Neurological complications associated with emerging viruses in Brazil

Abstract Objective To test the hypotheses that emerging viruses are associated with neurological hospitalizations and that statistical models can be used to predict neurological sequelae from viral infections. Methods An ecological study was carried out to observe time trends in the number of hospitalizations with inflammatory polyneuropathy and Guillain‐Barré syndrome (GBS) in the state of Rio de Janeiro from 1997 to 2017. Increases in GBS from month to month were assessed using a Farrington test. In addition, a cross‐sectional study was conducted analyzing 50 adults hospitalized for inflammatory polyneuropathies from 2015 to 2017. The extent to which Zika virus symptoms explained GBS hospitalizations was evaluated using a calibration test. Results There were significant increases (Farrington test, P<0.001) in the incidence of GBS following the introduction of influenza A/H1N1 in 2009, dengue virus type 4 in 2013, and Zika virus in 2015. Of 50 patients hospitalized, 14 (28.0%) were diagnosed with arboviruses, 9 (18.0%) with other viruses, and the remainder with other causes of such neuropathies. Statistical models based on cases of emerging viruses accurately predicted neurological sequelae, such as GBS. Conclusion The introduction of novel viruses increases the incidence of inflammatory neuropathies.

Cytomegalovirus, and influenza A, the disorder has also been associated with dengue virus and Zika virus. 3 During the Zika virus epidemic in Rio de Janeiro, Brazil, during the period 2015-2016, it was shown that neuropathies increased. 4 However, until now there has not been a baseline of GBS cases prior to the Zika virus epidemic to provide a comparison with GBS rates reported during the epidemic. To assemble such a baseline, we investigated the extent to which the increase in GBS cases in 2015-2016 was greater or less than other increases over the two previous decades.
We compared a historical time series of inflammatory polyneuropathies, which included GBS, to time series of influenza, dengue virus, and Zika virus in the state of Rio de Janeiro, Brazil. Further, we analyzed medical charts of arbovirus infections that resulted in hospitalization with neurological disease. Our objectives were to explore whether epidemics of emerging viruses are associated with neurological hospitalizations in Brazil and whether models could be used to predict neurological sequelae from viral infections.

| MATERIALS AND METHODS
We carried out an ecological study in which we observed time trends To quantify trends in GBS from 1997 to 2017, we compared the number of cases reported to SIH in a given month to the number the following month. We assessed whether the difference in the number of cases from 1 month to the next was significant using Farrington's outbreak rate test, which is a regression model that predicts the expected number of cases in a given month based on the history of cases in previous months. 5 Next, we evaluated the extent to which cases of GBS could be explained by cases of Zika virus and Zika + dengue virus. We used the hhh4 model for areal time series of counts in the statistical software R version 3.5. 1. 6 hh4 is a multivariate model in which the number of cases of a disease in a geographic region depends on the number of cases in the region in previous months and a neighborhood matrix that represents the probability of disease spread between regions. To prepare the data for the model, we tabulated the number of cases of GBS, Zika virus, and dengue virus per month in Rio de Janeiro's nine health districts. The number of cases of GBS in district i in month t depended on the population of i, cases of dengue virus and Zika virus in month t, the number of GBS cases in district i in month t-1, and the neighborhood matrix (supporting information Tables S1 and S2). The model included all months of the year and had sine-cosine effects of time to reflect seasonal variation in GBS incidence.
Since Zika virus and dengue virus have similar clinical presentations such as febrile illness and rash, 7 we pooled the Zika + dengue virus cases into the category arbovirus cases. We also analyzed clinical cases of Zika virus separately from the Zika + dengue virus cases.
While the analysis of syndromic surveillance data was an ecological study in which we observed temporal trends in GBS, we also carried out a cross-sectional study to assess the cause of GBS on clinical grounds. The  inflammatory polyneuropathies to SIH. We requested authorization from the hospitals' research departments to review charts and received permission at eight hospitals that represented 81% of admissions for inflammatory polyneuropathies. We visited these hospitals and used the authorization number from SIH to identify each patient's chart (the same patient had more than one authorization number if hospitalized repeatedly). We transcribed the physicians' notes about clinical signs and symptoms reported by the patient. Based on this information, we used the classification system proposed by Martyn and Hughes 8 to categorize the inflammatory polyneuropathy as either diabetes, genetic disorder, infectious disease, vaccine-related, alcohol, or other etiology. We also made note of whether the patient had a lab test for chikungunya, dengue, or Zika virus. There was seasonal variation in GBS incidence with 15% fewer cases during winter. Zika virus cases were a significant variable for explaining cases of GBS (hhh4 model, t=2.23, P=0.04; supporting information Table S3). The data were compatible with the null hypothesis that Zika virus cases provided a good fit to the GBS cases (calibration test z=0.836, n=189, P=0.403). As noted above, due to the difficulty of distinguishing Zika virus from dengue virus, we also pooled the data into the category arboviruses. The pooled data were not significant for explaining GBS cases (hhh4 model, t=1.59, P>0.05; supporting information Table S4). When we compared the goodness of fit of the Zika virus and Zika + dengue virus models, the former had lower errors (mean difference=0.044, permutation test P=0.0049).

| RESULTS
The greatest increase in GBS from 1997 to 2017 coincided with the introduction of Zika virus in 2015-2016 (Fig. 3). A predictive model based on Zika virus cases explains the temporal pattern of GBS cases better than a model based on Zika + dengue virus (Fig. 4). An example is the metropolitan area of the city of Rio de Janeiro (Fig. 4B). The model based on Zika virus predicts a major increase in GBS cases, while the model based on Zika + dengue virus predicts a smaller increase.
Dengue virus did not improve the predictive accuracy of the model because there were more than 200 000 dengue virus cases from 2001 to 2010 (supporting information Fig. S2), but few GBS cases were reported during this period. The fit of the Zika virus model to the GBS cases is poorer for the entire state of Rio de Janeiro (Fig. 4A) than for the metropolitan area of the city of Rio de Janeiro (Fig. 4B).
We reviewed charts at eight hospitals, which encompassed 80 (82.3%) of the 97 admissions for inflammatory polyneuropathies. These 80 admissions corresponded to 50 patient charts because some patients were admitted repeatedly. The most frequent causes of inflammatory polyneuropathy were arboviruses (n=14, 28.0%), followed by other viruses (18.0%) ( Table 1). Table 2  No patient was treated with plasma exchange. Five (55.6%) of the nine treated with intravenous immunoglobulin were transferred to other hospitals because the treatment was not available at the first hospital to which they were admitted. As of 2018, 12 (85.7%) of the 14 had no motor deficits, one still had facial paralysis, and one remained quadriplegic.

| DISCUSSION
While it is well established that infection with pathogens that cause inflammation can result in neuropathy, 8  accompanied by an increase in GBS, and the incidence of dengue virus was low in 2015-2016. 22 The predictive model was more accurate in the metropolitan area of the city of Rio de Janeiro, which is primarily urban, than across the whole state of Rio de Janeiro, which encompasses urban and rural areas. Poorer accuracy in rural areas could be due to less intensive epidemiological surveillance or fewer Aedes mosquitos in these areas.

SUPPORTING INFORMATION
Additional supporting information may be found online in the Supporting Information section at the end of the article.   Table S1. Notation for the hhh4 time series model to predict the number of cases of GBS per health district in each month. Table S2. Formulation of the hhh4 model. α ν , β ν , γ, δ, α λ , β λ , ϕ it e ψ i were estimated from the data using the library SURVEILLANCE in R. 6