Clinical and Laboratory Diagnosis of Dengue Virus Infection

Dengue virus (DENV) infection is responsible for the most significant mosquito-borne viral disease
in the world today. Like other flaviviruses, its genome comprises a single strand of positive-sense
RNA encoding 3 structural and 7 nonstructural (NS) proteins [1]. Unlike the other flaviviruses,
there are 4 serotypes, referred to as DENV1–4, that are genetically similar but antigenically distinct
[2], defined by the inability of individually elicited antibodies to cross-neutralize. Dengue is spread
primarily by the female Aedes aegypti mosquito, a vector that can be found throughout the
tropical and subtropical regions of the world [3, 4]. However, an increasing number of outbreaks
have been attributed to transmission by the temperate climate mosquito Aedes albopictus [5],
presenting the possibility of further geographical incursions. Mass global travel has resulted in
many regions of the tropical world displaying hyperendemic dengue activity, with multiple
serotypes circulating at any one time [6–9]. It is estimated there are up to 390 million DENV
infections annually, with more than 500 000 hospitalizations and 25 000 deaths [10]. Infection
with any of the 4 DENV serotypes can result in a range of clinical outcomes, with the majority of
infections (70%–80%) being asymptomatic [10]. Clinical presentation can range from a mild fever
to classical dengue fever with hemorrhage (DHF) and/or shock (dengue shock syndrome [DSS])
[11]. Classical dengue fever is an acute infection presenting clinically 4–10 days following the bite
of an infected mosquito. The disease is characterized by elevated temperature (up to 40°C), severe
headache, retro-orbital pain, malaise, severe joint and muscle pain, nausea, and vomiting, with a
rash appearing after 3–4 days after fever onset [12]. Following a primary infection, the patient is
immunologically protected from disease caused by that particular dengue serotype [13]. The
severe forms of dengue disease are seen primarily in individuals experiencing a secondary
infection with a different dengue serotype [13]. However, primary infection in young infants may
also be associated with severe disease outcome [14]. Early in the acute febrile period of disease,
dengue fever presents with the same clinical symptoms as primary dengue. Later, during
defervescence, patients can rapidly deteriorate, progressing to hemorrhage with or without
vascular leak. During this period, patients can experience bleeding, thrombocytopenia with <100
000 platelets/uL, ascites, pleural effusion, increased hematocrit concentrations, severe abdominal
pain, restlessness, vomiting, and sudden reduction in temperature with profuse perspiration and
adynamia [11]. Currently, there Is no therapeutic option for dengue, with treatment being purely
supportive. Nevertheless, the symptoms of DHF and DSS can be effectively managed in most cases
by fluid replacement [15]. In addition to early and accurate diagnosis, early markers of progression
to severe disease are urgently needed.

CLINICAL DIAGNOSIS

Clinical diagnosis of dengue can be challenging, depending largely on what stage in the infection
process a patient presents. Depending on the geographic region of the world, there can be a
number of disease-causing pathogens or disease states that can mimic the disease spectrum
arising from dengue infection (Supplementary Table 1). In the early stages of clinical disease,
dengue can present as a mild undifferentiated “flu-like” fever with symptoms similar to those of
other diseases such as influenza, measles, Zika, chikungunya, yellow fever, and malaria [16].
Correct diagnosis of the pathogen responsible for the later manifestation of shock is of particular
importance, as treatment for dengue-induced shock vs that arising from sepsis traditionally
requires different approaches [16]. However a potential, paradigm-shifting observation that DENV
infection activates similar innate immune pathways as those induced in sepsis may suggest
alternative, common targets for treatment [17]. Because the clinical symptoms of dengue are so
diverse, accurate clinical diagnosis is challenging. As such, it is essential that laboratory or point-of-
care diagnostics be used in conjunction with assessment of clinical presentation.

Clinical Presentation

In 2009 a working group coordinated by the World Health Organization (WHO) set out a series of
guidelines for clinical management of dengue [16]. They also modified the existing dengue disease
classifications from dengue fever, DHF, and DSS to dengue (with or without warning signs) and
severe dengue (Figure 1). The aim of these guidelines was to establish uniform and simpler clinical
criteria that provided a standardized global approach to disease classification. Dengue virus
infection can result in either asymptomatic or symptomatic infection [18]. Roughly 20% of all
infections are symptomatic, with individuals experiencing disease symptoms that cover a broad
clinical spectrum of nonsevere to severe clinical manifestations [19]. Illness caused by dengue has
an abrupt onset with 3 broadly identifiable phases: febrile, critical, and recovery [16]. Appropriate
viral diagnosis and evaluation of warning signs of progression to severe disease are critical for
effective patient management. The initial febrile phase is characterized by rapid onset, initially
with sudden high-grade fever [18]. This phase lasts between 2 and 7  days, with the febrile phase
of the disease being characterized by a facial flushing skin erythema, generalized body ache,
myalgia, arthralgia, retro-orbital eye pain, photophobia, rubeliform exanthema, and headache
[20]. Sore throat, anorexia, nausea, and vomiting are also common [20]. During this phase, a
positive tourniquet test is able to differentiate dengue from other diseases presenting with similar
symptoms [21]. The acute febrile phase may also be accompanied by hemorrhagic symptoms
ranging from a positive tourniquet test and petechiae to spontaneous bleeding from the
gastrointestinal tract, nose, gums, and other mucosal sites [16]. The severity of symptoms during
this phase is not a predictor of progression to severe dengue; therefore, monitoring of early
warning signs needs to be undertaken during the critical phase of disease [16]. The majority of
DENV-infected patients make a full recovery after the febrile period and do not enter the critical
phase of disease. However, patients that do enter the critical phase may develop warning signs
that indicate increased capillary permeability leading to plasma leakage. Generally, patients
worsen at the time of defervescence (from illness day 4)  when their temperature drops to 37.5°C–
38°C [22], and it is during this period that early symptoms of vascular leakage may be seen.

Leukopenia followed by a rapid drop in platelet count generally leads to plasma leakage [20].
Coincident with dropping platelet counts is a corresponding rise in hematocrit level. Leakage can
last 24–48 hours and, during this time, hematocrit levels need to be closely monitored as an
indicator of the need for intravenous fluid adjustment [20, 23]. During this time, and unless
profound leak is clinically apparent, other methods such as ultrasound detection can be employed
to detect free fluid in the chest or abdominal cavities [24]. Warning signs (Figure   1) are almost
always present in patients before the onset of shock [24]. Shock occurs when a patient loses a
critical volume of plasma through vascular leakage. The hypoperfusion that occurs during
profound/prolonged shock results in metabolic acidosis that can lead to progressive organ
impairment and eventual intravascular coagulation [25]. Once patients pass through the critical
24–48 hour period, disease recovery can be remarkably rapid. Reabsorption of extravascular fluids
accompanies an increase in the general well-being of the patient, and appetite returns along with
cessation of other symptoms [25]. Patients can develop what has been referred to as a “recovery
rash” with characteristic patches of normal skin likened to “isles of white in a sea of red” that
develop on the trunk, then spread to the head and extremities [26]. Patient blood counts stabilize
and return to normal during this recovery phase. The severe forms of disease are defined as a
patient that has dengue with one of the following: severe plasma leakage that leads to shock
and/or fluid accumulation with respiratory distress; severe bleeding, and severe organ impairment
(Figure 1) [16]. As noted above, dengue-induced shock occurs at defervescence and at a time
when viral levels are falling (Figure 2), indicating likely immune-mediated pathology [27]. The
hypovolemic shock that occurs is a result of prolonged increased vascular permeability causing
plasma leakage [28]. Patients with DSS initially suffer from asymptomatic capillary leakage
progressing to compensated shock to hypotensive shock, eventually leading to cardiac arrest [29,
30]. Dengue shock patients need to be closely monitored, as the time between warning signs and
the development of compensated shock and hypotensive shock may only be a matter of hours
[28]. Only minutes may separate hypotensive shock and cardiorespiratory collapse and cardiac
arrest [28]. For a more in-depth review of the clinical presentation of severe dengue and patient
management, consult the WHO Handbook for clinical management of dengue [16].

Predictive Algorithms

Given the wide presentation and dynamic nature of DENVrelated illness, a number of studies have
been performed to assess the potential of using an algorithm approach to the prediction of patient
progression from dengue fever to severe dengue disease [19, 24, 25]. These studies have looked at
a combination of clinical and laboratory markers, both viral and host derived, that have previously
been shown to vary significantly between dengue fever and severe dengue patients [31–34].
Virological markers such as the virus or viral genome and the secreted NS1 protein along with host
factors including virus-specific immune responses, liver enzymes (aspartate aminotransferase and
alanine aminotransferase) [35], and hematological factors such as platelet and hematocrit counts
have all been considered [32, 33, 36]. While some of these studies have shown promise, a
comprehensive multicenter clinical study is yet to be evaluated and reported [34, 37–39]. Such a
study has been under way since 2011 [40] and its findings, due in early 2017, are keenly awaited.
However, the development of a universal predictive algorithm for severe disease progression
presents numerous challenges given the significant variables introduced by local virus evolution,
the subtleties of virus–host interactions, geographic spread of the disease, and regional host
genetics. Nevertheless, the current WHO guidelines aid clinicians in providing a clear set of clinical
warning signs to help predict severe disease onset. Although not all the warning signs appear early
in disease, when appropriately implemented in the clinic, along with accurate laboratory
diagnosis, the WHO guidelines provide a strong framework for effective monitoring of severe
disease.

LABORATORY DIAGNOSIS

Biomarkers that have been targeted for diagnosis include the virus itself (virus isolation in culture
or mosquitoes or the direct detection of viral genomic RNA), viral products (capture and detection
of the secreted NS1 protein), or the host immune response to virus infection (through
measurement of virus-specific immunoglobulin M [IgM] and immunoglobulin G [IgG]). The timing
of the appearance and duration of these biomarkers in both primary and secondary dengue
infection is graphically presented in Figure 2. In the following, we briefly review both traditional
and more recent approaches to measuring the presence of these markers of infection.

Virus Isolation

Virus isolation has been the traditional diagnostic method for detecting DENV infection. However,
it has gradually been replaced by reverse-transcription polymerase chain reaction (RT-PCR) and,
more recently, by NS1 antigen-capture enzymelinked immunosorbent assays (ELISAs) for more
rapid diagnosis [41]. For virus isolation, clinical samples taken from patients are cultured in a
variety of cell lines of either mosquito (AP61, Tra-284, AP64, C6/36, and CLA-1 cells) or mammalian
(LLCMK2, Vero, and BHK-21 cells) origin or in live mosquitoes [41, 42]. Blood samples taken from
infected patients experiencing febrile illness up to 5 days after the onset of disease yield the most
successful results. However, virus isolation from secondary infected patients is made more difficult
by the rapid anamnestic production of cross-reactive antibodies early during the acute phase of
disease that form immune-complexes with circulating virus [43]. Although detection of DENV by
virus isolation is definitive, it is not particularly practical, as isolation can take days to weeks to
perform [44].

RT-PCR

Molecular methods such as RT-PCR and nucleic acid hybridization have been used to great effect in
successfully diagnosing DENV infection. PCR-based methods provide same- or next-day diagnosis
of DENV during the acute phase of disease. Lanciotti et  al [44] originally reported a 2-step
heminested RT-PCR assay that was highly sensitive. This method was then modified to a single-
step multiplex real-time RT-PCR assay, which was adopted worldwide. A  major advantage of PCR-
based techniques is that viral RNA can be detected from the onset of illness and is sensitive,
specific, fast, less complicated, and cheaper than virus isolation methods [45]. Although PCR-based
methods are fast and accurate, they require a laboratory with specialized equipment and trained
staff to perform the analysis. These are not always an option in resource-poor remote settings
where dengue is endemic. Furthermore, despite the availability of commercial kits, the bulk of
reported RT-PCR methods are developed in-house and lack center-to-center standardization [46].
Non-PCR-based methods that mimic in vitro nucleic acid amplification, such as isothermal
amplification (eg, single-tube reverse transcription–loop-mediated isothermal amplification), have
shown high levels of sensitivity and specificity when used alongside other diagnostic methods [47].

NS1 Antigen Capture

The viral protein NS1 is an ideal diagnostic target because it is secreted from infected cells, is
found at high levels circulating in the blood of infected individuals, and can be detected from the
onset of symptoms through to 9 days or longer after disease onset, at least in primary infections.
NS1 can be detected at the same time as viral RNA and before an antibody response is mounted in
primary infections. It can be viewed as a surrogate marker for viremia, with the level of NS1 shown
to correlate with viral titer [48, 49]. Detection of NS1 in patient blood using an antigen-capture
ELISA approach was first described in 2000 [48]. Using quantitative-capture ELISA, it has been
found that NS1 is secreted at high levels, within the range of low nanograms per milliliter to
micrograms per milliliter, with up to 50 µg/mL found circulating in some infected individuals.
Subsequent studies investigating the kinetics of NS1 in secondary infections found that NS1 levels
≥600 ng/ mL within the first 72 hours of disease was a strong predictor of progression to more
severe disease [27]. These early reports led to the commercial development of NS1 capture ELISAs
and rapid strip tests [48, 49]. The commercial development of NS1 as a diagnostic tool has
revolutionized dengue diagnosis as it has provided simple and low-technology assays that have
high sensitivity and specificity. These detection assays have now become the new standard for
dengue diagnostics [14, 27, 50, S51–S57], allowing early diagnosis and more effective patient
management. Despite the suggested predictive value of NS1 as a marker of disease progression,
the required quantitation still remains the province of academic research, with all commercial
tests only providing qualitative positive/negative readout. A limitation of NS1 detection for
patients experiencing a secondary infection is the rapid anamnestic rise in NS1 cross-reacting
antibodies during the acute phase of disease. These antibodies sequester NS1 in immune
complexes that cannot be readily detected in capture assays. Consequently, the kinetics of NS1
detection over the course of disease during secondary infections is shorter than that for primary
infections (Figure 2).

Serology

There are numerous approaches to serological diagnosis available, including hemagglutination

inhibition (HI) assays, complement fixation tests, dot-blot assays, Western blotting, indirect
immunofluorescent antibody tests, and plaque reduction neutralization tests, as well as IgM and
IgG antibody-capture ELISAs [S58–S64]. HI assays along with IgM and IgG antibody-capture ELISAs
have proven to be the most useful serological diagnostic methods for routine DENV detection. The
HI test has been applied to dengue diagnosis for many years, with most laboratories developing in-
house methodologies, although commercial kits are also available [S65]. As with all assays based
on antibody detection, the early acute disease period usually presents a negative window of
detection, given the need for the relevant antibody response to be elicited. Nevertheless, high-
throughput IgM and IgG antibody-capture ELISAs have become relatively routine, particularly
following assay automation. IgM can appear as early as day 3–5 in primary infection, peaking
several weeks after recovery and remaining at detectable levels for several months [41, S66, S67].
IgG does not generally appear during the acute phase of primary disease. However, during
secondary infection, there is a rapid anamnestic IgG response to shared epitopes on multiple viral
proteins between the first and second infecting serotypes, with IgG appearing as early as 3 days
after onset of illness [S66]. Consequently, when performed in parallel, IgM and IgG detection can
provide a diagnostic indication of primary or secondary infection based on the ratio of IgM and IgG
during the acute phase of disease [S66, S68]. Detection of DENV infection by serology is
complicated in areas of the world where >1 flavivirus is circulating (eg, yellow fever, Japanese
encephalitis, and, more recently, Zika virus), because of shared cross-reactive epitopes on the
flavivirus E protein, and hence cross-reactivity of the antibody response. This is particularly
problematic in the current Zika virus epidemic in Brazil, which is occurring in a background of
complicating DENV infection and serology. Antibodies directed against these flaviviruses can cross-
react in DENV serology assays, leading to false-positive results. To reduce these false-positive
results, IgM and IgG serology should be paired with NS1 antigen capture. Commercial dengue NS1
antigen-capture ELISA and immunochromatographic (Rapid) strip tests have been designed to be
highly specific, with no demonstrable cross-reactivity with other flavivirus NS1 species [S57, S69,
S70]. However, with the possible cross-reactivity detected in a recent Zika virus–infected patient,
the specificity of these assays will need to be revisited [S71].

Combined Approaches

Given the varying kinetics of each biomarker, no single assay can be used to definitively diagnose
DENV in patients who may present at a different stage in their infection. Detection of NS1 antigen
is perhaps the most robust of all the DENV diagnostic methods with a relatively long detection
window, particularly in patients with primary infection. However, as noted above, NS1 detection
can be obscured by immune complex formation in secondary infections, thereby shortening its
window of detection (Figure 2). In this case, combining NS1 detection with IgM and/or IgG
detection has been shown to dramatically improve positive dengue diagnosis [S72, S73]. There are
several diagnostic kits currently available that take advantage of this approach in rapid point-of-
care devices, including those developed by SD Bioline Dengue Duo (NS1 Ag + Ab Kit). Using this
combination approach, detection sensitivities nearing 100% have been reported from the onset of
illness through recovery [S69].

FUTURE TESTS

There are many new approaches for rapid dengue diagnosis currently under development. These
include micro/paper fluidics, in vivo micropatches [S74], isothermal PCR [47], and electrochemical
[S75, S76] and piezoelectric [S77] detection. All of these technologies are in the early stage of
development, requiring continued refinement to make them practical solutions in real-world
settings. In our view, the ideal goal for dengue diagnosis would be a test that differentiates
primary from secondary dengue infection with IgM and IgG capture, with quantitative serotype-
specific NS1 detection.

CONCLUSIONS

Early and accurate laboratory diagnosis of DENV infection is critical to effective patient
management. While we already have the tools that allow us to reliably determine if a patient is
suffering from dengue infection (Table  1 summarizes the diagnostic methods available, and a
representative diagnostic algorithm is shown in Figure 3), we are still lacking effective predictive
biomarkers of progression to severe disease. Early studies have suggested that elevated levels of
NS1 and viremia may have predictive value [27]; however, in the absence of readily available
quantitative NS1 assays, few validation studies have been performed. Beyond DENV biomarkers,
host responses provide potential predictive markers of severe disease progression. Several
candidate proteins [S78] have been identified; however, their predictive value is yet to be
validated. To determine a set of predictive biomarkers, a large multicenter study is currently under
way across multiple countries, with the aim of identifying a set of predictive parameters for severe
disease. Results from this study are expected to be published in the early part of 2017 [40].
Combinatorial approaches employing virus and/or viral product detection along with serology
currently yield the most reliable diagnosis of dengue infection and should be coupled with close
clinical monitoring of warning signs to identify patients at risk of progression to severe dengue
disease.

https://sci-hub.tw/10.1093/infdis/jiw649