Silent transmission of the dengue virus in southern Taiwan

Copyright C 1996 by The American Society of Tropical Medicine and Hygiene

SILENT TRANSMISSION

OF THE DENGUE

VIRUS IN SOUTHERN

TAIWAN

WEI-JUNE CHEN, SU-LIH CHEN, LI-JUNG CHIEN, CHI-CHEN CHEN,

CHWAN-CHUEN KING, MING-RONG HARN, KAO-PIN HWANG, @DJIH-HUA FANG

Department of Parasitology, Chang Gung College of Medicine and Technology, Kwei-San, Tao Yuan, Taiwan

Abstract. Immunoglobulin M (1gM) antibody to dengue virus was examined from a total of 3,099 serum samples

collected in southern Taiwan. Of 1,232 sera collected from a junior high school and four elementary schools in Liu Chiu, 35 were 1gM-positive, demonstrating that the dengue virus has been circulating on the island, despite the fact that no epidemic has been reported in the past 10 years. Sixteen of 925 sera collected from three elementary schools

in Thng-Kang in 1991 were found to be 1gM-positive and two of 192 sera from adults in the local community were

positive. The 1gM-positive subjects tended to be aggregated around a port. Fishing boats that had stopped in neigh boring endemic countries were presumed to have introduced the virus periodically, causing a low level of inapparent infections. In the Kaohsiung area, two of 108 suspected clinical cases and four of 642 community-based sera were 1gM-positive. Rapid urbanization has provided appropriate circumstances for vector breeding in this area and the high population density has also increased contact frequency between humans and mosquito vectors. This has, in turn, increased the possibility of silent transmission of the dengue virus via either intermittent reintroduction of the virus or continuation of inapparent infections or both. Establishment of a early warning system using the 1gM antibody capture—enzyme-linked immunosorbent assay is suggested for effective monitoring of the disease.

Dengue fever has been seen as an important public health problem in southeast Asia and Latin 2 In Taiwan, no outbreak was reported after 1945 until an epidemic oc

curred on Liu-Chiu island in 198 1, resulting in 80% of the

population becoming 34 Another major epidemic,

primarily caused by dengue 1 virus, occurred from late 1987

through 1988 in southern Taiwan.5 Since then, sporadic cases have occasionally been 67 Recently, traffic between Taiwan and neighboring countries, either by sea or by air, has been quite busy. In addition, there have been more than

100,000 non-Taiwanese southeast Asians working annually

in Taiwan since 1992, increasing the possibility of dengue being introduced.

Dengue fever is transmitted by Aedes mosquitoes among susceptible human hosts.8 In southern Taiwan the Breteau index has remained > 5,9 making it difficult to avoid an outbreak should the virus be introduced. Moreover, silent

transmission of dengue infections has recently been recog

nized. For the purpose of this study, silent transmission re fers to inapparent or asymptomatic with dengue virus. These infections are common in Taiwan.'° Therefore, a long-termed active surveillance has become necessary in any place at risk.

For infected patients, dengue virus immunoglobulin M (1gM) antibodies usually appear earlier than IgG antibodies and do not decrease until 30—60 days te'2 Thus, an assay detecting 1gM antibody was recently used to confirm flaviviral infections such as Japanese encephalitis (JE) vi

ms'3 and dengue virus,'4 for which 1gM antibody detection

has become a useful technique for gn'6 and epide miologic surveys.'7 Recently, an enzyme-linked immunosor bent assay (ELISA) has been used extensively for the detec tion of 1gM antibody due to the rapidity, high sensitivity, and high specificity of this technique.'4 As a result, we have applied the 1gM antibody-capture ELISA (MAC-ELISA) to investigate whether or not transmission of dengue virus is continuing in southern Taiwan. Ideally, the MAC-ELISA can also be a good warning system to provide information that

will help prevent an outbreak from spreading.'8

MATERIALS AND METHODS

Collection of serum specimens. This study covered two

cities, Kaohsiung and Ping-Thng. Three districts of Kao hsiung, Feng-San, Kang-San, and San-Mm, and two in Ping Tung, Liu-Chiu and Tung-Kang, were included. All study sites are located in the southern part of Taiwan. Of them, Liu-Chiu is a small island about 15 km from Tung-Kang.

A total of 642 serum specimens were collected from res idents in July—September 1992, from the Kaohsiung area, including 203 in Kang-San, 234 in Feng-San, and 205 in

San-Mm. In addition, clinical samples (n = 108) were col

lected from suspected cases in Ham's Clinic and Kaohsiung Medical College Hospital in 1991—1993 (January—July).

The survey in Liu-Chiu was based on 1,232 serum sam ples collected from four elementary schools and one junior high school in January—February 1990. Liu-Chiu island was geographically divided into eight villages. Serum samples of

Tung-Kang were collected from three elementary schools. A

total of 925 blood samples were taken from school children in January—February 1991, and another 192 sera were col lected from the community in July—September 1992.

Cell culture and antigen preparation. The C6/36 cells

used in the present study were maintained in Eagle's mini

mal essential medium supplemented with 10% fetal bovine serum at 28°C.Four prototype dengue viruses (Hawaii, New Guinea C, H87, and H24l), which were obtained from the

Centers for Disease Control and Prevention San Juan Lab

oratory in Puerto Rico, were used for antigen production. To prepare antigen, confluent cells were inoculated with viruses at a multiplicity of infection (number of infectious virus par tides/number of cells inoculated) of 0. 1—1. Inoculated cells were then harvested to verify the infection using a direct immunofluorescence antibody test after an incubation period of seven days. The cells were subsequently frozen and thawed several times and centrifuged at 3,000 rpm for 15 mm at 4°C.Discarded cell debris and the supernatant, which would be used as the antigen, was stored at —70°Cuntil use. 12

antibody was determined to be the HI antibody titer. Each run had two positive and two negative controls.

RESULTS

Seroepidemiology of dengue virus infection in Liu

Chiu. The geographic distribution of eight villages on the island and their positive rates is shown in Figure 1. The MAC-ELISA data showed that 35 (2.80%) of 1,232 blood samples were 1gM-positive (Table 1). Statistical analysis did not show a significant correlation between the 1gM positive rate and geographic distribution (P > 0.05, by chi-square test), and their prevalence odds ratios were inconsistent (Ta ble 1).

Analysis of age distribution of 1gM-positive individuals showed that the highest prevalence odds ratio was only 2.05 among the 15-year olds and the lowest was among the 13-year olds. We did not see the correlation of age with the prevalence rate (P > 0.05, by chi-square test) probably be cause all sera were collected from persons with a very nar row age range (8—15years of age).

The HI antibody test with 35 1gM-positive individuals showed that 20 had titers < 1:10 (57.14%), with the others having HI titers ranging from 1: 10 to > 1:80 (Table 2).

Seroepidemiology in Tung-Kang. Sixteen of 925 sera (1.73%) collected in 1991 from three elementary schools in Tung-Kang were 1gM-positive (Table 3). The prevalence rates among three schools were significantly different (P < 0.01, by chi-square test). However, only two of another 192 community-based serum specimens were 1gM positive (Ta ble 3).

Seroepidemiology in the Kaohsiung area. Fifty-seven

blood samples were collected from suspected cases in Ham's Clinic and at the Kaohsiung Medical College Hospital. Those collected in 1991 were all negative while one of 34 sera collected in 1992 and one of 17 collected in 1993 (Jan uary—July) were positive.

A total of 642 serum samples were collected from three districts in the Kaohsiung area, i.e., Kang-San, Feng-San, and San-Mm. Four (0.62%) were determined to be 1gM positive (Table 4). Statistical analysis showed that location was not sigificantly correlated with the prevalence rate in this area (P > 0.05, by chi-square test).

DISCUSSION

In the present study, 1gM antibody was detected in both clinical and nonclinical community sera, indicating that the MAC-ELISA can be a useful tool for antibody screen

12, 19, 22-25 The technique has also been proven to be su

perior to the HI test not only in sensitivity but also in spec

4 It was even more useful with outpatients in the com

munity because it does not require a second blood specimen as with the HI test. This versatile serologic technique has been extensively used in detecting dengue antibodies both

in epidemic and nonepidemic areas.26

In this study, more than half of the 1gM positive samples were HI-negative. Because most dengue cases in Taiwan are primary infections, the HI antibody (IgG) titer of some Se rum samples seems to have been too low to be detected by The antigen for the assay was prepared from the cell culture

fluid, and its final dilution was based on the virus titer in the supernatant. The working antigen preparation was actually a mixture of the four prototypes of the virus.

Immunoglobulin M capture ELISA. The MAC-ELISA

for 1gM detection followed the method described by Kuno and

@ Briefly, Linbro microplates (Flow Laboratories, McLean, VA) were coated with 100 pA of goat anti-human 1gM diluted to 1:200 with 0. 1 M carbonate buffer. The coat ed plates were stored at 4°Cuntil use. Before testing the specimen, the plates were washed with phosphate-buffered saline (PBS, pH 7.4). Four percent bovine serum albumin (BSA) was used for the blocking reaction. The plates were incubated for 15 mm at 37°C and washed five times with PBS. Then 0.05 ml of the serum specimen diluted to 1:10

in 0.05% BSA-PBS was added and the plates were incubated

at room temperature for 2 hr. They were then washed again

and 0.5 ml the prepared antigen was added. After overnight incubation at 4°C,they were washed again and 0.025 ml of horseradish peroxidase—conjugated 6B6C- 1 monoclonal an tibody (kindly provided by Dr. D. J. Gubler, Centers for Dis ease Control and Prevention, Fort Collins, CO) diluted 1: 6,000 in 20% normal human serum was added to each well. The plates were then incubated at 37°C for 1 hr, washed again, and 0.1 ml of ABTS (Kirkegaard and Perry Labora tories, Inc, Gaithersburg, VA) substrate solution was added to each well. After incubation at 37°Cfor 30 mm and stand ing at room temperature for 2 hr. the optical density (OD) value was read with the Anthos (Salzburg, Austria) htll ELI SA reader at a wavelength of 405 nmn.

Evaluation of ELISA reults. For quality control, one positive control and three negative controls were always in cluded in each plate. The positive designation was made when the OD value of the sample was higher than the av erage of three negative controls plus three standard errors. All data related to 1gM prevalence rates were analyzed using the chi-square test.

Direct immunofluorescence antibody (DFA) test. The

DFA technique followed the method described by Kuberski and Rosen.2° Briefly, harvested cells were smeared on a te flon-coated slide. After fixation in cold acetone, a drop of fluorescein isothiocyanate—conjugated antibody against the dengue virus from human hyperimmune serum was added to the cells. The smears were then incubated at 37°Cfor 30 mm in a humid chamber, washed in PBS for 10 mm, and mounted with a mixture of glycerol and PBS (1 :9). They were then examined under a fluorescence microscope (Op tiphot-2; Nikon, Tokyo, Japan). Positive and negative con trols were included in each experiment.

Hemagglutination inhibition (HI) test. This test fol lowed the procedure described by Clarke and Casals.2' Brief ly, 0.025 ml of acetone-extracted serum was added to each well of a 96-well U-shaped plate. Two-fold serial dilutions were made with 0.4% bovine albumin in borate saline, and 0.025 ml of antigen (8 hemagglutination units) (acetone-ex tracted from mouse brain) was added to each well, mixed thoroughly, then incubated at 4°C for 18—21 hr. Finally, 0.025 ml of 0.33% red blood cell suspension was added and the mixture incubated at 37°Cfor 1 hr. The result was con sidered positive when all red blood cells precipitated on the bottom of the well. The end point dilution of the positive

villageNo. positive/no. examined (%)Prevalence odds ratiotNan-Fu3/170 (1 .76)2.38Ta-Fu5/169 (2.96)4.00Yu-Fu7/101 (6.93)9.36Chung-Fu2/108(1.85)2.50Pen-Fu6/262 (2.29)3.09Shang-Fu3/173(1.73)2.34San-Fu3/1 13 (2.66)3.59Tien-Fu1/136 (0.74)1 HI titerNo. 1gM positive (%)<1:1020 (57.14)1:102(5.71)1:204(11.43)1 :408 (22.86)>1:801 (2.86)Total35 (100)

14

CHEN AND OTHERS

San-Fu

(31113: 2.66%)

port

FIGuan 1. Map and 1gM antibody distribution among eight villages in Liu-Chiu.

@ = Kaohsiung; 0 = Thng-Kang; @‘Liu-Chiu.

Ta-Fu

(5/169;296%)

port

the HI test at the time when 1gM antibody appears in the

blood.

In southern Taiwan, dengue fever has appeared intermit tently since the 1987—1988 epidemic. Two outbreaks causing

149 and 220 confirmed indigenous cases were reported in 1991 and 1994, respectively.27 Inapparent or subclinical in

T@ut@n1

Geographic distribution of 1gM antibody in Liu-Chiu, Taiwan, 1990*

fection seems to be more common than was

@ The

risk of emerging dengue hemorrhagic fever/dengue shock syndrome (DHFIDSS) prompted us to conduct this investi gation.

The continued transmission of dengue virus was observed in Liu-Chiu. Although prevalence rates among eight villages did not show a significant difference, the prevalence-odds ratio in Yu-Fu was as high as 9.36 when compared with

TABLE 2

Relationships between hemagglutination inhibition (HI) antibody ti ters and 1gM-positive samples among sera collected from Liu Chiu, Taiwan, 1990

ax2 _ ll.35,dcgreesoffreedom 7,P>0.05.

SchoolNo. positive examined (%)Thng-Kang3/316(0.95)Hai-Pin7/144(4.86)Thng-Long6/465

(1.29)a

@2 13.98, degrees of freedom = 2, P < 0.01.

DistrictNo. positive/no. examined

(%)San-Mm2/205 (0.98)Kang-San1/203 (0.49)Feng-San1/234 (0.43)a @2= @ degrees of freedom —2, P > 0.05. Ti@nu@3

1gM antibody distribution among three elementary schools in Tung Kang, Taiwan, 1991*

TABLE 4

1gM antibody distribution of community-based sera collected from

three districts in the Kaoshiung area of Taiwan, 1992*

Tien-Fu, in which the prevalence rate was the lowest, sug gesting that the dengue virus was relatively active in Yu-Fu. Villages located in areas close to ports have usually shown higher prevalence rates. Fishing boats were believed to be associated with introduction of the virus from neighboring

endemic countries. In addition, crowded living conditions

was also considered a risk factor for a higher rate of dengue infections.28 Thus, it was not surprising that Yu-Fu had a

higher prevalence rate since this village was the most crowd

ed locality on the island. The virus may subsequently spread to adjacent agricultural villages. However, a vector control campaign has been conducted on the island by health au thorities since 1987,29 and eventually limited the spread of the virus. Illegal fishermen, mostly from adjacent countries, may in part account for the introduction of dengue virus. It was noted that Chung-Fu had a lower 1gM antibody preva lence rate even though it is located adjacent to Yu-Fu village, but it had a higher HI antibody titer rate (Chien U, National Yang-Ming Medical college, unpublished data).

The Tung-Kang survey revealed that the 1gM prevalence rate is higher in school children than in adults from the com munity and was presumably associated with the high HI an tibody rate in the adult population (Chien Ii, National Yang Ming Medical College, unpublished data). During the inter view, many residents reported that they had had dengue fe ver in the previous years. Among three elementary schools in Thng-Kang, the school in Hai-Pin had the highest preva lence rate, probably because it was located in the proximity of the fishing port and most children from this school were from families with fishing as an occupation. The virus may have been introduced in the same way as that in Liu-Chiu.

The low 1gM antibody prevalence rate in the Kaohsiung area could be the result of a campaign on vector control after the outbreak occurring in 1991 . However, it was not possible to exclude the possibility of persistent virus transmission even though prevalence rate was so low. Three districts in the Kaohsiung area showed no difference in their prevalence rates for IgM antibody, probably because all environments

in theses areas were similar.

The existence of 1gM antibody in all five areas has dem onstrated that the dengue virus may have established a means of silent transmission in southern Taiwan. Asympto

matic infections or symtomatic infections that are not diag

nosed as dengue fever must have been occurring in these areas. The occurrence of intermittent outbreaks such as those in 1991 and 1994 are not surprising. The most important concern is the emergence of DHF/DSS, as has been reported in Sri Larka, India, New Caledonia, and Tahiti.@°A con firmed fatal case of DHF occurred in l994@' This was the only case of DHF in Taiwan thus far. Therefore, it would be

worthwhile to establish a surveillance system for early de

tection as an indicator of epidemic potential. Active screen ing using the MAC-ELISA is suggested to be the best tech nique because of its high sensitivity and versatility.

Acknowledgment: We are indebted to Dr. Duane J. Gubler for kindly providing some of the reagents for the assays.

Financial support: This work was supported by a grant from Dc partment of Health, Republic of China (D0H82-TD-059).

Authors' addresses: Wei-June Chen, Su-Lih Chen, and Jih-Hua

Fang, Department of Parasitology, Chang Gung College of Medicine and Technology, Kwei-San, Tao-Yuan 333, Taiwan. Li-Jung Chien, Chi-Chen Chen, and Chwan-Chuen King, School of Public Health, National Taiwan University, Taipei 100, Taiwan. Min-Rong Ham, Ham's Clinic, Kaohsiung 807, Taiwan. Kao-Pin Hwang, Department

of Pediatrics, Kaohsiung Medical College, Kaoshiung 807, Taiwan.

Reprint requests: Wei-June Chen, Department of Parasitology, Chang Gung College of Medicine and Technology, Kwei-San, Tao Yuan 333, Taiwan.

REFERENCES

1. Halstead SB, 1980. Dengue haemorrhagic fever-a public health

problem and a field for research. Bull World Health Organ

58: 1—21.

2. Goh KT, Ng 5K, Chan YC, Lim SJ, Chua EC, 1987. Epide

miological aspects of an outbreak of dengue fever/dengue haemorrhagic fever in Singapore. Southeast Asian J Trop

Med Public Health 18: 291—294.

3. Wu YC, 1986. Epidemic of dengue 2 on Liou-Chyou Shiang, Pingtung, Taiwan, in 1981. Chin J Microbiol Immunol 9: 203—211.

4. Hsieh WC, Chen ME Lin KT, Hsu ST. Ma CI, Wu SS, 1982.

Study of outbreak of dengue fever on Liou-Chyou Shiang, Ping-Thng county in 1981. J Formos Med Assoc 81: 1388—

1395.

5. Ko YC, 1989. Epidemiology of dengue fever in Taiwan. Kao hsiung I Med Sci 5: 1—11.

6. Department of Health, Executive Yuan, ROC, 1987. Prelimi nary report of the dengue investigation in Taiwan area. Epi

demiol Bull 3: 93—99.

7. Department of Health, Executive Yuan, ROC, 1989. Epidemi

ology of dengue in Taiwan area. Epidemiol Bull 5: 1—3.

8. Gubler DJ, 1987. Current research on dengue. Harris KF, ed.

Current Topics in Vector Research. Volume 3. New York:

Spnnger-Verrlag, 37—56.

9. Huang JS, Hsu EL, 1984. Investigations on the distribution and breeding habitats of dengue vectors in Kaohsiung city. Chin

J Entomol 14: 233—244.

10. Malison MD, Waterman SH, 1983. Dengue fever in the United States. JAMA 249: 496—500.

11. Nogueira RMR, Schatzmayr HG, Miagostovich MP, Fatima M,

Fatima DB, Filho JCE 1988. Virological study of a dengue type 1 epidemic at Rio de Janeiro. Mem Inst Oswaldo Cruz

23: 219—225.

12. Innis BL, Nisalak A, Nimmannitya 5, Kusalerdchariya 5, Chongswadi V, Suntayakorn 5, Puttisri P. Koke CH, 1989. An enzyme-linked immunosorbent assay to characterize den gue infection where dengue and Japanese encephalitis co-cir culate. Am J Trop Med Hyg 40: 418—427.

16

CHEN AND OTHERS 13. Burke DS, Nisalak A, 1982. Detection of Japanese virus im

munoglobulin M antibodies in serum by antibody capture ra dioimmunoassay. J Clin Microbiol 15: 353—361.

14. Chen WJ, 1989. Serological and virological diagnosis for den gue infection. Kaohsiung J Med Sci 5: 66—71.

15. Botros BAM, Watts DM, Soliman AK, Salib AW, Moussa MI, Mursal H, Douglas C, Farah M, 1989. Serological evidence of dengue fever among refugees, Hargeysa, Somalia. J Med

Virol 29: 79—81.

16. Chow L, Hsu St 1989. MAC-ELISA for the detection of 1gM

antibodies to dengue type 1 virus. Chin J Microbiol Immunol 22: 278—285.

17. Chen Wi, Fang AH, Liaw YH, Lo FL, Wei YY, Ko YC, Liu

WE Chow HR. King CC, Kao CL, Yang flY, Chuang CH,

Wu YC, Chou MJ, Chen Ci, 1990. Antibody response by

MAC-ELISA to dengue viruses during the epidemic in south

em Taiwan, 1988. Arbovirus Res Aust 5: 35—36.

18. Chow L, Sun HC, Chen HY, Lin SY, Wu JS, 1992. Detection and differentiation of dengue- 1 from Japanese encephalitis vi n's infections by ABC MAC-ELISA. Chin J Microbiol im

munol 25: 172—180.

19. Kuno G, Gomez I, Gubler DJ, 1987. Detecting artificial anti dengue 1gM immune complexes using an enzyme-linked im munosorbent assay. Am J Trop Med Hyg 36: 153—159.

20. Kuberski IT, Rosen L, 1977. A simple technique for the detec

tion of dengue antigen in mosquito by immunofluorescence.

Am J Trop Med Hyg 26: 533—537.

21. Clarke DH, Casals J, 1958. Techniques for hemagglutination and hemagglutination-inhibition with arthropod-borne viruses.

Am J Trop Med Hyg 7: 561—573.

22. Chen Wi, Hwang KR Fang AR, 1991. Detection of 1gM anti bodies from cerebrospinal fluid and sera of dengue fever pa

tients. Southeast Asian J Trop Med Public Health 22: 659—

663.

23. Bundo K, Igarashi A, 1985. Antibody capture ELISA for de tection of immunoglobulin M antibodies in sera from Japa nese encephalitis and dengue hemorrhagic fever patients. J

Virol Methods 11: 15—22.

24. Nogueira RMR, Miagostovich MR Cavalcanti 5MB, Marzochi

KBE Schatzmayr HG, 1992. Levels of 1gM antibodies against

dengue virus in Rio de Janeiro, Brazil. Res Virol 143: 423—

427.

25. Zeller HG, Traore-Lamizana M, Monlum E, Hervy JP, Mondo M, Digouue JP, 1992. Dengue-2 virus isolation from humans

during an epizootic in southeastern Senegal in November,

1990.Res Virol143:101—102.

26. Gubler DJ, 1989. Surveillance for dengue and dengue hemor

rhagic fever. Bull Pan Am Health Organ 23: 397—404.

27. Lee SE Chou HC, Wu PH, Wu YC, Chou L, 1995. Investigating a dengue fever outbreak in Fenshan city of Kaohsiung Coun ty. Epidemiol Bull 11: 49—57.

28. Morens DM, Rigau-Perez JG, Lopez-Correa RH, Morre CG, Ruiz-Tiben EE, Sather GE, Chiriboga J, Eliason DA, Casts Velez A, Woodall IF, 1986. The dengue outbreak investiga tion group: dengue in Puerto Rico, 1977: public health re sponse to characterize and control an epidemic of multiple

serotypes. Am J Trop Med Hyg 35: 197—211.

29. Wang CH, Hwang JS, Lay JR. 1990. Preliminary study on the biological control of dengue vectors by fish in Liou-Chyou prefecture, Ping-fling County, Taiwan. Kaohsiung J Med Sci 6: 382—388.

30. Lam 5K, 1990. Dengue and dengue hemorrhagic fever. South

east Asian J Trop Med Public Health 21: 520—521.

31. Hsu YT, Huang PJ, Chen KH, Su CY, Chuang CH, Hung HY, Fang RN, Lu PH, 1994. Dengue hemorrhagic fever: a case report. Epidemic Rep 10: 259—260.