INTRODUCTION
Dengue viruses (DENVs) are a member of the Flavivirus group of the virus family Flaviviridae.1 The four serotypes of DENV (DENV-1, DENV-2, DENV-3, and DENV-4) are transmitted primarily by Aedes aegypti (Stegomyia aegypti)2 mosquitoes and are the etiologic agents of dengue fever (DF) and dengue hemorrhagic fever (DHF). DF typically presents as a febrile illness with headache, eye pain, myalgia, and arthralgia, which can progress to DHF with plasma leakage that can lead to shock (dengue shock syndrome) and death. There is an increasing incidence of DF and DHF throughout tropical regions. Because the risk of DHF relative to DF is greater with a secondary antibody response pattern versus a primary antibody response, the rise in DHF may reflect greater sequential exposure to more than one virus serotype.3 The primary, but not exclusive, hypothesis for this phenomenon is antibody-dependent enhancement (ADE). Studies in support of ADE include seroepidemiologic studies and prospective studies from Thailand,4,5 Indonesia,6 and Nicaragua and seroepidemiologic and clinical observations in Cuba.7,8
Generally, the greatest risk for DHF is thought to be with the second virus serotype exposure (although this risk may also be affected by the particular serotype and sequence of infection). Monkey studies showed that multivalent neutralizing antibodies may reduce the risk of detectable dengue viremia with heterologous challenge and have suggested that after the second DENV infection enough cross-immunity may be present to prevent significant disease.9–12 The risk for DHF with a third or fourth DENV infection relative to a first or second exposure is not known. To estimate the risk of DHF during repeated DENV infections, we performed an analysis of our dengue diagnostics database for children admitted to the Queen Sirikit Institute of Child Health (formerly the Bangkok Children’s Hospital) and the Kamphaeng Phet Provincial Hospital for suspected dengue.
MATERIALS AND METHODS
The Armed Forces Research Institute of Medical Sciences (AFRIMS) routinely performs dengue diagnostic testing on serum samples from children with suspected dengue for the provincial hospital in Kamphaeng Phet Province (located in lower northern Thailand) and the Queen Sirikit National Institute of Child Health in Bangkok. For the time interval of January 1994 through February 2005, we screened our database for dengue-positive diagnostic samples by name, birth date, hospital number, admission number, age, sex, and address to identify sequential dengue admissions. Researchers had no access to personal identifiers, and the protocol was approved by ethical review boards at the Thai Ministry of Public Health and the Walter Reed Army Institute of Research.
Clinical diagnoses, if available, were recorded from the final diagnosis on the patient’s hospital chart (DF, DHF, or a specific grade of DHF according to World Health Organization [WHO] criteria).13 For the purposes of analysis, all stages of DHF were categorized simply as DHF. Japanese encephalitis (JE) vaccination status, if available, was recorded from the patients’ history along with the age of the patient when the vaccine was given. Unfortunately, no record of the number of doses received was available.
All acute and convalescent samples were serologically tested using combined dengue/JE enzyme immunoassay (EIA). Serologic results indicating acute primary dengue virus infection or acute secondary dengue virus infection were used in this analysis; all other results (i.e., recent flavivirus infection or acute JE virus infection) were excluded from the analysis. Dengue EIA results were classified as primary if anti-dengue IgM was ≥ 40 units and anti-dengue IgM:anti-IgG ratio was ≥ 1.8. EIA results were classified as secondary if anti-dengue IgM was ≥ 40 units and anti-dengue IgM:anti-IgG ratio was < 1.8.14 For specimens with subdiagnostic levels of IgM and elevated IgG, a hemagglutination inhibition assay was performed and interpreted according to WHO guidelines.13,15,16
The infecting DENV serotype was identified by virus isolation in live Toxorhynchitis splendens mosquitoes or by reverse transcriptase-polymerase chain reaction (RT-PCR). Viruses isolated in mosquitoes were serotyped using an antigen capture EIA.17–19 Molecular identification of DENV serotype in acute serum was performed using RT-PCR modified after Lanciotti and others.20 There was one sample where the RT-PCR and typing EIA were in conflict; the PCR was DENV-3 and the EIA was DENV-2. An admission 335 days later was PCR positive for DENV-2 (EIA was negative), so the first admission was considered to be caused by DENV-3.
χ2 with Yates correction was used for categorical comparisons, except if expected cells were less than five; then, a Fisher exact test was used. A two-sided t test was used to compare means. Statistical comparisons were made using SPSS version 10.1 and Epi Info version 3.3.2.
RESULTS
General descriptive results.
We found 191 patients with two admissions for laboratory-confirmed dengue from a total of 15,825 patients with laboratory-confirmed dengue (15,634 patients had a single admission for dengue). Thus, 191 (1.2%) of 15,825 patients had two dengue admissions identified. No children were admitted with a diagnosis of dengue for a third or fourth time.
One hundred one (52.9%) of the 191 patients with multiple dengue admissions were female. The mean age at the first admission was 6.9 years (range, 0.33–19 years), and the mean age at the second admission was 10.4 years (range, 1.6–22 years). The mean time between admissions was 3.5 years (range, 0.33–8.8 years); limiting this to isolation or RT-PCR–confirmed pairs, the mean interval was 3.5 years (range, 0.59–7.7 years).
There were 125 (65%) repeat dengue admission patients that had complete data regarding JE vaccination and serologic responses at both admissions. Based on this data, we categorized the patients and assigned a plausible DENV infection sequence (Table 1). We found that 36 (29%) of paired admissions were caused by a first and second DENV infection (primary response followed by a secondary response), 13 (10%) could be attributed to a third or fourth infection (those with sequential secondary responses and no history of JE vaccination), and 76 (61%) could be caused by several possibilities (first followed by second, second by third, and third by fourth infection). Thus, at a minimum, 0.08% (13/16,016) of admissions were caused by third or fourth dengue virus infections. Assuming the same fraction of possible third or fourth infections in those without JE vaccination history available, the total caused by third or fourth infections would be 0.12% (20/16,016). At the maximum, 0.6% (89/16016) were caused by third or fourth dengue virus infections. Assuming the same fraction in those without JE vaccination history, it would be 0.8% (136/16,016).
JE vaccination before the first admission correlated with a secondary antibody response pattern (76/98 versus 13/27; OR = 3.7 [1.4, 9.99], P = 0.006). For those with a primary antibody response at the first admission, we failed to show a statistically significant difference in age at first dengue admission between those who were previously JE vaccinated (N = 22) compared with those that were not JE vaccinated (N = 14; 6.1 versus 4.1 years, P = 0.13). This finding carried through the second dengue admission (9.4 versus 7.2 years, P = 0.13). Considering only those patients with secondary antibody response patterns at the time of first admission, there was no difference in age at first admission for those with a history of JE vaccination versus those without such a history (N = 76 and N = 13 and 6.5 and 7.1 years, respectively, P = 0.52).
Evaluations of associations with DHF versus DF.
Table 2 classifies the clinical diagnosis by the serologic response and admission sequence. At the first admission, there were 54 primary and 135 secondary DENV infections; at the second admission, all 191 dengue patients had secondary antibody response patterns.
The risk of DHF relative to DF at the first admission was significantly less for those with a primary antibody response compared with those with a secondary response (18 of 38 versus 72 of 96 DHF cases, OR = 0.30 [0.13, 0.71], P = 0.004). JE vaccination before the first admission was not significantly correlated with a diagnosis of DHF versus DF (48 of 71 versus 15 of 24 DHF cases, OR = 1.25 [0.43, 3.63], P = 0.8) at first admission.
There was no significant difference in age between DHF and DF cases for the first (6.5 versus 6.9 years, P = 0.5) and second (10.4 versus 10.6 years, P = 0.7) admissions. The time between admissions was not different between DHF and DF cases at the second admission (3.6 versus 3.6 years, P = 0.8). A diagnosis of DHF at the first admission did not significantly decrease the likelihood of a diagnosis of DHF with the second admission (55 of 79 versus 26 of 37 DHF cases, OR = 0.79 [0.29, 2.11], P = 0.8).
Among these hospitalized patients, there was no difference in the risk of DHF versus DF between a second dengue virus infection (primary response followed by secondary response) and third or fourth dengue virus infection (secondary response in those without prior JE vaccination followed by secondary response; 18 of 27 versus 9 of 12 DHF cases, OR = 0.67 [0.11, 3.78], P = 0.7).
Based on virus isolation and RT-PCR data, the following sequences led to a diagnosis of DHF: 1 followed by 2, 3, or 4; 2 followed by 1, 3, or 4; 3 followed by 1, 2, or 4; and 4 followed by 2 (Table 3). When we looked at just those pairs where the first admission showed a primary antibody response (excluding the possibility of undocumented first dengue infection), there remained sequences resulting in DHF at the second infection that have not previously been documented in the literature (two patients with a serotype 1 followed by a serotype 4 infection, one patient with a serotype 2 followed by a serotype 3 infection, one patient with a serotype 3 followed by a serotype 4 infection, and two patients with a serotype 3 followed by a serotype 1 infection) in addition to the prior documented sequences of 1–2, 1–3, 2–1, and 3–2.4,6,21–23 No sequence could be statistically shown to be a higher risk for subsequent DHF than other sequences.
DISCUSSION
These data suggest that, for greater than second DENV infections, 1) DHF does occur; 2) the risk for DHF is probably low based on the observation that there were few multiple admissions for dengue illness; and 3) the risk for DHF once presenting to the hospital is similar for second and subsequent DENV infections. In addition, we confirmed that multiple serotype sequences may result in DHF.
An earlier unpublished study looking at data from 1974–1990 found that 53 (0.5%—compared with 1.2% for this analysis) out of 10,446 admissions were second admissions. The rate of second admissions may have increased in our study due to more intense transmission of all four DENV serotypes. The older analysis found that 41% of children with first dengue admissions had primary antibody response patterns compared with 29% for the more recent study. This may reflect more frequent JE vaccination as well as more intense DENV transmission. The mean time between admissions for both studies was similar (3.9 and 3.5 years).
In this study, JE vaccination did not increase the risk of DHF.24,25 Based on our comparison of first DENV infections, JE vaccination was not statistically associated with a delay in the first dengue admission (P = 0.13).
Our finding of decreased risk of DHF compared with DF for primary compared with secondary responses is consistent with other studies.26 The odds ratio for protection from DHF during a primary infection was 0.30 (0.13, 0.71) (Table 2). The degree of protection increased to 0.22 (0.09, 0.57) when infants experiencing a primary infection were removed from the analysis. While the infants experienced first dengue virus infections, the presence of maternal antibody likely increases the risk of severe disease.24,25
Including this report, we believe that all sequences of dengue serotype infection have been shown to lead to DHF except serotype 4 followed by serotype 1 and serotype 4 followed by serotype 3. In addition, this study is the first to virologically (PCR or virus isolation) document the sequences.
Limitations.
The data are limited to hospitalized dengue patients only, clinical diagnoses were not confirmed by chart review, and JE vaccination history was based on parental recall. It is also possible that some cases identified as second DENV infections on the first admission could be caused by naturally acquired JE virus infection rather than vaccination followed by DENV infection.
Despite these significant limitations, the age data of our DENV infection numbers (Table 1) suggest that the data are reliable. Considering those without a history of JE vaccination, there was a significant difference in the ages at first admission between those with a primary and secondary responses (4.1 versus 7.1 years, P = 0.03, reflecting that first DENV infections occur at a younger mean age than second infections), whereas second DENV infections occurred at essentially the same age (7.1 and 7.2 years). Likewise, considering those with a history of JE vaccination, there was a difference between the age at first DENV infection and what we considered to be the second, third, or fourth infections (6.1 and 10.2 years, P < 0.001). In addition, our data reaffirm other studies showing the risk for DHF is greater for secondary versus a primary response.
Conclusions.
Although no confirmed third dengue admissions were found, our results suggest that 0.08% to 0.8% of dengue admissions may be caused by third or fourth infections. Though the risk for a subsequent hospitalization caused by dengue after one hospitalization for dengue seems to be low, once admitted, the risk for DHF versus DF is similar. Finally, we document new dengue serotype infection sequences leading to DHF of 1–4, 2–3, 3–1, and 3–4.
Patients categorized by JE vaccination history and first admission serologic response pattern
| JE vaccination before first admission | First admission serologic response* | Presumed dengue virus infection at first admission (mean age in years) | Presumed dengue virus infection at second admission (mean age in years) | n |
|---|---|---|---|---|
| * All second admission responses were secondary. | ||||
| No | Primary | First (4.2) | Second (7.2) | 14 |
| Secondary | Second or third (7.1) | Third or fourth (11.2) | 13 | |
| Yes | Primary | First (6.1) | Second (9.4) | 22 |
| Secondary | First, second, or third (6.5) | Second, third, or fourth (10.2) | 76 | |
DF and DHF patients by antibody response pattern (number with virus isolation or RT-PCR confirmed dengue for both admissions shown in parentheses.
| Response | DF | DHF | Unknown | OR (CT) |
|---|---|---|---|---|
| * Calculation based on serology-confirmed cases. If the seven children < 1 year of age at the time of first admission are removed from the analysis, the OR with 95% CI is 0.22 (0.09, 0.57). | ||||
| First admission | ||||
| Primary 54 (18) | 20 (11) | 18 (4) | 16 (3) | 0.30 (0.13, 0.71)* |
| Secondary 135 (36) | 24 (9) | 72 (18) | 39 (9) | Reference |
| Unknown 2 (2) | 1 (1) | 1 (1) | ||
| Second admission | ||||
| Secondary 54 (18) | 10 (2) | 28 (12) | 16 (4) | 1.29 (0.53, 3.20) |
| Secondary 135 (36) | 35 (11) | 76 (21) | 24 (4) | Reference |
| Unknown 2 | 1 | 1 | ||
Sequence of DENV infection by serotypes and clinical outcome: dengue fever DF versus DHF
| DENV serotype sequence* | Number (%) | Second diagnosis was DF | Second diagnosis was DHF |
|---|---|---|---|
| * RT-PCR or virus isolation positive pairs. | |||
| † Sequence still documented even if considering only those with where the first admission showed a primary antibody response. | |||
| ‡ Dengue serotype sequences leading to DHF not previously reported. | |||
| § Does not add up to 56 because the clinical diagnosis was not known for all cases. | |||
| 1–2†4,21 | 15 (27) | 4 | 10 |
| 1–3†4,6,22 | 4 (7) | 0 | 1 |
| 1–4†‡ | 7 (13) | 3 | 2 |
| 2–3†‡ | 4 (7) | 1 | 3 |
| 2–46 | 3 (5) | 1 | 2 |
| 2–1†6 | 4 (7) | 0 | 4 |
| 3–4†‡ | 3 (5) | 1 | 2 |
| 3–1†‡ | 7 (13) | 2 | 3 |
| 3–2†23 | 6 (11) | 1 | 5 |
| 4–24,6 | 3 (5) | 1 | 2 |
| Total | 56 (100) | 14§ | 34§ |
Address correspondence to Robert V. Gibbons, APO-AP, Bangkok 96546, Thailand. E-mail: Robert.gibbons@afrims.org
Authors’ addressses: Robert V. Gibbons, USAMC-AFRIMS APO AP, 96566, Telephone: 662-6444674, Fax: 662-6444760, E-mail: robert.gibbons@afrims.org. Siripen Kalanarooj, Queen Sirikit National Institute of Child Health, 420/8 Rajvithi Road, Bangkok, Thailand 10400, Telephone: 662-354-8434, E-mail: ph sirip@health.moph.go.th. Richard G. Jarman, USAMC-AFRIMS APO AP, 96566, Telephone: 662-6444674, Fax: 662-6444760, E-mail: richard.jarman@afrims.org. Ananda Nisaluk, AFRIMS, 315/6 Rajvithi Road, Bangkok, Thailand 10400, Telephone: 662-6444674, Fax: 662-6444760, E-mail: Ananda.nisaluk@afrims.org. David W. Vaughn, GlaxoSmithKline, 2301 Renaissance Boulevard, RN0220, King of Prussia, PA 19406, Telephone: 610-787-3907, Fax: 610-787-7057, E-mail: d.w.vaughn@usa.net. Timothy P. Endy, WRAIR, Division of Communicable Diseases and Immunology, 503 Robert Grant Ave., Silver Spring, MD 90210, Telephone: 301-319-3053, Fax: 301-319-3053, E-mail: endyt@upstate.edu. Mammen P. Mammen Jr, Product Manager, Pharmaceutical Systems Division, US Army Medical Material Development Activity (USAMMDA), 1430 Veterans Drive, Ft. Detrick, MD 21702, Telephone: 301-619-2069, Fax: 301-619-2304, E-mail: mammen.mammen@amedd.army.mil. Anon Srikiatkhachorn, Center for Infectious Disease and Vaccine Research, University of Massachusetts Medical School, 55 Lake Avenue, North Worcester, MA 01557, Telephone: 508-856-4182, Fax: 508-856-4890, E-mail: anon.srikiatkhachorn@umassmed.edu.
Acknowledgments: The authors thank Tipawan Thipwong for outstanding database support, making this research and publication possible. We also acknowledge the assistance of Panor Srisongkram, Napaporn Latthiwongsakorn, Sumitda Narupiti, Vipa Thirawuth, and Butsaya Thaisomboonsuk; without their assistance, this study could not have been completed. Last, we thank all of our collaborators at QSNICH, especially Suchitra Nimmannitya.
Financial support: Funding was partially provided by the US Military Infectious Diseases Research Program, Fort Detrick, MD, and the National Institutes of Health (NIH-P01-AI34533).
Disclosure: The opinions or assertions contained herein are the personal views of the authors and are not to be construed as official or reflecting the views of the Armed Forces Research Institute of Medical Sciences, the United States Department of the Army, or the United States Department of Defense.
REFERENCES
- 1↑
Calisher CH, Mahy BW, 2003. Taxonomy: get it right or leave it alone. Am J Trop Med Hyg 68 :505–506.
- 2↑
Reinert JF, Harbach RE, 2005. Generic and subgeneric status of aedine mosquito species (Diptera: Culicidae: Aedini) occurring in the Australasian Region. Zootaxa 887 :1–10.
- 3↑
Halstead SB, Nimmannitya S, Cohen SN, 1970. Observations related to pathogenesis of dengue hemorrhagic fever. IV. Relation of disease severity to antibody response and virus recovered. Yale J Biol Med 42 :311–328.
- 4↑
Sangkawibha N, Rojanasuphot S, Ahandrik S, Viriyapongse S, Jatanasen S, Salitul V, Phanthumachinda B, Halstead SB, 1984. Risk factors in dengue shock syndrome: a prospective epidemiologic study in Rayong, Thailand. I. The 1980 outbreak. Am J Epidemiol 120 :653–669.
- 5↑
Russell PK, Yuill TM, Nisalak A, Udomsakdi S, Gould DJ, Winter PE, 1968. An insular outbreak of dengue hemorrhagic fever. II. Virologic and serologic studies. Am J Trop Med Hyg 17 :600–608.
- 6↑
Graham RR, Juffrie M, Tan R, Hayes CG, Laksono I, Ma’roef C, Erlin, Sutaryo, Porter KR, Halstead SB, 1999. A prospective seroepidemiologic study on dengue in children four to nine years of age in Yogyakarta, Indonesia I. studies in 1995–1996. Am J Trop Med Hyg 61 :412–419.
- 7↑
Guzmán MG, Kouri G, Valdes L, Bravo J, Alvarez M, Vazques S, Delgado I, Halstead SB, 2000. Epidemiologic studies on Dengue in Santiago de Cuba, 1997. Am J Epidemiol 152 :793–799.
- 8↑
Guzman MG, Kouri G, Valdes L, Bravo J, Alvarez M, Vazques S, Delgado I, Halstead SB, 2000. Dr. Guzman et al. response to Dr. Vaughn. Am J Epidemiol 152 :804.
- 9↑
Halstead SB, Shotwell H, Casals J, 1973. Studies on the pathogenesis of dengue infection in monkeys. I. clinical laboratory responses to primary infection. J Infect Dis 128 :7–14.
- 10
Scherer WF, Breakenridge FA, Dickerman RW, 1972. Cross-protection studies and search for subclinical disease in new world monkeys infected sequentially with different immunologic types of dengue viruses. Am J Epidemiol 95 :67–79.
- 11
Whitehead RH, Chaicumpa V, Olson LC, Russell PK, 1970. Sequential dengue virus infections in the white-handed gibbon (Hylobates lar). Am J Trop Med Hyg 19 :94–102.
- 12↑
Price WH, 1968. Sequential immunization as a vaccination procedure against dengue viruses. Am J Epidemiol 88 :392–397.
- 13↑
Anonymous, 1997. Dengue Haemorrhagic Fever: Diagnosis, Treatment, Prevention and Control. Second edition. Geneva: World Health Organization.
- 14↑
Innis BL, Nisalak A, Nimmannitya S, Kusalerdchariya S, Chong-swadsi V, Suntayakorn S, Puttisri P, Hoke CH, 1989. An enzyme-linked immunosorbent assay to characterize dengue infections where dengue and Japanese encephalitis co-circulate. Am J Trop Med Hyg 40 :418–427.
- 15↑
Clarke DH, Casals J, 1958. Techniques for hemagglutination and hemagglutination inhibition with arthropod-borne viruses. Am J Trop Med Hyg 7 :561–573.
- 16↑
Vaughn DW, Green S, Kalayanarooj S, Innis BL, Nimmannitya S, Suntayakorn S, Rothman AL, Ennis FA, Nisalak A, 1997. Dengue in the early febrile phase: viremia and antibody responses. J Infect Dis 176 :322–330.
- 17↑
Rosen L, Gubler DJ, 1974. The use of mosquitoes to detect and propagate dengue viruses. Am J Trop Med Hyg 23 :1153–1160.
- 18
Kuberski TT, Rosen L, 1977. A simple technique for the detection of dengue antigen in mosquitoes by immunofluorescence. Am J Trop Med Hyg 26 :533–537.
- 19↑
Rosen L, Shroyer DA, 1985. Comparative susceptibility of five species of Toxorhynchites mosquitoes to parenteral infection with dengue and other flaviviruses. Am J Trop Med Hyg 34 :805–809.
- 20↑
Lanciotti RS, Calisher CH, Gubler DJ, Chang GJ, Vorndam AV, 1992. Rapid detection and typing of dengue viruses from clinical samples by using reverse transcriptase-polymerase chain reaction. J Clin Microbiol 30 :545–551.
- 21↑
Guzmán MG, Kouri G, Bravo J, Soler M, Martínez E, 1991. Sequential infection as risk factor for dengue hemorrhagic fever/dengue shock syndrome (DHF/DSS) during the 1981 dengue hemorrhagic Cuban epidemic. Mem Inst Oswaldo Cruz 86 :367.
- 22↑
Alvarez M, Rodriguez-Roche R, Bernardo L, Vázquez S, Morier L, Gonzalez D, Castro O, Kouri G, Halstead SB, Guzman MG, 2006. Dengue hemorrhagic Fever caused by sequential dengue 1–3 virus infections over a long time interval: Havana epidemic, 2001–2002. Am J Trop Med Hyg 75 :1113–1117.
- 23↑
Deparis X, Murque B, Roche C, Cassar O, Chungue E, 1998. Changing clinical and biological manifestations of dengue during the dengue-2 epidemic in French Polynesia in 1996/97—description and analysis in a prospective study. Trop Med Int Health 3 :859–865.
- 25↑
Hoke CH, Nisalak A, Sangawhipa N, Jatanasen S, Laoraka-pongse T, Innis BL, Kotchasenee S, Gingrich JB, Latendresse J, Fukai K, et al., 1988. Protection against Japanese encephalitis by inactivated vaccines. N Engl J Med 319 :608–614.
- 26↑
Burke DS, Nisalak A, Johnson DE, Scott RM, 1988. A prospective study of dengue infections in Bangkok. Am J Trop Med Hyg 38 :172–180.
- 27
Kliks SC, Nimmanitya S, Nisalak A, Burke DS, 1988. Evidence that maternal dengue antibodies are important in the development of dengue hemorrhagic fever in infants. Am J Trop Med Hyg 38 :411–419.
- 28
Halstead SB, Lan NT, Myint TT, Shwe TN, Nisalak A, Kalyana-rooj S, Nimmannitya S, Soegijanto S, Vaughn DW, Endy TP, 2002. Dengue hemorrhagic fever in infants: research opportunities ignored. Emerg Infect Dis 8 :1474–1479.