Introduction

Although dengue infection results in substantial morbidity and mortality [1], there have been no population-based studies in young children aged three years, and as a consequence there are no clear definitions of dengue immunoglobulin G (IgG) serological thresholds to define prior exposure. However, knowledge of cases of definite prior clinical infection with dengue virus may help identify these threshold values, and also permit comparison with a control group with no prior infection, hence allowing estimation of the number of sub-clinical cases of dengue infection.

We have used data from a community-based study of healthy three-year old children living in Havana, Cuba, to explore the association of increased titres of dengue IgG immunoglobulin for those with a definite prior clinical diagnosis of dengue infection, compared to those with no history of dengue infection. We also assessed the association of increased titres of dengue IgG immunoglobulin with blood measurements of inflammation as previous studies have demonstrated that leucocytes are decreased during acute dengue virus infection [2–9].

Methods

Statistical analysis

The cases with a definite prior medical diagnosis of dengue infection were compared with those with no prior medical diagnosis of dengue infection. The exposure of dengue IgG levels were initially plotted as a histogram to explore the distribution. This demonstrated that a substantial proportion of the population has no serum dengue IgG response (Fig 1). A binary outcome measure of an elevated serum IgG titre was then generated by calculating the median value. An elevated dengue IgG was defined as above the median value, and compared with those values that were equal to or below the median.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Histogram of dengue IgG levels.

https://doi.org/10.1371/journal.pone.0207391.g001

Logistic regression was used to explore the association of the blood exposures of interest with an increased dengue IgG titre with the various outcome measures using Stata v14 statistical software (Texas, USA).

Results

1543 individuals provided data for the third year of the study, and of these 865 (56%) donated a blood sample that was analysed to generate the dengue IgG measurements (Table 1). 439 individuals (51%) were male and the median age was 39.6 months (interquartile range 37.2 to 42.0). The range of dengue IgG titres for the study population is presented in the Fig 1, and had a median value of 1.3 (interquartile range of 0.37 to 2.6).

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Study population with associations with increased dengue IgG titres.

https://doi.org/10.1371/journal.pone.0207391.t001

There were 14 (1.6%) individuals with a prior medical diagnosis of dengue infection, and these cases had a median dengue IgG titre of 1.26 (interquartile range IQR 0.60 to 3.10). There were 851 individuals with no prior medical diagnosis of dengue infection, and this group had a median IgG titre of 1.30 (0.34 to 2.57). There was no statistical difference between the IgG levels for these two groups (Fig 2, Mann-Whitney test, p = 0.49, S1 Data). Four hundred and thirty one individuals (50.6%) with no previous medical diagnosis of dengue infection had IgG titres above the median value for those who had a prior medical diagnosis of dengue infection, while only 131 individuals (15.4%) of this group had no dengue IgG response at all.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Comparison of the distribution of IgG dengue serology stratified by cases with medically confirmed dengue infection and controls with no medical diagnosis of dengue infection.

The box plot shows the median value (central line in box) and the 25^th and 75^th percentiles (horizontal lines of box). The individual horizontal lines demonstrate the limits of adjacent data, and the dots represent the outlying values. Statistical comparison used the Mann-Whitney test.

https://doi.org/10.1371/journal.pone.0207391.g002

The total white cell count was linearly negatively associated with a dengue IgG value above the median value with an odds ratio (OR) of 0.87 (95% confidence intervals CI: 0.78 to 0.96) per unit increase in white cell count. Blood neutrophil and eosinophil counts were also linearly negatively associated with dengue IgG, with ORs of 0.90 (95% CI: 0.85 to 0.95) and 0.70 (95% CI: 0.61 to 0.78) respectively. There was no association of blood lymphocyte counts, log serum IgE, or log serum hsCRP with increased blood dengue IgG levels

Discussions

This is the first population-based study in three-year old children to explore the differential serological response between cases with medically confirmed prior infection and those with no previous infection. There was no significant difference between the two groups, and there was a substantial proportion of the non-infected control group with IgG titres above those of the infected group, suggesting possible sub-clinical infection. We also identified that an elevated dengue IgG level was associated with decreased blood leucocyte levels, lower neutrophil and eosinophil levels, although our cross-sectional study design does not allow any causal inferences to be drawn.

The strengths of this analysis include the population-based nature of the cohort, and the use of objective measures of both the exposures and the outcome measures to minimise the risk of observer bias. The study was based in policlinics that are the first point of contact for Cubans with the national health system. The study was designed to study allergic disease and hence there was no sensitisation or risk of increased awareness of the risk of dengue infection above that expected in the general Cuban population, who were experiencing an increase in the incidence of dengue infection cases at the time [14]. All children with suspected dengue infection in Cuba are referred to the hospital for investigation including serological (IgM) testing, and hence we are confident that a positive medical diagnosis of dengue is correct. In addition, the absence of alternative health providers in Cuba makes it unlikely that a diagnosis of acute dengue infection would be made elsewhere. As the dengue IgG assay that we used has not been validated as a measure of previous infection in three-year old children, a binary outcome variable was generated based on whether an individual’s dengue IgG response was above or below the median value for the whole population. This is a pragmatic but relatively simple statistical test that demonstrates that it is unlikely that these associations occurred by chance, but do not explain the natural history of any biological processes that may be involved. It is well recognised that the antibody response to external exposures in young children below the age of five years old is limited, increasing as the child gets older [15], and hence this binary outcome measure of dengue IgG may not necessarily equate to prior clinical infection [11]. It is important to clarify that our data may not be generalizable to other different populations, especially those that are older and hence are likely to have more mature immune systems, or that are living different in other geographical locations where the prevalence of dengue and other associated environmental factors will differ to those in Cuba.

The comparison of the serological IgG titres between the diagnosed cases of prior dengue infection and the control group with no prior infection is interesting. There is no difference between the serum IgG levels between the two groups, and although the number of cases is relatively low, we would anticipate that these numbers would be sufficient to observe a difference in serum dengue IgG titres if the control group has had no prior exposure at all to dengue virus. This raises the possibility that the control group have had substantial sub-clinical dengue virus exposure. Quantifying this population is challenging in the absence of a threshold for the IgG serological assay, but 51% of the control population had a serum IgG level above the median value of those who have had medically confirmed dengue infection and 15% had no detectable serum IgG to dengue. An alternative explanation for the absence of a difference between IgG titres between the two groups is that in the context of a maturing immune system, some of the children who had definite prior IgG infection were not able to generate a robust immune response. However, as age was positively associated with risk of increased dengue IgG and is also a measure of the duration of risk of exposure to dengue infection, the former explanation is more plausible.

The observation that dengue virus exposure is often a sub-clinical event has been made previously in a large epidemiological study of children and adults living in Santiago de Cuba, in 1997. This analysis estimated that approximately 3% of primary dengue infections were clinically overt, with the remaining 97% being subclinical [16]. Some of these subclinical cases may constitute children who experienced a low-grade febrile episode that was independently managed by the carers, and did not come to the attention of the medical authorities. Infection with a variety of viruses including arboviruses are present in the tropics, and are understandably not always are bought to the attention of the medical authorities. Our data are consistent with this hypothesis, and suggest that there is no significant difference in the residual IgG immune response between those who had clinically overt dengue infection and those who did not. The immune response may be of potential clinical importance, as a secondary dengue infection after a primary infection generates a more symptomatic dengue infection [17, 18]. This is an area with limited epidemiological data, and hence continued observation of similar cohorts with good baseline data will help increase understanding of the natural history of clinical response dengue infection over the life course.

In conclusion, data from a population-based study of healthy young children living in Havana, Cuba has demonstrated that cases with prior serologically confirmed dengue infection have similar dengue serological IgG levels to those with no prior medical diagnosis of dengue infection. This may represent a substantial prevalence of sub-clinical dengue infection.

Supporting information

Acknowledgments

We would like to thank the children and their parents and guardians for participating in this study. We would like to thank all of the staff at the municipalities and policlinics who helped in many ways with data collection and sample analysis. HINASIC study group (Historia Natural de la Sibilancia en Cuba/National History of Wheezing in Cuba) consists of the following individuals:

National Institute of Hygiene, Epidemiology and Microbiology

Silvia J Venero-Fernández (Leader), Ramón Suárez-Medina, Lenina Menocal-Heredia, Yuria Isabel Caraballo-Sánchez, Félix Manuel Rosado-García, Reina Amelia Quintana, Lourdes Batista-Gutiérrez. Patricia Varona-Pérez, María del Carmen Hinojosa.

Municipality of Arroyo Naranjo

Esperanza de la C. Mora-Faife, María Teresa Méndez-Ratger, Amed Fernández-Casamayor, Gisela Álvarez-Valdez, Anadelis Alfonso-Hernández, Roberto Hidalgo-Mederos, Nereida Calderin-Martínez, .Jorge Antonio Febles-del Toro, Danay Silva, Violeta Suarez-Angelo, Noelvis Zayas-Mompie, Grisel M. Esquivel-Barrios, Zoe de los Angeles Figueroa-Barreto, Olga Lidia Negrin-Molina, Odalis Kessell- Díaz, Mariela de la Caridad Hernández-González, Dulcima Casanave-Guarnaluce, Gretel Comas-Fonseca, Vilma Álvarez-Valdez, María Engracia Báez-Rodríguez, Felicia Sánchez-Cardentey, Nieves Sardinas-Báez, Roberto Esteban Márquez-Solís, Marlene Flores-Carballosa, Nagaby Gómez-Baro, Ivette Castillo-Aguilar.

Municipality of La Lisa

Gladys García-García, María de Lourdes Ortiz-Hernández, María Antonia Betancourt-López, Marlén Batista-Cedeño, Iris Alfonso-Castellanos, Leticia Gómez-García, Ernesto Rafael Gutiérrez-Mendoza, María Luisa Loynaz-González, Nibenia Rodríguez-Trujillo, Yanet Pozo-Herrera, Víctor Manuel Montejo-Guerra, Julia Urbina-Reynaldo, Valentina Gómez-Suliman, Caridad Alicia Rodríguez-Aragón.

Municipality of Cerro

Ileana del Valle-Infante, Martha Betancourt-Orue, Oscar Alba-Monteagudo, Yuderkis Ferrer-Ceruto, Aída Damas-Martínez, Mercedes Peñalver-Pérez

Municipality of Habana del Este

Liem Gómez-Marrero, Sarahí Castillo-Martínez, Amor de los Angeles Castaño-Vega, Norberto Torriente-Barzaga, Ileana Ávila-Rodríguez, Magalys Navarro-Ruiz, Kirenia

Díaz-Hernández, Iluska De La Torre Suarez, Gilberto Roque-Pereira, Yamilet Corona-Carnero, Idania Gonzalaz-Fernandez, Fidelia Romeu-Ravelo, Regla Hernandez-Ponce,

Teresa Serrano-González, Dulce Romeo-Cepero, Caridad Gonzalez Leiva, Teresa De Jesus Cobas Espino, Nuris-Fajardo, Midiala Perez-Arcia, Sarahy Diaz-Araujo, Yanet

Medina- Lescay, Sandra Collazo-Rodriguez, Julia Amparo Griñán-Ramos, Teresa Serrano-Gonzalez, Beatriz Lazo-Vazquez, Tania Pupo-Portal, Nidia Leyva-Porra, Odalys Pacheco-Mesa, Martha Rizo-Ramos, Yaneysi Vallafuerte- Perez, Aliniuska De La Paz-Arias, Maite B Garcia-Sotolongo, Yusimí Calzado-Herrera, Martha Nidia Rizo-Ramos, Guillermo Verdecia, Mayté B García-Sotolongo, Juana F Abreu-Quijano, Fidelia Romeo-Ravelo.

Thanks also to all Municipality Directors and the laboratory workers who have also supported the study.