Fatores associados à Síndrome Respiratória Aguda Grave em uma Região Central do Brasil Factors associated with Severe Acute Respiratory Syndrome in a Brazilian central region

Resumo A Síndrome Respiratória Aguda Grave (SRAG) deve ser notificada e investigada. O objetivo do estudo foi analisar o perfil epidemiológico e fatores associados à internação e óbitos por SRAG notificados em Goiás. Estudo de coorte retrospectiva, com dados das fichas de investigação do Sistema de Informação de Agravos de Notificação Influenza Web. Métodos de análise multivariada foram realizados para verificar a associação entre variáveis de exposição com os desfechos internação em UTI e óbito. Entre 2013 e 2018 foram notificados 4.832 casos de SRAG em Goiás. O principal diagnóstico etiológico foi influenza A (22,3%) com o subtipo A (H1N1pdm09) predominante, seguido pelo Vírus Sincicial Respiratório. 34,6% dos pacientes necessitaram de internação em UTI e 19% evoluíram para o óbito. Maior tempo de início do tratamento com antiviral foi associado à maior chance de internação em UTI, enquanto a não vacinação prévia contra a influenza, maior tempo para início do antiviral e idade mais avançada foram associados à maior chance de óbito. O estudo mostrou uma elevada frequência de doenças respiratórias provocadas pelo vírus Influenza no estado de Goiás e que a gravidade da síndrome, caracterizada pela internação em UTI e óbitos, está associada com o tempo de início do tratamento com o antiviral, o status vacinal e a idade do paciente. Palavras-chave Influenza Humana, Vigilância, Epidemiologia, Síndrome Respiratória Aguda Grave Abstract Severe Acute Respiratory Infection (SARI) is a notifiable syndrome that must be investigated. This study aimed to analyze the epidemiological profile and factors associated with SARI-related hospitalization and deaths reported in Goiás. Retrospective cohort study, with data from the investigation files of the Notifiable Diseases Information System’s Influenza Web. Multivariate analysis methods were employed to verify the association between exposure variables with the outcomes of ICU admission and death. A total of 4,832 SARI cases were reported in Goiás from 2013 to 2018. The primary etiological diagnosis was Influenza A (22.3%) with the predominant subtype A (H1N1pdm09), followed by the Respiratory Syncytial Virus. A total of 34.6% of the patients required ICU admission, and 19% died. A longer time to start treatment with antivirals was associated with a higher likelihood to have an ICU admission, while a previous non-vaccination against Influenza, longer time to start treatment, and older age were associated with a higher likelihood to suffer death. The study showed a high frequency of respiratory diseases caused by the Influenza virus in Goiás and that the severity of the syndrome, characterized by ICU admission and deaths, is associated with the start of antiviral treatment vaccine status, and patient’s age.


introduction
Severe Acute Respiratory Infection (SARI) is an infectious viral respiratory syndrome caused by influenza viruses and other etiological agents, such as the respiratory syncytial virus (RSV), parainfluenza, and adenovirus, which infect the upper respiratory tract 1 . Patients have a flu-like condition associated with dyspnea or tachypnea (respiratory rate equal to or above 20 incursions per minute) or hypoxemia, with oxygen saturation (SpO 2 ) < 95% in room air 2 .
It is a disease of epidemiological relevance that must be notified and investigated. The epidemiological monitoring of SARI in Brazil occurs through the epidemiological surveillance of the disease by completing notification forms through the Notifiable Diseases Information System (SINAN), and the operation of sentinel units. These units monitor SARI-related hospitalized cases and deaths to identify circulating respiratory viruses 2,3 . Thus, all suspected SARI cases are hospitalized, notified, and investigated for the disease 1 .
Epidemiological surveillance of SARI allows preventing and monitoring severe cases of the syndrome, identifying circulating viral types and subtypes, performing antigenic and genetic analysis of viruses, identifying new viral subtypes, monitoring resistance to antivirals and demand for care by disease, and controlling possible outbreaks, aiming at public health decision-making for its prevention and control 4 .
Globally, the circulation of influenza viruses remains high, and seasonal A virus is the most prevalent 5 . According to the epidemiological bulletin of the Ministry of Health 6 , Brazil has the same circulation profile, with the highest prevalence of the Influenza A H1N1 pandemic virus 2009 (H1N1pdm09). This scenario has recently changed with the new coronavirus pandemic, which started in late 2019 in China, with first Brazilian cases in early 2020 7 .
SARI can progress to complications, which lead to an increased risk of ICU admissions, mechanical ventilation, and even death 8 . Some factors are associated with a worse prognosis, such as the presence of chronic diseases, age extremes, viral strain, no previous vaccination, and use of antivirals 72 hours after the onset of symptoms, characteristics that should be investigated 9,10 .
Thus, this study aimed to analyze the epidemiological profile and the factors associated with hospitalization and deaths from SARI in a central Brazilian region from 2013 to 2018.

Methods
This is a retrospective cohort study with data analysis regarding notified SARI cases and their progress during hospitalization in Goiás. Goiás is located in the Midwest region, with an area of 340,125,715 km², and is the seventh-largest Brazilian state in territorial extension. Its HDI ranks eighth in the country, at 0.735 11 .
According to the definition of the Ministry of Health 2 , all reported SARI cases in SINAN's Influenza Web, residing in the state of Goiás, regardless of age, were considered eligible for the study. The Ministry of Health defines SARI as the individual of any age with a flu-like syndrome (fever, cough, or sore throat, and at least one of the following symptoms: headache, myalgia, or arthralgia) associated with dyspnea or oxygen saturation SpO 2 < 95% 2 .
Data was collected on April 10, 2019. In 2012, a change was implemented in the notification form, and epidemiological surveillance directed the monitoring of Influenza by a new pandemic subtype for SARI. We decided to use 2012 data to ensure the performance of analyses with the same variables homogeneously. Thus, the data reported between 2013 and 2018 were collected. This project was approved by the Research Ethics Committee linked directly to the "Leide das Neves Ferreira" Center for Excellence in Education, Research and Projects (GCEEPP-LNF) and the Research Ethics Committee linked to the Hospital das Clínicas of the Federal University of Goiás (UFG), and met the conditions established by Resolution MS/CNS 466/2012.

Data statistical analysis
Absolute and relative frequencies were obtained in the descriptive analysis according to each category under analysis (sociodemographic characteristics and data on the prevention, treatment, and development of the disease). Multivariate data analysis methods were used to verify the association between exposure variables and outcomes (deaths and ICU admission), and previous vaccination, use of antivirals, presence of the Influenza A/B virus, RSV, other respiratory viruses, treatment time in days, and age group, were considered variables of exposure.
Initially, a bivariate analysis was performed, testing the association of exposure variables with outcomes by calculating Odds Ratio (OR) and p-value. The variables whose association with the outcome was p ≤ 0.20 were included in the mul-tivariate logistic regression model. This model was progressively saturated with each variable's addition, observing its effects on the precision of the final model on the adjustment of the other variables for inclusion in the definitive model. The variables associated with the outcome finally showed a p-value < 0.05. IBM SPSS Statistics software version 25.0 was used for statistical data analysis.

results
A total of 4,832 SARI cases were reported in Goiás from 2013 to 2018. The highest percentage of notifications occurred in females (53.4%) and the 20-59 years' age group (45.8%). Other sociodemographic data are shown in Table 1.
Regarding the prevention, treatment, and development of the syndrome, only 20.7% of patients were previously vaccinated against Influenza, and 70.7% of them used antivirals. ICU admission was found in 34.6% of the cases, while 19% died (Table 1).
A predominance of Influenza A virus was observed throughout the study, followed by RSV. Influenza A virus (H1N1pdm09) was the predominant virus subtype. As for cases without etiology, defined as unspecified SARI, we observed 2,796 patients, which corresponds to approximately 58% of notifications, with an increase in notifications in 2016 and 2018. These same years recorded a significant increase in cases compared to other seasonal factors and, consequently, increased ICU admissions and deaths ( Figure 1). Table 2 shows the results of the bivariate and multivariate analyses between the exposure variables with the death outcome. In the bivariate analysis, a higher likelihood of death was observed in individuals who did not use antivirals, whose SARI was caused by Influenza A or B virus, with a higher number of days to start treatment, and older age. In the multivariate analysis, in turn, the variables use of antivirals, Influenza A or B infection, and RSV, lost association with the outcome, keeping the association between the time of onset of treatment and age group. A higher likelihood of death was also observed in the multivariate analysis in individuals who had not previously been vaccinated against Influenza.
For the outcome of ICU admission, in the bivariate analysis, the non-use of antivirals, infection by Influenza A or B, RSV, and other respiratory viruses reduced the likelihood of ICU admission. At the same time, the longer time to start treatment increased the likelihood of this complication. In the multivariate analysis, the variables infection by Influenza A or B and RSV continued to be negatively associated with the outcome, while the time to start treatment remained positively associated with the outcome (Table 3).

Discussion
This study allowed observing SARI notifications in the state of Goiás for six years. Analyzing the confirmed cases' sociodemographic characteristics, we observed no significant difference regarding gender, unlike studies carried out in New Zealand, the U.S., and France, in which females were predominant [12][13][14][15][16] . Regarding the age group, the largest number of people affected was aged between 20 and 59 years, a group that is not part of the indication for vaccination against Influenza by the Brazilian Ministry of Health. A high frequency of cases was also observed in groups considered to be at higher risk, children under five years old (28.6%) and older adults above 60 years (15.0%), as seen in other studies 12,17,18 . In the literature, individuals at extreme ages are at higher risk for respiratory conditions, such as Influenza and pneumonia 19,20 .
Brazil has one of the most extensive immunization programs globally, with a considerable volume of doses administered against various diseases, and is provided free of charge 21 . Among these vaccines is the one against Influenza, incorporated in the National Vaccination Program (PNI) in 1999 22 , which is made available free of charge to risk groups nationwide, among them: children aged six months to under six years of age, pregnant women, puerperae, and patients with NCDs. In this study, we observed that 59.1% of the SARI individuals had not received previous vaccination against Influenza. It is worth mentioning that the largest number of cases in our study is in the age group not covered by vaccination provided by the PNI. While not provided to the entire population through the SUS, influenza vaccination is available in the private health network. It is essential to highlight that influenza vaccination, as a public health measure, reduces hospitalizations, especially in people at risk (children under five, older adults, and pregnant women) 23 . However, several factors influence the vaccine's effectiveness, such as the compatibility of the circulating strains with the vaccine strains, the individual's immune response, and previous vaccination by the same strain of the current vaccine 24,25 .
With a suspected disease, all patients should start antiviral treatment as early as possible, even without the etiological identification 2 . Antiviral administration can reduce the duration of symptoms, viral load, and the transmission capacity 26 . Failure to use antivirals or inappropriate use (more than 48 hours after the onset of symptoms) is associated with virus infection complications, such as aggravation followed by ICU admission and progression to death 8 . The nonuse of the medication was observed in 24.9% of patients in this study. Thus, it is necessary to strengthen actions for the proper management of patients with SARI as a strategy to reduce severe cases of this disease.
We observed that the influenza virus is the most prevalent etiological agent, followed by RSV. A higher frequency of virus A (H1N1pdm09) was noted among the influenza viruses, followed by A (H3N2) and the influenza B virus. Year-toyear variations were observed for both pathogens, which may be due to different strains and circulation patterns of viruses at every season 27 , and highlights the importance of estimating the disease burden associated with virus infection and strengthening surveillance actions in the epidemiological investigation of the syndrome, which corroborates public health measures 4,28 . This same viral circulation pattern is observed in other studies, where A viruses (H1N1pdm09) represent the majority of identified viruses [29][30][31] .
It is interesting to note a significant increase in SARI cases in the state of Goiás   erage for some specific groups, such as patients with comorbidities, was below the 90% target in previous years (2015 and 2017), favoring higher viral circulation in the following seasons. Identifying the influenza virus in symptomatic patients is associated with an increased risk of respiratory complications, represented by deaths and ICU admission, besides other contributing factors, such as the patient's age, viral strain, vaccine status, and antiviral use 27 . Our study observed that 19% of patients died, and 34.6% were admitted to the ICU, as seen in other studies 8,32 . Also, we observed that the likelihood of having a severe outcome (death and ICU admission) was higher in young adults or older adults than children, in those who had not previously been vaccinated against Influenza and with a higher number of days to start antiviral treatment. According to Naudion et al. 33 , this virus is the most associated with the complications of the syndrome and, likewise, the non-use of antivirals and the absence of immunization are related to the complications 33 . Other studies indicate that patients who used antivirals and had previous vaccination for Influenza had a significant reduction in the risk of complications 19,22 . Regarding the age group, studies in the literature have also shown a higher likelihood of complications at older ages 8,33 .
As a limitation of the study, we can highlight that the frequency of cases and deaths due to SARI can be underestimated due to their passive notification and the sensitivity of health professionals in identifying suspected cases. Underreporting of cases in the SINAN should not occur, as SARI is mandatory for health surveillance 34 . The number of unidentified cases, but capable of transmitting the disease, is a critical epidemiological characteristic that modulates the potential for the spread of the virus 35 . Another downside is the use of secondary data, which can sometimes be incomplete and with limitations on variables. Thus, local surveillance levels should be trained and make extensive use of data from information     systems to identify weaknesses and inconsistencies in the data, improve the quality of the systems, and make them effective 34 . SINAN and other Health Information Systems of the Brazilian Ministry of Health are a valuable source of health information that assists in planning health policies and programs, collaborating in the decision-making process, besides allowing assessment of the impact of interventions. Thus, even in the face of limitations, they are essential tools for public health and, especially, for disease epidemiological surveillance. Also, with the new coronavirus pandemic in the country in 2020, the SARI surveillance system is being strengthened and is an essential step for public health concerning the epidemiological and laboratory surveillance of respiratory agents.
conclusion Several cases of respiratory diseases caused by the Influenza virus and other etiological agents occur in the state of Goiás, and we observed that the severity of the syndrome, characterized by ICU admission and deaths, is associated with the patient's age, start time to treatment with antivirals, and vaccination status.

collaborations
KLR Araujo and YMF Ternes participated in the conception, design and writing of the article. EC Aquino, LLS Silva and YMF Ternes participated in the statistical analysis and interpretation of the data. All authors participated in the critical review and approval of the version to be published.