Influenza‐associated severe acute respiratory infections in 2 sentinel sites in Lebanon—September 2015 to August 2016

Background Given the sparse information on the burden of influenza in Lebanon, the Ministry of Public Health established a sentinel surveillance for severe acute respiratory infections (SARI) to identify the attribution of influenza to reported cases. We aim to highlight the proportion of influenza‐associated SARI from September 1st, 2015 to August 31st, 2016 in 2 Lebanese hospitals. Methods The study was conducted in 2 sentinel sites located in Beirut suburbs and southern province of Lebanon. WHO's 2011 standardized SARI case definition was used. Data from September 1, 2015 to August 31, 2016 were reviewed, and all‐cause hospital admission numbers were obtained. Nasopharyngeal swabs were collected and tested by RT‐PCR. Descriptive and bivariate analyses were conducted using STATA 13. Results The 2 sentinel sites reported 746 SARI cases during the studied time frame: 467 from the southern province site and 279 from the Beirut suburbs site. SARI reports peaked between January and March 2016. All, except 4, cases were sampled, and a co‐dominance of influenza B (43%) and influenza A (H1N1) (41%) was evident. A high proportion of cases was reported in children <2 years 274 (37%). The proportional contribution of influenza‐associated SARI to all‐cause hospital admissions was high in children <2 years in the south (4.5% [95% CI: 3.1‐6.5]) and in children <5 years in Beirut (0.7% [95% CI: 0.6‐0.8]). Conclusion This is the first study to highlight the proportion of influenza‐associated SARI in 2 hospitals in Lebanon. The findings will be beneficial for supporting respiratory prevention and immunization program policies.


| INTRODUCTION
Following the 2009 H1N1 pandemic, the importance of having appropriate surveillance systems to monitor influenza trends became evident globally. 1 However, there remains scarce information about the burden of influenza in the Eastern Mediterranean Region (EMR). 2 Lebanon is divided into 6 governorates, constituting a wide range of private and public hospitals. 3 In the past, the Ministry of Public Health (MOPH) relied on several sources to identify acute respiratory infections such as intensive care unit-based surveillance, school absenteeism system, and admissions covered by the MOPH health insurance data. All these programs relied solely on passive surveillance, and most often had no laboratory confirmation. To fill this knowledge gap and understand better the circulation of influenza viruses in Lebanon, the World Health Organization (WHO) recommended the establishment of a severe acute respiratory infections (SARI) sentinel surveillance, under its "Pandemic Influenza Preparedness Framework" (PIP). 4 The Epidemiologic Surveillance Unit (ESU) at the MOPH initiated the SARI sentinel surveillance system in December 2014 in collaboration with the National Influenza Center (NIC) located at Rafik Hariri University Hospital in Lebanon.
In this study, we aim to highlight the results of estimating the proportion of influenza-associated SARI from September 1st, 2015 to August 31st, 2016 in 2 sentinel sites located in Beirut suburbs and southern governorate of Lebanon for evidence-based interventions and control programs such as immunization and awareness campaigns.

| SARI surveillance
Data on SARI cases were collected in 2 pilot sentinel sites that are public general hospitals covering Beirut suburbs and the southern province of Lebanon. At the start of this pilot, a total of 11 general hospitals were selected for the SARI surveillance. However, this study will present data for the 2 pilot sites as they had complete seasonal data.
A focal point appointed by the sentinel site was responsible for case finding and reporting since February 2015. The SARI case definition was as per WHO guidelines 1 : any patient having an acute respiratory infection with (i) fever or history of fever ≥38°C); (ii) cough; (iii) illness within 10 days; and (iv) requiring hospitalization.
The focal point collected data through passive and active surveillance: passively by receiving notification of cases from medical teams within the hospital and actively by visiting all hospital wards searching for cases. The same method of case finding was applied at both sites.
Specimens were collected using nasopharyngeal or oropharyngeal swabs and stored in viral transport media (VTM) at the sentinel site laboratory for 48-72 hours at 2-4°C. If the storage time exceeded 72 hours, the specimens were stored at −80°C. The specimens were transferred to the National Influenza Center (NIC) for testing in a cold box with frozen ice packs. At the NIC, specimens were tested by RT-PCR for influenza types A and B, and type A was further subtyped.

| Influenza-associated SARI study
To ensure that a complete influenza season is demonstrated, SARI data from September 1st, 2015 to August 31st, 2016 were used for the analysis.

| DISCUSSION
Two sentinel sites were piloted for influenza-associated SARI testing and reported 746 SARI cases during the studied time frame. The Severe acute respiratory infections surveillance systems have proved to be a good source for obtaining estimates of influenza trends as well as estimating the proportion of influenza-related SARI. 6 We defined the proportional contribution of influenza-associated SARI to allcause hospital admissions from all wards rather than using catchment population; hence, incidence rates were not obtained. Calculating the catchment population in Lebanon is challenging due to its decentralized system; however, this may be overcome by specific catchment estimates in the future.
While reviewing the data at the sentinel sites, the possibility of using ICD-10 codes in addition to reported SARI cases was explored; however, this was finally disregarded as codes used in WHO's manual did not correspond to what codes were used at the hospital admis-   . [1][2]). An explanation for these slight differences might be the hospital profile and differences in admission patterns in particular. The southern district had higher admissions in pediatric wards (80%) compared to the Beirut suburb site (30%). These proportions among children, however, might still be considered small given other pathogens might be causing increased admissions of children <5 years such as respiratory syncytial virus (RSV) and others [6][7][8] . Routinely testing for a range of non-influenza respiratory pathogens to find out what other pathogens might be the cause of SARI admissions is recommended.
Comparing our data with the literature, the influenza-associated SARI seemed to be low, yet the fact that children less than 5 years of age have the highest risk is commonly reported. Data from 2008 to 2014 in Jordan revealed that 57% of influenza positive SARI cases are children aged less than 5 years old. 9 In Oman, children less than 5 years had the highest influenza-associated SARI incidence (32-42 per 10 0000). 10 Another study conducted in 3 selected provinces in Iran also revealed the highest risk groups among SARI cases were children less than 5 years of age with an overall influenza-associated SARI of 29/10 0000 [95% CI: 16.8-43.8/10 0000]. 11 In Kenya, most SARI cases reported were children less than 5 years of age with influenzaassociated hospitalized SARI for the same age group of 2.7/1000 T A B L E 1 Proportion of influenza-associated severe acute respiratory infections (SARI) and influenza-associated SARI to all hospital admissions from September 1st, 2015 to August 31st, 2016 With the further development of sentinel SARI surveillance in Lebanon by including other sentinel sites and determination of the catchment population, we will have better estimates of influenza-associated SARI in the future.
The limitations of the study are many, and 1 major limitation is that only 2 sentinel sites were selected for estimating influenza-associated SARI. The sentinel sites were selected given their complete seasonal data, yet this can lead to difficulty generalizing the findings. In addition, a relatively low number of influenza positive cases was observed.
Therefore, the data should be interpreted carefully and may not be directly compared to data from the southern province. The other aspect is that SARI surveillance only covers severe cases admitted to the hospital, and hence, the estimates here include only severe influenza cases.
A general understanding of influenza burden in Lebanon can be developed once the results in SARI surveillance system are combined with an influenza-like illnesses (ILI) sentinel surveillance system once initiated.
This is the first study to highlight the proportion of influenzaassociated SARI cases using data from 2 governmental hospitals in Beirut suburbs and the southern governorate in Lebanon and using a national surveillance system. The study is a basis for replication in the other SARI sentinel sites to be able to have national findings on influenzaassociated SARI. This study is the first step in better understanding severe influenza in Lebanon, and this information can be used for better control programs such as immunization and awareness campaigns.