Incidence and etiology of hospitalized acute respiratory infections in the Egyptian Delta

Introduction Acute Respiratory Infections (ARI) are responsible for nearly two million childhood deaths worldwide. A limited number of studies have been published on the epidemiology of viral respiratory pathogens in Egypt. Methods A total of 6113 hospitalized patients >1 month of age with suspected ARI were enrolled between June 23, 2009 and December 31, 2013. Naso‐ and oropharyngeal specimens were collected and tested for influenza A and B, respiratory syncytial virus, human metapneumovirus, adenovirus, and parainfluenza viruses 1–3. Blood specimens from children 1–11 months were cultured and bacterial growth was identified by polymerase chain reaction. Results from a healthcare utilization survey on the proportion of persons seeking care for ARI was used to calculate adjusted ARI incidence rates in the surveillance population. Results The proportion of patients with a viral pathogen detected decreased with age from 67% in patients age 1–11 months to 19% in patients ≥65 years of age. Influenza was the dominant viral pathogen detected in patients ≥1 year of age (13.9%). The highest incidence rates for hospitalized ARI were observed in children 1–11 months (1757.9–5537.5/100 000 population) and RSV was the most commonly detected pathogen in this age group. Conclusion In this study population, influenza is the largest viral contributor to hospitalized ARIs and children 1–11 months of age experience a high rate of ARI hospitalizations. This study highlights a need for surveillance of additional viral pathogens and alternative detection methods for bacterial pathogens, which may reveal a substantial proportion of as yet unidentified etiologies in adults.


Incidence and etiology of hospitalized acute respiratory infections in the Egyptian
The World Health Organization estimates that in 2013, over 8% of all deaths in the Eastern Mediterranean Region were attributable to ARIs. 1 However, regional data on morbidity and etiology of these infections are lacking. Certain risk factors for severe respiratory disease are widely prevalent in Egypt, including diabetes (9.3%) 3 and liver disease (specifically hepatitis C, 14.7%). 4,5 Additionally, nearly 20% of adult Egyptians and 38% of adult males regularly use tobacco products. 6 Data on the burden and etiologies of ARI are especially pertinent for Egypt, a nation with a high birth rate and 11% of its population consisting of children <5 years of age 7  This manuscript presents the findings of the first three and a half years of population-based ARI surveillance from IEIP Egypt's site in Damanhour. because physicians were not comfortable performing nasopharyngeal swabs on ill patients in this age group. After consent was obtained from the patient or their legal guardian, study personnel conducted patient interviews and chart reviews to complete standardized data collection forms detailing demographic variables, relevant medical history, and symptomology of the patients' current illness. Oxygen saturation was measured at enrollment using a finger pulse oximeter. Chest X-rays were performed when clinically indicated by the attending physician, Specimens were tested by real-time reverse-transcription PCR (rRT-PCR) for a panel of viral respiratory pathogens, including influenza A and B, respiratory syncytial virus (RSV), human metapneumovirus (hMPV), adenovirus, and parainfluenza viruses (PIV) 1-3, using CDCdeveloped assays and testing protocol. [14][15][16][17] Blood specimens were collected from enrolled children ≥31 days to <5 years of age by drawing 3-5 mL of blood directly into BACTEC™ blood culture bottles. Blood culture bottles were stored at room temperature in hospital laboratories, transferred to the IEIP Damanhour laboratory daily for incubation in the BACTEC™ blood culture system at 35-36°C, and observed daily for signs of microbial growth.

| Study population
Specimens with growth were examined by gram stain, and cultured onto chocolate, blood, and MacConkey agars for further identification.
After five days of incubation in the IEIP Egypt Damanhour laboratory, 1 mL aliquots from the BACTEC™ blood culture bottles were transferred into a cryovial and sent in a liquid nitrogen tank to NAMRU-3 for PCR analysis for Streptococcus pneumonia. were included for analysis. Incidence rates for ARI were calculated using methods described previously. 9 Numerators were the total number of patients eligible for ARI enrollment, stratified by sex, residence (rural vs urban), and age group (1-11 months, 1-4, 5-19, 20-49, 50-64, and ≥65 years). Age groups were based on recommendations from the WHO 19 and modified due to availability of population data age groupings from CAPMAS. Gender data were not collected from eligible patients who did not enroll; these patients were assumed to be proportionally equal in gender to enrolled patients stratified by age group and residence (rural vs urban). Unadjusted age-specific hospitalized ARI rates were calculated by dividing the number of enrolled ARI patients in each age category by the corresponding age-specific population denominator. Rates were adjusted for patients eligible but not enrolled, stratified by sex, age, and residence (rural vs urban). No adjustment for healthcare utilization was necessary due to the near universal use of surveillance facilities for hospitalization reported by healthcare utilization survey respondents. ARI hospitalization rates were calculated for all years separately and combined, and for rural and urban locations separately. Urban and rural rates of ARI were ageadjusted to the total Damanhour population demographics.

| Enrollment and demographics
Between June 23, 2009, and December 31, 2013, a total of 11 929 patients with suspected moderate or severe ARI were screened for enrollment. Of these patients, 6684 (56%) were determined to be eligible by symptomology, residency, and age requirements and 6113 (91.5%) were enrolled. Over one-third of all enrolled ARI patients were 1 month to <5 years of age, and nearly three quarters were rural area residents ( Table 2)

| Signs, symptoms, and clinical characteristics of enrolled patients
Nearly all enrolled ARI patients presented with cough (97%) and history of fever with the current illness (98%) ( Table 2). Less than half of these patients had evidence of current fever ≥38°C, tachypnea, dyspnea, or sputum production. Nearly three quarters of patients had abnormal breathing sounds and approximately one-quarter reported symptom onset ≥7 days prior to enrollment.
Almost 20% of patients reported a comorbid chronic condition. The proportion of patients reporting a chronic condition increased with age, with 29% of patients age 20-49, 41% of patients age 50-64, and 43% of patients over age 65 reporting ≥1 chronic condition.
Among these patients with a chronic condition, hypertension, diabetes (each 31%), liver disease (including hepatitis C) (16%), and asthma (12%) were most commonly reported. These chronic conditions were reported exclusively in patients >20 years of age. In children aged 1 month-11 years reporting a chronic condition, the most commonly reported chronic condition was congenital heart defects (24 patients, 53%), followed by asthma (5, 11%). Among older children and adolescents reporting a chronic condition, 33% of patients aged 1-4 years and 32% of patients aged 5-19 years reported asthma.
A small percentage (2.4%) of patients were admitted to the ICU during hospitalization (Table 2), and even fewer were placed on mechanical ventilation. Nearly 10% of the patients with an oxygen saturation reading had levels <90%. About half of enrolled patients had a chest X-ray performed. A higher percentage of adults 50-64 (64%) and ≥65 (63%) years of age had a chest X-ray performed. Age groups with the highest percentage of abnormal chest X-rays were adults aged ≥65 years (52%), adults aged 50-64 years (47%), and infants aged 1-11 months (41%). Of those with abnormal chest X-ray results, the most common finding was pulmonary consolidation, followed by infiltrates ( Table 2). Nearly half of the 50 patients who died in hospital were ≥65 years of age. Patients who died in hospital were significantly more likely to report a chronic condition one or more of the following comorbidities: liver disease, hypertension, asthma, and diabetes (60% vs 19%, P<.01). The case fatality rates for age groups were as follows: There were some differences between rural and urban patients with respect to symptomology. A lower proportion of rural patients had dyspnea, tachypnea, abnormal breathing sounds, and sputum production at admission (Table 2). Additionally, a lower proportion of rural patients reported a chronic condition. The mean length of hospitalization was shorter for rural patients compared with urban patients (5.2 vs 6.7 days).

| Diagnostic results
Results of the rRT-PCR viral pathogen panel on NP/OP specimens are presented in Table 3. At least one viral pathogen was detected in 36% of specimens collected. Pathogen detection was highest in patients whose samples were taken 3-5 days after onset of symptoms (42%).
Pathogen detection was 32% in patients presenting 11-14 days after symptom onset and decreased to 24% in patients presenting >2 weeks after symptom onset. Overall, the most commonly detected pathogen was influenza, followed by RSV, adenovirus, and hMPV. The profile of pathogens detected in different age groups is presented in Fig. 1.
The largest proportion of RSV-positive patients occurred among

| Seasonality of pathogens
The seasonality of viral pathogens is shown in Fig. 2  The incidence of hospitalized ARI varied markedly between years of data collection, age groups, and geographic location. The overall incidence ranged from a high of 315.9 cases per 100 000 population in 2010 to a low of 219.6 per 100 000 population in 2011. The highest incidence rates for hospitalized ARI were consistently observed in children aged 1-11 months (Table 4), which ranged from 1758 to 5538/100 000 population during 2009-2012. Incidence rates for hospitalized ARI were comparable between patients aged 1-4 years and those aged ≥65 years. The age-adjusted rate of hospitalized ARI in the rural population was consistently higher than that of the urban population.

| DISCUSSION
Hospitalization of patients for moderate or severe acute respiratory infection remains an important and complex public health concern in the Egyptian Delta. The incidence of patients hospitalized for moderate or severe ARI during our study period was 290/100 000 population, with higher rates in very young children aged 1-11 months (5135/100 000 population overall for the period 2009-2012). Our incidence estimates of hospitalized ARI were comparable to previously published data from several similar studies in Asia and Central America. The very high estimated incidence rate for children aged 1-11 months in our study population (5135/100 000) is similar to estimates from Thailand (5772/100 000). 20 The incidence for all age groups combined was similar to published estimates from Guatemala (128/100 000), 21 although our age-specific estimates were much higher. This may be explained by differences in healthcare utilization and hospitalization rates between the two populations. Published incidence rates from a population-based surveillance site in Kenya were much higher than our estimates (699.8/100 000), 22 but did not include fever as enrollment criteria, likely increasing their overall capture of respiratory disease due to non-infectious etiologies. Compared with the above studies, very few ARI deaths occurred during hospitalization in our study population, and an even smaller proportion of these deaths were associated with an identified viral pathogen. Seven percent of enrolled ARI patients either left against medical advice or were transferred to another hospital; therefore, we have no information on their outcome after they left the study hospitals. Transferred cases may have been more severe and in need of specialized care, and it is possible that rates of death may have been higher in these patients than those who remained hospitalized in the study hospitals. in children aged 1-4 years; however, they were seen relatively infrequently in older age groups. This is consistent with a wide body of published data on viral etiologies in young children. 10,[23][24][25] High incidence of RSV infections has been reported previously in this population, 9 and this study confirms RSV as the predominant cause of viral ARI, and likely all ARI, in children aged 1-11 months.
RSV was identified in nearly 40% of all children aged 1-11 months hospitalized with ARI. Our data demonstrate that RSV causes a much higher burden of hospitalized ARI in children compared with influenza, similar to studies from other countries. 9,14,20,26 This study does not capture the high burden of RSV in children who may not require hospitalization, but experience substantial morbidity. Additionally, children <31 days were not included in this study, but have been shown to have high rates of RSV infection. 27,28 With the majority of ARI hospitalizations occurring in this age group, and the disproportionate number of these hospitalizations due to RSV, the burden of RSV in this population cannot be understated.
In this study population, influenza is the largest viral contributor to hospitalized ARIs. This is inconsistent with similar surveillance data from Guatemala in the same time period, during which influenza was identified in a much smaller proportion of patients compared with RSV and adenovirus. 21 In that study, however, over 40% of enrolled patients were under 5 years of age, and consistent with our data, RSV was the most common etiology in that age group. Relatively few children under Our data demonstrated a markedly higher incidence of hospitalized ARI in rural population. Some locations have demonstrated lower incidence rates of respiratory infections in rural populations due to increased distance from hospital facilities. 35,36 Distance and associated travel time is not a barrier to hospital access in this population, as 96% of respondents in the Damanhour Healthcare Utilization Survey reported travel time to one of the surveillance facilities as less than one hour. 9 Higher incidence rates in the rural locality were also demonstrated in Guatemala, 21 possibly indicating that another aspect of rural residence is associated with higher rates of disease and/or hospitalization.
One potential explanation for the higher rate of hospitalization in the rural population is that admitting physicians are more likely to hospitalize rural patients presenting with even moderate or possibly progressive respiratory symptoms. Our data showed that a lower  therefore, future studies on bacterial pathogens may require additional serology and PCR detection.
Previous data have demonstrated that adult ARI and pneumonia may be less commonly due to viral etiologies 50 and our data seem to support this. However, the diagnostic panels used for this study did not include all common and possible causes of severe acute respiratory infection. Additional diagnostic methodologies on older children and adults, including blood cultures and sputum testing, may have yielded additional pathogens. The population-based surveillance site in Damanhour is currently participating in a multisite study to determine the etiology of community-acquired pneumonia. The study will test NP/OP swabs for an additional 19 bacterial and viral pathogens and provide a more complete picture of the epidemiology of ARI in the study population. Healthy controls will also be tested to determine proportional etiology for pathogens which may have some degree of baseline carriage or previous exposure in the study population.

| CONCLUSION
This paper presents the first population-based incidence rates of hos- Government as part of that person's official duties.