The neurological complications of chikungunya virus: A systematic review

Summary We performed a systematic review on the neurological complications of chikungunya virus. Such complications are being reported increasingly, owing primarily to the scale of recent epidemics but also to a growing understanding of the virus' neurovirulence. We performed a thorough literature search using PubMed and Scopus databases, summating the data on all published reports of neurological disease associated with chikungunya virus. We appraised the data for each major condition in adults, children, and neonates, as well as evaluating the latest evidence on disease pathogenesis and management strategies. The review provides a comprehensive summary for clinicians, public health officials, and researchers tackling the challenges associated with this important emerging pathogen.

diseases, and indeed certain diseases can be caused by more than 1 pathogen (postulates 1 and 2). Furthermore, modern technologies such as polymerase chain reaction (PCR) assays have increased our detection rate of certain pathogens over isolation in culture (postulate 3).
We therefore adapt these postulates today in studies regarding causality. Neurological disorders associated with arbovirus infection have an added layer of complexity, owing to the different samples used for testing for the presence of the viruses. The strongest evidence of causality comes from demonstrating that the virus is in the central nervous system (CNS), which is most often shown by detecting viral RNA in the CSF by PCR; alternatively, the virus may be cultured. In fatal cases, autopsy material may be positive by PCR. For many patients, the virus has cleared from the CSF by the time they present; in which case, the detection of CSF IgM antibody by enzyme-linked immunosorbent assay is considered diagnostic. Interpretation is complicated for flavivirus infections because a positive Zika-IgM test can result from crossreactivity of serum containing antibodies against dengue virus. 21 Because chikungunya is an alphavirus, there is no serological crossreactivity with the flaviviruses, making diagnosis more straightforward (in areas where other alphaviruses are not circulating). It is not known for how long virus, RNA, or IgM remains detectable in chikungunyaassociated neurological disease, and whether this differs for the different neurological disorders. By analogy with similar arboviruses, we might expect the virus to be detectable for the first few days of illness, at which point it is replaced by antibody, which remains for weeks to months. Chikungunya virus is more readily detected by PCR or culture in the blood, because of its long and high viraemia; IgM antibody can also be detected in the blood. However, a positive blood test in a patient with neurological disease does not necessarily mean the virus caused the disease; infection may be coincidental, and care must be taken to exclude other possible causes. The virus can also be detected in urine, saliva, semen, and milk, [22][23][24] but the same caveats apply.

| CLINICAL FINDINGS
Seroprevalence studies have reported a range of asymptomatic rates of chikungunya infection, from 3% to 47%. 25 In acute symptomatic infection, following an incubation period of approximately 3 days, 26 there is an abrupt onset of fever, headache, rash, arthralgia, and myalgia, which typically last for 1 to 2 weeks. 27 After this, seroconversion likely confers lifelong immunity. 28 As well as neurological manifestations, chikungunya virus is associated with complications of the cardiovascular, renal, respiratory, hepatic, gastrointestinal, and adrenal systems, sometimes collectively referred to as "atypical features." [29][30][31] However, a disorder of the nervous system appears to be the most common severe complication of chikungunya infection (Table 1). In 2 studies investigating manifestations of chikungunya in patients requiring intensive care, a neurological disorder was the primary issue in 61% 32 and 79% 33 of chikungunya-infected patients. In chikungunya-associated neurological disease, the clinician must be vigilant for other complications in the same patient, a pattern also described in dengue but rarely in Zika infection. Amongst 99 cases of chikungunya-associated neurological disease described in a study in India, 69 also had other complications involving, for example, the renal, hepatic, and respiratory systems. 30 These patients should be managed using a multidisciplinary approach.  Figure 1); autopsy data were included in one report found. 34 Table 2 provides a summary of all cases of reported chikungunya-associated neurological disease. In total, we found 856 cases of chikungunya-associated neurological disease in the literature; 796 (93.0%) were in adults and children infected directly via mosquito, 60 (7.0%) were in neonates infected vertically from mother to child ( Figure 2). Because patients were investigated to variable extents, we have categorised them according to the presenting clinical syndromes of encephalopathy, myelopathy, neuropathy, combinations of these, and neuro-ocular disease; we then provided diagnoses, treatment, and outcome where available. A limitation of our study was that we could not incorporate diagnostic criteria, given the range of information available amongst all the cases published. The most common clinical presentation of neurological disease associated with adult and child chikungunya infection was encephalopathy; it accounts for 322 (40.5%) of the 781 patients described. Excluding ocular disease, 474 (77.3%) of all adult and child cases had a pure CNS disorder, 82 (13.4%) had a pure peripheral nervous system disorder, and 57 (9.3%) had a disorder of both the CNS and peripheral nervous system.
Cases in which coinfection of chikungunya with another neurovirulent arbovirus was detected are not included in Table 2 but are further discussed below.
Neurological disease following chikungunya virus infection was first reported during an outbreak in 1964 in Madras, India. 65 Four cases with chikungunya confirmed serologically or by viral isolation were described. Two presented with a meningoencephalitic picture ("delirium or coma, and signs of meningeal irritation with nuchal rigidity and Kernig's sign, sluggish pupillary reaction etc"), one with acute flaccid paralysis and elevated CSF protein, suggestive of Guillain-Barré syndrome (GBS), and one with transient dysarthria. Since then, neurological manifestations have been reported throughout the Indian Ocean, South Asia, the Pacific islands, Southern Europe, the Caribbean, and South America, ranging from mild behavioural disorders to severe acute syndromes of both CNS and peripheral nervous system ( Figure 3).
Given the large spectrum of neurological disease and scarce epidemiological data, estimating the incidence of neurological disease amongst all systemically symptomatic chikungunya infections is difficult. In one study from the 2006 Indian outbreak, 39  It has been suggested that severe complications of chikungunya infection typically arise in those with co-morbidities. 96 The above epidemiological study describing 610 atypical cases of chikungunya infection showed that underlying respiratory disease, cardiac disease, and hypertension were all independently associated with severe complications, including neurological disease. 29 However, a study from India of 124 atypical chikungunya cases did not identify co-morbidity as a significant risk factor for systemic complications or fatality 30 ; similarly, in a case series of chikungunya-associated GBS, 6 (67%) of 9 cases did not have any co-morbidities. 70 (9) 2007, India 77 10 IgM ser 1-6 wk … ↓visual acuity (10), pain (10), bilateral (3), visual field defects (10), disc oedema (10), RAPD (7), delayed VEP Papillitis (7), retrobulbar neuritis (1) (3), children (9); associated with fever (12), focal (2),

| Myelopathy and myelitis
Chikungunya virus can cause myelopathy, symptoms of spinal cord disease, which may present with limb weakness, sensory changes, hyperreflexia, and bowel and bladder disturbances, depending on the level of the lesion and extent to which the cord is involved. If cord inflammation is confirmed by MRI, a CSF pleocytosis, or elevated CSF IgG index, showing local immunoglobulin production, then it is classified as myelitis. 106 The incidence of spinal cord disease after chikungunya infection is not known, but it is likely to be less than that of encephalopathy,

| Acute disseminated encephalomyelitis
Like other acute viral infections, chikungunya can trigger an acute inflammatory syndrome involving the brain parenchyma and spinal cord, which is thought to be an immune-mediated response to infection, rather than due to direct viral invasion. The diagnosis of this monophasic illness is usually based on finding focal or multifocal, poorly demarcated white matter demyelinating lesions on MRI. 107,108 Six cases of acute disseminated encephalomyelitis (5 adults and 1 child) have been described, with the disease starting 5 to 16 days after the initial fever-arthralgia-rash symptoms of chikungunya infection. Patients presented with a variety of neurological features, including headache; drowsiness; cranial nerve involvement such as facial nerve palsy, vertigo, nystagmus, and bulbar weakness; limb weakness; sensory disturbance; and urinary retention. MRI of the brain and/or spine suggested demyelinating pathology. All 6 were treated with intravenous methylprednisolone; the outcome varied between good clinical and radiological recovery 91 and permanent neurological disability with confinement to a wheelchair and longterm urinary catheterisation. 92

| Guillain-Barré syndrome
Chikungunya-associated peripheral neuropathy without CNS disease has been described for 72 patients in case reports or series, the majority of whom were described as having GBS. In one series of 4 patients with acute flaccid paralysis, no CSF or neurophysiology results were reported, making diagnosis difficult. 69 Other causes of acute flaccid paralysis, such as anterior myelitis, have not yet been reported in association with chikungunya virus.
Two studies from Réunion Island showed an increased incidence of GBS following a large outbreak of chikungunya. One from the 2014 to 2015 outbreak reported 9 patients with chikungunya-associated GBS, representing a 4-to 9-fold increase in the island's annual GBS incidence. 70 The increase in the incidence of GBS following the 2006 chikungunya virus outbreak on Réunion Island was estimated to be around 22% compared with the year before. 68 A study from Martinique and Guadeloupe also showed an increase in incidence of GBS following the 2014 chikungunya virus outbreak, but to a lesser extent (2-fold). 71 Clinically, chikungunya-associated GBS resembles GBS associated with other infections such as Campylobacter jejuni, presenting with symmetrical, bilateral flaccid weakness, often with paraesthesia and/ or cranial nerve palsy. 109 (Table 2). 67,68,73 Unlike infections such as C. jejuni, which is associated with a more severe pure motor variant of GBS, 110 chikungunya appears to be associated with the full range of GBS variants, as determined by neurophysiological studies, including disorders of motor and sensory axons, and myelin sheaths, sometimes in combination (see Table 2). Further investigation is required to determine risk factors and markers for the different GBS variants, including antiganglioside antibodies.

| Ocular complications
Although photophobia and conjunctivitis are associated with the acute phase of chikungunya infection, 111 many later ocular complications have been described up to 12 weeks after infection, which may require emergency management. These include disease of the uvea, retina, and optic nerve. As well as inflammation, other pathologies have been described, including retinal detachment, intraretinal haemorrhage, and branch retinal artery occlusion (Table 2). 38 Disease relapse has been reported in one report, which described bilateral uveitis and retinal detachment, with loss of visual acuity starting 4 days after symptoms of chikungunya infection. 82 Having received a week's course of oral and topical steroids and recovered within 6 weeks, the patient re-presented 3 months later with floaters and keratic precipitates; clinicians should be vigilant for relapse in such cases.

| Disease affecting multiple components of the nervous system
As well as the distinct syndromes described above, chikungunya infection is associated with complex diseases involving multiple parts of the nervous system causing, for example, encephalomyelopathy (23 patients identified in our literature review), myeloneuropathy (24), and encephalomyeloneuropathy (24). Where follow-up data were available, unlike in pure myelopathy, more of these patients had an unfavourable outcome (8 deaths, 4 no improvement) compared with those who improved (8).

| Other
A handful of other neurological disorders have been associated with chikungunya, albeit in smaller numbers. A study from the Réunion outbreak described 32 patients presenting with behavioural changes including attention disorders, irritability, and memory issues. 37 Other reports have described febrile seizures, isolated cranial nerve palsies, stroke, and hearing loss (Table 1), and one report (in press) describes the possibility of an association with chronic fatigue syndrome.
Although it is difficult to definitively associate these isolated disorders with chikungunya infection, the full spectrum of chikungunya-associated neurological diseases appears to be broad.

| Perinatally acquired neurological disease
Most of the evidence on chikungunya causing neonatal disease relates to transmission in the intrapartum period, rather than earlier in pregnancy. A wide range of severe manifestations has been described affecting neonates whose mothers had acute chikungunya infection near the time of delivery (Table 3). A case series from Colombia from 2014 to 2015 described 8 infants who required admission to an intensive care unit after contracting chikungunya infection perinatally. 31 All mothers and neonates were PCR and IgM positive for chikungunya in serum; the neonates presented with severe diseases including meningoencephalitis, respiratory distress, sepsis, necrotising enterocolitis, myocarditis, and pericarditis.
One study from Réunion Island described 739 mothers who experienced symptoms of chikungunya infection during pregnancy, 39 of whom were symptomatic in the intrapartum period (between 2 d before and 2 d after delivery). 105 Of these 39 mothers, all infants were asymptomatic at birth, but 19 developed acute disease a median 4 (range 3-7) days after delivery, giving a vertical transmission rate of approximately 50% for mothers symptomatic in the intrapartum period. The initial clinical features in the affected neonates included fever, distress, poor feeding, petechiae, and a maculopapular rash. Nine cases (47%) were reported to have developed encephalopathy. Cerebrospinal fluid for all 9 showed normal biochemistry and white cell counts, and chikungunya PCR was detected in 5. Magnetic resonance imaging data from this study, combined with a later follow-up study, 120 show that amongst neonates developing encephalopathy or encephalitis after perinatally acquired chikungunya, severe white matter injury is well characterised in a 3-stage pattern: cytotoxic brain oedema (ischaemia), vasogenic oedema (reperfusion), and mass reduction (demyelination). 105,120 In addition to these severe features, hypotonia has been described in 17 neonates with chikungunya infection from Réunion Island, 97 as well as intracerebral haemorrhage secondary to clotting abnormalities. 105,112 Interestingly, neonatal spinal cord and peripheral nervous system diseases have not been reported.
With regard to prevention of transmission, no studies have found a protective effect of caesarean section. 105,114 The risk factors for vertical transmission are not understood, although in one study, the viral load of chikungunya in the placentas of the 19 transmitters was found to be significantly higher than in 13 nontransmitters. 105 Interestingly, one of the transmitters gave birth to dizygous twins, of whom one acquired infection but the other did not.
In addition to overt diseases in some perinatally infected neonates soon after delivery, there is also evidence for impacts on longer-term development. In one study comparing the neurocognitive function at approximately 2 years of age for 33 children with and 135 without perinatal chikungunya infection, significant differences in development quotients were found, including movement, coordination, language, and sociability. 120 Importantly, even those infected at birth without obvious clinical features of neurological disease, such as encephalopathy, had significantly worsened neurocognitive function than uninfected children. Thus, the neurological effects of vertically transmitted chikungunya may not be obvious at birth, emphasising the importance of follow-up of this cohort. The 12 cases with encephalopathy at birth showed still more severe developmental deficit, including cerebral palsy and microcephaly.
Maternal chikungunya virus infection earlier in pregnancy does not appear to affect the fetus; in the above study investigating 739 mothers with chikungunya infection, 700 were symptomatic outside of the intrapartum period, and none of these infants developed symptoms of chikungunya. 105 Of note, in 3 of 7 miscarriages occurring before 22 weeks, chikungunya RNA was detected in amniotic fluid for all 3 and placenta and fetal brain for 2. However, no significant increase in antepartum fetal deaths was seen during the chikungunya outbreak as compared with previous years. Another study of 1400 mothers from Réunion Island 121 found no effect of antepartum chikungunya infection on pregnancy outcomes. Together, these data would suggest that although antepartum congenital infection has been detected in miscarried fetuses, given the lack of epidemiological evidence for a causal association between infection and miscarriage, this may have been an incidental finding.
Zika virus is also now known to cause devastating neurological disease in neonates, which is of global concern. 122 Whereas chikungunya appears to be most damaging around the time of birth, Zika has been associated with neurological sequelae in infections at all stages of pregnancy. 101  A study from India compared the cytokine profile for patients with and without neurological complications following chikungunya infection. 128 Of those with neurological disease, 4 had encephalitis and 1  129,130 In vivo, subcutaneous inoculation in macaques resulted in morphological changes in astrocytes, including cell body hypertrophy and alteration in the pattern of branching of their primary processes. 131 In relation to the latter, immune-mediated hypothesis, this study also showed upregulation of TLR2 in grey matter astrocytes, a gene associated with the innate immune response. The clinical neurological state of the macaques was not reported, which adds to the uncertainty of whether the immune response was protective or pathogenic in these cases. Another study that subcutaneously inoculated mice with chikungunya virus detected upregulation of TLR3 in the brain, a gene that is also associated with the innate immune response. 132 Amongst other clinical signs, these mice developed hind-limb paralysis, dehydration, and weight loss, and 25% of them died after 1 week. Interestingly, pretreatment with polyinosinic: polycytidylic acid (a TLR3 agonist and interferon inducer), which caused upregulation of proinflammatory cytokines, chemokines, antiviral genes, and IFN-β, was protective clinically and promoted viral clearance from the brain, arguing for a protective innate and adaptive immune response, at least in CNS disease. In concordance, faster viral clearance after chikungunya infection was seen in wild-type mice compared with a TLR3-knockout model; this was thought to be secondary to increased antibody-neutralising activity in the wild-type mice. 133 Chikungunya-infected TLR3-knockout mice had increased viral dissemination throughout the viscera, including the brain. Along with a U-shaped pattern of age-specific incidence, this critical role of TLR3 is reminiscent of susceptibility to HSV encephalitis. Neurological disease secondary to chikungunya has been reported in areas with both ECSA (or ECSA-diverged Indian Ocean lineage) and Asian strains, but whether these strains have differing neurovirulence is unknown. One study compared the effect of intracerebral inoculation of Asian and ECSA-diverged strains in mice. 134 Both spread within the brain to a similar extent, but the Asian strain was associated with higher mortality than the ECSA-diverged strain. Upregulation of a gene associated with apoptosis was seen in the former, whilst antiapoptosis, antiviral, and CNS protective gene upregulation were seen in the latter.
This potentially suggests a higher neurovirulence of the Asian strain, and comparative clinical data from countries such as Brazil, where both strains are circulating, will be useful.
On neonatal neurological disease, given that caesarean section is not protective, vertical transmission is unlikely to occur via the birth canal, as is the case in other neonatal infections such as herpes simplex. 135 Furthermore, the placenta seems to act as a barrier to transmission, as one study reported (as unpublished data) that placental cells from infected neonates were negative when labelled with antichikungunya antibody. 105 One hypothesis raised by the authors is that uterine contractions result in breaches of this placental barrier, allowing passive passage of the virus.

| Management
There are currently no specific antiviral agents or vaccines for chikungunya virus. 136 Various in vitro compounds active against chikungunya virus have been reported, including both direct-acting and host-targeting antivirals; however, most of these compounds have yet to find their way into in vivo models and clinical trials. 137 Two of the few that have been tested clinically, chloroquine and ribavirin, are already widely used in the treatment of other diseases and have a known safety profile. However, chloroquine was found not to have any benefit for arthritic chikungunya when compared with a nonsteroidal anti-inflammatory drug in a randomised clinical trial. 138 Ribavirin, on the other hand, had promising results in a small case series of 10 patients with severe arthritis post-chikungunya infection. 139 To the best of our knowledge, no antiviral has been evaluated in the management of chikungunya-associated neurological disease, which therefore remains the same as that of neurological disease without associated chikungunya infection: for patients with encephalitis, those with a reduced Glasgow coma score (GCS) require assessment by intensive care specialists and may need intubation, ventilatory support, correction of electrolyte abnormalities, management of raised intracranial pressure, and enhancement of cerebral perfusion pressure. 140 In patients with myelitis, corticosteroids are the standard first-line treatment, despite the lack of trial evidence for their use in this scenario. 141 The management of GBS focuses on immunotherapy with intravenous immunoglobulin or plasma exchange, and ventilatory support if the innervation of respiratory muscles is affected. 109 Although not yet commercially available, it is hoped that a vaccine for chikungunya is on the horizon. Two phase 1 clinical trials have shown a good safety and immunogenicity profile to date. 142   Patients in areas endemic for chikungunya, Zika, or dengue presenting with an acute neurological disorder should be investigated for all 3 arboviruses Neonates born to mothers experiencing symptoms of chikungunya infection near the time of delivery require admission and observation for signs of vertical transmission for at least 7 d postpartum, as they may be asymptomatic for the first few days of life Encephalitis is the most commonly reported neurological complication associated with chikungunya; encephalitis has a worse prognosis than encephalopathy alone; a CSF pleocytosis is not always seen Neonates born to mothers infected outside of the peripartum period are usually unaffected by chikungunya virus In myelitis associated with chikungunya, CSF pleocytosis and magnetic resonance imaging changes are not always seen Caesarean section does not appear to be protective in vertical transmission of chikungunya Guillain-Barré syndrome associated with chikungunya follows a similar course compared with other infections such as Campylobacter jejuni; most patients recover after immunomodulatory treatment Neonates infected with chikungunya should be followed up for at least 2 y, regardless of symptoms in the first week of life; the neurodevelopment of those without clinical encephalopathy at birth can still be affected Disease of both the central and peripheral nervous systems in the same patient can be seen in association with chikungunya infection Ophthalmological complications associated with chikungunya have been reported both at the time of infection and up to 12 wk after; some reports describe treating with steroids, recovery is variable Following chikungunya infection, complications of other organs can also occur at the same time as disease of the nervous system; such cases should be managed using a multidisciplinary approach There is currently no available antiviral treatment or vaccine for chikungunya