Estimation of seasonal influenza disease burden using sentinel site data in Pakistan 2017–2019: A cross‐sectional study

Abstract Background The influenza A(H1N1)pdm09 pandemic highlighted the need for reliable disease burden estimation from low‐ and middle‐income countries like Pakistan. We designed retrospective age‐stratified estimation of influenza‐related severe acute respiratory infections (SARIs) incidence in Islamabad Pakistan 2017–2019. Materials and Methods The catchment area was mapped on SARI data from one designated influenza sentinel site and other healthcare facilities in the Islamabad region. The incidence rate was calculated as per 100,000 for each age group with 95% confidence interval. Results The catchment population for the sentinel site was 0.7 million against the total denominator of 1.015 million, and incidence rates were adjusted. During January 2017 to December 2019, among 13,905 hospitalizations, 6715 (48%) patients were enrolled; 1208 of these (18%) were positive for influenza. During 2017, influenza A/H3 dominated with 52% detections followed by A(H1N1)pdm09 (35%) and influenza B (13%). Furthermore, elderly 65+ years age group had highest hospitalizations and influenza positive. The incidence rates of all cause respiratory and influenza‐related SARI were highest among children >5 years; highest incidence was found in 0 to 11 month/year group with 424/100,000 cases and lowest in 5–15 years 56/100,000. The estimated average annual influenza‐associated hospitalization percentage was 29.3% during the study period. Conclusion Influenza accounts for a significant proportion of respiratory morbidity and hospitalization. These estimates would enable governments for evidence‐based decisions and priority allocation of health resources. It is necessary to test for other respiratory pathogens for more clear disease burden estimation.


| INTRODUCTION
Influenza virus infections pose substantial risk of morbidity and mortality globally, particularly in young children and elderly people. 1 The World Health Organization (WHO) estimates that the influenza virus infection alone results in between 290,000 and 650,000 annual deaths, 2 with 36% of these deaths taking place in low-and middleincome countries (LMICs) like Pakistan. The epidemiology and disease burden of influenza is poorly understood in the region. In Pakistan, with a temperate climate and a population of about 22.5 billion people influenza virus peaks during winter months (October-February). 3 To assess the severity of the influenza disease, it is crucial to track influenza incidence and hospitalization.
According to WHO, there is need for influenza virus associated disease burden estimates, particularly from low-and middle-income nations. 4 With the aid of these projections, governments would be able to allocate limited resources and devise intervention efforts to lessen the impact and spread of the disease. National estimates would also help to clarify how widespread influenza-associated illness is and would help to guide global public health priorities.
Severe acute respiratory infections (SARIs), including influenza, constitute a major cause of morbidity and mortality globally. 5 Influenza A(H1N1)pdm09 pandemic emphasized reliable influenza disease burden estimates from LMICs like Pakistan to better recognize the impact of this vaccine preventable disease. Although global data on SARI is present, reliable information on disease burden estimates due to Influenza are still not available for Pakistan.
Seasonal fluctuations in influenza virus infection are known, and vaccination campaigns are conducted in many countries at the beginning of influenza season to reduce morbidity and mortality. 6 However, in Pakistan, immunization against influenza is not being given high attention. Currently, the Ministry of National Health Services, Regulations and Coordination (MoNH&RC) advises that high-risk populations, such as healthcare workers, pregnant women, and persons with chronic health conditions, receive an annual influenza vaccination.
Children below 5 years and adults 65 years and above, the seasonal influenza vaccine is recommended but MoNH&RC did not include in routine immunization program for these age groups. 7 Governments will be able to make evidence-based decisions for preventative and control measures, such the adoption of the seasonal influenza vaccine in public health programs, after having these age-specific burden estimates. We designed a retrospective study to generate a preliminary estimate for age-specific incidence of influenza in the Islamabad region during 2017-2019.

| Study population
The National Influenza Center (NIC) of the NIH in Islamabad serves as the national coordination hub for the laboratory-based influenza virus surveillance network that the National Institute of Health (NIH) developed. 8

| Laboratory testing
Respiratory specimens were tested for influenza viruses using realtime reverse-transcription polymerase chain reaction (rRT-PCR) (CDC protocol). From 2017 to 2019, we collected five to 10 nasopharyngeal swabs per week. 10 The data were gathered all year round to show influenza virus circulation pattern. By influenza virus type (A or B), influenza A subtype (H1N1pdm09 or H3N2), and overall number of specimens analyzed, we were able to determine the number of specimens that tested positive each week. Viral surveillance data were stratified by age group and utilized in models due to the lack of viral surveillance data for the age groups in this investigation.

| Disease burden estimation
Incidence rates were estimated according to WHO manual for disease burden estimation of influenza ( Figure 1). 4 The catchment area was mapped (GIS) based on SARI data collected from designated influenza F I G U R E 1 Methodology for disease burden estimation.
sentinel site and other healthcare facilities in the Islamabad region.
The catchment population was estimated based on data obtained from FGSH, all SARI cases reported from both urban and rural areas of Islamabad region were mapped, and all cases outside the catchment area were excluded ( Figure 2).
From the calculated number of influenza-associated SARI cases from the designated sentinel hospital, the proportion of lab confirmed influenza cases was inferred for the catchment population (denominator) after adjustment of sampling fraction. The lab confirmed influenza cases in the catchment population was estimated by adjusting the proportion for the catchment population. The proportion of SARI-associated hospitalizations reported outside the catchment area was determined for the total population under demographic surveillance and incidence rates were adjusted accordingly.
To estimate the proportion of influenza SARI cases, the numerator is total positive cases annually and denominator is total hospitalized cases annually. The incidence rate was calculated as per 100,000 population for each age group with 95% confidence interval. The cases were stratified into different age groups on the basis of age and compared through analysis of variance (ANOVA) under crosssectional study.

| Catchment population/demographics
The catchment population for the sentinel site is 7,000,000 against the total denominator of 1,200,000 and incidence rates were accord-

| Incidence rates by age groups
In comparison with 2017-2018, the proportion of influenza positive hospitalized patients who also had an acute medical condition increased significantly in 2019 (p = 0.0001). The age of the patients ranged from 0 to 90 years with a mean age of 32 (SD ± 22) years; most of those tested positive was from the age group 15-50 years (27%) ( Table 2).
Over a 3-year period, the rates of influenza-related hospitalization were highest among those aged above 65 years followed by children under 5 years of age and lowest among patients 5-15 years. The estimated average annual influenza-associated hospitalization percentage was 29.3% during the study period (  Note: Red colour shows that the highest percentage of influenza-associated SARI cases were in elderly group 50-64 years of age followed by >65 years of age during 2017. Green colour shows that the highest percentage of influenza-associated cases was in children <1 year of age group during 2018-2019.
period. Our findings are in contrast with those from Iran, where the overall incidence of SARI caused by influenza was estimated to be 29 per 100,000 people. 11 On the other hand, Khanh et al., 12   Incidence rate due to influenzaassociated SARI over estimated or proportion of population/100,000 population. The red colour text shows highest incidence rate/100,000 population of influenza-associated SARI cases were in <1 year of age followed by 1-2 years of age group.
the need for prevention and control measures, such as vaccination, in these at-risk populations.
Although different years had variation in influenza transmission patterns, influenza viruses circulate all year round, with September in every year seeing a decline in influenza activity. Influenza virus subtypes A (H3) was the predominant strain during 2017 which was similar to findings of Iran. 13 During 2018 and 2019 seasons, influenza H1N1pdm09 was the dominant strain followed by A (H3) and influenza B; this similar pattern of subtype circulation was observed in Egypt. 14 The incidence rate of influenza-associated hospitalizations was highest among those 65 years and above followed by children under 5 years. Our findings are similar with other countries, such as China, 15 India, 16 Oman, 17 and Tunisia. 18 Factors that may contribute to the hospitalization rate include access to medical care and antiviral drugs, influenza vaccination coverage, and cocirculating bacterial and viral pathogens. 19 Hirve et al. study showed higher rate of hospitalization in adults aged above 60 years, followed by above 5 years of age. 16

| DATA LIMITATIONS
The  writing-review and editing.