Streptococcus pneumoniae is a major pathogen, causing otitis media, sinusitis, pneumonia, bacteremia, and meningitis in children worldwide. It also offers some threat of mortality, mostly in developing countries1,2). The recent spread of antibiotic-resistant S. pneumoniae strains has caused considerable concern3,4). Pneumococcal conjugate vaccines (PCVs), used since 2000, have been shown to reduce the incidence of vaccine-serotype invasive pneumococcal disease (IPD) and nasopharyngeal (NP) colonization of pneumococcus by the vaccine serotypes3,5-11). In spite of the availability of these efficient vaccines, pneumococcal infections remain a global problem due to the emergence of resistance and replacement by nonvaccine-serotype strains12,13).

Pneumococcal infections are traditionally diagnosed using bacterial culture combined with biochemical or immunochemical identification tests. However, conventional culture methods take at least 48 hours to obtain results, and antibiotic treatment prior to obtaining patient specimens or autolysis of pneumococci during transportation of samples may cause false-negative culture results. In addition, the current determination of serotypes by serological methods requires isolation of pneumococci, is time consuming and expensive, and the results are difficult to interpret. Rapid and precise methods of detection and serotyping of S. pneumoniae are needed to improve the diagnosis of IPD, but they are also essential for predicting the consequences of vaccinations on NP carriage. In this study we aimed to compare PCR with traditional culture isolation methods for detecting S. pneumoniae in NP aspirates and to improve the design of previously published protocols for sequential multiplex PCR-based serotyping of pneumococci according to the predicted distribution of pneumococcal serotypes in Korea.

In this study, multiplex PCR was used to detect pneumococci and to determine pneumococcal serotypes among 842 total cases. Of the total samples, 793 underwent simultaneous bacterial culture and PCR analysis, with a concordance of 85.6%. The detection rate of pneumococcal strains by multiplex PCR (19.3%) was higher than that by bacterial culture (10.2%). PCR analysis based on the amplification of cpsA or other pneumococcal genes has been shown to be sensitive and efficient in detecting pneumococcus in clinical samples16,20-23). Virolainen et al.20) found 18% of middle ear fluid specimens to be positive by culture and 30% by culture and PCR. Saha et al.16) showed that sequential multiplex PCR of 127 culture-negative cerebrospinal fluid samples yielded serotypes for 51 additional cases. Our results are consistent with those of previous studies in this field. In the present study, although 2.6% of samples showed positive results in culture but negative by PCR, these are thought to be the result of operator error and will likely be reduced by increased operator experience.

In addition to detection of pneumococci, accurate serotype determination is important for vaccine development. Presently the choice of vaccine strains depends on serotype prevalence patterns. Although serological determination by a quelling reaction of pneumococcal cells with antipolysaccharide sera is the currently used method for pneumococcal serotyping24,25), the high cost of antisera, the requirements for technical expertise, and subjectivity in interpretationof results are major shortcomings of the procedure14,16). Multiplex PCR has the potential to overcome these drawbacks and has been reported to have higher sensitivity and specificity than traditional quelling reaction methods14,16,26,27).

Because the serotype distribution among NP carriage isolates varies by geographic region, time, and age-group, we designed primers and modified the sequential multiplex PCR scheme based on the distribution of serotypes that were frequently isolated in Korea17-19). Among the serotypes identified in our study, the most common serotype was 19A, followed by: 6A/B, 19F, 15B/C, 15A, 11A, and in a small minority 35B, 23F, 14, 9V, 45 and 33. In Korea, PCV7 has been available since November 2003 and PCV10 and PCV13 were introduced in May 2010. Thus, our results, obtained from March 2009 to June 2010, represent the impact of PCV7 on NP carriage of pneumococcus. Several studies conducted in Korea prior to the use of PCV7 demonstrated that the proportion of PCV7 serotypes in invasive and noninvasive isolates, including carried strains, was 54 to 72% and 57 to 83%, respectively17-19,28,29). Our study revealed that PCV7 serotypes decreased to 30.5% of carried serotypes over the course of this study period, indicating a replacement phenomenon in NP carriage after the introduction of PCV7.

As PCV10 or PCV13 have only recently been introduced, there will likely be changes in serotype distribution among pneumococcal carriage isolates in the future. Surveillance of pneumococcal serotypes should be continued in order to monitor changes in serotype distribution necessary for the pneumococcal vaccine policy. To that end, our multiplex PCR system can be used as an accurate and rapid way of isolating and serotyping S. pneumoniae compared to conventional microbiology culturing. However, our study had an important limitation; 26% of total isolates were non-typeable because only 35 primer sets for pneumococcal serotyping were used in the protocol. In the future, new serotype-specific primers will need to be added, so that more serotypes may be detected. This will enhance detection and serotyping of S. pneumoniae when using multiplex PCR.