Neurological Manifestations of Dengue Infection

Dengue counts among the most commonly encountered arboviral diseases, representing the fastest spreading tropical illness in the world. It is prevalent in 128 countries, and each year >2.5 billion people are at risk of dengue virus infection worldwide. Neurological signs of dengue infection are increasingly reported. In this review, the main neurological complications of dengue virus infection, such as central nervous system (CNS), peripheral nervous system, and ophthalmic complications were discussed according to clinical features, treatment and possible pathogenesis. In addition, neurological complications in children were assessed due to their atypical clinical features. Finally, dengue infection and Japanese encephalitis were compared for pathogenesis and main clinical manifestations.


INTRODUCTION
Dengue counts among the most common arboviral illnesses, representing the fastest spreading tropical disease in the world (Lundberg, 1957). It is considered the second leading cause of acute febrile disease in travelers (Freedman et al., 2006). Four different serotypes (DENV-1,−2,−3, and−4) (Guzman et al., 2010) cause dengue fever, with various infectious outcomes (asymptomatic to severe hemorrhagic fever).
Dengue is prevalent in 128 countries (Brady et al., 2012;Lorenzi et al., 2013;Teixeira et al., 2013), and more than 2.5 billion individuals are in danger each year of contracting dengue virus worldwide. According to some estimates, almost 400 million individuals are infected annually, with ∼96 million showing clinical relevance. About 2.5% of all diseased people die (Guzman et al., 2010;Bhatt et al., 2013). In recent years, neurological manifestations of dengue infection have been increasingly reported; however, their precise incidence rates remain undefined.

NEUROPATHOGENESIS
Neuropathogenesis is likely associated with direct invasion of the CNS by the virus, autoimmune reactions, and metabolic alterations. The dengue virus is considered to be nonneurotropic. However, recent reports associating dengue with neurological complications have changed this view. This virus was described in cerebro-spinal fluid (CSF) more than two decades ago (Lum et al., 1996;Thisyakorn et al., 1999). Chaturvedi et al. demonstrated that the bloodbrain barrier (BBB) is damaged during infection by the dengue virus in experimental animal experiments, indicating viral invasion (Chaturvedi et al., 1991). Meanwhile, immunoreactive neurons, astrocytes, microglia, and endothelial cells were found in cerebral tissues of a fatal case with hemorrhagic dengue fever in 1998 (Ramos et al., 1998). Domingues et al. proposed that dengue virus can actively enter the CNS (Domingues et al., 2008). Data obtained in Vietnam also support direct invasion by dengue virus as pathologically important (Solomon et al., 2000b). Autoimmune reactions and metabolic alterations have been demonstrated in most neurological complications of dengue fever cases (Seet and Lim, 2007;Basu and Chaturvedi, 2008;Jha and Ansari, 2010;Murthy, 2010;Sharma et al., 2011;Verma et al., 2011b;Weeratunga et al., 2014b).
Most neurological manifestations of dengue virus infection have been reported in case reports or short series, and its spectrum is diverse; thus, the classification of neurological manifestations is difficult to apply in practice. The new (2009) World Health Organization (WHO) classification groups dengue infection into three categories, including dengue with no warning signs, disease with warning signs, and severe dengue (WHO, 2009). Different from the traditional system, the revised classification includes CNS involvement as severe dengue. However, neurological complications are not well described and very little is known about these manifestations.
The main neurological complications of dengue virus infection are discussed below.

CENTRAL NEUROLOGICAL SYSTEM COMPLICATIONS
CNS complications are diagnosed by assessing anti-DENV immunoglobulin (Ig)M, detecting viral RNA or non-structural protein 1 (NS1)in the CSF, isolating the virus from the CSF, and after excluding other causative agents of viral brain diseases (Sahu et al., 2014;Solbrig and Perng, 2015). Misra et al. found that CNS involvement reflects severer disease with poorer recovery (Misra et al., 2015). Encephalitis and encephalopathy are the most common neurological presentations of dengue infection (World Health Organization, 1997;Pancharoen and Thisyakorn, 2001;Gupta et al., 2013; Table 1).

Encephalitis
Encephalitis is considered a severe manifestation of dengue virus infection, and found in the three classical disease groups. Diagnosis of dengue encephalitis is based on criteria proposed by Soares and Marzia (2014).
However, normal CSF cellularity cannot exclude dengue encephalitis. According to Soares et al., dengue is the first reason for encephalitis with normal CSF cellularity in 75% of patients with viral meningitis and encephalitis in a dengue endemic region, followed by Herpes Simplex Virus 1 (HSV1) and mild encephalitis (Soares et al., 2011). Neuroimaging of dengue encephalitis yields divergent data, with normal findings in most cases (Baldaçara et al., 2013;Madi et al., 2014). In case of abnormal neuroimaging findings, magnetic resonance imaging (MRI) has advantages over cranial computed tomography (CT) in revealing cerebral lesions in dengue encephalitis. However, changes are usually non-specific (Garg et al., 2017). Decisive characterizations of MRI properties in dengue encephalitis remain undefined (Mathew et al., 2015). Treatment is nonspecific, with mostly symptomatic treatment provided. Most patients have good recovery.

Encephalopathy
Encephalopathy caused by dengue fever can be reflected by reduced sensitivity, cognitive impairment, convulsions, and personality and behavior disorders, including acute mania, depression, emotional lability, anxiety, psychosis, and agoraphobia (Rittmannsberger et al., 2010;Baldaçara et al., 2013;Srivastava et al., 2013). In a review reporting cases of dengue fever associated with neurological disorders in 2012, encephalopathy was considered by far the most encountered complication (Oehler et al., 2012). Most encephalopathy cases occur in children of developing countries, and do not show CSF abnormalities (Angibaud et al., 2001). Dengue associated encephalopathy is generally very serious, with around 50% of the affected patients succumbing (Angibaud et al., 2001).

Meningitis
Dengue infection associated meningitis is rarely encountered (Soares et al., 2006). In the related literature, three reports about dengue meningitis are found (Soares et al., 2010(Soares et al., , 2011Mamdouh et al., 2013). Soares et al. described a 24-year-old woman with the symptoms of fever, headache, and nuchal rigidity, in whom diagnosis was confirmed by polymerase chain reaction (PCR) and positive results in CSF evaluation (Soares et al., 2010). Two additional cases of dengue meningitis were reported by Mamdouh in 2013. Both cases presented with fever, headache, neck rigidity, and low platelet count in a dengue endemic area. Diagnosis was confirmed by positive IgM in the CSF in the absence of the typical clinical spectrum of infection. CSF cellularity was normal in one case. Brain CT or MRI scan was normal. Dengue meningitis might cause migraine such as headaches with poor response to anti-migraine therapy and common analgesics. The two patients achieved full recovery after several months without any residual neurological deficit (Mamdouh et al., 2013).

Stroke
Ischemic and hemorrhagic strokes have been reported in a few cases. Liou et al. reported a dengue fever patient who showed thrombocytopenia and ischemic stroke in 2008 (Liou et al., 2008a). In 2013, Rajesh Verma et al. reported a 68-year-old male patient presenting with moderate grade, continuous fever for 15 days; the patient also had sudden onset of weakness of the left half of his body as well as facial asymmetry, 10 days prior to hospital admission. Dengue infection was confirmed and brain MRI revealed acute infarction in the right parietal lobe. After 2 months of medication and physiotherapy, a partial improvement was observed in limb weakness (Verma et al., 2013).In cases with ischemic stroke, meningovasculitis, or a transient hypercoagulable state during dengue infection was postulated as the pathogenetic mechanism (Liou et al., 2008a;Verma et al., 2013).
A summary of intracerebral hemorrhage cases associated with dengue virus infection in various countries from 2001 to 2014 was reported by Vargas-Sánchez et al. (2014). In the above study, patient age ranged from 9 to 68 years, and there were 5 females and 7 males. Outcome was unspecified in one patient, while 7 and 5 patients recovered and died, respectively. Meanwhile, DENV2 was detected in 5 patients; one patient was infected with DENV3 and precision was not provided for the remaining seven. The involved regions included the pontine, basal gangliar, cerebellar, parietal, temporal, and frontal lobes of the brain (Vargas-Sánchez et al., 2014).
Current studies suggest that in dengue virus infection, cytokine overproduction results in immune-mediated endothelial cell damage (Seet and Lim, 2007;Basu and Chaturvedi, 2008). Hemorrhage may be caused by elevated vascular permeability, plasma leakage, and vasculitis due to dengue (Seet and Lim, 2007;Basu and Chaturvedi, 2008).

Cerebellar Syndrome
Cerebellar syndrome was mentioned by Weeratunga et al. (2014b). They described three patients diagnosed with dengue infection based on combined clinical and laboratory findings, fulfilling the WHO criteria in an endemic area. Within 2 weeks of diagnosis, they developed cerebellar symptoms. The manifestations of cerebellar syndrome include bilateral vertical and horizontal nystagmus, dysarthria, bilateral limb, and gait ataxia. Other causes of cerebellar dysfunction were excluded, and all cases were self-limiting. MR brain scans showed cerebellar T2 hyperintense lesions in one case, while the remaining two showed normal signals. The CSF in all cases had normal protein levels and cell count. All cases showed antibodies for dengue in blood and CSF samples. A low-grade inflammatory process was the proposed mechanism (Weeratunga et al., 2014b).
However, the spectrum of TM could be broad, and more attention should be paid to atypical cases. For example, Mota reported a patient confirmed with acute TM without paraparesis following a dengue virus infection last year (Mota et al., 2017).
Treatment with intravenous steroids is useful, and individuals with parainfectious-dengue TM show satisfactory recovery (Fong et al., 2016). Fong et al. reported that a previously healthy 12year-old girl with parainfectious-dengue TM and concomitant spinal epidural haematoma had good clinical recovery without surgical intervention after 6 months (Fong et al., 2016). It was shown that post-infection autoimmune reactions and direct infection are associated with transverse myelitis (Solomon et al., 2000b;Sindic et al., 2001;de Sousa et al., 2014;Mo et al., 2016).
Dengue related ADEM results from immune reactions (Gupta et al., 2013). In Murthy's study, its pathophysiology was considered a transient autoimmune reaction to myelin or unknown self-antigens (Murthy, 2010). There is no established treatment for ADEM, but the use of steroids is effective during its active phase (Carod-Artal et al., 2013;Gupta et al., 2013;Domingues and Kuster, 2014).
Neurological signs develop at 1-19 days after the onset of dengue. Simon et al. found GBS occurs early in DF with an unusual progression; in their report, average time elapsed from fever to neurological signs was 2 days, ranging between 1 and 3 days (Simon et al., 2016). A patient with Miller Fisher syndrome caused by dengue has been described (Gaultier et al., 2000). Umapat et al. reported that asymptomatic dengue infection may also trigger Guillain-Barré syndrome, and found four GBS cases after dengue diagnosis, including 1 and 3 with acute motor axonal neuropathy and acute inflammatory demyelinating polyneuropathy, respectively (Umapathi et al., 2016).
The exact mechanism of dengue associated GBS remains unclear. It is considered a neurological disease modulated by immunocytes (Chew et al., 1998;Gupta P. et al., 2009). Pro-inflammatory substances such as tumor necrosis factor, interleukins, and the complement may play important roles in GBS pathogenesis (Shah, 2007). Immune reactions induced by dengue virus infection might involve peripheral nerve constituents; for example, myelin or axon may be targeted by the immune response (Carod-Artal et al., 2013).
Plasma exchange seems to be more effective than supportive treatments in the treatment of GBS in randomized clinical trials (Comi et al., 1985;Qureshi et al., 2012). Immunoglobulins administered intravenously show similar effectiveness to plasma exchange, and might be more advantageous (Chew et al., 1998;Gupta P. et al., 2009;Qureshi et al., 2012). Corticosteroids alone do not make a difference, and there is not enough data supporting their usefulness (Comi et al., 1985;Qureshi et al., 2012). Other treatment options, e.g., CSF filtration, are still being assessed (Qureshi et al., 2012).

Myositis
Dengue myositis diagnosis is based on clinical manifestations of dengue virus infection, positive serum IgM for the dengue virus, high creatine phosphokinase levels, normal CSF, and exclusion of other causes (Paliwal et al., 2011). The clinical spectrum is broad, from mild asymmetrical weakness of lower extremities to sudden progressive severe limb and trunk weakness, and even lung failure (Paliwal et al., 2011). However, muscle weakness is distinctly uncommon in dengue virus infection (Malheiros et al., 1993;Sangle et al., 2010). Doctors should pay more attention to patients suffering from early pulmonary impairment, elevated creatine phosphokinase amounts, and serious myalgia (Paliwal et al., 2011).
The pathogenesis of myositis remains unclear. The proposed mechanisms comprise direct muscular invasion by the virus and immunity associated destruction of muscle fibers, particularly by tumor necrosis factor (Malheiros et al., 1993;Paliwal et al., 2011). Dengue myositis is reflected by perivascular infiltration of mononuclear cells, mitochondrial proliferation, fat accumulation, nuclear centralization, fiber-type grouping, and/or myonecrosis foci on muscle biopsy (Malheiros et al., 1993;Misra et al., 2012).
Normal motor unit potentials are of reduced duration and amplitude, and polyphasic on electromyographic examination. Fibrillations, sharp waves, and complex repetitive discharges do not occur (Kalita et al., 2005;Misra et al., 2012).

2011
). In adult patients, dengue myositis is often more severe, even leading to severe rhabdomyolysis (Kalita et al., 2005;Finsterer and Kongchan, 2006;Siriyakorn and Insiripong, 2015). Sangle et al. reported a 16-year-old girl with dengue shock syndrome showing myositis and myocarditis. Symptomatic treatment and rehabilitation were provided, and she had recovered well at discharge after 1 month of hospitalization (Sangle et al., 2010). Finsterer et al. found dengue can cause persistent and serious myositis, which is resolved after administration of corticosteroids (Finsterer and Kongchan, 2006).

Hypokalemic Paralysis
Dengue-associated hypokalemic paralysis was reported by several authors, but most studies were based on isolated case reports and short series (Esack et al., 1999;Santos et al., 2004;Kalita et al., 2005;Jha and Ansari, 2010;Roy et al., 2011;Jain et al., 2014;Maurya et al., 2016). Maurya and colleagues assessed 12 dengue patients with hypokalemic paralysis, who had complete and rapid recovery after potassium supplementation (Maurya et al., 2016). It has been suggested that elevated creatine phosphokinase amounts result from vasoconstriction and muscular ischemia caused by hypokalemia (Comi et al., 1985). In a study of epidemic, 10 individuals showed hypokalemia, while Guillain-Barré syndrome and myositis were found in one patient each. Of the latter 2 patients, the one with myositis showed slow improvement and normal serum potassium levels, and the other with Guillain-Barré syndrome recovered within 6 weeks (Hira et al., 2012). In a retrospective study of 7 patients with dengue myositis who had generalized weakness, a mild serum creatin phosphokinase (CPK) elevation was noted in 3 individuals with hypokalemia, whereas three cases with fulminant presentation and respiratory muscle involvement had markedly elevated CPK levels (16,200 U/L) with normal potassium amounts. All patients completely recovered within 4 weeks (Paliwal et al., 2011).
The severity of weakness may not correlate with potassium levels in Maurya's report (Maurya et al., 2016). Hypokalemia does not always lead to paralysis. In 1,342 patients with dengue fever assessed in China, hypokalemia occurred in up to 28% individuals but weakness was not mentioned (Ying et al., 2007).
Hypokalemic paralysis differs from dengue myositis and idiopathic hypokalemic paralysis according to clinical, biochemical features, and consequences (Maurya et al., 2016). It represents a systemic complication of dengue fever (Murthy, 2010;Jain et al., 2014). The pathogenesis of hypokalemic paralysis in dengue remains obscure. A probable explanation for hypokalemia in dengue is serum potassium redistribution in cells and transient kidney tubular dysfunction increasing urinary potassium elimination (Jha and Ansari, 2010;Jain et al., 2015;Maurya et al., 2016). In addition, infection-related stress may lead to catecholamine or insulin release, which results in an intracellular shift of potassium (Jha and Ansari, 2010;Hira et al., 2012). Supplementation of potassium can achieve satisfactory recovery in patients with hypokalemia paralysis (Jha and Ansari, 2010;Kayal et al., 2013). Jain et al. reported a 30-year-old man with dengue and hypokalemic paralysis alongside hypomagnesemia. Weakness persisted after potassium supplementation, while low serum magnesium levels were detected. However, he completely recovered within 48-72 h, with a normalization of serum potassium and magnesium levels. The possible mechanism may be that muscle Na+ and K+-ATPase activity is inhibited by hypomagnesemia, leading to a reduced ion influx into muscle fibers and secondary kaliuresis (Jain et al., 2015).
Its diagnosis is established, after other reasons for neuritis, such as tumor and demyelinating diseases, infections, trauma, and stroke, are accordingly excluded. Dengue-associated nerve palsy is mostly managed by supportive treatment. Some cases improve even without specific treatment, e.g., with steroids or intravenous immunoglobulins (Peter et al., 2013;Biswas and Pal, 2014), while others have a good response to steroids (Verma et al., 2011b).
The pathogenesis of dengue-associated neuritis is likely related to immune reactions, although the mechanisms remain largely unclear (Carod-Artal et al., 2013). Most patients achieve a remarkable recovery.
The mechanisms underlying dengue infection-related ocular signs remain unclear, but could involve immune processes with possible association with dengue serotyping Bacsal et al., 2007;Su et al., 2007;Yip et al., 2012).

NEUROLOGICAL COMPLICATIONS IN CHILDREN
Dengue causes high morbidity and mortality in children living in tropical and subtropical areas of the world. Around 95% of patients with serious disease are below 15 years old (Bhattacharya et al., 2013). All four dengue virus serotypes can be detected in infectious children, and clinical symptoms range from mild fever to fatal dengue shock syndrome (Verhagen and de Groot, 2014).
National surveillance in Asia showed that individuals below 1 year old and those between 4 and 9 years of age are most likely to develop severe dengue infection (Kongsomboon et al., 2004;Huy et al., 2010).It is worth noting that compared with older children, infants often show a higher frequency of plasma leakage and shock in dengue (Nguyen et al., 2004;Hammond et al., 2005).Therefore, the management of infants with dengue infection is important because they sometimes present with unusual manifestations, and early diagnosis is very challenging (Kalayanarooj and Nimmannitya, 2003). Whether severe dengue disease occurs or not largely depends on factors such as virus properties, host immunity, age, and genetic makeup. Female children may be associated with severe dengue (Anders et al., 2011). Prognosis of dengue hemorrhagic fever and dengue shock syndrome is affected by prevention, early detection, and timely therapy; mortality ranges between 2.5 and 5.0% (Alejandria, 2009). In case of shock, mortality can approach 12-44% (Rigau-Pérez et al., 1998). Patients usually have satisfactory recovery after optimal fluid/electrolyte supplementation.
Hospitalization may not be necessary for children with mild dengue infection (Nguyen et al., 2004;Hammond et al., 2005). There are no specific therapeutic agents for dengue, but fluid replacement is immediately required in pediatric cases with haemorrhagic fever or shock syndrome, to expand the plasma volume. Crystalloids are as potent as colloids in children with moderately severe and severe dengue shock syndrome (Alejandria, 2009). For children with suspected DHF, attentive clinical monitoring and supportive care are critical measures for reducing fatality rates (Tantawichien, 2012). Preventive transfusions are not recommended (Tantawichien, 2012), and steroid administration is not considered a beneficial option in dengue shock syndrome (Smart and Safitri, 2009).

THERAPY AND PREVENTION
Currently, no definite effective antiviral agents are available for dengue infection treatment. General supportive therapy prevails, emphasizing on intense hematological monitoring, fluid-replacement, and/or blood transfusion if needed. Nonsteroidal anti-inflammatory drugs may worsen gastritis or cause bleeding.
A safe and efficacious dengue vaccine is considered a great hope for preventing and controlling this disease. Currently, prevention is only achieved by vector control. However, several vaccine preparations are under investigation (Simmons et al., 2012). Notably, it is especially important for children to avoid Aedes mosquito bites in dengue-endemic regions (Elling et al., 2013).

COMPARSION WITH JAPANESE ENCEPHALITIS VIRUS
Dengue is considered a non-neurotropic virus. However, neurotropism and neuro-invasion in dengue have been reported (Chardboonchart et al., 1990;Despres et al., 1998;Solomon et al., 2000b;Cam et al., 2001). DF is usually self-limiting, and death is rather uncommon (Tantawichien, 2012). In vitro experiments in 1998 indicated DENV directly infects neurons, which results in permanent damage (Despres et al., 1998). In 2000, Jan and colleages demonstrated that phospholipase A2 (PLA2) activation, cytochrome C release from the mitochondria, superoxide anion production, and nuclear factor-kB translocation could lead to neuronal apoptosis (Jan et al., 2000).
The major pathogenetic mechanisms include endothelial cell disfunction and development of coagulation disorders. Therefore, the clinical signs of dengue fever mostly comprise endothelial cell damage, enhanced vascular permeability, and increased plasma leakage (Halstead and Cohen, 2015). The dengue virus shows tropisms mainly for monocytes, macrophages, and dendritic cells (Jessie et al., 2004;Martina et al., 2009).
Though belonging to the same flavivirus group, Japanese encephalitis virus (JEV) is a proven neurotropic virus, and mainly targets neuronal cells (Kimura-Kuroda et al., 1993). Nearly 75% of symptomatic patients show manifestations of encephalitis, which often leads to various neurological complications or patient fatality Sarkari et al., 2012a,b). JEV infection is considered a leading cause of pediatric encephalitis (Thongtan et al., 2012).
Little is known about the mechanism by which the virus spreads to the CNS as well as its brain tropism (Myint et al., 2007). Ghosh Roy et al. assumed that direct spread to the CNS occurs when mosquitos directly bite into blood vessels (Ghosh Roy et al., 2014).
Apoptosis induced by JEV involves the following three related mechanisms: direct neuron infection, infection of other CNS cells (e.g., microglia and astrocytes), and inflammatory reactions (Chen et al., 2004;Raung et al., 2007;Das and Basu, 2008;Swarup et al., 2008). As a result, CNS cell infection by JEV causes considerable neuronal apoptosis (Chen et al., 2004;German et al., 2006;Ghoshal et al., 2007). Serious vascular congestion, cerebral edema, neuronal death, astrocyte activation, and microglial proliferation are found in fatal cases of JEV infection (Misra and Kalita, 2010). Therefore, it is relatively easy to understand why up to one-third of JE cases hospitalized die while about one-half of all survivors show permanent neurological sequelae, even individuals with apparent recovery (Kumar et al., 1993;Solomon et al., 2000a). However, most patients even with neurological manifestations in dengue infection have an unexpected good recovery with no obvious sequelae.
Dengue is gradually becoming a main public health problem worldwide. A growing number of related studies demands increased awareness and understanding of the neurological complications of dengue infection. Physicians, especially neurologists, will continue to play important roles in its diagnosis and treatment. Pathogenesis, detection, antivirals, vaccines, environmental risk reduction, and vector control still deserve further studies.

AUTHOR CONTRIBUTIONS
GL: Designed the study, collected data, wrote the manuscript. YL: Designed the study, reviewed the manuscript. XL: Designed the study, reviewed the manuscript. ZN: Collected data and edited the manuscript.