, large scale, nationwide surveillance using the molecular typing technique should be urgently performed in Japan.