SEARCH

SEARCH BY CITATION

Keywords:

  • dengue;
  • transmission;
  • serial interval;
  • space-time;
  • clustering;
  • Thailand
  • dengue;
  • transmission;
  • intervalle de séries;
  • spatio-temporel;
  • regroupement;
  • Thaïlande
  • dengue;
  • transmisión;
  • intervalo seriado;
  • espacio-temporal;
  • conglomerado;
  • Tailandia

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Dengue transmission cycle
  5. Space-time analysis of DENV transmission
  6. Materials and methods
  7. Results
  8. Discussion
  9. Acknowledgements and disclaimer
  10. References

Objective  To determine the temporal intervals at which spatial clustering of dengue hospitalisations occurs.

Methods  Space-time analysis of 262 people hospitalised and serologically confirmed with dengue virus infections in Kamphaeng Phet, Thailand was performed. The cases were observed between 1 January 2009 and 6 May 2011. Spatial coordinates of each patient’s home were captured using the Global Positioning System. A novel method based on the Knox test was used to determine the temporal intervals between cases at which spatial clustering occurred. These intervals are indicative of the length of time between successive illnesses in the chain of dengue virus transmission.

Results  The strongest spatial clustering occurred at the 15–17-day interval. There was also significant spatial clustering over short intervals (2–5 days). The highest excess risk was observed within 200 m of a previous hospitalised case and significantly elevated risk persisted within this distance for 32–34 days.

Conclusions  Fifteen to seventeen days are the most likely serial interval between successive dengue illnesses. This novel method relies only on passively detected, hospitalised case data with household locations and provides a useful tool for understanding region-specific and outbreak-specific dengue virus transmission dynamics.

Objectif:  Déterminer les intervalles temporels au cours desquels survient le regroupement spatial des hospitalisations pour dengue.

Méthodes:  L’’analyse spatio-temporelle de 262 personnes hospitalisées et confirmées par la sérologie pour des infections au virus de la dengue à Kamphaeng Phet, en Thaïlande a été réalisée. Les cas ont été observés entre le 1er janvier 2009 et le 6 mai 2011. Les coordonnées spatiales du domicile de chaque patient ont été capturées à l’aide du Système de Positionnement Global. Une nouvelle méthode basée sur le test de Knox a été utilisée pour déterminer les intervalles temporels entre les cas au cours desquels le regroupement spatial a eu lieu. Ces intervalles sont des indicateurs de la durée de temps entre les maladies successives dans la chaîne de transmission de virus de la dengue.

Résultats:  Le plus important regroupement spatial a eu lieu dans l’intervalle de 15 à 17 jours. Il y avait également un regroupement spatial significatif sur des intervalles courts (2 à 5 jours). Le risque en excès le plus élevé a été observéà endéans 200 m d’un cas précédent hospitalisé et un risque significativement élevé a persisté dans cette zone durant 32 à 34 jours.

Conclusions:  L’intervalle de séries de 15 à 17 jours est le plus probable entre les maladies successives de la dengue. Cette nouvelle méthode repose uniquement sur les données de cas hospitalisés, détectés passivement avec les emplacements des ménages et constitue un outil utile pour comprendre la dynamique de la transmission du virus de la dengue spécifique à la région et spécifique à l’épidémie.

Objetivo:  Determinar los intervalos temporales en los que ocurren conglomerados espaciales de hospitalización por dengue.

Métodos:  Se realizó un análisis espacio temporal a 262 personas hospitalizadas y con confirmación serológica de infección por el virus del dengue en Kamphaeng Phet, Tailandia. Los casos se observaron entre el 1 de Enero del 2009 y el 6 de Mayo del 2011. Las coordenadas espaciales del hogar de cada paciente fueron capturadas utilizando un Sistema de Posicionamiento Global. Se utilizó un nuevo método, basado en la prueba de Knox, para determinar los intervalos temporales entre los casos en los que había conglomerados espaciales. Estos intervalos son indicativos del lapso de tiempo entre enfermedades sucesivas en la cadena de transmisión del virus del dengue.

Resultados:  El mayor conglomerado espacial ocurrió en un intervalo de 15 a 17 días. También había un conglomerado espacial significativo en intervalos cortos (2–5 días). El mayor exceso de riesgo se observó dentro de un radio de 200 m de un caso previamente hospitalizado y un riesgo significativamente elevado persistía, dentro de esta distancia, durante 32–34 días.

Conclusiones:  Los 15–17 días son el intervalo más probable entre dos casos sucesivos de dengue. Este nuevo método depende solamente de la disponibilidad de datos de casos hospitalizados, detectados de forma pasiva y con hogares localizables, y brinda una herramienta útil para entender la dinámica de transmisión del virus del dengue, región-específica y específica de brotes.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Dengue transmission cycle
  5. Space-time analysis of DENV transmission
  6. Materials and methods
  7. Results
  8. Discussion
  9. Acknowledgements and disclaimer
  10. References

Dengue fever and dengue haemorrhagic fever are important and expanding public health problems in tropical and subtropical regions (Endy et al. 2010; Gubler 2002). Dengue has spread from being endemic in just nine countries in 1970 to more than 100 countries as of 2009, with an estimated 2.5 billion people at risk of dengue virus (DENV) infection annually (WHO 2009). There are no licensed vaccines or drugs for dengue, and current prevention and control efforts focus on vector population reduction. More accurate characterisation of the spatial and temporal pattern of DENV transmission can inform improved surveillance and efficient control programmes. It can also help to provide data for models that are key tools for evaluating the cost and effectiveness of proposed control efforts, whether the approach is based on vaccination alone or a combined vaccine and vector control effort. Characterising the variation in the length of time between successive infections in the chain of transmission would help to better understand region-specific transmission dynamics. This study examined the spatial and temporal pattern of hospitalised dengue patients in Kamphaeng Phet, Thailand. The results indicate that the most likely interval between successive dengue illnesses can be estimated using only the severe cases that reach healthcare facilities.

Dengue transmission cycle

  1. Top of page
  2. Abstract
  3. Introduction
  4. Dengue transmission cycle
  5. Space-time analysis of DENV transmission
  6. Materials and methods
  7. Results
  8. Discussion
  9. Acknowledgements and disclaimer
  10. References

A cycle in the chain of DENV transmission (Figure 1) is initiated when an infective mosquito takes a blood meal from a susceptible person, injecting virus in the process. The latent intrinsic incubation period is thought to range from 2 to 12 days and is most commonly between 4 and 6 days (Sabin 1952; Siler et al. 1926; Nishiura & Halstead 2007). The infectious period in the host varies from 1 to 7 or more days, and viraemia is most often detected for a duration of 3–5 days (Gubler et al. 1981; Vaughn et al. 2000, 1997). Siler et al. (1926) reported that infectiousness preceded the onset of symptoms by 6–18 h in a group of volunteers challenged with near wild-type DENV. During the infectious period, a mosquito that feeds on the host may become infected. The extrinsic incubation period varies with ambient temperatures and can be as short as 8 or as long 20 days (Lambrechts et al. 2011; Siler et al. 1926; Watts et al. 1987). After the extrinsic incubation period, the infectious mosquito may transmit the virus with each subsequent blood meal (Putnam & Scott 1995), and female Aedes aegypti bite humans frequently, almost daily (0.76–0.63 human blood meals per day) (Scott et al. 2000). The serial interval between successive dengue illnesses along the chain of transmission is at least as long as the sum of the intrinsic and extrinsic incubation periods (Fine 2003). In the explanation of the DENV transmission cycle, it should also be noted that the duration of the extrinsic incubation period is long relative to the estimated life expectancy of A. aegypti (Bartley et al. 2002). There is considerable uncertainty associated with the estimates of daily survival, but a majority of adult A. aegypti are expected to die before they are able to transmit DENV (Harrington et al. 2001; Sheppard et al. 1969).

image

Figure 1.  The DENV transmission cycle. Human hosts are represented by timelines A, B1 and B2. The vector is represented by timeline V. The cycle begins when susceptible host A acquires an infection from a vector (not shown) at the upper left. The vector, V, takes a blood meal from host A during the infectious period. After the extrinsic incubation period, the cycle is completed when DENV is transmitted to susceptible hosts B1 and B2. The serial interval between successive illnesses is depicted with the thick grey arrows.

Download figure to PowerPoint

Space-time analysis of DENV transmission

  1. Top of page
  2. Abstract
  3. Introduction
  4. Dengue transmission cycle
  5. Space-time analysis of DENV transmission
  6. Materials and methods
  7. Results
  8. Discussion
  9. Acknowledgements and disclaimer
  10. References

The spatial and temporal dynamics of DENV transmission are governed by the dispersal of infected vectors and hosts. Release-recapture studies have shown that A. aegypti do not generally disperse over long distances (Harrington et al. 2005; Sheppard et al. 1969; Trpis & Hausermann 1986). Harrington et al. (2005) observed that most recaptured A. aegypti were located within the house where they were released or in a neighbouring house. Garcia-Rejon et al. (2008) reported that DENV-infected A. aegypti were collected from dengue patients’ homes up to 27 days after the onset of symptoms, indicating that infected mosquitoes do not disperse far from the likely site of transmission. It is thought, therefore, that human movement is responsible for most DENV transmission over long distances and between human settlements. The extent and role of human movement in dengue transmission is not yet well characterised, but modelling results indicate that movement is important in understanding individual risk of infection (Stoddard et al. 2009).

Clustering of dengue cases in space and time has been observed in a variety of settings. Early epidemiological studies described the clustering of cases within households and an apparent diffusion through communities from first observed cases (Halstead 1966; Neff et al. 1967). More recent examinations have employed spatial statistics to identify significant clusters and describe the evolution of patterns through time (Kan et al. 2008; Vazquez-Prokopec et al. 2010). With the focal nature of DENV transmission as a guide, several cluster sampling strategies have been employed to detect dengue infections across the clinical spectrum (Beckett et al. 2005; Mammen et al. 2008; Reyes et al. 2010). In the same region under study here, Mammen et al. (2008) found that DENV infection risk surrounding an index case identified through a school-based surveillance programme was clustered within 100 m radius sampling areas. Morrison et al. (1998) used the Knox test to examine clinically apparent case data from an outbreak in Puerto Rico. They determined that there was significant clustering of cases within 5 m and 3 days of each other and also within 35 m and 3 days of each other. Tran et al. (2004) employed the Knox test in an exploratory mode and examined the space-time clustering of cases over a large range of distances and times during an outbreak in French Guiana. Their study also identified focal dengue risk, with the highest risk within 15 m and 6 days of a previously identified case. Additionally, the Knox test results showed a temporal periodicity in risk at 3-, 6- and 9-day intervals. The authors suggest that this pattern reflects the feeding interval of infectious A. aegypti.

Observations and analysis of the space-time pattern of dengue cases have provided guidance for vector control and disease surveillance efforts (Scott & Morrison 2010). For the most part, however, existing examinations of case data have provided little information on the timing and spatial extent of successive illnesses along the chain of transmission. This is in part a limitation of existing space-time clustering methods (Jacquez et al. 2007). The techniques employed in the studies outlined previously are designed to test hypotheses about clusters within a cumulative space-time window (i.e. cases within 5 m and 3 days of each other). The incremental Knox test (IKT) was introduced to overcome this limitation and examine the clustering of cases at a series of time intervals (Aldstadt 2007). The IKT can be used to examine the clustering of cases that occur at a given time interval apart. For example, the proximity of cases occurring 7 days apart could be examined. The interval could also be defined as a range of values to account for uncertainty in the reporting of onset of illness and effectively smooth the data. Given the typically limited flight range of A. Aegypti, we would expect that successive cases in the chain of transmission would be clustered in space. Those time intervals between cases when significant spatial clustering occurs likely indicate the most frequent lengths of intervals in the dengue transmission cycle. The case study included in Aldstadt (2007) found the most frequent serial interval between cases during an outbreak in Puerto Rico to be approximately 18–19 days.

Useful estimates of the serial interval between consecutive infections are important for understanding DENV transmission dynamics. It has been hypothesised that variations in the extrinsic incubation period are responsible for the timing of seasonal patterns and its importance is borne out by modelling efforts (Bartley et al. 2002; Focks et al. 1993; Lambrechts et al. 2011). Estimates of the serial interval and the distribution of spatial distances between cases occurring at that interval can be used to inform the proper extent and duration of vector control in response to an outbreak. Similarly, an accurate characterisation of the interval between successive illnesses can help to improve surveillance and sampling techniques designed to capture DENV infections across the clinical spectrum of disease and to estimate the capacity for virus population expansion; that is, Ro or basic reproductive rate. The ability to detect differences in the distribution of serial intervals between outbreaks could provide answers to other open questions in dengue research. For example, variability in extrinsic incubation period could be a component of viral fitness that drives replacement of one virus genotype by another within a serotype; that is, clade replacement (Anderson & Rico-Hesse 2006; Hanley et al. 2008; Lambrechts et al. 2012). In this research, the IKT methodology was used to examine the space–time interactions of people with hospitalised, serologically confirmed dengue illnesses.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Dengue transmission cycle
  5. Space-time analysis of DENV transmission
  6. Materials and methods
  7. Results
  8. Discussion
  9. Acknowledgements and disclaimer
  10. References

Study site and sample

The study site was the Muang District of Kamphaeng Phet Province in north-central Thailand (Figure 2). There were 206 456 registered residents of the district in 2009 (Department of Provincial Administration 2010). The district is served primarily by Kamphaeng Phet Provincial Hospital in the Nai Muang Subdistrict (capital subdistrict). Dengue illness occurs year round; however, there is a regular seasonal pattern with higher rates of illness in the wet months from June through September. The local Thai Ministry of Public Health performs insecticide spraying on notification of a dengue illness and again after 7 days (Mammen et al. 2008). The policy requires spraying houses within 100 m of the infected household.

image

Figure 2.  The study site and household locations of DENV-infected, hospitalised individuals. The black line is the boundary of the Mueang District of Kamphaeng Phet, Thailand. The grey shaded area is the Nai Mueang Subdistrict, and the Ping River is depicted in light blue.

Download figure to PowerPoint

The space-time analysis was performed using data on subjects admitted to the hospital with serologically confirmed DENV infections and virus serotype identifications by nested PCR (Lanciotti et al. 1992; Klungthong et al. 2007). Over the period from 1 January 2009 through 6 May 2011, 262 cases were identified for analysis. There were 42 DENV1 infections, 163 DENV2 infections, 53 DENV3 infections and 4 DENV4 infections (Figure 3). The household of each study participant was located using the Global Positioning System (GPS) and entered into a geographic information system (GIS). The reported date of onset of symptoms was used as the time for each case.

image

Figure 3.  The times series plot for hospitalised dengue patients included in this research. The asterisk denotes that the count for April 2011 includes five cases with onset of illness between 1 and 3 May 2011.

Download figure to PowerPoint

Statistical analysis

An incremental Knox test was carried out to examine the level of spatial clustering at sequential overlapping time intervals. This test is a modification of the widely used Knox (1964) test for space–time interaction. The Knox test statistic, X, is the number of pairs of cases that are close in both space and time. The statistic is calculated as

  • image(1)

where N is the number cases, inline image is equal to 1 if cases i and j are close in space, and 0 otherwise, inline image is equal to 1 if cases i and j are close in time and 0 otherwise, and s and t represent pre-specified spatial and temporal distances. The Knox test statistic amounts to the number of cases occurring within a given distance and time interval apart. For example, when s is equal to 100 m and t is = 7 days, the resulting value would indicate the number of pairs of cases occurring within 100 m and 7 days of each other. The resulting statistic is useful for testing the null hypothesis of no space–time interaction vs. the alternative hypothesis of a contagious or focal process. The test does not, however, indicate which time intervals within t that the spatial clustering occurs. The significance of the observed value is most often determined using the randomisation procedure suggested by Mantel (1967). In this Monte Carlo significance test, the N occurrence times are randomly permuted among the cases to estimate the distribution of the statistic under the null hypothesis of no space–time interaction. The interval Knox statistic is formulated as

  • image(2)

where inline image is equal to 1 if cases i and j occur t units apart, and 0 otherwise. This modification considers only pairs of cases that occur at a given time interval, for example, 7–9 days. When calculated for a series of temporal intervals, the IKT results indicate the intervals when spatial clustering of cases occurs.

For this analysis, we employed a series of overlapping time intervals, t, to account for uncertainty in date of illness onset. The first test examined cases occurring on the same day or 1 day apart. The second case interval is 0–2 days apart, and the series continued with 3-day overlapping intervals to cases 39–41 days apart. Based on previous observations of the focality of DENV transmission in the study area (Mammen et al. 2008) and to ensure that the results would be robust to choice of distance cut-off, the analysis was performed for distances of 100, 200, 300, 400 and 500 m. Only pairs of hospitalisations resulting from infection with the same serotype were considered. Simulation studies have shown that space–time interactions tests are liberal in rejecting the null hypothesis (Kulldorff & Hjalmars 1999), and therefore, the conservative Bonferroni adjustment was employed to correct for multiple testing (Shaffer 1995). The test results are also reported as the epidemiological meaningful notion of excess risk; in this case, owing to the space–time interaction present in the observed data set (Diggle et al. 1995). Excess risk is calculated as the ratio of the observed statistic divided by the permutation mean inline image.

  • image(3)

Excess risk was computed for the same time intervals and distances as above. A total of 100 000 permutations under the null hypothesis were completed to estimate inline image and critical values of the distribution for each test.

Ethical approval

The study protocol and consent forms were approved by Upstate Medical University’s Institutional Review Board, the Thailand Ministry of Public Health Ethics Committee, and the Walter Reed Army Institute of Research’s Institutional Review Board.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Dengue transmission cycle
  5. Space-time analysis of DENV transmission
  6. Materials and methods
  7. Results
  8. Discussion
  9. Acknowledgements and disclaimer
  10. References

The results of the IKT analysis are displayed in Figure 4. The temporal intervals with the strongest spatial clustering are 2–5 days and 15–17 days. These results are consistent for all five distances. For the shortest distances, 100 and 200 m, there are no pairs of hospitalisations observed at the 6–8-day interval, which has the lowest excess risk at all distances examined. There is also clustering observed, although to a lesser extent, at intervals of 9–13, 19–22, 23–27 and 30–34 days. The excess risk of hospitalised dengue because of the space–time interaction of the transmission process is shown in Figure 5. This graph shows that the temporal pattern of clustering observed at 200 m is similar to the pattern observed at 100 m. The largest estimated excess risk attributable to space–time interaction of hospitalisations was 15.2 and occurred at the 200 m distance and 15–17-day interval. This means that the probability of being hospitalised for dengue at this temporal interval owing to spatial proximity is 15.2 times that which would occur at a randomly selected case location in the study area.

image

Figure 4.  Space-time clustering results. The bars indicate the number of hospitalised dengue patients with the same infecting serotype that resided within the stated distance of each other. The results are shown for distances of 100, 200, 300, 400 and 500 m. The black dashed line indicates the expected number of cases under the null hypothesis of no space–time interaction. The dashed red line indicates the 99.98th percentile of the permutation distribution. Bars exceeding this line have a nominal significance level <0.0002, and a Bonferroni adjusted significance level <0.05.

Download figure to PowerPoint

image

Figure 5.  The excess risk of dengue patient hospitalisation because of the space–time interaction between cases. Spatial and temporal distances are from a previous hospitalised case.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Dengue transmission cycle
  5. Space-time analysis of DENV transmission
  6. Materials and methods
  7. Results
  8. Discussion
  9. Acknowledgements and disclaimer
  10. References

This research demonstrates that spatiotemporal methods can be used to detect temporal patterns of excess risk within a focus of DENV transmission. These patterns are detectable using only hospitalised illnesses, which represent a small proportion of total infections and are indicative of important temporal intervals in the diffusion of DENV. The spatial clustering of hospitalisations over short-time intervals (2–5 days) confirms the results of previous studies. There are several mechanisms that are potentially responsible for this pattern. The first is that multiple mosquitoes became infected at approximately the same time by feeding on one or more infectious hosts. They have survived the extrinsic incubation period and infected subsequent individuals that have onset of symptoms a few days apart. Second, A. aegypti often take multiple blood meals from multiple hosts with each egg laying cycle (Scott et al. 1993a,b; Gould et al. 1970; Macdonald 1956). This multiple feeding behaviour increases the likelihood that multiple people are infected by a single infectious vector over a short-time period. Third, the gonotrophic cycle of A. aegypti can be as short as 3 days (Christophers 1960), although it is assumed to last longer (Watts et al. 1987). This means that cases occurring within a 3–5-day period could be the result of the same infected mosquito(s) feeding on nearby hosts during consecutive egg laying cycles. The combination of these factors is likely responsible for the focal and sometimes explosive nature of DENV transmission, but the relative contribution of these three aspects of DENV transmission cannot be untangled as part of the current analysis.

The strongest spatial clustering occurred at the 15–17-day interval, and this likely indicates the most frequent serial interval between successive human DENV infections. This estimate fits with prior knowledge regarding the length of the intrinsic and extrinsic incubation periods (Lambrechts et al. 2011; Nishiura & Halstead 2007). We would also expect that the strongest clustering signal is associated with the most direct link in the chain of transmission. The spatial extent of the clustering is consistent with the normally limited migration of female A. aegypti, although we cannot determine from our analysis the relative contribution of virus movement in flying infected mosquitoes vs. movements by humans. Of the 14 pairs of hospitalisations within 500 m of each other at the 15–17-day interval, 13 were within 200 m and 8 were within 100 m of each other. The significant result was not because of temporally or spatially isolated events. The first case for each of the pairs observed at this interval occurred from March through October, and the pairs of infections resulted from the three predominant serotypes (4 DENV1, 7 DENV2, and 2 DENV3). The general trend of declining risk held for nearly all time intervals; however, the number of pairs observed at the 3–5-day interval continued to increase at larger distances (from 10 pairs at 200 m to 17 pairs at 500 m). The significant excess risk at longer intervals shows that transmission may persist within a small area. The high excess risk at the 32–34-day interval seems to indicate that two complete transmission cycles can be completed within a small area.

There are several limitations to this study. The first is that we used residential distance as a proxy for relatedness of hospitalisations. Previous work in the same region has demonstrated the focal nature of transmission, but it is possible that geographically proximate infections, even owing to the same serotype, were not closely linked in the chain of transmission. DENV infections may also be acquired at locations far from the place of residence (Mondini et al. 2009). Data on patient movement patterns were not collected, and so there is no way to link cases associated with transmission at schools, workplaces or other activity spaces. With the advent of inexpensive full genome sequencing and deep sequencing, it may soon be possible to use genetic similarity of viruses on its own or in concert with geographic information as a more reliable measure of the relatedness of infections. Another concern is that the current analysis employed only hospitalised illnesses. Studies have indicated that virulence varies by serotype and strain of infecting virus (Hesse 2007; Nisalak et al. 2003). Host-virus interactions have also been implicated in disease severity (Endy et al. 2004). If the lengths of incubation periods are correlated with clinical severity, then the results presented here may not hold for milder DENV infections. Hospitalised cases comprise a small proportion of total DENV infections (Endy et al. 2011), and very few nearby pairs were observed in the cooler portion of the year. A more sensitive surveillance system might be required to perform a similar analysis in periods of low transmission or low virulence.

The strength of space-time permutation tests is that they require few assumptions other than that nearby cases are more likely to be related than cases that occur far away. Special populations and resource intensive techniques are not required, nor is it assumed that cases are directly linked in the chain of transmission. One drawback of this type of test is that the results may be subject to population shift bias (Kulldorff & Hjalmars 1999). The effect of population shift was limited here by examining only temporal intervals that are short relative to the overall study period. The results of simulation studies indicate that the inclusion of data on the susceptibility of individuals within a cohort improves the power to detect significant interaction (Aldstadt 2007). This type of information is, however, rarely available. Despite reduced power, the same simulation studies found that the permutation approach employed here provided unbiased estimates of the serial interval between illnesses even when only a small proportion of total infections were observed.

This novel approach to space-time analysis revealed temporal intervals at which the homes of people hospitalised because of DENV infection were spatially clustered. These patterns in the diffusion of the infection are linked to the temporal intervals in the DENV transmission cycle. The most likely serial interval was observed as the 15–17-day period, and significant excess risk of dengue illness persisted as long as 32–34 days. The IKT methodology is a tool that can be used to better understand region-specific and outbreak-specific transmission dynamics. Systematic characterisation of differences between settings can lead to a more complete understanding of the variation of the force of transmission, guide locally appropriate dengue control efforts and constitutes valuable new information for construction and parameterisation of DENV transmission models.

Acknowledgements and disclaimer

  1. Top of page
  2. Abstract
  3. Introduction
  4. Dengue transmission cycle
  5. Space-time analysis of DENV transmission
  6. Materials and methods
  7. Results
  8. Discussion
  9. Acknowledgements and disclaimer
  10. References

This research would not have been possible without the cooperation of the administration and staff of the Kamphaeng Phet Provincial Hospital, especially Dr. Kamchai Rungsimunpaiboon, Hospital Director. We would like to acknowledge the efforts of Ms. Wilaiwan Sridadeth, Ms. Rattiya Wannawong and the staff of the Kamphaeng Phet-AFRIMS Virology Research Unit for their thorough fieldwork and careful data management. This research was funded by National Institutes of Health grants R01 GM083224 and P01 AI034533 and U.S. Military Infectious Diseases Research Programme. This research benefitted from discussions with working group members in the Research and Policy for Infectious Disease Dynamics (RAPIDD) programme of the Science and Technology Directorate, Department of Homeland Security, and the Fogarty International Center, National Institutes of Health. The opinions or assertions contained herein are the private views of the authors and are not to be construed as reflecting the official views of the National Institutes of Health, the United States Army or the United States Department of Defense.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Dengue transmission cycle
  5. Space-time analysis of DENV transmission
  6. Materials and methods
  7. Results
  8. Discussion
  9. Acknowledgements and disclaimer
  10. References
  • Aldstadt J (2007) An incremental Knox test for the determination of the interval between successive cases of an infectious disease. Stochastic Environmental Research and Risk Assessment 21, 487500.
  • Anderson JR & Rico-Hesse R (2006) Aedes aegypti vectorial capacity is determined by the infecting genotype of dengue virus. The American Journal of Tropical Medicine and Hygiene 75, 886.
  • Bartley L, Donnelly C & Garnett G (2002) The seasonal pattern of dengue in endemic areas: mathematical models of mechanisms. Transactions of the Royal Society of Tropical Medicine and Hygiene 96, 387397.
  • Beckett CG, Kosasih H, Faisal I et al. (2005) Early detection of dengue infections using cluster sampling around index cases. The American Journal of Tropical Medicine and Hygiene 72, 777782.
  • Christophers SR (1960) Aëdes aegypti (L.) the Yellow Fever Mosquito: its Life History, Bionomics and Structure. Cambridge University Press, London.
  • Department of Provincial Administration (2010) Population Statistics. Department of Provincial Administration, Available at: http://203.113.86.149/xstat/xstat.html (accessed August 31 2011).
  • Diggle P, Chetwynd A, Haggkvist R & Morris S (1995) Second-order analysis of space-time clustering. Statistical Methods in Medical Research 4, 124136.
  • Endy TP, Nisalak A, Chunsuttitwat S et al. (2004) Relationship of preexisting dengue virus (DV) neutralizing antibody levels to viremia and severity of disease in a prospective cohort study of DV infection in Thailand. Journal of Infectious Diseases 189, 9901000.
  • Endy TP, Weaver SC & Hanley KA (2010) Dengue virus: past, present and future. In: Frontiers in Dengue Virus Research (KA Hanley & SC Weaver) Caister Academic Press, Norfolk, pp. 265297.
  • Endy TP, Anderson KB, Nisalak A et al. (2011) Determinants of inapparent and symptomatic dengue infection in a prospective study of primary school children in kamphaeng phet, Thailand. PLoS Neglected Tropical Diseases 5, e975.
  • Fine PEM (2003) The interval betwen successive cases of an infectious disease. American Journal of Epidemiology 158, 10391047.
  • Focks D, Haile D, Daniels E & Mount G (1993) Dynamic life table model for Aedes aegypti (Diptera: Culicidae): analysis of the literature and model development. Journal of Medical Entomology 30, 10031017.
  • Garcia-Rejon J, Loroño-Pino MA, Farfan-Ale JA et al. (2008) Dengue virus–infected Aedes Aegypti in the home environment. The American Journal of Tropical Medicine and Hygiene 79, 940950.
  • Gould DJ, Mount GA, Scanlon JE, Ford H & Sullivan MF (1970) Ecology and control of dengue vectors on an island in the Gulf of Thailand. Journal of Medical Entomology 7, 499508.
  • Gubler DJ (2002) Epidemic dengue/dengue hemorrhagic fever as a public health, social, and economic problem in the 21st century. Trends in Microbology 10, 100103.
  • Gubler D, Suharyono W, Tan R, Abidin M & Sie A (1981) Viraemia in patients with naturally acquired dengue infection. Bulletin of the World Health Organization 59, 623630.
  • Halstead SB (1966) Epidemiological studies of Thai haemorrhagic fever 1962–1964. Bulletin of the World Health Organization 35, 8081.
  • Hanley K, Nelson J, Schirtzinger E, Whitehead S & Hanson C (2008) Superior infectivity for mosquito vectors contributes to competitive displacement among strains of dengue virus. BMC Ecology 8, 1.
  • Harrington LC, Buonaccorsi JP, Edman JD et al. (2001) Analysis of survival of young and old Aedes aegypti (Diptera: Culicidae) from Puerto Rico and Thailand. Journal of Medical Entomology 38, 537547.
  • Harrington LC, Scott TW, Lerdthusnee K et al. (2005) Dispersal of the dengue vector Aedes aegypti within and between rural communities. The American Journal of Tropical Medicine and Hygiene 72(20), 9.
  • Hesse RR (2007) Dengue virus evolution and virulence models. Clinical Infectious Diseases 44, 14621466.
  • Jacquez GM, Meliker J & Kaufmann A (2007) In search of induction and latency periods: space-time interaction accounting for residential mobility, risk factors and covariates. International Journal of Health Geographics 6, 35.
  • Kan C-C, Lee P-F, Wen T-H et al. (2008) Two clustering diffusion patterns identified from the 2001–2003 dengue epidemic, Kaohsiung, Taiwan. American Journal of Tropical Medicine and Hygiene 79, 344352.
  • Klungthong C, Gibbons RV, Thaisomboonsuk B et al. (2007) Dengue virus detection using whole blood for reverse transcriptase PCR and virus isolation. Journal of Clinical Microbiology, 45, 2480.
  • Knox EG (1964) Detection of space-time interactions. Applied Statistics 13, 2529.
  • Kulldorff M & Hjalmars U (1999) The Knox method and other tests for space-time interaction. Biometrics 55, 544552.
  • Lambrechts L, Paaijmans KP, Fansiri T et al. (2011) Impact of daily temperature fluctuations on dengue virus transmission by Aedes aegypti. Proceedings of the National Academy of Sciences 108, 7460.
  • Lambrechts L, Fansiri T, Pongsiri A et al. (2012) Dengue-1 virus clade replacement in Thailand associated with enhanced mosquito transmission. Journal of Virology 86, 18531861.
  • 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. Journal of Clinical Microbiology 30, 545.
  • Macdonald W (1956) Aedes aegypti in Malaya. II. Larval and adult biology. Annals of Tropical Medicine and Parasitology 50, 399.
  • Mammen MP Jr, Pimgate C, Koenraadt CJM et al. (2008) Spatial and temporal clustering of dengue virus transmission in Thai villages. PLoS Medecine 5, e205.
  • Mantel N (1967) The detection of disease clustering and a generalized regression approach. Cancer Research 27(20), 9220.
  • Mondini A, Bronzoni RVDM, Nunes SHP et al. (2009) Spatio-temporal tracking and phylodynamics of an urban dengue 3 outbreak in São Paulo, Brazil. PLoS Neglected Tropical Diseases 3, e448.
  • Morrison AC, Getis A, Santiago M, Rigau-Perez JG & Reiter P (1998) Exploratory disease analysis of reported dengue cases during an outbreak in Florida, Puerto Rico 1991–1992. American Journal of Tropical Hygiene and Medicine 58, 287298.
  • Neff JM, Morris L, Gonzalez-Alcover R, Coleman PH, Lyss SB & Negron H (1967) Dengue fever in a Puerto Rican community. American Journal of Epidemiology 86, 162184.
  • Nisalak A, Endy TP, Nimmannitya S et al. (2003) Serotype-specific dengue virus circulation and dengue disease in Bangkok, Thailand from 1973 to 1999. The American Journal of Tropical Medicine and Hygeine 68(19), 1202.
  • Nishiura H & Halstead SB (2007) Natural History of Dengue Virus (DENV)—1 and DENV—4 Infections: Reanalysis of Classic Studies. Journal of Infectious Diseases 195, 1007.
  • Putnam JL & Scott TW (1995) The effect of multiple host contacts on the infectivity of dengue-2 virus-infected Aedes aegypti. The Journal of Parasitology 17, ???174.
  • Reyes M, Mercado JC, Standish K et al. (2010) Index cluster study of dengue virus infection in Nicaragua. The American Journal of Tropical Medicine and Hygiene 83, 683689.
  • Sabin AB (1952) Research on dengue during world war II. The American Journal of Tropical Medicine and Hygiene 1, 30.
  • Scott TW & Morrison AC (2010) Vector dynamics and transmission of dengue virus: implications for dengue surveillance and prevention strategies. In: Dengue Virus (ed. AL Rothman) Springer, Berlin Heidelberg.
  • Scott TW, Chow E, Strickman D et al. (1993a) Blood-feeding patterns of Aedes aegypti (Diptera: Culicidae) collected in a rural Thai village. Jounal of Medical Entomology 30, 922927.
  • Scott TW, Clark GG, Lorenz LH, Amerasinghe PH, Reiter P & Edman JD (1993b) Detection of multiple blood feeding in Aedes aegypti (Diptera: Culicidae) during a single gonotrophic cycle using a histologic technique. Journal of Medical Entomology 30, 9499.
  • Scott TW, Amerasinghe PH, Morrison AC et al. (2000) Longitudinal studies of Aedes aegypti (Diptera: Culicidae) in Thailand and Puerto Rico: blood feeding frequency. Journal of Medical Entomology 37, 89101.
  • Shaffer JP (1995) Multiple hypothesis testing. Annual Review of Psychology 46, 561584.
  • Sheppard P, Macdonald W, Tonn R & Grab B (1969) The dynamics of an adult Aedes aegypti in relation to dengue hemorrhagic fever in Bangkok. The Journal of Animal Ecology 38, 661702.
  • Siler JF, Hall MW & Kitchens A (1926) Dengue: its history, epidemiology, mechanism of transmission, etiology, clinical manifestations, immunity and prevention. Philippine Journal of Science 29, 1304.
  • Stoddard ST, Morrison AC, Vazquez-Prokopec GM et al. (2009) The role of human movement in the transmission of vector-borne pathogens. PLoS Neglected Tropical Diseases 3, e481.
  • Tran A, Deparis X, Dussart P et al. (2004) Dengue spatial and temporal patterns, French Guiana 2001. Emerging Infectious Diseases 10, 615621.
  • Trpis M & Hausermann W (1986) Dispersal and other population parameters of Aedes aegypti in an African village and their possible significance in epidemiology of Vector-Borne diseases. American Journal of Tropical Medicine and Hygiene 35, 12631279.
  • Vaughn DW, Green S, Kalayanarooj S et al. (1997) Dengue in the early febrile phase: viremia and antibody responses. Journal of Infectious Diseases 176, 322.
  • Vaughn DW, Green S, Kalayanarooj S et al. (2000) Dengue viremia titer, antibody response pattern, and virus serotype correlate with disease severity. Journal of Infectious Diseases 181, 2.
  • Vazquez-Prokopec GM, Kitron U, Montgomery B, Horne P & Ritchie SA (2010) Quantifying the Spatial Dimension of Dengue Virus Epidemic Spread within a Tropical Urban Environment. PLoS Neglected Tropical Diseases 4, e920.
  • Watts D, Burke DS, Harrison BA, Whitmire RE & Nisalak A (1987) Effect of temperature on the vector efficiency of Aedes aegypti for dengue 2 virus. American Journal of Tropical Hygiene and Medicine 36, 143152.
  • WHO (2009) Fact sheet no.117, Dengue and dengue haemorrhagic fever. http://www.who.int/mediacentre/factsheets/fs117/en/ (accessed September 1 2011).