Dengue type 4 in Rio de Janeiro, Brazil: case characterization following its introduction in an endemic region

Dengue is one of the most important arboviruses in the world and the World Health Organization (WHO) estimates that between 70 and 500 million people are annually infected worldwide [1]. Dengue viruses (DENV 1–4) belong to the family Flaviviridae, genus Flavivirus [2] and based on the analysis of the envelope gene sequences [3, 4] or complete sequencing of the viral genome [5], four genotypes are characterized for DENV-4: Genotype I represented by strains from Thailand, Philippines, Sri Lanka and Japan (from Southeast Asia); Genotype II represented by strains from Indonesia, Malaysia, Tahiti, the Caribbean and the Americas; Genotype III represented by recent strains in Thailand that were different from the others; and Genotype IV represented by wild strains of Malaysia [3‐5].

After 63 years without reported cases in Brazil, reinfestation by vector Aedes aegypti in the 70’s led to epidemics in Boa Vista, Roraima in 1981–1982 [6]. However, dengue became a public health problem in the country when the DENV-1 was identified in the serum of patients in an epidemic in the state of Rio de Janeiro (RJ) in 1986 [7]. The introduction of DENV-2 in 1990, also in the state of RJ [8], led to an increase in the disease severity and the first dengue hemorrhagic fever (DHF) cases were reported in the country [9]. DENV-3 introduction in RJ and in the country in 2000 caused one of the most severe epidemics reported in the country in 2002 [10, 11]. In 2007–2008, the country experienced the most severe epidemic in terms of morbidity and mortality, in addition to severe cases in children due to the DENV-2 re-emergence. RJ alone was responsible for 255,818 cases [12‐14]. In 2009, DENV-1 re-emerged in the southeast region of the country and it was the serotype detected in 50.4% of the viral isolations, displacing DENV-2 and DENV-3 [15]. In July of 2010, DENV-4 was isolated in Roraima [16], 28 years after its first detection in that same State and soon this serotype spread other States, including RJ [17]. Despite the epidemic caused by DENV-1, DENV-4 could be isolated during the disease surveillance supported by the performed laboratorial diagnosis. Due to the populations’ susceptibility to this newly introduced serotype, explosive epidemics in the country were a real threat. Furthermore, despite being known as a mild serotype, the impact of the emergence of DENV-4 in an endemic region where the other three serotypes were circulating was unknown. This study aimed to analyze the epidemiology of DENV-4 cases occurring in Rio de Janeiro after this serotype introduction, the role of the different laboratory methodologies for case confirmation, the association of viremia and NS1 antigenemia with disease severity and the surveillance of the circulating genotype.

A total of 3727 dengue suspected cases from the state of RJ was analyzed from January 2011 to December 2013. After the confirmatory laboratorial diagnosis, 1593 cases were confirmed as dengue positive, and 705 (44.2%) due to DENV-4 infections. Not all analysis, such as those by gender, age, immune response and disease severity were performed in all 705 DENV-4 due to the lack of information in some cases.

The highest percentage of the cases received and confirmed in this study were from the Metropolitan Region of RJ (89.9%; 634/705), and the most affected municipalities by DENV-4 in 2011–2013 were Niterói and Rio de Janeiro, followed by Campos dos Goytacazes, located in the North region of the state, where 44.9% (317/705), 36.8% (260/705) and 7.5% (53/705) of infections by this serotype were confirmed, respectively.

The RT-PCR failed to identified DENV-4 in only one sample (Table 2), which was confirmed by virus isolation, that identified the serotype in 48.7% (298/612) of the samples tested. The detection of anti-DENV IgM, considering both acute and convalescent samples, was possible in 39.4% (104/264) of the cases. However, considering cases with ≥4 days of illness, this marker was detected in 57.0% of the cases. DENV NS1 antigen was identified in 41.5% (231/557) of the DENV-4 cases by using the NS1 antigen capture ELISA according to the manufacturer’s instructions, however, after the addition of an antigen-antibody complex dissociation step, the assay sensitivity significantly increased and confirmed 87.6% (488/557; p = 0.001) of the samples tested, Table 2.

DENV-4 cases were more frequently observed in females (n = 396) than males (n = 305), and a higher number of cases were observed on children aged 11–15 year old, and was no significant differences were observed between age groups (Fig. 1).

The patient’s immune response characterization in primary or secondary infections was possible in 508 DENV-4 cases, due to the lack of information in some cases allowing this analysis. Overall, primary cases were more frequently observed (259/508; 51.0%) than secondary ones (249/508; 49.0%), and despite the lower number of cases in 2011, no significant differences were observed, Table 3. DENV-4 cases (n = 668) were also characterized according to the disease severity. Most of the cases were classified as dengue cases (n = 560; 83.8%), 12.8% (n = 86) as DWAS and 3.2% (n = 22) as severe dengue according to the 2009 WHO criteria, Fig. 2. The clinical manifestations observed in most cases were fever (n = 640; 95.8%), followed by headache (n = 574; 85.9%), independent of the disease severity. Among the most common symptoms observed on dengue, no differences were observed in relation to the disease severity (Fig. 3a). When observing the warning symptoms, a higher proportion of abdominal pain in DWAS cases was observed (n = 75; 87.2%). Despite this, the proportion of abdominal pain in severe dengue cases was considerably lower (n = 3; 13.6%), Fig. 3b.

The NS1 antigenemia on DENV-4 cases (n = 138) showed no differences between NS1 levels from dengue, DWAS or severe dengue cases (Fig. 4a). The same was observed on DENV-4 cases previoulsly negative by NS1 ELISA capture and that resulted positive after heat-mediated immune complex dissociation (Fig. 4b). Viremia was also quantified on dengue (n = 8) and severe dengue (n = 8) cases, however no differences were observed (5.95 log10 RNA copies/mL and 5.83 log10 RNA copies/mL, respectively (p = 0,001).

A relationship between the disease severity and the type of infection (primary or secondary) was shown. There was a greater proportion of a more severe disease among individuals 65 years old and older, however a paired analysis by age showed no significant difference in risk of more severe disease (p > 0.05) nor for secondary infections (p > 0.05), Table 4.

Fatal cases due to DENV-4 (n = 14) were received and analyzed during the studied period, 42.8% (6/14) occurred in 2012 and 57.1% (8/14) in 2013. Immune response was characterized on 7 cases and, 42.8% (3/7) occurred during a primary infection and 57.1% (4/4), due to secondary ones. Age information was available on 12 cases and, 7 occurred on patients’ 16 to 65 years old, Table 5.

The phylogeny based on the partial genome sequencing of representative DENV-4 strains (n = 12) isolated in State of Rio de Janeiro, represented by dengue and one fatal case, demonstrated the circulation of DENV-4 Genotype II (Fig. 5).

The introduction of DENV-4 in the state of Roraima, North Brazil in 2010 warned for the possibility of a new dengue epidemics in Brazil [16], as the main vector, Aedes aegypti is widespread in the country. In the disease surveillance, the identification of the vectors has important implications for the disease outbreak control, especially with the rapid disease spread when a new serotype is introduced.

In 2011, the first isolation of DENV- 4 in RJ from cases occurred in the city of Niteroi resulted in the emergence of this serotype in the state in 2012, characterized by isolation of this serotype in 48.7% of confirmed cases. That same year showed for, the first time, the co-circulation of the four DENV serotypes in the state. However, in 2013, DENV-4 was responsible for the largest number of cases in the state.

The Metropolitan region of the state was responsible for most cases occurred during the studied period, mainly in the municipalities of RJ and Niterói. This region has a high population density and high-traffic, which partly explains the concentration of the cases studied [12]. Furthermore, as the Flavivirus Laboratory, FIOCRUZ acts as a Regional Reference Laboratory for dengue diagnosis in Brazil, samples from other municipalities were also received and analyzed.

Although viral isolation is considered the “gold standard” method for dengue viral detection, the sensitivity of the molecular methods in detecting the viral genome offers a great advantage during an epidemic period, as shown by the results presented here. However, it is noteworthy that the high percentage of positive results obtained in RT-PCR may be the result of the selection criteria for the analysis in this study, as known DENV-4 cases were selected. Despite this, previous studies have demonstrated the usefulness of RT-PCR to confirm the cases and identification of the infecting serotype, when the virus isolation was not possible [25, 26].

Serological methods are still the most useful tool for the diagnosis of the disease during epidemics, as observed since its establishment in 1986 in Rio de Janeiro [27]. In this study, the low detection rate of anti-DENV IgM by ELISA (39.4%) was probably due to the acute nature of the cases selected for analysis, as previously described. When it has been described that patients with a primary infection have evidently undetectable IgM antibody titers and there is an increase in detection from a paired sample in the later disease [9]. Usually, higher anti-DENV IgM detection rates are observed after the sixth day after the onset of symptoms. Therefore, considering cases with ≥4 days of symptoms, the detection rate significantly increases.

The usefulness of NS1 Ag capture tests for the early diagnosis of dengue, even in limited resourced setting has been shown [28] and evaluated to confirm primary and secondary acute dengue infections [28‐31]. In this study, initially, the NS1 capture ELISA detection rate was 41.5% on previously confirmed DENV-4 cases. However, after using a heat immune complex dissociation step, the NS1 detection rate significantly increased to 87.6%, as previously described.

In this study, hemorrhage, shock, increased transaminases levels and central nervous system involvement were considered for severe cases classification. In the study, 3.2% of DENV-4 confirmed cases were classified as severe dengue and 12.8% as dengue with warning signs. We observed that abdominal pain is an important warning sign regarding the evolution of the disease and which can help in the correct patient management and thus, avoiding increased severity and fatal outcome. However, due to the small number of severe cases available in this study, no statistical analysis on the disease severity was performed. Furthermore, during the study period, no increase in the disease severity was observed and this is in agreement with previous observations in the same period in the state of RJ [32]. The lower severity on DENV-4 cases is known [33] and here, 14 deaths due to DENV-4 were confirmed in the period studied.

No significant differences between the viremia on mild dengue and severe cases were observed, which does not corroborate with previous studies analyzing other serotypes [34‐36]. However, a limitation of the present study was the small sample size of severe cases. Moreover, the sampling collection represents the patient’s first visit to the health unit and the information on their progression to a more severe disease is not available.

DENV-4 strains isolated from 2011 to 2013, following this serotype introduction and spread in RJ were partially sequenced and the phylogenetic analysis demonstrated that those belong to genotype II. Moreover, based on previous studies, the sampling size sequenced here is reliable for this analysis considering the epidemic duration, site and the low-mutation characteristic of DENV [16, 37]. The strains of genotype II first isolated during the outbreak of 1982 in Roraima, North Brazil [38] did not spread to other states at that time. However, this genotype was once again detected in Roraima in 2010, being circulated through Central America, Northeastern South America and the Caribbean [39]. According to Temporão et al. (2011) [16], these strains of genotype II were genetically different from those isolated in the 1980s in Roraima and demonstrated a probable origin from Venezuela and Colombia. However, Nunes et al. [40] complete and partial DENV-4 sequences isolated in Brazil identified the circulation of two genotypes (I and II). Genotype II is the most commonly one observed circulating in South America and the Caribbean. Genotype I, one the other hand is represented by a strain isolated in the Northeast of Brazil (Bahia), which may have some evolutionary relationship with Asian strains, according to phylogenetic analysis performed by Campos et al. [41].

After the DENV-4 isolation in 2011 in RJ, this serotype spread in the following two years and, the co-circulation of the four DENV was evidenced for the first time in the state. The DENV-4 diagnosis by NS1 ELISA was significantly improved by using an immune complex dissociation step. We observed a low percentage of clinical manifestations related to a more severe disease, and despite the occurrence of secondary infections, immune status and age were not related to a more severe disease. Although two distinct DENV-4 genotypes were already reported circulating in Brazil, in this study, genotype II was the only identified during the period in RJ. The impact of the introduction and emergence of DENV-4 in the state of RJ was not well understood and studies accessing a new DENV serotype circulation in an endemic scenario are needed.