Burden of influenza among hospitalized febrile children in Ghana

Background Influenza surveillance data from Africa indicate a substantial disease burden with high mortality. However, local influenza data from district hospitals with limited laboratory facilities are still scarce. Objectives To identify the frequency and seasonal distribution of influenza among hospitalized febrile children in a rural hospital in Ghana and to describe differential diagnoses to other severe febrile infections. Methods Between January 2014 and April 2015, all children with a temperature of ≥38°C admitted to a district hospital in Ghana were screened for influenza A and B by RT‐PCR and differentiated to subtypes A(H1N1)pdm09 and A(H3N2). Malaria microscopy and blood cultures were performed for each patient. Results A total of 1063 children with a median age of 2 years (IQR: 1‐4 years) were recruited. Of those, 271 (21%) were classified as severe acute respiratory infection (SARI) and 47 (4%) were positive for influenza, namely 26 (55%) influenza B, 15 (32%) A(H1N1)pdm09, and 6 (13%) A(H3N2) cases. Influenza predominantly occurred in children aged 3‐5 years and was more frequently detected in the major rainy season (OR = 2.9; 95% CI: 1.47‐6.19) during the first half of the year. Two (4%) and seven (15%) influenza‐positive children were co‐diagnosed with an invasive bloodstream infection or malaria, respectively. Conclusion Influenza contributes substantially to the burden of hospitalized febrile children in Ghana being strongly dependent on age and corresponds with the major rainy season during the first half‐year.


| INTRODUCTION
Globally influenza has long been regarded as a major public health concern. Not only the elderly but also very young children have been identified as a vulnerable group for influenza infections. A recent meta-analysis estimated 90 million cases of influenza and 20 million episodes of influenza-associated acute lower respiratory infections (ALRI) annually worldwide in children below 5 years. 1 As a result, influenza is the second most common pathogen identified in children with ALRI after respiratory syncytial virus (RSV) and followed by parainfluenza virus. 2 About 870 000 annual hospitalizations in children below 5 years of age have been attributed to influenza with influenza-associated hospitalization rates being three times higher in developing than in industrialized countries. 3 It has been assumed that 99% of inpatient deaths from severe acute respiratory infections (SARI) 4 and 70% of all deaths attributable to influenzalike illnesses (ILI) among children below 5 years of age occurred in developing countries. 5 The extent of influenza infections in Africa is now slowly being recognized due to strengthened national influenza surveillance systems across the continent. 6 Most countries have had inadequate data on the influenza disease burden until the influenza A(H1N1) pandemic in 2009. Since then, many countries implemented hospital-based influenza surveillance among SARI patients. 3,7 In Ghana, ILI sentinel surveillance within outpatient facilities across all regions was established in 2007, which screened 2357 samples in 2014 (https://www.ghanahealthservice.org). The burden of influenza disease among hospitalized children in Ghana has so far only been studied in large tertiary hospitals. 8,9 However, data on the burden of influenza are particularly important for healthcare workers in small rural or district hospitals with no or limited laboratory facilities, considering that influenza infections are not easily distinguishable from other febrile infections such as malaria and may therefore lead to false treatment decisions. 10 This study aims to identify the proportion and seasonal distribution of influenza infections among hospitalized febrile children in a rural district hospital in Ghana to inform healthcare workers on the contribution of influenza as a differential diagnosis to other severe febrile infections.

| Study site and sample collection
Study participants were recruited at the pediatric ward of the Agogo Presbyterian Hospital (APH), a district hospital with 250 beds, situated in the Asante Akim North municipality of the Ashanti Region in Ghana.
The climate is tropical with two rainy seasons from March to June and from September to October. 11 The study area is located in a holoendemic malaria region with perennial malaria transmission.
Oropharyngeal swabs (Copan, Italy) were taken from all children aged between 1 month and 15 years with a tympanic temperature of ≥38°C between January 2014 and April 2015. Swabs were taken at admission and immediately transferred in viral transport medium and stored at −20° until RNA extraction. For each patient, two independent slide readers conducted malaria microscopy on Giemsa-stained thick and thin smears and a blood culture was performed on standard media (Oxoid, Basingstoke, UK). The following respiratory signs and symptoms were assessed by the study physician: abnormal lung auscultation, breathing difficulties, chest indrawing, chest pain, coryza, cough, intercostal retractions, nasal flaring, sore throat, and stridor.
Repeated visits of study children were considered as new visits if they were at least 30 days apart. SARI was diagnosed according to the World Health Organization (WHO) case definition in a hospitalized patient with fever of ≥38°C or history of fever, with cough and an onset of illness within the last 10 days. 12

| Sample processing and virus detection
RNA was extracted from oropharyngeal swabs with the RTP Pathogen Kit (Stratec biomedical, Birkenfeld, Germany) and eluted in 120 μL.

| Epidemiological analysis
Categorical variables were described as frequencies and percentages and continuous variables as medians and their corresponding interquartile ranges (IQRs). All data analyses were performed with Stata 14 (StataCorp LP, College Station, TX, USA).
Children were grouped by age into the categories <3, 3-5, and >5 years. To quantify the association between a given exposure and outcome, odds ratios (OR) with their respective 95% confidence intervals (CI) were calculated.

| Ethical considerations
The Committee on Human Research, Publications and Ethics, School of Medical Sciences, Kwame Nkrumah University of Science and Technology, Kumasi, Ghana and the "Ethikkommission der Ärztekammer Hamburg," Germany provided ethical approvals for this study. All participants were informed about the study's purpose and procedures. Written informed consent was obtained from the parents or the guardian on behalf of the study children prior to study enrollment.

| RESULTS
A total of 1063 hospital visits were made from 991 study children. Sex distribution was slightly unbalanced with 586 male cases (55%) aged In total, 274 (26%) children were diagnosed with SARI, which was predominantly found in the age group <3 years (194/618; 31%), with a median of 2 years (IQR: 1-3) ( Table 1)

| DISCUSSION
Within our study group, SARI cases made up more than a quarter of all febrile pediatric hospital admissions. Of the SARI cases, 8% are diagnosed with influenza A/B, which is in line with a large African multicountry surveillance project, which identified influenza in 5%-26% of SARI cases. 6 From Ghana, similar SARI rates (8%-9%) have been reported. 8,14 Although the SARI case definition showed a good association with influenza in this study, it is noteworthy that influenza was still identified in 3% of non-SARI cases. It has been shown before that influenza case definitions are highly age-dependant and performances vary due to the unspecific clinical picture of influenza. 15 Hence, laboratory-based surveillance systems using a SARI definition will probably underestimate the true burden of influenza infections.
The median age of children with influenza was higher than the median age of SARI cases. This age distribution is probably due to the T A B L E 1 Influenza and SARI cases stratified by sex and age high frequency of other serious respiratory pathogens, notably respiratory syncytial virus (RSV), among very young children. 16 The outpatient department of the same study hospital reports human parainfluenza viruses 1-4, enteroviruses and adenoviruses as the predominant respiratory pathogens in children below 5 years of age, while influenza A/B was the most frequently detected respiratory virus in older children (5-12 years). 11 However, in the present study, samples were not tested for other respiratory viruses; therefore, this assumption could not be further investigated. The higher age of children with influenza at the outpatient department compared to hospitalized children in the present study can be explained by the more severe disease presentation among young children, which is illustrated by high death rates in this age group. 3 Influenza-related deaths were not reported during the present study, although high case fatality ratios, more than seventeen times as high as in industrialized countries, were estimated for developing countries. 1 This study's data are in line with findings from three Ghanaian tertiary hospitals, which also did not report any influenzaassociated deaths during a 30 months surveillance period. 8  This study has a few limitations. Oropharyngeal swabs were used for all patients, although the combination of nasopharyngeal and oropharyngeal swabs is reported to be more sensitive for virus detection. 22 The study was conducted throughout a 16-month period. Therefore, temporal patterns from this study can give a rough estimate, however, to time public health interventions, such as vaccinations, longer studies over a period of at least 2 years are required. Finally, this study aimed to assess the burden of influenza infections among hospitalized children.
As hospitalization rates surely depend on the proximity to a healthcare facility and health-seeking patterns, these findings are not suitable to estimate influenza incidences in the general population. Another limiting factor of hospital-based surveillance is the time delay from symptom onset to specimen collection, which might underestimate influenza prevalence in patients living in very distant communities.
F I G U R E 1 Influenza and Severe acute respiratory infection (SARI) cases per study month. Proportions of both influenza and SARI cases were calculated using the total number of recruited patients per study month. Influenza was further defined on subtype level: namely influenza B, influenza A(H1N1)pdm09, and influenza A(H3N2)

| CONCLUSION
SARI and influenza contribute substantially to the burden of hospitalized febrile children in Ghana. In comparison with SARI, which is most frequently found in children aged <3 years, influenza predominantly occurs in children aged 3-5 years and is associated with the major annual rainy season during the first half-year. During this time period, the use of rapid diagnostic tests may be considered on the pediatric ward, taking into account the test's low sensitivity. Distinguishing influenza from other non-specific febrile diseases, such as malaria and invasive bloodstream infections, could help to reduce the unnecessary application of antimicrobial and antimalarial drugs.