Infants born to dengue immune mothers acquire maternal antibodies to dengue. These antibodies, though initially protective, decline during the first year of life to levels thought to be disease enhancing, before reaching undetectable levels. Infants have long been studied to understand the interaction between infection and disease on an individual level.
Considering infants (cases <1 year old) as a unique group, we analyzed serotype specific dengue case data from patients admitted to a pediatric hospital in Bangkok, Thailand. We show differences in the propensity of serotypes to cause disease in individuals with dengue antibodies (infants and post-primary cases) and in individuals without dengue antibodies (primary cases). The mean age of infant cases differed among serotypes, consistent with previously observed differential waning of maternal antibody titers by serotype. We show that trends over time in epidemiology of infant cases are consistent with those observed in the whole population, and therefore with trends in the force of infection.
Infants with dengue are informative about the interaction between antibody and the dengue serotypes, confirming that in this population DENV-2 and DENV-4 almost exclusively cause disease in the presence of dengue antibody despite infections occurring in others. We also observe differences between the serotypes in the mean age in infant cases, informative about the interaction between waning immunity and disease for the different serotypes in infants. In addition, we show that the mean age of infant cases over time is informative about transmission in the whole population. Therefore, ongoing surveillance for dengue in infants could provide useful insights into dengue epidemiology, particularly after the introduction of a dengue vaccine targeting adults and older children.
Infants born to dengue immune mothers acquire maternal dengue antibodies. These antibodies, though initially protective, decline during the first year of life to levels thought to be disease enhancing, before reaching undetectable levels. We show that in this population, DENV-2 and DENV-4 almost exclusively cause disease in the presence of dengue antibody, despite infections occurring in others. We also observe serotype-specificity in the mean age of infant cases, consistent with differential waning of antibody to each serotype. These results highlight serotype-specificity in the way the immune response interacts with infection to cause disease. In addition, we show that the mean age of infant cases over time is informative about transmission in the whole population. Therefore, ongoing surveillance for dengue in infants could provide useful insights into dengue epidemiology, particularly after the introduction of a dengue vaccine targeting adults and older children.
Citation: Clapham H, Cummings DAT, Nisalak A, Kalayanarooj S, Thaisomboonsuk B, Klungthong C, et al. (2015) Epidemiology of Infant Dengue Cases Illuminates Serotype-Specificity in the Interaction between Immunity and Disease, and Changes in Transmission Dynamics. PLoS Negl Trop Dis 9(12): e0004262. doi:10.1371/journal.pntd.0004262
Editor: Aravinda M. de Silva, University of North Carolina at Chapel Hill, UNITED STATES
Received: July 21, 2015; Accepted: November 3, 2015; Published: December 11, 2015
Copyright: © 2015 Clapham et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited
Data Availability: Data are available from the WRAIR IRB Institutional Data Access / Ethics Committee for researchers who meet the criteria for access to confidential data. Please contact firstname.lastname@example.org.
Funding: The views expressed in this article are those of the authors and do not represent the official policy or position of the US Department of the Army, Department of Defense, or US Government. DATC received funding from the National Institute of General Medical Sciences of the National Institutes of Health under Award Number U54 GM088491 and received funding from grant R01Al114703. HC, JL and DATC received funding from the National Institutes of Health under Award Number 5R01AI102939-03. AS received funding from the National Institute of Allergy and Infectious Diseases under Award Number 5P01 AI034533-22. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
DENV is a flavivirus that exists as four serotypes. Infection with one serotype leads to long-term immunity to that serotype. There is also a short-term period of cross-protection to other serotypes [1, 2] followed by an indeterminate period during which infection by another serotype may lead to more severe disease . One theory for this increased severity is antibody dependent enhancement, whereby non-neutralizing antibodies bind to the virus and facilitate viral entry into cells and increased viral replication . The overwhelming majority of hospitalized cases in regions where all four serotypes circulate are due to post-primary infections . Infants born to dengue-immune mothers receive dengue antibodies, and, over the first year of life, experience an accelerated version of the susceptibility pattern that individuals experience during a lifetime in endemic areas: there is a short period of universal protection lasting a few months after birth, followed by a period also lasting a few months in which infections are more likely to be severe possibly through the action of antibody dependent enhancement .
Infant cases of dengue have been an important group for studying dengue immunopathogenesis. Previous studies have described the disease presentation and age distributions of infants in Thailand, Vietnam, Indonesia and the Philippines [7–9], as well as considering the interaction between antibody titres and disease [6, 10–12]. Infant cases may also be an important group for understanding other aspects of the epidemiology of dengue at population scales. There are two main advantages to evaluating infant cases for studying the interaction between immunity and disease. First, at a population scale and even at individual scales, infants have fairly uniform antibody titers across serotypes and, thus, eliminate the uncertainty of timing and nature of past exposures that exists when considering serotype differences in disease severity among older children. Second, the time period that infants are at high risk of infection with severe outcome is relatively short, thus providing information on forces of infection in the population at this time.
In the current study, we analyzed dengue case data from Queen Sirikit National Institute of Child Health (QSNICH) from 1973–2012 to investigate dengue in infants (cases <1 year old). We sought to elucidate intrinsic differences in the propensity for different DENV serotypes to cause disease among patients with pre-existing antibodies by examining serotype distributions in hospitalized infants, compared to other age and immunity groups. We also examined possible relationships between antibody levels and disease outcome by examining the age of severe cases among infants. Finally, we considered changes in dengue case numbers and mean age of infant cases over time and what these changes revealed about the force of infection (FOI) of dengue and population level transmission. This work is important for the study of dengue pathogenesis and epidemiology and is particularly relevant to the development of vaccines. An understanding of potentially protective antibody titers could inform vaccine immunogenicity targets and could clarify the interactions among serotype, immunity and disease outcome when interpreting population level vaccine trial results. Since no vaccine approaching licensure currently plans to target those under 1 year of age, this group could also be an important resource in Phase 4 studies as they will be easily identified as non-vaccinees, and can be used to characterize indirect effects of vaccine campaigns as well as characterize temporal patterns in population level transmission after the introduction of vaccines.
Materials and Methods
The analyzed data were obtained from public health samples collected during passive surveillance of hospitalized dengue cases from 1973 to 2012 at QSNICH, a 420-bed tertiary care pediatric hospital located in Bangkok, Thailand, that serves as a Thailand Ministry of Public Health (MOPH) dengue referral center for Bangkok. Ninety-nine percent of the dengue cases were ≤15 years of age. Acute and convalescent blood samples from clinically suspected dengue inpatients at QSNICH were tested for evidence of DENV infection at the Armed Forces Research Institute of Medical Sciences (AFRIMS) laboratory in Bangkok. The case data up until 1999 have been presented previously  and an updated analysis is in submission (Nisalak et al., submitted to AJTMH). Techniques used for measuring antibody titers and detecting virus have changed over the years (see  Table 2). In brief, acute blood samples were tested by viral isolation and/or hemi-nested reverse transcriptase polymerase chain reaction (RT-PCR) as previously described [14–18]. Acute and convalescent blood samples were tested by dengue serological assays as previously described [19–22]. Primary infection refers to the first DENV infection in an individual and was determined serologically by dengue hemagglutination inhibition assay (HAI) and/or dengue IgM/IgG capture enzyme-linked immunosorbent assay (ELISA) according to published criteria . Post-primary infection refers to any DENV infection subsequent to primary infection and was also determined serologically . The retrieval and analysis of coded pre-existing data in this study was approved by the QSNICH and Walter Reed Army Institute of Research Institutional Review Boards. Blood samples from passive surveillance were originally collected at QSNICH for public health purposes. All data analyzed were anonymized.
For the purpose of this analysis, cases were grouped into three groups: 1) primary cases aged less than 1 year old, which we refer to as infant primary cases, 2) primary cases aged ≥1 year old which we refer to as non-infant primary cases, and 3) post-primary cases of all ages. Only 40 of 21,090 post-primary cases were <1 year of age; these cases were included in the post-primary group, but their inclusion in this or the infant primary group did not alter the results. For each group, we calculated the proportion of cases that were of each serotype, and for each serotype, the proportion of cases that were in each group. Using Pearson correlations, the correlation between the annual numbers of cases in each group for each serotype was assessed.
For infant primary cases, we calculated the mean age for cases of each serotype over all years and the mean age for all serotypes for each year. Using generalized linear models, we assessed trends over time and the relationship between annual mean age and annual proportion of all cases in infants. Analysis was performed using R version 3.0.2 (R Foundation for Statistical Computing, Vienna, Austria) .
The serotype distribution in primary infant cases was more similar to the post-primary cases than to primary cases in non-infants (Table 1 and Fig 1). For both post-primary and infant primary cases, around 35% of cases were DENV-1 and and 31% of cases were DENV-2. This is significantly different to primary non-infant cases where a much greater 57% of cases were DENV-1 and only 5% were DENV-2. For DENV-3 there are slight, non-significant differences between post-primary and infant primary with 22% of post-primary and 27% of primary infant cases due to DENV-3, and these are both significantly less than the 37% of primary non-infant cases that were due to DENV-3. For DENV-4 there are significant differences across all 3 groups with 12% of post-primary cases, 4% of infant primary cases and only 1% of primary non-infant cases due to DENV-4 (Table 1 and Fig 1). Although disease was still more common in post-primary compared to primary cases for DENV-1 and DENV-3 (73% of cases were post-primary), primary infections with these two serotypes caused a substantial amount of disease in dengue naïve individuals. For DENV-2 and DENV-4 however, almost all of the cases of these serotypes were post-primary cases (92% and 96%, respectively). The percentage of cases that were untyped was the same for the primary infant and non-infant cases (both 40%), compared to post-primary cases (53%) (Table 2 and Fig 2).
Serotypes DENV-1 to -4 are in purple, red, green and blue, respectively, untyped is in yellow.
Cases are shown on a log10 scale. Primary cases in infants are in dark blue, post-primary cases are in mid blue, and primary cases in non-infants are in light blue.
There were positive, significant correlations within each serotype each year between the number of infant primary cases, non-infant primary cases, and post-primary cases in each year, i.e., the number of primary non-infant cases of a serotype each year was associated with the number of infant primary cases and, separately, associated with the number of post-primary cases of that serotype each year (see Fig 2 and Table 3). For DENV-1 and DENV-3, the correlations between the numbers in each group were between 0.78 and 0.95, while for DENV-2 and DENV-4, these correlations were lower between 0.40 and 0.60.
Finally, we considered the trends over time in the mean age of infant primary cases and the infant primary cases as a proportion of all cases (Fig 4). There was a significant positive correlation between mean age and year (correlation 0.31, p-value: < 0.05). This increase in mean age was clear from 1990 to around 2007, but there was a suggestion of a decrease after 2007. Age distributions by decade are shown in Fig 3B. Using a linear model, there was a significant relationship between proportion of cases that were in infants and the mean age, year and the interaction between mean age and year (p-values all <0.01, coefficients: mean age: 18, year: 0.005 and mean age*year: -0.009). This relationship showed that during the years of increasing mean age of infant cases (1990–2007), there was a decrease in the proportion of all cases that were in infants; and during the years of decreasing mean age (after 2007), there was an increase in the proportion of cases that were in infants (Fig 4).
The serotype distributions of hospitalized dengue cases in different immune groups, as presented in this paper, add to the evidence that differences in the outcome of infection by each serotype depends on immune status. The paucity of primary cases of DENV-2 and DENV-4 has been shown in previous studies in Thailand . The presence of infant primary cases with these serotypes, suggests that dengue naive individuals ≥1 year of age are exposed to these serotypes, but that these exposures do not result in hospitalized disease. There are two non-mutually exclusive ways to interpret these findings: (1) DENV-1 and DENV-3 were more likely to cause disease in non-immune individuals compared to DENV-2 and DENV-4, or (2) DENV-2 and DENV-4 were more likely to cause disease in an enhanced post-primary infection than DENV-1 and DENV-3. These results from the infants suggest that the former is the most likely explanation, with the correlations between the annual case numbers in each group (infant primary, primary non-infant and post-primary) being lower for DENV-2 and DENV-4, than for DENV-1 and DENV-3, suggesting the immune status of the population plays a larger role in the dynamics of DENV-2 and DENV-4. These differences may also explain previously seemingly contradictory results where a relationship was shown between infant DHF and increased levels of enhancing activity in sera for DENV2 , but not DENV3 . In addition, there is a suggestion that DENV-4 is under represented in the infant primary cases compared to the secondary cases. This could be explained by a lower force of infection for DENV-4 (consistent with the fewer observed cases) leading to fewer exposures in this early time period. Differences between the serotypes in the propensity to cause disease in immune and non-immune individuals should be considered in the context of vaccination trial results with seemingly differential efficacy across serotypes, as with the Sanofi Phase 2b results . Could the observed effect of the vaccination for DENV2 be because of the differential outcomes of DENV2 exposure in naïve and non-naïve individuals?
The mean age of infant primary cases was similar to previous studies [7, 8]. The observed differences in the mean age in infants by serotype (highest for DENV-1 and DENV-3 followed by DENV-2 then DENV-4) could be due to two non-mutually exclusive reasons. Firstly, the force of infection could have been higher for the serotypes with the lower mean ages (so infections occurred on average earlier). Secondly, differential waning of antibody titers by serotype could lead to potential enhancement occuring at different ages for the different serotypes. The first explanation would suggest a higher force of infection for DENV-2 and DENV-4 compared to DENV-1 and DENV-3. Previous work has suggested R0 and thus the FOI may be higher for DENV-2 and the considerable numbers of DENV-2 cases  would be consistent with a high FOI for DENV-2. However, DENV-4 had lower incidence overall and serological studies do not suggest that DENV-4 has a higher R0 than other serotypes . For antibody waning, previous work by van Panhuis et al.  indicated the fastest antibody waning to be for DENV-4 (with a mean titer at 6 months of 17 [95% CI: 12, 25]). The next fastest waning serotype was DENV-2 (with a mean titre at 6 months of 25 (95% CI: 21, 31), followed by DENV-3 and DENV-1 (mean titer at 6 months of 35 (95% CI: 29, 43)). This order of antibody waning is consistent with the observed serotype-specific infant mean ages, though these results may be specific to the assays performed. There could, of course, also be differences between the serotypes in the antibody response required for protection or to lead to enhanced infections, as observed in older individuals in a study in Kamphaeng Phet, Thailand . With a population such as Bangkok, we would expect broad antibody responses in maternal antibody, however further study of maternal over time and in multiple populations would be of great interest.
The small but significant increase in mean age of infant cases over time (particularly in the 1990s to 2007) is consistent with the increase in mean age of dengue seen in the general population in Thailand . One of the leading hypotheses for this increase is a reduction in the force of infection . Both the trends in mean age of infant primary cases and the relative proportion of all cases that were in infants, are consistent with a decrease in FOI during this period (as the FOI drops we would expect to see fewer cases in infants as the chance of being infected in this first year of life drops). It is interesting that after 2007, a decrease in the mean age of infant cases, and an increase in the proportion of cases that were in infants was observed, suggesting an increase in FOI in this period. This decrease in mean age was also observed in primary cases of older age groups (Nisalak et al., submitted to AJTMH). Further years of data will be needed to determine whether this a persistent trend or a transient fluctation. We show a change in the mean age of infants over time and by serotype, however the numbers of cases in each year are too small to determine the changes over time for each serotype.
Our study suggests that infants, in addition to being informative about immune-mediated pathogenesis, could act as a sentinel population for understanding population-level transmission. Vaccines currently in development are unlikely to be given to infants [29, 30], and therefore, this age group will still be largely susceptible to infection even after vaccine introduction. With this lack of change in susceptibility, surveillance of this population could help determine whether transmission or simply disease presentation has altered due to vaccination. Whether infant susceptibility changes due to vaccine-derived maternal immunity will also be an important question to address in follow up studies. One would hope that ultimately, infants will be protected by herd immunity, however, if protection is suboptimal for some serotypes, is greater against severe disease as opposed to just infection, or vaccine coverage is low, the force of infection could be maintained at close to pre-vaccination levels. Therefore, infants will still be at substantial risk for infection and be an important population to observe to understand immunity and transmission.
S1 Checklist. STROBE Checklist.
The authors wish to acknowledge the clinical staff at QSNICH, and the clinical, laboratory and data management staff at AFRIMS. We thank Dr. Siraporn Sawasdivorn, Director, QSNICH, for her support. We thank three anonymous reviewers for their helpful comments on the manuscript. The views expressed in this article are those of the authors and do not represent the official policy or position of the US Department of the Army, Department of Defense, or US Government.
Conceived and designed the experiments: HC DATC AN IY. Performed the experiments: SK BT CK SF AS LRM AN. Analyzed the data: HC AN DATC. Contributed reagents/materials/analysis tools: SK BT CK SF AS LRM JR. Wrote the paper: HC DATC JL AN SF AS LRM IY.
- 1. Sabin A.B., Research on dengue during World War II. Amer Jour Trop Med and Hyg, 1952. 1((1)): p. 30–50.
- 2. Anderson K.B., et al., A shorter time interval between first and second dengue infections is associated with protection from clinical illness in a school-based cohort in Thailand. J Infect Dis, 2014. 209(3): p. 360–8. doi: 10.1093/infdis/jit436. pmid:23964110
- 3. Halstead S.B., Nimmanni S, and Cohen S.N., Observations related to pathogenesis of dengue hemorrhagic fever .4. Relation of disease severity to antibody response and virus recovered. Yale Journal of Biology and Medicine, 1970. 42(5): p. 311–&. pmid:5419206
- 4. Morens D.M., Antibody-dependent enhancement of infection and the pathogenesis of viral disease. Clinical Infectious Diseases, 1994. 19(3): p. 500–512. pmid:7811870 doi: 10.1093/clinids/19.3.500
- 5. Vaughn, D.W., et al., Dengue viremia titer, antibody response pattern, and virus serotype correlate with disease severity. 2000. 181(1): p. 2–9.
- 6. Kliks S.C., et al., Evidence that maternal dengue antibodies are important in the development of dengue hemorrhagic-fever in infants. American Journal of Tropical Medicine and Hygiene, 1988. 38(2): p. 411–419. pmid:3354774
- 7. Capeding R.Z., et al., The incidence, characteristics, and presentation of dengue virus infections during infancy. Am J Trop Med Hyg., 2010. 82(2): p. 330–6. doi: 10.4269/ajtmh.2010.09–0542. pmid:20134013
- 8. Halstead S.B., et al., Dengue hemorrhagic fever in infants: Research opportunities ignored. Emerging Infectious Diseases, 2002. 8(12): p. 1474–1479. pmid:12498666 doi: 10.3201/eid0812.020170
- 9. Nguyen T.H., et al., Dengue hemorrhagic fever in infants: a study of clinical and cytokine profiles. J Infect Dis, 2004. 189(2): p. 221–32. pmid:14722886 doi: 10.1086/380762
- 10. Libraty D.H., et al., A prospective nested case-control study of Dengue in infants: rethinking and refining the antibody-dependent enhancement dengue hemorrhagic fever model. PLoS Med, 2009. 6(10): p. e1000171. doi: 10.1371/journal.pmed.1000171. pmid:19859541
- 11. Chau T.N.B., et al., Dengue virus infections and maternal antibody decay in a prospective birth cohort study of Vietnamese infants. J Infect Dis, 2009. 200(12): p. 1893–900. doi: 10.1086/648407. pmid:19911991
- 12. Simmons C.P., et al., Maternal antibody and viral factors in the pathogenesis of dengue virus in infants. Journal of Infectious Diseases, 2007. 196(3): p. 416–424. pmid:17597456 doi: 10.1086/519170
- 13. Nisalak A., et al., Serotype-specific dengue virus circulation and dengue disease in Bangkok, Thailand from 1973 to 1999. Am J Trop Med Hyg., 2003. 68(2): p. 191–202. pmid:12641411
- 14. Yuill T.M., et al., Dengue-virus recovery by direct and delayed plaques in LLC-MK2 cells. Am J Trop Med Hyg, 1968. 17(3): p. 441–448. pmid:4385131
- 15. Kuberski T.T. and Rosen L., A simple technique for the detection of dengue antigen in mosquitoes by immunofluorescence. Am J Trop Med Hyg, 1977. 26(3): p. 533–537. pmid:326071
- 16. Watts D.M., et al., Evaluation of Toxorhynchites splendens (Diptera: Culicidae) as a bioassay host for dengue viruses. J Med Entomol, 1982. 19(1): p. 54–59. pmid:6126591 doi: 10.1093/jmedent/19.1.54
- 17. Klungthong C., et al., Dengue Viral Detection using Whole Blood for RT-PCR and Viral Isolation. J Clin Microbiol, 2007.
- 18. Jarman R.G., et al., Factors influencing dengue virus isolation by C6/36 cell culture and mosquito inoculation of nested PCR-positive clinical samples. Am J Trop Med Hyg, 2011. 84(2): p. 218–23. doi: 10.4269/ajtmh.2011.09-0798. pmid:21292887
- 19. Clarke D. and Casals J., Improved methods for hemagglutination studies with arthropod-borne viruses. Proc Soc Exp Biol Med, 1955. 88: p. 96–99. pmid:14357353 doi: 10.3181/00379727-88-21502
- 20. Russell P.K., et al., A plaque reduction test for dengue virus neutralization antibodies. J Immunol, 1967. 99(2): p. 285–290. pmid:6031202
- 21. Kuno G., Gubler D.J., and Santiago de Weil N.S., Antigen capture ELISA for the identification of dengue viruses. J Virol Methods, 1985. 12: p. 93–103. pmid:2867103 doi: 10.1016/0166-0934(85)90011-4
- 22. Innis B.L., et al., An enzyme-linked immunosorbent assay to characterize dengue infections where dengue and Japanese encephalitis co-circulate. Am J Trop Med Hyg, 1989. 40(4): p. 418–427. pmid:2540664
- 23. R Development Core Team, R: A language and environment for statistical computing. 2012, R Foundation for Statistical Computing: Vienna, Austria.
- 24. Sabchareon A., et al., Protective efficacy of the recombinant, live-attenuated, CYD tetravalent dengue vaccine in Thai schoolchildren: a randomised, controlled phase 2b trial. Lancet., 2012. 380(9853): p. 1559–67. doi: 10.1016/S0140-6736(12)61428-7. Epub 2012 Sep 11. pmid:22975340
- 25. Rodriguez-Barraquer I., et al., Revisiting Rayong: shifting seroprofiles of dengue in Thailand and their implications for transmission and control. Am J Epidemiol., 2014. 179(3): p. 353–60. doi: 10.1093/aje/kwt256. Epub 2013 Nov 5. pmid:24197388
- 26. van Panhuis W.G., et al., Decay and Persistence of Maternal Dengue Antibodies among Infants in Bangkok. Am J Trop Med Hyg, 2011. 85(2): p. 355–62. doi: 10.4269/ajtmh.2011.11-0125. pmid:21813859
- 27. Buddhari D., et al., Dengue virus neutralizing antibody levels associated with protection from infection in thai cluster studies. PLoS Negl Trop Dis., 2014. 8(10): p. e3230. doi: 10.1371/journal.pntd.0003230. eCollection 2014 Oct. pmid:25329173
- 28. Cummings D.A.T., et al., The Impact of the Demographic Transition on Dengue in Thailand: Insights from a Statistical Analysis and Mathematical Modeling. PLoS Med, 2009. 6(9): p. e1000139. doi: 10.1371/journal.pmed.1000139. pmid:19721696
- 29. Capeding R.Z., et al., Live-attenuated, tetravalent dengue vaccine in children, adolescents and adults in a dengue endemic country: Randomized controlled phase I trial in the Philippines. Vaccine, 2011. 29(22): p. 3863–3872. doi: 10.1016/j.vaccine.2011.03.057. pmid:21477675
- 30. Capeding M.R., et al., Clinical efficacy and safety of a novel tetravalent dengue vaccine in healthy children in Asia: a phase 3, randomised, observer-masked, placebo-controlled trial. Lancet., 2014. 384(9951): p. 1358–65. doi: 10.1016/S0140-6736(14)61060-6. Epub 2014 Jul 10. pmid:25018116