Dengue infections in India: A meta-analysis

TL;DR

This paper analyzes dengue infection trends in India from 2014 to 2023 using 127 studies to understand its health impact and suggest early detection strategies.

Contribution

The study provides a comprehensive meta-analysis of dengue trends in India over a 10-year period.

Findings

Significant heterogeneity was found in reported dengue outcomes across studies.

Early detection of dengue is recommended to improve public health strategies in India.

Abstract

The escalating impact of dengue infection on health and mortality is a critical global issue. Therefore, it is of interest to assess
the current trends of dengue infection in India. We searched through a wide range of internet databases to gather comprehensive studies
on the incidence, prevalence, sero-prevalence, cost effectiveness and mortality rate of dengue infection in India from 2014 to 2023 (10
years) in a total of 127 studies. Analysis shows significant heterogeneity (diversity) in reported outcomes (p-values < 0.001). Thus,
public health strategies should include early detection of dengue infection in our country.

Full text

Background:

Dengue is caused by an arbovirus of the Genus Flavivirus and Family Flaviviridae, is one of the most prevalent, fast-spreading vector-borne diseases impacting people [1]. As a result, research has shown that dengue disease may be clinically characterized as either mild dengue, dengue with or without warning signals, or severe dengue [1, 2]. According to a study, an estimated 105 million infections occur worldwide every year, only 51 million of which are symptomatic, making it a major public health issue [3]. Due to increasing worldwide travel and the geographical expansion of the Aedes vector mosquitoes, dengue virus are transmitted on all major continents, with new cases occurring and spreading to formerly non-endemic locations [4]. The primary dengue infection is presumed to provide permanent sterilizing immunity against homologous serotypes; however, exceptions exist in human and animal experimental investigations [5, 6]. Secondary infection (SC) with an un-encountered serotype often leads to classical dengue fever (fever) and is linked to a heightened risk of severe sequelae [7, 8]. This is a significant risk factor for the heightened severity of dengue fever via the antibody-dependent enhancement (ADE) pathway [9]. A second dengue fever occurring within two years after the first infection is likely to be an asymptomatic infection, as shown by the neutralizing antibody titer [10]. Therefore, it is of interest to assess dengue fever in India with the help of systematic review (SR) and meta-analysis (MA).

Material and Method:

Protocol development:

In the present manuscript, written according to the PRISMA checklist, [11] only the scientific evidence of dengue infection current Trent in India was investigated. This SR protocol was a priori registered in The International Prospective Register of SR (Registration No: CRD42024552341).

Search strategy, Databases & Selection criteria:

We have searched in electronic databases such as Cochrane Library, Medline, Web of Science (WoS), PubMed, Scopus & Google Scholar for publications published between January 2014 and December 2023. Appendix I: Search Strategy contains all of the search strategy's details. We have specifically used date/year as a filter to search three databases i.e. (PubMed, Scopus/Elsevier and Embase) from May 24-27, 2024. The Covidence application was used to screen abstracts.

Inclusion criteria:

[1] All studies conducted in India on this topic regardless of their design, purpose or population.

[2] Incidence

[3] Prevalence

[4] Number of cases

[5] Mortality

[6] Burden

[7] Complications

[8] Virus serotype details / seroprevalence

2 reviewers independently collected data from selected papers using a predefined data extraction form. Any discrepancies in it were resolved through consensus. The information that was extracted from studies includes year of publication, study setting, location, period, laboratory investigations, number of suspected patients tested & found positive, the age distribution of cases and details of dengue serotypes as shown in Table 1, Table 2, Table 3, Table 4, Table 5 to Table 6 (dataset I -VI).

Data extraction & Review synthesis:

3 reviewers carried out the initial screening. The collected literature was first searched to remove duplicates before being entered into Rayyan software [132]. After that, the titles and abstracts were screened. In 2nd screening phase, 3 reviewers evaluated the selected papers based on their compliance with the eligibility standards. While the 2, independently shortlisted studies that met the design, participant and result requirements. Disagreements were resolved by discussion and, if necessary, the involvement of a 3rd reviewer. Using a pre-designed data extraction form in Microsoft Excel, 3 reviewers independently gathered details from the selected research. Initially, the search results were imported into Mendeley software (Version 1.19.6) where duplicate records were removed.

The outcome measures were:

[1] The prevalence of laboratory-confirmed dengue infection among clinically suspected patients in the research area, as reported in hospital/laboratory or community-based investigations during outbreaks.

[2] Seroprevalence of dengue in the study population dengue fever conditions, dengue severity and Mortality rate among dengue patients those were confirmed in labs.

[3] Primary and secondary infections present.

[4] Cost of illness/burden which included reported direct and indirect costs associated with dengue hospitalization.

[5] The non-structural protein-1 (NS1) antigen, immunoglobulin M (IgM) antibodies against dengue virus, haem-agglutination inhibition (HI) antibodies against dengue virus, RT-PCR positivity, or virus isolation was used to diagnose acute dengue infection in the clinically suspected patients. The measurement of IgG or neutralizing antibodies against the dengue virus was used to determine the seroprevalence of dengue.

Quality/Risk of bias assessment:

We utilized a modified version of the Joanna Briggs Institute (JBI) appraisal checklist for assessing prevalence data [133], along with key components from the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist [134] to gauge potential bias. Our primary criteria for bias assessment included outcome variables, laboratory testing procedures and participant selection strategies (refer to Supplementary file S2 Appendix). 2nd reviewers independently evaluated bias risk, resolving any disagreements through discussion. In cases of unresolved disputes, the perspective of a 3rd reviewer was sought and any disagreements were resolved. When needed, the viewpoint of the 3rd reviewer was sought.

Statistical analysis:

Using the single user licenced version of STATA 18.5 StataCorp LLC, Texas, USA, software and R-Studio analysis was carried out. The proportions from the combined data were shown along with their 95% confidence intervals (CI). Heterogeneity was assessed using an I2-test, where values below 25% indicated mild heterogeneity, values between 25 and 75% indicated moderate heterogeneity and values over 75% indicated significant heterogeneity [15, 16]. Based on the inverse variance approach for weighting, the Der-Simonian-Laird method for a random-effects model was used to compute the total pooled prevalence. Both the pooled estimates for the general and subgroup analyses and the study-specific estimates for each participant were shown using forest plots. To further demonstrate publication bias, a funnel plot was made.

Results:

Initially, we searched 6582 published articles in various electronic databases such as PubMed-2281, Ovid/Medline-47, Web of Science-4037 and Google Scholar-217 published. This was later on narrowed down to 999 unique articles after duplicate removal over the last 10 years. Following titles and abstracts screening, 613 articles were excluded, leaving 386 articles for full-text evaluation. This resulted in 127 studies being selected for analysis [17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139-140] as shown in (Figure 1 see PDF).

Prevalence/proportion of laboratory dg cases & outbreak:

The clinically suspected patients are provided by 78 out of the 127 published studies included in this synthesis. This comprised 8 studies reporting outbreak investigations and 71 studies conducted in hospital or laboratory settings. A proportion of the studies that the hospital validated were conducted at the time; that the affected areas were going through an outbreak. The data of laboratory-confirmed cases by month were supplied by 32 research (40.5%) out of the 79 studies that reported a proportion of dengue cases; the majority of these studies (n = 53, 67%) indicated increased dengue positivity throughout the rainy seasons, particularly from July to October. The majority of the forty-seven investigations identified acute dengue infection using a single test, as follows: detection of the NS1 antigen = 1, virus isolation = 1, RT-PCR (Real-Time Reverse Transcription - Polymerase Chain) = 7, haem-agglutination inhibition antibodies = 2 and IgM antibodies = 36. The other studies employed multiple tests.

Case definitions used:

While discussed about case studies their; we took assistance of WHO (World Health Organizations), NVBDCP (National Center for Vector Borne Diseases Control) & AFI (Acute Febrile Illness) case definitions. Out of 79 studies during hospital settings majority n=53 were clinical suspected dengue followed by n=13 WHO case definition, n=9 AFI case definition and the remaining studies n=4 were used NVBDCP case definitions respectively. Both hospital confirmed dengue study and showed similarly, among 71 hospital confirmed dengue cases n=51 were clinical suspected dengue followed by n=9 WHO case definition, n=9 AFI case definition and the remaining studies n=2 were used NVBDCP case definitions respectively. Among the reported outbreaks, investigators used n=4 WHO case definition, n=2 AFI case definition and the remaining studies n=2 were used NVBDCP case definitions respectively.

Dengue proportion in India:

Based on testing of 206783 clinically suspected individuals from 78 studies, the overall estimate of the prevalence of laboratory-confirmed dengue infection in the random effects model was 39.4% (95% CI: 35.6%-44.67%) as shown in (Figure 2 see PDF). The heterogeneity was assessed by Hedge g statistics. The heterogeneity overcomes by using random effect model as shown in (Figure 3 see PDF). The publication biased(PB) was assessed by using funnel plot, some asymmetry observed because individual study had different proportion and this was directly impacts on shifting the points on funnel to outside but the both the side almost normality hence in our study there was no publication bias was reporting as shown in. The prevalence reported by the 79 studies showed significant heterogeneity (LRT p<0.001). In comparison to hospital-based surveillance (HBS) studies (40%, 95% CI: 35-44), the prevalence of laboratory-confirmed dengue infection was nearly identical in studies reporting outbreaks (OB) or hospital-based surveillance studies during outbreaks (39%, 95% CI: 34-44).

Age distribution:

Data was available for 30 out of 127 studies on laboratoryconfirmed DG cases. The overall average age of confirmed DG patients in this study was 24.47 years; with a standard deviation of 9.22 years with age range was 7 to 36 years as shown in (Figure 4 - see PDF).

Dg-FV & Dg-S proportion:

31 studies provided information on dengue fever, while 32 studies provided information on dengue symptoms. The majority of the research (n = 19, 59.38%) utilized the WHO 1997 classification, while the remaining studies (n = 3, 9.38%) employed the WHO 2007 classification. Additionally, for dengue fever condition and severity, (n = 6, 18.75%) used the WHO 2009 classification, whereas 4 studies (12.5%) used the WHO 2012 classification. It was reported that between 31% and 100% of laboratory-confirmed patients had dengue fever. According to the random effect model, 75% (95% CI: 67-82) of laboratory-confirmed studies exhibited dengue fever overall. The Hedges g-Method (HD-M) was used to estimate the random effect model, indicating no heterogeneity as shown in (Figure 5 see PDF). Bias in publications observed and depicted that higher prevalence publications were more side. On the other hand, among patients with laboratory-confirmed, the reported percentage of dengue symptoms cases varied from 2% to 69%. In the random effect model (REM), the total percentage of dengue symptoms across laboratory-confirmed studies was 25% (95% CI: 19-31). The data on dengue symptoms showed no evidence of heterogeneity as shown in (Figure 6 see PDF).

DG Mortality (MT) in India:

In the provided research, 48 provided information on MT rate of DG, It was reported that between 0% and 9% of LB-CN patients had DG-FV. According to the REM, 1% (95% CI: 1-2) of LB-CN studies exhibited DG-FV overall. The HD-M was used to estimate the REM, indicating no HTG. Bias in publications observed and depicted that higher prevalence publications was more side, The removal of the study with greatest weight in each LB-CN test of DG disease did not change the results.

Pm-if & SC among dg cases in India:

A comprehensive analysis of 15 studies [31, 37, 48, 59-60, 71, 78, 81- 82, 89, 104- 105, 115, 124] enabled the categorization of LB-CN-DG-IF into PM and SC. The prevalence of initial DG-IF varied widely ranges from 32% to 97% across the studies. Overall, PM-DG-IF accounted for 77% of LB-CN cases (95% CI: 65-87). Meanwhile, SC-DG-IF occurred in 23% of LB-CN cases (95% CI: 13-35), with a range of 3% to 68% across the studies.

PB-BA & sensitivity statistics (SS-ST):

There was no indication of publication bias in the dengue prevalence estimates from hospital-based studies with LB-CN cases, outbreaks & SP according to analysis utilizing funnel plots and the HD approach. The estimates of dengue severity and fatality did, however, reveal a substantial publication BA, with publications demonstrating higher prevalence being more likely to be published. However, sensitivity analysis showed that the pooled percentages of research results held steady, suggesting the estimates' resilience. The removal of the study with greatest weight in each dengue cases LB-CN did not change the results.

Discussion:

The analysis primarily drew on data from HB and laboratorybased surveillance studies, as well as reports from investigations into dengue outbreaks. There have been more than 10 million reported cases of dengue along with over 5,000 dengue-related deaths across 80 countries. The Pan American Health Organization (PAHO) region has reported the majority of cases, with over nine million cases. Within the PAHO region, Brazil has reported the highest number of cases (over eight million), followed by Argentina, Paraguay, Peru and Colombia. In Europe, imported cases from endemic areas have been reported in Germany, Italy and France, but no locally acquired cases have been reported.

Dengue circulation has also been reported in the Southeast Asia and Western Pacific regions, as well as in Africa. It concentrated on their operations, implementation and structure. The WHO had set aggressive goals to cut dengue-related mortality by 50% and morbidity by 25% along with burden by 2020 [135-136]. A recent study in Brazil found a significant disparity in the infection rates between wealthy and disadvantaged youth. Specifically, the study revealed that 60% of young people from disadvantaged backgrounds were infected, which is three times the rate of their wealthier peers and our study also found similar kind of results where average age was 24.4 years [137]. Overall, 127 studies with a total of 3Lacs population were covered for study of dengue disease in our country. Viral assays are used in laboratories to confirm dengue infection (RNA detection by RTPCR, NS1 antigen detection by ELISA) [138]. The overall prevalence of dengue disease in our India based on testing of 206783 clinically suspected individuals from 79 studies, the overall estimate of the prevalence of laboratory-confirmed - dengue infection in the random effect model was 39.4% (95% CI: 35.6%-44.67%) According to a study, the overall prevalence of dengue in country like India based on testing 206783 clinically suspected individuals from 79 different studies was 39.4% [139].

There are also research gaps in India's understanding of dengue epidemiology and the fact that different types of the dengue virus are still being spread. These factors show that dengue is still a major public health issue in India. The high percentage of dengue-positive cases, severity and case mortality in India are all indicators that dengue continues to be a significant public health concern in the country. As a consequence of this, it is required to undertake community-based cohort studies that are wellstructured and cover a variety of geographical locations of the country in order to offer reliable estimates of the age-specific incidence of dengue fever in India [140].

Conclusion:

DG-FV remains a pressing public health issue in India, as indicated by its high incidence, severity and mortality rates, as well as the circulation of multiple virus serotypes. To better comprehend the epidemic, we suggest conducting in-depth research, including community-based studies across various regions to determine age-specific incidence rates. Alternatively, a nationwide survey could be undertaken to determine age-specific seroprevalence rates, which also includes targeted studies in different geographic areas in India.

Limitation:

[1] We have restricted our search to quantitative sides which might be neglected towards qualitative enrichment of variables

[2] We considered peer-reviewed journals database from certain articles, which lead to exclusion of government registries data as a grey literature that could provide other aspects of the picture too.

Future research:

We should implement active surveillance systems, scaling up vector control measures, enhance more public awareness & education and finally, strengthen the prevention strategies.