Association of BAK1 single nucleotide polymorphism with a risk for dengue hemorrhagic fever

BACKGROUND

Dengue hemorrhagic fever (DHF) is a severe life-threatening form of dengue infection. Low platelet count is one of the characteristic clinical manifestations in patients with severe dengue. However, little is known about genetic factors in the host that cause low platelet count in patients with dengue.

METHODS

A previous genome-wide association study of hematological and biochemical traits identified single nucleotide polymorphisms (SNPs) associated with low platelet count in healthy subjects. To examine the possible association of these SNPs with DHF, 918 Thai patients with dengue [509 patients with DHF and 409 with dengue fever (DF)] were genotyped for five SNPs: rs5745568 in BAK1, rs6141 in THPO, rs6065 in GP1BA, rs739496 in SH2B3, and rs385893 in RCL1. In addition, rs4804803 in CD209, that has been reported to be associated with dengue infection, was also genotyped to examine if rs4804803 affects the association detected in this study.

RESULTS

The allele frequencies of each SNP were compared between the DHF and DF groups. Among the five SNPs, the G allele of rs5745568 in BAK1 was significantly associated with a risk for DHF [P = 0.006 and crude odd ratio (95 % confidence interval) = 1.32 (1.09-1.60)]. The association of this allele with DHF was also significant in a logistic regression analysis adjusted for age, sex, hospital (i.e., geographic region), immune status (i.e., primary or secondary infection), and virus serotype [P = 0.016 and adjusted odd ratio (95 % confidence interval) = 1.29 (1.05-1.58)]. The result was not influenced by rs4804803 [P = 0.0167 and adjusted OR (95 % CI) = 1.29 (1.05-1.58)]. No other SNPs including rs4804803 showed significant association.

CONCLUSIONS

The low-level constitutive production of platelets caused by the G allele of rs5745568 seems to increase the risk of bleeding in dengue infection. Our results suggest that BCL-2 homologous antagonist/killer (BAK) protein, encoded by BAK1, plays a crucial role in the pathogenesis of DHF.

Sequence variation of dengue type 2 virus isolated from clinical cases in Thailand

Dengue fever (DF) and dengue hemorrhagic fever (DHF) are caused by mosquito-borne dengue virus (DENV) infection leading to death in tropical and subtropical countries. In Thailand, all 4 serotypes of DENV are circulating. The most severe cases of DF and DHF are primarily introduced by secondary infections. Epidemiological studies have demonstrated that approximately 20% of the primary infection cases were caused by DENV-1 and -3, while the cases of DENV-2 or -4 accounted for less than 3%. For this reason, DENV-2 and -4 from primary infections have not been well studied. In this study, the sequence diversity of the envelope gene of 8 DENV-2 clinical isolates from primary/secondary infections was analyzed. DENV-2 from primary infections were highly heterogeneous in individual patients, whereas those from secondary infections were homogeneous. Phylogenetic analysis demonstrated that the heterogeneous population of DENV-2 from primary infections was composed of closely related quasispecies. Homogenous DENV-2 could be derived from selection of a particular viral population in secondary infections. The degree of sequence diversity of DENV-2 varied, and thus quasispecies may be involved in the progression of DENV infection.

Characterization of chikungunya virus-like particles

Chikungunya virus (CHIKV) is becoming a global concern due to the increasing number of outbreaks throughout the world and the absence of any CHIKV-specific vaccine or treatment. Virus-like particles (VLPs) are multistructured proteins that mimic the organization and conformation of native viruses but lack the viral genome. They are noninfectious and potentially safer vaccine candidates. Recent studies demonstrated that the yield of CHIKV VLPs varies depending on the strains, despite the 95% amino acid similarity of the strains. This might be due to the codon usage, since protein expression is differently controlled by different organisms. We optimized the region encoding CHIKV structural proteins, C-E3-E2-6k-E1, inserted it into a mammalian expression vector, and used the resulting construct to transfect 293 cells. We detected 50-kDa proteins corresponding to E1 and/or E2 in the cell lysate and the supernatant. Transmission electron microscopy revealed spherical particles with a 50- to 60-nm diameter in the supernatant that resembled the native CHIKV virions. The buoyant density of the VLPs was 1.23 g/mL, and the yield was 20 µg purified VLPs per 108 cells. The VLPs aggregated when mixed with convalescent sera from chikungunya patients, indicating that their antigenicity is similar to that of native CHIKV. Antibodies elicited with the VLPs were capable of detecting native CHIKV, demonstrating that the VLPs retain immunogenicity similar to that of the native virion. These results indicated that CHIKV VLPs are morphologically, antigenically, and immunologically similar to the native CHIKV, suggesting that they have potential for use in chikungunya vaccines.

A replication study confirms the association of GWAS-identified SNPs at MICB and PLCE1 in Thai patients with dengue shock syndrome

Background

Dengue shock syndrome (DSS), a severe life-threatening form of dengue infection, mostly occurs in children. A recent genome wide association study (GWAS) identified two SNPs, rs3132468 of major histocompatibility complex class I polypeptide-related sequence B (MICB) and rs3765524 of phospholipase C, epsilon 1 (PLCE1), associated with DSS in Vietnamese children. In this study, to examine whether an identical association is found in a different population, the association of these two SNPs with DSS was assessed in Thai children with dengue.

Methods

The rs3132468 and rs3765524 SNPs were genotyped in 917 Thai children with dengue: 76 patients with DSS and 841 patients with non-DSS. The allele frequencies were compared between DSS and non-DSS groups by one-sided Fisher’s exact test. The association of rs3132468 and rs3765524 with the mRNA expression levels of MICB and PLCE1 were assessed in EBV-transformed lymphoblastoid cell lines.

Results

The reported DSS-risk alleles were significantly associated with DSS in Thai patients with dengue (one-sided P = 0.0213 and odds ratio [OR] = 1.58 for rs3132468-C and one-sided P = 0.0252 and OR = 1.49 for rs3765524-C). The rs3132468-C allele showed a significant association with lower mRNA level of MICB (P = 0.0267), whereas the rs3765524-C allele did not. These results imply that the MICB molecule may play an important role in the prevention of DSS in dengue infection.

Conclusions

Together with previous association studies, we conclude that rs3132468-C at MICB and rs3765524-C at PLCE1 confer risk of DSS in Southeast Asians.

Association of IL1B -31C/T and IL1RA variable number of an 86-bp tandem repeat with dengue shock syndrome in Thailand

BACKGROUND

Dengue patients present a range of symptoms: dengue fever (DF), dengue hemorrhagic fever (DHF), and dengue shock syndrome (DSS). It is not clear whether this variability is due to their genetic background. Here we tested polymorphisms of interleukin 1 beta (IL1B) and interleukin 1 receptor antagonist (IL1RA) genes for association with DSS in the Thai population.

METHODS

Polymorphisms of IL1B -31C/T (rs1143627) and IL1RA 86-base-pair tandem repeat were analyzed in 871 patients (DF = 384, DHF = 413, and DSS = 74).

RESULTS

IL1B -31C and IL1RA 2/4 genotype were associated with DSS (IL1B -31C: DSS vs DHF: P = .0061, odds ratio [OR, 95% confidence interval {CI}], 3.49 [1.36-8.95]; DSS vs DF: P = .027, OR [95% CI], 2.81 [1.12-7.06]; IL1RA 2/4: DSS vs DHF: P = .017, OR [95% CI], 1.94 [1.12-3.40]; DSS vs DF: P = .024, OR [95% CI], 1.90 [1.07-3.4]). No difference was found between DF and DHF. Logistic regression analysis revealed that IL1B -31C and IL1RA 2/4 genotypes were each independently associated with DSS.

CONCLUSIONS

Patients with IL1B -31C carrier, or IL1RA 2/4 genotype carry a risk for DSS, implying that IL1B may play a role in pathogenesis of DSS.

Prevalence and characteristics of dengue and chikungunya infections among acute febrile patients in Nong Khai Province, Thailand

We conducted a cross sectional study at three hospitals of Nong Khai Province, Thailand to determine the prevalence and characteristics of dengue and chikungunya infection among patients who sought care. The study population was acute febrile patients who visited these hospitals during 1 August -31 October, 2010 who were aged 2-60 years and had clinical symptoms compatible with the case definition. Dengue and chikungunya cases were confirmed by an ELISA IgM titer or RT-PCR. We also reviewed surveillance data of dengue and chikungunya infections from 2003-2009. Of the 200 participants recruited into the study, 103 patients (51.5%) were confirmed to have acute dengue infection; dengue serotype 2 was the most prevalence serotype. The ages of confirmed dengue cases ranged from 2-37 years old. The distribution of cases showed that dengue morbidity tended to be clustered in adjacent areas, particularly in Mueang District. Only a small proportion of the patients uses mosquito repellant and had screens on their windows. One patient (0.5%) had laboratory confirmed chikungunya infection. She was from Rattanawapi District, an area where no chikungunya had been reported before. Since the disease varies by age and geographic location, increased awareness of health care workers and public health officers about the diseases in the area is needed for early detection of cases and to promote early prevention and control measures.

Chikungunya virus induces a more moderate cytopathic effect in mosquito cells than in mammalian cells

OBJECTIVES

Chikungunya virus (CHIKV) is an alphavirus belonging to the Togaviridae family. Alphaviruses cause a chronic non-cytopathic infection in mosquito cells, while they develop a highly cytopathic infection in cells originating from various vertebrates. In this study, we compared the cytopathic effect (CPE) induced by CHIKV in Vero cells and a mosquito cell line, C6/36 cells.

METHODS

CPE and the virus titers were compared between the CHIKV-infected C6/36 and Vero cells. Apoptosis was measured by TUNEL assay, and the differences between the C6/36 and Vero cells were compared.

RESULTS

CHIKV infection induced strong CPE and apoptosis in the Vero cells, but light CPE in the C6/36 cells. The virus titers produced in the C6/36 cells were much higher than those produced in the Vero cells.

CONCLUSIONS

The reason CHIKV induced strong CPE is that this virus triggers strong apoptosis in Vero cells compared with C6/36 cells. CHIKV established a persistent infection in C6/36 cells after being passaged 20 times. CHIKV infection in mosquito cells was distinct from that in Vero cells. The cell and species specificity of CHIKV-induced cell death implies that the cellular and viral regulators involved in apoptosis may play an important role in determining the outcome of CHIKV infection.

Poly (I:C), an agonist of toll-like receptor-3, inhibits replication of the Chikungunya virus in BEAS-2B cells

Background

Double-stranded RNA (dsRNA) and its mimic, polyinosinic acid: polycytidylic acid [Poly (I:C)], are recognized by toll-like receptor 3 (TLR3) and induce interferon (IFN)-β in many cell types. Poly (I:C) is the most potent IFN inducer. In in vivo mouse studies, intraperitoneal injection of Poly (I:C) elicited IFN-α/β production and natural killer (NK) cells activation. The TLR3 pathway is suggested to contribute to innate immune responses against many viruses, including influenza virus, respiratory syncytial virus, herpes simplex virus 2, and murine cytomegalovirus. In Chikungunya virus (CHIKV) infection, the viruses are cleared within 7–10 days postinfection before adaptive immune responses emerge. The innate immune response is important for CHIKV clearance.

Results

The effects of Poly (I:C) on the replication of CHIKV in human bronchial epithelial cells, BEAS-2B, were studied. Poly (I:C) suppressed cytopathic effects (CPE) induced by CHIKV infection in BEAS-2B cells in the presence of Poly (I:C) and inhibited the replication of CHIKV in the cells. The virus titers of Poly (I:C)-treated cells were much lower compared with those of untreated cells. CHIKV infection and Poly (I:C) treatment of BEAS-2B cells induced the production of IFN-β and increased the expression of anti-viral genes, including IFN-α, IFN-β, MxA, and OAS. Both Poly (I:C) and CHIKV infection upregulate the expression of TLR3 in BEAS-2B cells.

Conclusions

CHIKV is sensitive to innate immune response induced by Poly (I:C). The inhibition of CHIKV replication by Poly (I:C) may be through the induction of TLR3, which triggers the production of IFNs and other anti-viral genes. The innate immune response is important to clear CHIKV in infected cells.

Protection from arthritis and myositis in a mouse model of acute chikungunya virus disease by bindarit, an inhibitor of monocyte chemotactic protein-1 synthesis

Chikungunya virus (CHIKV) is associated with outbreaks of infectious rheumatic disease in humans. Using a mouse model of CHIKV arthritis and myositis, we show that tumor necrosis factor-α, interferon-γ, and monocyte chemotactic protein 1 (MCP-1) were dramatically induced in tissues from infected mice. The same factors were detected in the serum of patients with CHIKV-induced polyarthralgia and polyarthritis, with MCP-1 levels being particularly elevated. Bindarit (MCP inhibitor) treatment ameliorated CHIKV disease in mice. Histological analysis of muscle and joint tissues showed a reduction in inflammatory infiltrate in infected mice treated with bindarit. These results suggest that bindarit may be useful in treating CHIKV-induced arthritides in humans.

Japanese encephalitis virus remains an important cause of encephalitis in Thailand

BACKGROUND

Japanese encephalitis virus (JEV) is endemic in Thailand and prevention strategies include vaccination, vector control, and health education.

METHODS

Between July 2003 and August 2005, we conducted hospital-based surveillance for encephalitis at seven hospitals in Bangkok and Hat Yai. Serum and cerebrospinal (CSF) specimens were tested for evidence of recent JEV infection by immunoglobulin M (IgM) enzyme-linked immunosorbent assay (ELISA) and a plaque reduction neutralization test (PRNT).

RESULTS

Of the 147 patients enrolled and tested, 24 (16%) had evidence of acute flavivirus infection: 22 (15%) with JEV and two (1%) with dengue virus. Of the 22 Japanese encephalitis (JE) cases, 10 (46%) were aged ≤ 15 years. The median length of hospital stay was 13 days; one 13-year-old child died. Ten percent of encephalitis patients enrolled in Bangkok hospitals were found to have JEV infection compared to 28% of patients enrolled in hospitals in southern Thailand (p < 0.01). Four (40%) of the 10 children with JE were reported as being vaccinated.

CONCLUSIONS

JEV remains an important cause of encephalitis among hospitalized patients in Thailand. The high proportion of JE among encephalitis cases is concerning and additional public health prevention efforts or expanded vaccination may be needed.

Highly efficient rescue of dengue virus using a co-culture system with mosquito/mammalian cells

The production rate of dengue viruses (DENVs), especially low-passage virus isolates, is low, and, therefore, the isolates are generally used only after several passages. However, in vitro passages could induce mutation(s). In this study, we established a system for the characterization of low-passage viral isolates using an infectious cDNA clone. We used R05-624, a plaque derived from type 2 (DENV-2) Thai strain, for the construction of the cDNA clone, named pmMW/R05-624. We found that transfection of both of mammalian Vero cells and mosquito C6/36 cells with viral RNA derived from the cDNA clone produced a significant amount of progeny virus: 3.2x10(6) focus-forming units (FFU) production per ml of cultured fluid only 3days after transfection with 2 microg RNA. Conversely, no detectable level of viruses was produced by conventional methods using a single cell line, Vero or C6/36. When this system was applied for the characterization of eight low-passage clinical viral isolates by placing their 5'-half or 3'-half in the above cDNA clone, we found that all the isolates, except for L04-225, produced similar levels of progeny virus. Among a total of eight cDNA clones reconstructed with the NS4A-3'NCR region derived from L04-225, one clone carried an insertion and produced a low level of progeny virus. Thus, our system to efficiently rescue clinical samples or low-passage viral isolates could be useful for assessing the virological and molecular characteristics of DENV that could be related to disease pathogenesis.

Outbreak of chikungunya fever in Thailand and virus detection in field population of vector mosquitoes, Aedes aegypti (L.) and Aedes albopictus Skuse (Diptera: Culicidae)

We investigated chikungunya fever outbreak in the southern part of Thailand. Human plasma specimens obtained from suspected patients and adult wild-caught mosquitoes were detected for chikungunya virus employing reverse transcriptase-polymerase chain reaction technique. Chikungunya virus was detected in about half of the blood specimens whereas a range of 5.5 to 100% relative infection rate was found in both sexes of the vector mosquitoes, Aedes aegypti (L.) and Ae. albopictus Skuse. The infection rate in Ae. albopictus was higher than in Ae. aegypti, with relative infection rate in male of both species being higher than in female. The appearance of chikungunya virus in adult male mosquitoes of both species reveals a role of transovarial transmission of the virus in field population of the mosquito vectors. These findings have provided further understanding of the relationship among mosquito vectors, chikungunya virus and epidemiology of chikungunya fever in Thailand.

Dengue virus cross-reactive hemagglutination inhibition antibody responses in patients with primary dengue virus infection

Acute and convalescent plasma samples were obtained from 101 confirmed primary dengue cases: 48 cases infected with dengue virus type 1, 10 cases with type 2, 42 cases with type 3 and one case with type 4. The hemagglutination inhibition (HI) titers of individual samples were at levels similar to each of the 4 dengue viruses at both the acute and convalescent stages, irrespective of the dengue virus that infected the patients. The results indicate that HI antibodies to dengue viruses are cross-reactive. When an HI test is used as a diagnostic test for dengue virus infection, the cross-reactive nature needs to be considered when interpreting the results.

Secreted complement regulatory protein clusterin interacts with dengue virus nonstructural protein 1

Vascular leakage and shock are the major causes of death in patients with dengue hemorrhagic fever (DHF) and dengue shock syndrome (DSS). It has been suggested that patients with an elevated level of the free soluble form of dengue virus (DV) nonstructural protein 1 (sNS1) are at risk of developing DHF. To understand the role of sNS1 in blood, we searched for the host molecule with which NS1 interacts in human plasma by affinity purification using a GST-fused NS1. Complement inhibitory factor clusterin (Clu), which naturally inhibits the formation of terminal complement complex (TCC), was identified by mass spectrometry. A recombinant sNS1 produced from 293T cells and sNS1 from DV-infected Vero cells interacted with human Clu. Since an activated complement system reportedly causes vascular leakage, the interaction between sNS1 and Clu may contribute to the progression of DHF.

Double infection of heteroserotypes of dengue viruses in field populations of Aedes aegypti and Aedes albopictus (Diptera: Culicidae) and serological features of dengue viruses found in patients in southern Thailand

In order to understand more about the epidemiology of DHF, a study of the type of dengue viruses and vectors under natural conditions was carried out. Mosquito vectors in the field and the serum of DHF patients in southern Thailand were examined. The two mosquito species are abundant and DHF incidence remains high in this region. Dengue viruses were examined in field-caught mosquitoes by RT-PCR technique. The mosquitoes were caught in 4 provinces: Krabi, Phuket, Phang-Nga and Surat Thani during the late dry season until the early rainy season in 2005. Three dengue serotypes (DEN-2, DEN-3, DEN-4) were detected in Ae. aegypti males and females, and 2 (DEN-2, DEN-3) were detected in Ae. albopictus females. Double infection with 2 serotypes of dengue viruses (DEN-2 and DEN-3) were detected in Ae. aegypti males and females and Ae. albopictus females. DEN-2 and DEN-1 were the most prevalent serotypes found in the serum of the patients in this area, followed by DEN-4 and DEN-3. The prevalence of the predominant dengue serotype varied from province to province. Detection of viruses in adult male mosquitoes reveals the role of transovarial transmission of dengue viruses in field populations of DHF vectors and elucidates circulation of dengue viruses in vectors in the natural environment of endemic areas. The incidence of multiple serotypes of dengue virus in Ae. aegypti and Ae. albopictus in the same area points toward a high risk for an epidemic of DHF. These findings provide greater understanding of the relationship among mosquito vectors, virus transmission and DHF epidemiology in endemic areas.

Specific IgM and IgG responses in primary and secondary dengue virus infections determined by enzyme-linked immunosorbent assay

IgM- and IgG-capture ELISAs are widely used as diagnostic tests for confirmation of dengue virus infection. The positive rate of anti-dengue IgM and IgG detection was examined in primary and secondary dengue virus infections in the setting of a provincial hospital using IgM- and IgG-capture ELISAs. Disease day 1 was defined as the day of onset of symptoms. In total, 232 plasma samples were collected from 106 confirmed dengue cases consisting of 12 primary and 94 secondary infections. In primary infection, anti-dengue IgM was detected in 4 out of 5 samples collected on disease day 5 and in all the 21 samples collected on disease day 6 or later. Specific IgG was detected in 2 out of 5 samples collected on day 12, and in 5 out of 6 samples collected on disease days 13-15, but was not detected in samples collected on disease day 10 or earlier. In secondary infection, IgM was not detected in the samples on disease days 2 and 3, but detected in 20 out of 79 samples collected on days 4-6, in 44 out of 65 on disease days 7-11 and in 40 out of 51 samples on disease days 12-14. In contrast, specific IgG was detected in 21 out of 60 samples on disease days 4 and 5, in 13 out of 19 on disease day 6, in 62 out of 65 on disease days 7-11 and in all the samples collected on disease day 12 or later. The result indicate that seroconversion rates of IgM and IgG are different between primary and secondary infections, and suggest that detection of specific IgM and IgG is necessary for determining dengue virus infection and for differentiating primary and secondary dengue infections.

Detection of Japanese encephalitis (JE) virus-specific IgM in cerebrospinal fluid and serum samples from JE patients

Detection of Japanese encephalitis virus (JEV)-specific IgM by IgM-capture enzymed-linked immunosorbent assay (IgM-capture ELISA) has been accepted as the standard for serological diagnosis. In the present study, we analyzed the time course of the positive rate of JEV-specific IgM in serum and cerebrospinal fluid (CSF) specimens from confirmed JE patients. Serum and CSF samples were obtained from 155 JE cases for diagnostic purposes at hospitals in Thailand from 2002 to 2004. The levels of specific IgM were assessed by IgM-capture ELISA in the 171 serum and 156 CSF samples. Anti-JEV IgM was detected in 26 of 44 serum samples collected on days 1-4 of the disease period, in 31 of 44 samples collected on days 5-8, in 23 of 26 samples collected on days 9-12, and in all the samples collected on day 13 or later. Specific IgM was detected in 60 of 66 CSF samples collected on days 1-4 of illness, and in all the CSF samples but one collected on day 7 or later. The results indicate that the detection of JEV-specific IgM in CSF by IgM-capture ELISA is a reliable laboratory diagnostic method for confirmation of JE throughout the disease period, while the detection of IgM in serum samples is a reliable method on day 9 or later.

Serological and virological features of dengue fever and dengue haemorrhagic fever in Thailand from 1999 to 2002

Serological and virological features of dengue fever (DF) and dengue haemorrhagic fever (DHF) in Thailand were analysed in 2715 patients from 1999 to 2002. The illness was caused by DEN-1 in 45%, DEN-2 in 32%, DEN-3 in 18% and DEN-4 in 5% of patients. Almost all of the DHF cases caused by DEN-2 and DEN-4 were in secondary infection, while approximately 20% of the DHF cases caused by DEN-1 and DEN-3 were in primary infection. Male:female ratio and age distribution were not different among four serotypes in primary and secondary infections. These results indicate that DEN-1 and DEN-3 induce DHF in both primary and secondary infections, and suggest that DEN-2 and DEN-4 in Thailand are less likely to cause DHF in primary infections.

Analysis of specific IgM responses in secondary dengue virus infections: levels and positive rates in comparison with primary infections

BACKGROUND

Dengue viruses are a serious cause of illness in tropical and subtropical areas of the world. Laboratory diagnosis is essential for confirmation of dengue virus infections. Detection of specific IgM by IgM-capture enzymed-linked immunoassay (ELISA) has been widely used as a main serological diagnostic technique.

OBJECTIVES

The levels of specific IgM in secondary dengue virus infections were compared with those in primary infections.

STUDY DESIGN

A total of 1780 samples collected from 924 confirmed dengue cases were tested for anti-dengue IgM by IgM-capture ELISA.

RESULTS AND CONCLUSIONS

Specific IgM was detected in all the cases with primary dengue virus infection on disease day 9 or later. However, specific IgM cannot be detected in 28% (204/716) of the cases in secondary infections. The average titers of IgM were higher in primary infections than in secondary infections. The results confirmed that IgM detection is a reliable serological diagnostic test in primary dengue virus infections. Although IgM detection is also a useful test, other serological diagnostic tests or tests for dengue virus detection are necessary for confirmation of all the secondary dengue virus infections.

Evaluation of RT-PCR as a tool for diagnosis of secondary dengue virus infection

Dengue fever and dengue hemorrhagic fever are serious illnesses in many tropical and subtropical countries. Laboratory tests are essential for the confirmation of dengue virus infection. In the present study, we examined the reliability of reverse transcriptase polymerase chain reaction (RT-PCR) in the laboratory diagnosis of dengue, especially in secondary dengue virus infections. We defined the day when fever subsided as fever day 0. In primary dengue virus infection, the dengue viral genome was detected in all of the 7 samples which were collected on fever day -1 or earlier, in 3 of 4 samples on fever day 0, and in 1 of 2 samples on fever day 1. None of the samples collected on fever day 2 or later were positive by RT-PCR. In secondary dengue virus infection, the dengue viral genome was detected in all of the 28 samples which were collected on fever day -2 or earlier, in 25 of 26 on fever day -1, in 29 of 34 on fever day 0, and in 5 of 10 on fever days 1-2. None of the samples collected on fever day 3 or later were positive. Virus isolation and direct titration were attempted using the plasma samples. When the data of secondary infection cases were analyzed based on fever day, dengue viruses were isolated from all of the 5 samples which were collected on fever day -2 or earlier, in 5 of 13 samples on fever day -1, and in 4 of 22 on fever day 0, but were not isolated from any of the 4 samples collected on fever days 1-2. Viruses were directly detected in 7 of 11 samples on fever day -2 or earlier, in 4 of 13 on fever day -1, and in 1 of 16 on fever day 0. These results indicate that RT-PCR is more sensitive than virus isolation and direct virus titration for determining secondary dengue virus infection. The results also suggest that RT-PCR is a useful diagnostic test for confirmation of dengue virus infection in secondary infection as well as in primary infection, especially when plasma samples are collected before the fever subsides.