To the Editor: Streptococcus suis is an important zoonotic pathogen for swine and humans. Among 33 serotypes, serotype 2 is more frequently isolated from diseased pigs than other serotypes (1). However, not all serotype 2 strains are virulent, and degree of virulence varies among strains (2). Previous studies have reported several S. suis putative virulence factors, including the polysaccharide capsule, the muramidase-released protein, the extracellular factor, and suilysin (3–5). Some of these factors have been used as virulence-associated markers, and the association of the factors of S. suis isolates with virulence or clinical background has been suggested in Europe (2,5). However, because many virulent isolates lacking these factors have also been isolated from clinical cases in Canada (6), they cannot be used as virulence markers in North America. Recent analysis of S. suis isolates by multilocus sequence typing (MLST) suggested the association of some clonal groups with particular clinical manifestations. That is, most invasive isolates belonged to the sequence type (ST) 1 complex, while the ST27 and ST87 complexes were found to include a higher proportion of lung isolates (7). Although S. suis has been prevalent worldwide, the geographic location of the isolates used so far was mainly Europe, North America, and East Asia (7–9). Moreover, the clonal association with virulence of S. suis has been discussed mainly on the basis of clinical and experimental data in swine (7). In this report, to broaden understanding of the population structure of S. suis as a zoonotic agent, we characterize 20 S. suis isolates (Table) recovered from humans in Thailand in 1998–2002. Table Epidemiologic data of Streptococcus suis isolates from patients in Thailand, 1998–2002* Serotyping by coagglutination tests showed that 19 of the 20 isolates belonged to serotype 2, while the remaining 1 (MNCM07) was serotype 14. MLST analysis resolved the 20 isolates into 8 STs (Table). By using eBURST (http://eburst.mlst.net), we assigned 4 isolates (MNCM01, MNCM06, MNCM07, and MNCM16) from 1 case of endocarditis and 3 cases of meningitis to the ST1 complex. The remaining isolates were assigned to the ST27 complex with a less-stringent group definition (Table), although ST101 (MNCM21) and ST104 (MNCM50) shared only 2 alleles with ST27 and were incorporated into this complex by a chaining effect. Regarding the clinical cases from which the ST27 complex isolates were recovered, the patients had meningitis, endocarditis, septicemia, septic shock, diarrhea, and respiratory involvement. The 2 ST complexes both contained isolates from deceased patients (Table). All the isolates assigned to the ST1 complex were positive for the suilysin gene sly, the extracellular factor gene epf or its variant, and the muramidase-released protein gene mrp or its variant. With the exception of MNCM21 and MNCM50, which had only sly, all isolates classified into the ST27 complex were negative for sly and epf but positive for mrp or its variant. These results showed the congruence between STs and the virulence-associated gene profiles and further support the usefulness of MLST for epidemiologic studies of S. suis. Of the 3 major clonal complexes identified so far in S. suis (ST1, ST27, and ST87), the ST1 complex particularly attracts considerable public attention as a clonal group that may have the potential for a higher degree of virulence than the others (7), and most (96%) of the human isolates investigated so far, including ST7 isolates, which caused the largest outbreak in China, belong to the ST1 complex (7–9). In this study, although no ST7 isolate was found, 4 isolates were assigned to the ST1 complex. This further confirmed the gravity of the ST1 complex not only for swine industries but also for public health. In contrast to the ST1 complex, only 4 human clinical isolates have so far been reported to belong to the ST27 complex. Three of the 4 are isolates from Canada that belong to ST25 (7). The remaining 1 is from Japan and assigned to ST28 (8). Unlike in previous reports, 80% of the human clinical isolates (16 isolates) characterized in this study were assigned to the ST27 complex. Although previous studies suggested that members of the ST27 complex may have lower potential to cause invasive diseases in swine (7), all the isolates were isolated from blood or cerebrospinal fluid of the patients, suggesting a high degree of invasiveness (Table). Because it is unknown whether the ST27 complex is also dominant among isolates from diseased pigs in Thailand, future surveillance will be necessary to know the situation in pigs. However, our data indicate that the ST27 complex is another clonal group that should be assessed for its importance for human infection. Because mrp, epf, and sly are not appropriate as virulence markers for the ST27 complex members, development of novel virulence markers will be needed for efficient discrimination of S. suis strains virulent for humans.