Virological interpretations of dengue disease spectrum in infants in Chennai, Tamil Nadu, India, need reevaluation.

We read with interest the study of Kabilan et al. (2) regarding the dengue disease spectrum in infants in Chennai, Tamil Nadu, India. However, certain points in this study regarding the virological investigations and their interpretations are not clear. As specified in the paper, the authors used denguespecific immunoglobulin M (IgM) and IgG microenzymelinked immunosorbent assay (microELISA) reagents for the detection of the respective antibodies. As a considerable number of cross-reactions occur between members of the flavivirus group, i.e., those causing dengue, Japanese encephalitis, and West Nile viral infections (1), it is worthwhile to look for specific IgM antibody against dengue viruses by the -capture technique. We have been working in the field of flaviviruses for the last decade (4, 5) and are using Pan-Bio -capture IgM microELISA reagents for the diagnosis of suspected dengue cases. In our experience, the -capture technique has been found to be better (94.7% sensitivity and 98.9% specificity, according to the manufacturer) than the non-capture microELISA system so far as diagnosis of dengue virus infection is concerned. Hence, it would have been more useful if the authors had used the -capture technique for the serological confirmation of dengue virus infection in their study. As is evident from Table 2 in the original study (2), two cases of dengue shock syndrome were diagnosed as acute dengue virus infections based upon the dengue virus IgG antibody only. In areas where dengue is endemic, it appears to be difficult even among infants to exclude the presence of either passively transferred maternal IgG antibodies or IgG antibodies acquired due to past exposure in the form of clinical or subclinical infection. The first problem would be easily eliminated by determining the absence of dengue virus IgG antibodies in the maternal serum. Otherwise, the role of dengue virus IgG antibody in the diagnosis of acute infection lies with either the demonstration of seroconversion or a significant rise in antibody titers in follow-up samples (1). Since Kabilan et al. reported that the epidemic was due to dengue virus serotype 4, we would like to clarify that crossreactions between flaviviruses as well as between dengue virus serotypes have been observed. It is difficult to type dengue viruses serologically by using the cross-reactive antigens that are available in many microELISA kits; i.e., the Pan-Bio dengue IgM kit utilizes dengue virus serotype 2 as the coating antigen, which cross-reacts with other dengue virus serotypes (serotypes 1, 3, and 4), whereas the Pan-Bio -capture IgM microELISA kit utilizes a pool of dengue virus serotypes 1 to 4 as the source of common dengue virus antigen. None of the kits mentioned above have the ability to differentiate between the serotypes unless serotype-specific antigenic determinants of dengue viruses are used individually in separate kits. At best, this type of kit can be used only for diagnosing acute dengue virus infection, not for serotyping. Conventionally, serotyping is done by neutralization, immunofluorescence tests using type-specific antisera (1), or reverse transcriptase PCR using type-specific primers (3). Hence, it would be worthwhile to know the type of antigens the authors used to determine the infection to be caused by dengue virus serotype 4. This information would be of much use to the researchers working in similar fields.