Geographic information system-aided evaluation of epidemiological trends of dengue serotypes in West Bengal, India

Dengue fever (DF) is an acute viral infection re-emerging rapidly, resulting in increased mortality and morbidity in tropical as well as subtropical regions worldwide. In the last two decades, an 8-fold increase in dengue cases with approximately 100-400 million active infections has been reported globally1. Of these, an estimated 96 million dengue infections manifest clinically, putting 3.9 billion world’s population at risk1. The first widespread epidemics of dengue cases occurred in Delhi and Lucknow, India, in 1996, before spreading across the rest of the country23. Dengue virus (DENV) is a single-stranded positive-sense RNA virus and belongs to the family of Flaviviridae, characterized by four antigenically distinct serotypes (DENV-1 to 4)4. DENV is mainly transmitted by the bites of female Aedes aegypti mosquito and Aedes albopictus (to a lesser extent), found in parts of southern and eastern India5. DENV infection causes a spectrum of diseases ranging from asymptomatic or symptomatic febrile DF to severe dengue (SD) (dengue shock syndrome dengue haemorrhagic fever)6. Infection with one particular serotype reportedly provides lifelong immunity against that serotype; however, protective immunity to other serotypes reportedly lasts only for 3-4 months7. Furthermore, secondary dengue infection with a different serotype is a major risk factor for causing SD8.

West Bengal, being the 4^th most populous State of India, encounters several epidemic outbreaks and deaths every year9. Reports from the National Vector Borne Disease Control Programme (NVBDCP) in 2022 recorded the highest number of dengue infection cases in West Bengal, amounting to 67271 and 30 death cases10. Co-circulation of different serotypes of DENV has been reported over the past 50 years from different parts of India11. Since the first reported DHF in Kolkata (erstwhile Calcutta) in 1963, the city has become a dengue-endemic zone with increasing dengue cases12. In recent years, the emergence of SD cases has intensified the public health challenge. SD can lead to life-threatening complications, including severe bleeding, plasma leakage and organ failure13. Considering the consequences of secondary dengue infection, this retrospective study was undertaken to identify the predominant serotypes co-circulating in West Bengal and address the changing serotypic shift of the DENV with the aim of understanding the transmission dynamics of dengue better, a geographic information system (GIS)-aided analysis was applied to observe the spatiotemporal distribution of DENV serotype in search of dengue endemic zones. The combination of DENV molecular serotyping and mapping their distribution can uniquely predict impending dengue epidemics and the burden of dengue virus on public health.

Material & Methods

This observational study was undertaken at the ICMR-NICED Virus Laboratory, Indian Council of Medical Research-National Institute of Cholera and Enteric Disease (ICMR-NICED), Kolkata after obtaining ethical clearance from the Institutional Ethical Committee. A written informed consent was obtained from all the study participants prior to sample collection. Individuals not providing consent or co-infected with other diseases were excluded from this study.

Study area: West Bengal (22.9868° N, 87.8550° E) (Fig. 1) is situated in eastern India along the Bay of Bengal and has over 91 million inhabitants14. The climate of West Bengal is tropical and humid and during monsoon season, heavy precipitation lasts from June to October. District and sub-divisional hospitals receive a large number of blood samples for dengue non-structural protein 1 (NS1) testing during this period. For this study, samples were received from 20 districts of West Bengal, i.e. North 24 Parganas, South 24 Parganas, Paschim Midnapore, Purba Midnapore, Hooghly, Howrah, Kolkata, Purba Bardhaman, Paschim Bardhaman, Bankura, Birbhum, Purulia, Nadia, Murshidabad, Malda, Jalpaiguri, Alipurduar, Uttar Dinajpur, Dakshin Dinajpur and Darjeeling. Kolkata is the capital of West Bengal and has the highest population density of 24,306/km² among the districts. A convenient sampling strategy was used for sample collection. As the extent of dengue outbreak varies from year to year, hence dengue samples were received accordingly.

Close

Fig. 1

Map of West Bengal showcasing the districts for dengue NS1-positive sample collection during the study period of 2015-2019.

Export to PPT

Sample & data collection: Two millilitres of blood samples were collected in ethylenediaminetetraacetic acid vial from dengue NS1-positive individuals with ≤5 days of fever. Sampling febrile illness and seroprevalence groups should ideally be representative of the resident population15. Therefore, demographic data including age, sex residential address etc., was obtained from the study participants at the time of taking consent. A total of 3915 samples were collected from different district and sub-divisional hospitals from July 2015 to December 2019. Samples were transported to the ICMR-NICED virus laboratory while maintaining a cold chain and subjected to dengue NS1 enzyme-linked immunosorbent assay (Pan Bio, Brisbane, Australia). On receiving, samples were rechecked for dengue NS1 sero-reactivity and confirmation for dengue early phase infection suitable for viral serotyping. Among the 3,915 NS1 sero-positive individuals, 3,384 samples were subjected to dengue molecular serotyping and 531 samples were excluded due to haemolysis or insufficient sample quantity.

Viral RNA was extracted from dengue NS1 sero-positive serum samples as per the manufacturer’s protocol using QIAamp Viral RNA Mini Kit (Qiagen, Hilden, Germany). RNA was stored at −80°C for molecular serotyping of the dengue virus and further use.

Molecular serotyping of dengue virus by reverse transcriptase PCR: Detection of DENV was done by targeting the C-prM region of the DENV genome by multiplex semi-nested RT-PCR, a DENV serotype detection method of Lanciotti et al16 in 1992. The initial RT-PCR round utilized universally conserved D1 (forward) and D2 (reverse) primers, demonstrating homology across all four serotypes. Subsequently, serotype-specific amplification of DENV was conducted using 1:100 times dilution of amplicons of first-round RT-PCR as the template. For the second-round semi-nested PCR, the primers employed included D1, TS1, TS2, TS3, and TS416. Expected PCR amplicon sizes were 485 bp, 119 bp, 290 bp and 392 bp for DENV-1, DENV-2, DENV-3 and DENV-4, respectively. The amplified nested PCR products were separated on 1.5 per cent agarose gel, stained with ethidium bromide and observed using Gel Documentation System (Bio-Rad, USA).

GIS mapping for heatmap analysis, Getis-Ord Gi^* hotspot cluster analysis and spatial distribution of dengue virus (DENV) serotypes: The latitude and longitude of each participant’s residential address were transformed into Universal Transverse Mercator (UTM) units for plotting on the map. As the study focuses on the shift of dominance of DENV serotypes, the maps were based on the received NS1 seropositive cases to visualize the serotype distribution in West Bengal from 2015 to 2019 and where most NS1-positive cases were centred. A heatmap was generated to depict the intensity of dengue NS1 infections over the five year study period. Getis-Ord Gi^* hotspot cluster analysis was performed using 500 m as a distance threshold to identify high concentrations of clusters. To locate which areas are prone to secondary infection because of the changing prevalent serotype, vector analysis was done to depict serotype distribution (data not shown). For the development of maps, ‘ESRI India’s ArcGIS ArcMap 10.7’ and ‘QGIS version 3.16.15’ were used. WGS 84/UTM zone 45N was used for Geo referencing and the map of West Bengal was obtained from Wikimedia Commons17.

Statistical analysis: Categorical variables were summarized as percentages for analysis of dengue NS1-positive cases among the study population. The significance of dengue NS1-positive infection in the male and female populations was calculated using Pearson’s Chi-square test. To evaluate the increasing or decreasing trend as well as the significance of dengue NS1 cases over time across different age groups, Chi-square for linear trend in proportion was applied. Before hotspot analysis, spatial autocorrelation (Moran’s I statistics) was carried out to denote if the nature of the pattern was uniformly or randomly clustered based on fixed distance band parameters.

Results

Demographic and seasonal distribution of dengue NS1-positive cases: Of the 3,915 dengue NS1 seroreactive serum samples received, 3,384 samples were processed and 3,249 tested positive for DENV RNA during the five years of the study period (Table I). Some NS1 samples were RNA negative due to poor maintenance of the cold chain, which resulted in RNA degradation and due to false NS1 positivity with other viral infections. Gender-wise distribution of the study population showed that the number of infected males was higher than females (M: F ratio 2:1) (P<0.05) (Fig. 2). Age-based distribution of dengue NS1 cases indicated that those adults belonging to the age group of 21-30 yr (28.63%) age groups were the most susceptible followed by adolescents belonging to the 11-20 yr (21.38%) age group. A significant association was observed between NS1 seropositivity and age groups. An increasing linear trend of NS1 seroprevalence till 21-30 yr age group (Z=18.999, P<0.0001) was followed by a decreasing trend in the older population (Z=−10.691, P<0.0001) (

Close

Fig. 2

Gender-wise distribution of dengue NS1 seroreactive cases among the study population during 2015-2019.

Export to PPT

Close

Fig. 3

Distribution of dengue NS1-positive cases across monsoon and post-monsoon seasons of 2015-2019.

Export to PPT

Prevalent dengue serotypes and co-circulating of multiple serotypes: Co-circulation of all four serotypes was observed during seasonal dengue infections in different districts of West Bengal from 2015-2019 (Fig. 4). In 2015, DENV-3 was the predominant serotype (63.54%) followed by DENV-2 (18.95%), DENV-1 (13.88%) and DENV-4 (3.63%), whereas, in 2016, the major circulating serotype was DENV-1 (52.79%) followed by DENV-2 (26.26%), DENV-4 (11.56%) and DENV-3 (9.39%). In 2017, a sudden emergence of DENV-2 (73.47%) occurred as a prevalent serotype, followed by DENV-4 (14.72%), DENV-1 (6.87%) and DENV-3 (4.94%). DENV-2 continued to circulate till 2018 all over the State with high prevalence (76.04%) along with other co-circulating serotypes, DENV-3 (15.88%), DENV-1 (4.19%) and DENV-4 (3.89%). In 2019, the scenario changed noticeably, along with DENV-2 infection, the percentage of other circulating serotypes increased DENV-3 (28.13%), DENV-1 (18.27%) and DENV-4 (6.45%), but the major circulating strain remained DENV-2 (47.15%) with increased morbidity and mortality.

Close

Fig. 4

Distribution of co-circulating dengue virus serotypes in this study during 2015-2019.

Export to PPT

GIS mapping of dengue NS1 cases and the co-circulating dengue virus (DENV) serotypes: Within the survey area, the heatmap (Fig. 5) correlates with the Getis-Ord Gi^* hotspot cluster maps of 2015-2019 (Fig. 6A-E) reasonably. The G Index of Getis-Ord Gi^* aided in identifying the calibre to which units of high values (hotspots) have grouped and it prioritized cluster formation. In this study, the highest Gi^* values were observed throughout five years for Kolkata and relatively high values for North and South 24 Parganas, Howrah and Hooghly. These were the same regions identified as the most intense for dengue NS1 cases with Kolkata being the most severely affected district, which are highlighted in red and yellow hue. Spatial autocorrelation indicated a significant clustering pattern for all study years (P values for 2015-2019 were <0.01, <0.05, <0.001, <0.05 and <0.001, respectively). Vector analysis (data not shown) helped to visualize the circulation pattern of each prevalent serotype in each year of the study period. Co-circulation of multiple DENV serotypes in each year (2015-2019) was recorded in West Bengal, with only one prevalent serotype being responsible for causing an outbreak in Kolkata or other parts of the state (Fig. 4). However, all districts did not showcase co-circulating serotypes. Plotting of DENV serotypes revealed that the maximum numbers of cases were reported from Kolkata, North 24 and South 24 Parganas, Hooghly and Howrah. During 2017, DENV-2 serotype was the most prevalent in both South and North Bengal and most of the dengue NS1 cases in 2018 and 2019 were caused by this strain. In 2019, a shift of dominance was detected in the northern part of Bengal, where DENV-3 became the predominant circulating strain. However, this pattern was altered when DENV-3 became the predominant circulating serotype in the northern part of the State (North Bengal), whereas the major circulating serotype in the southern part of the State (South Bengal) remained DENV-2 (data not shown).

Close

Fig. 5

Heatmap of dengue NS1-positive cases in West Bengal during the study period of 2015-2019.

Export to PPT

Close

Fig. 6

(A-E) Identified hotspots for West Bengal based on the received NS1 sero-reactive dengue samples during 2015-2019.

Export to PPT

Discussion

Dengue fever is an emerging and neglected tropical disease which is a major public health problem in India. Dengue infection was first reported from in Kolkata in 1824. Since then, during 1836, 1906, 1911 and 1972 major outbreaks took place, which affected about 40 per cent of the city’s population18. Due to the progressing geographic distribution of vector infestation, growing urbanization and climate change, West Bengal continues to face frequent outbreaks. In 2005, a major outbreak was caused by DENV-3 in West Bengal, after which no continuous monitoring was done19. In 2008, DENV-1 reportedly caused most dengue infections in Kolkata and its suburbs, which remained the main circulating strain till 2011. During this period, DENV-3 strains circulated in insignificant numbers until a sudden re-emergence was detected in 201220. This phenomenon resulted in a spike in hospitalization with the increased number of SD cases (DHF and or/and DSS)21. In 2013, dengue febrile cases drastically dropped, but a sudden escalation in dengue cases was reported again in 2014 and 201522. As per this study, DENV-1 replaced DENV-3 in 2016 and became a predominated circulating serotype all over West Bengal. Furthermore, the co-circulation of all four dengue serotypes was observed during the chosen study period. Although the circulation of DENV-2 was lower in comparison to DENV-3, infections caused by DENV-2 began rising from 2015-2016. It became the prevalent serotype with increased cases of hospitalizations in 2017. Since protective immunity to other DENV serotypes is considered as short-lived in humans, the proportions of co-circulating viral serotypes at any given point are highly dependent on the previous disease pattern and can influence the number of susceptible individuals to each serotype23. The prevalence of DENV-2 was again higher (76.04%) in 2018, which noteworthily dropped to 47.15 per cent in 2019 (Fig. 4). This prevalence drop can be explained by the studies by Nisalak et al24 which showed that an increase in the serotype diversity can subsequently increase disease prevalence, which can cause a reduction in the circulating serotypes. Another study by Balmaseda et al25 has shown that the predominance of clinical symptoms could also be influenced by the majorly circulating serotype. These studies, however, did not explain the risk of SD by a prior infecting serotype. A probable explanation for this might be that, after exposure to one serotype, protective immunity decreases the chances of symptomatic infection but increases the risk of severity as immunity weakens26.

The heatmap detected Kolkata and its surrounding districts (North and South 24 Parganas, Hooghly and Howrah) as the most severely affected areas. With increasing sample size in subsequent years, the picture of circulating serotypes became clearer with the aid of GIS mapping. Getis-Ord Gi^* hotspot cluster analysis confirms most dengue NS1-positive cases were clustered in Kolkata and identified it as a hotspot. This hotspot was consistent during the five years of the study period, which is in accordance with past studies referring to Kolkata as a hyperendemic zone27. As co-circulation of all serotypes has been observed in Kolkata, the city has a higher chance of dealing with DHF and/or DSS28. Another hike in DENV-1 infections could be detected in parts of South Bengal districts, i.e., Nadia, Kolkata, Hooghly and Purba Midnapore (data not shown) in 2019. A shift of dominance was also detected in the northern part of the State during 2019, where DENV-3 remained the main circulating serotype, especially in the districts of Darjeeling, Jalpaiguri and Cooch Behar despite DENV-2 being prevalent in the state. Earlier studies also reported co-circulation of all four DENV serotypes in the Northern part of the State, but no hotspot was observed29. Increased cases of secondary infections can result from the co-circulation of multiple serotypes for years and the suppression of pre-existing serotypes by the predominant serotype to re-emerge suddenly30. In support of this statement, a different study by Vicente et al31 proposed that serotype shift poses a possible threat as secondary dengue infections are more likely to occur. Contrary to several other studies from this part of India, no dual infection with multiple serotypes was observed in this study29. The major reason behind most dengue cases being reported from Kolkata is due to high population density, which also dominates the overall cases from the State each year. Previous studies from different parts of India present similar findings on the relationship between population density and dengue incidences32.

Within the viraemic population, overall males (60.21%) were majorly infected compared to females (39.79%) (P<0.05) and no linear trend was observed over five years (Fig. 2) which corroborates with past studies33. The most commonly afflicted age group was young adults of 21-30 yr (28.63%), followed by adolescents of 11-20 yr (21.38%)34 (Table II). We found an increasing linear trend in NS1 seroprevalence till the age group of 21-30, which declined significantly in the older population, indicating a negative linear trend (P<0.0001) (

The present study, over a five-year study period recorded re-emergence and co-circulation of all four DENV serotypes, with one serotype being prevalent during an outbreak all over the State. Although the reason for one dominating serotype is not yet fully understood which adds weight for more observational studies like this, we also report a change of trends in major circulating serotypes which is capable of causing frequent outbreaks in West Bengal. Along with monitoring, data representation by GIS map techniques can assist in plotting risk maps to predict future outbreaks and specific clusters of areas of the population, which are more prone to possible secondary dengue infections. Vaughn et al40 also studied that viral titre and infection with a different serotype can result in disease severity. Having prior knowledge of the locations of estimated outbreaks can help health administrators in preparedness during the dengue season. It will help them to optimize limited resources and workforce to fight the situation. Added preventive measures such as awareness campaigns and destroying vector breeding sites can be implemented before outbreak forecasting.

Despite several strengths, there were some limitations of this study. First, most of the samples in 2015 and 2016 were received from Kolkata and its adjoining districts as major dengue outbreaks were confined in this area and comparatively a smaller number of samples were received from other districts, especially from the Northern part of West Bengal. Second, to determine the epidemiological cause of rapid changes in DENV serotypes, climate change, population density, anthropogenic factors, vector capacity, entomological data and risk factors could be incorporated into the study. Including these to obtain results can improve the scope of forecast outbreaks and more predictive models can be explored.

Overall, dengue has emerged as a potential threat in West Bengal, India and the capital, Kolkata, has become an endemic centre for dengue. This study documented the co-circulation of all four DENV serotypes over five years using GIS mapping techniques. Kolkata and its adjoining districts are facing frequent dengue outbreaks and the chances of developing SD cases are increasing proportionally with changing trends in circulating prevalent serotypes. As the co-circulation of multiple dengue serotypes is responsible for the shift in dominance of circulating serotypes, it holds significant value in epidemiological studies. This information on regions where chances of DHF and/or DSS are comparatively higher can be used to develop models to predict future outbreaks. Since the therapeutics against DENV are not available in India, our best chance to combat the disease is through local surveillance of all dengue serotypes and organizing public awareness campaigns alongside cleaning of vector breeding sites.