Viral etiologies of influenza‐like illness and severe acute respiratory infections in Thailand

Background Information on the burden, characteristics and seasonality of non‐influenza respiratory viruses is limited in tropical countries. Objectives Describe the epidemiology of selected non‐influenza respiratory viruses in Thailand between June 2010 and May 2014 using a sentinel surveillance platform established for influenza. Methods Patients with influenza‐like illness (ILI; history of fever or documented temperature ≥38°C, cough, not requiring hospitalization) or severe acute respiratory infection (SARI; history of fever or documented temperature ≥38°C, cough, onset <10 days, requiring hospitalization) were enrolled from 10 sites. Throat swabs were tested for influenza viruses, respiratory syncytial virus (RSV), metapneumovirus (MPV), parainfluenza viruses (PIV) 1‐3, and adenoviruses by polymerase chain reaction (PCR) or real‐time reverse transcriptase‐PCR. Results We screened 15 369 persons with acute respiratory infections and enrolled 8106 cases of ILI (5069 cases <15 years old) and 1754 cases of SARI (1404 cases <15 years old). Among ILI cases <15 years old, influenza viruses (1173, 23%), RSV (447, 9%), and adenoviruses (430, 8%) were the most frequently identified respiratory viruses tested, while for SARI cases <15 years old, RSV (196, 14%) influenza (157, 11%) and adenoviruses (90, 6%) were the most common. The RSV season significantly overlapped the larger influenza season from July to November in Thailand. Conclusions The global expansion of influenza sentinel surveillance provides an opportunity to gather information on the characteristics of cases positive for non‐influenza respiratory viruses, particularly seasonality, although adjustments to case definitions may be required.


| 483
CHITTAGANPITCH eT Al. been limited before avian influenza reemerged in the late 1990s. 3 However, generally only 6%-21% of respiratory infections are associated with influenza viruses in the tropics, 4 although the overall number of sites had decreased to 9 by 2012. 8 The surveillance system was established to monitor the frequency of influenza virus infection, identify new strains, and describe seasonality. In 2008, testing for 6 additional respiratory viruses was initiated. We present results of this expanded testing and compare characteristics of cases with different respiratory viral infections.

| Surveillance sites, case enrollment, and sample collection
Thailand is an upper middle-income country located in Southeast Asia north of the Equator and South of the Tropic of Cancer. Annual mean temperature is 27.5°C, and annual rainfall total is approximately 1700 mm, with August and September recording the most rainfall. 9 Sentinel surveillance for influenza viruses was established in Thailand in 2004 as routine public health surveillance and has been described in more detail elsewhere. 8  During this period, cases were enrolled from 13 hospitals, but 3 were excluded from analysis because their participation ended by 2012, and they each enrolled fewer than 100 cases between 2010 and 2012 ( Figure 1). ILI was modified from the WHO case definition as documented fever (axillary temperature ≥ 38°C) or history of fever and cough and illness not requiring hospitalization. 10 SARI was defined as documented or history of fever and cough, onset within 10 days of presentation, and severity of illness requiring hospitalization. Each site was instructed to enroll up to 10 case-patients per week (5 persons <15 years and 5 persons ≥15 years).

| Sample collection and laboratory testing
Throat swab specimens were collected from enrolled patients and stored in viral transport media. The vials were kept at 2-4°C for up to 4 hours, then moved to a liquid nitrogen tank at the hospital laboratory and transported weekly to the Thai NIH outside of Bangkok.
Specimens were tested for influenza A and B viruses using the standard World Health Organization (WHO) and CDC protocol for realtime reverse transcription-polymerase chain reaction (rRT-PCR). 11 Samples were tested using PCR for adenoviruses and rRT-PCR for RSV, MPV, and PIV 1-3 using methods described for individual rRT-PCR assays in Kodani et al. 12 Mixed infections were classified in order of the most frequently identified virus families first, that is, any virus found in a specimen with influenza was classified as an influenza mixed infection, followed by RSV, adenovirus, MPV, PIV-1, PIV-3, and PIV-2.

| Data analysis
We explored differences in characteristics of cases of ILI and SARI, including the prevalence of specific viral respiratory pathogens, by age (<15 years and ≥15 years) using the chi-square test. Differences in the median number of days cases were ill before presenting to the health facility were compared using the Mann-Whitney U test.
Percentages are proportions of non-missing data. The proportion of specimens found positive for respiratory viruses tested each month was plotted over time to describe annual and seasonal patterns. The mean monthly proportions positive for respiratory viruses were calculated for each calendar month to provide a summary of the seasonal patterns. All analyses were conducted using sas (Statistical Analysis Software 9.3, Cary, NC, USA).

| Human subjects
Data were collected as part of routine public health surveillance, and written consent was not required.
We identified 8106 (53%) cases of ILI and 1754 (11%) cases of SARI; 5509 (36%) patients did not meet either case definition. Among the 4715 outpatients who did not meet an ILI case definition, 2484 (53%) did not have a documented or history of fever, and 2541 (54%) did not present with a cough. Among the 754 inpatients who did not meet a SARI case definition, 394 (52%) did not have a documented or history of fever, 294 (39%) did not have a cough, and 145 (19%) had an illness duration that was unknown or >10 days.

| Identification of respiratory virus infections by case definition and age category
The proportion of cases with any respiratory virus among those tested identified ranged from 26% (among SARI patients ≥15 years old) to 49% (among ILI cases <15 years old) (   (Table S1).   April, while a seasonal pattern for PIV-2 was not apparent. There were no differences in seasonal patterns by age grouping for any of the viruses (data not shown).

| D ISCUSS I ON
Influenza viruses were the most common respiratory viruses diagnosed among ILI and SARI cases enrolled in a sentinel surveillance system designed to capture influenza-positive patients in Thailand.
RSV, adenoviruses, MPV, and PIV 1-3 were also identified in ILI and SARI cases. Although the surveillance system was designed to detect influenza viruses, important information on the burden and seasonality of cases of ILI and SARI associated with non-influenza respiratory viruses can be obtained. The generalizability of these results to all of Thailand is dependent on the representativeness of the sites chosen for sentinel surveillance and of the patients who contributed data and specimens. The sites were chosen from the 5 regions of Thailand but were located in large urban centers and may not be representative of more rural settings. Thailand has a universal healthcare coverage scheme that ensures good access to care for a large proportion of the country, which strengthens the generalizability of the results to the general population. However, the sentinel sites did not select a probability sample of patients, and it is possible that there was some unmeasured bias in the convenience sample of patients participating in the surveillance.
However, we have no reason to believe that any systematic bias prevents these data from being generalizable to the country as a whole.
As has been found in other tropical countries, ILI and SARI in Thailand can be caused by a variety of respiratory viruses in addition to influenza. Influenza virus surveillance systems can be used to detect and describe the characteristics of other respiratory viruses. However, some adjustments may be required to calculate an accurate burden of non-influenza respiratory viruses due to differences in their clinical presentation. Given the global expansion of influenza surveillance systems, adapting ILI and SARI case definitions to include non-influenza respiratory viruses will be helpful in the future to measure the impact of new control measures for non-influenza respiratory viruses.

D I SCL A I M ER
The findings and conclusions in this report are those of the authors and do not necessarily represent official position of the US Centers for Disease Control and Prevention.