The clinical and etiological characteristics of influenza-like illness (ILI) in outpatients in Shanghai, China, 2011 to 2013.

Introduction Clinical and etiological characteristics of influenza-like illness (ILI) in outpatients is poorly understood in the southern temperate region of China. We conducted laboratory-based surveillance of viral etiology for ILI outpatients in Shanghai from January 2011 to December 2013. Materials and Methods Clinical and epidemiological data from ILI outpatients, both children and adults, were collected. A total of 1970 nasopharyngeal swabs were collected and tested for 12 respiratory viruses using multiplex RT-PCR, and the data were analyzed anonymously. Results All 12 respiratory viruses were detected in the specimens. At least one virus was detected in 32.4% of 1970 specimens analyzed, with 1.1% showing co-infections. The most frequently detected agents were influenza A (11.7%), influenza B (9.6%), and rhinoviruses (3.1%).Other viruses were present at a frequency less than 3.0%. We observed a winter peak in the detection rate in ILI patients during 3 years of surveillance and a summer peak in 2012. HCoV, HADV, and HMPV were detected more frequently in children than in adults. Patients infected with influenza virus experienced higher temperatures, more coughs, running noses, headaches and fatigue than patients infected with other viruses and virus-free patients (p<0.001). Conclusions The spectrum, seasonality, age distribution and clinical associations of respiratory virus infections in children and adults with influenza-like illness were analyzed in this study for the first time. To a certain extent, the findings can provide baseline data for evaluating the burden of respiratory virus infection in children and adults in Shanghai. It will also provide clinicians with helpful information about the etiological patterns of outpatients presenting with complaints of acute respiratory syndrome, but further studies should be conducted, and longer-term laboratory-based surveillance would give a better picture of the etiology of ILI.


Introduction
Influenza-like illness (ILI), a subset of acute respiratory infections (ARI), represents approximately 62% of acute respiratory infections [1]. Viral pathogens are the most significant contributors [2].ARI is caused by many viruses, including the following: influenza viruses (IAV, IBV and ICV), respiratory syncytial virus(RSV),rhinovirus(HRV), human parainfluenza viruses (HPIV)1-4, human metapneumovirus(HMPV), human coronaviruses (HCoV) NL63,229E, OC43, and HKU1, human enterovirus (HEV) and adenovirus(HADV) [3].The similar clinical presentations of patients infected by various respiratory viruses make etiological diagnoses difficult when physicians simply make decisions based only on physical symptoms. The etiology of influenza-like illness (ILI) has been well characterized in some parts of the world, especially in temperate regions of the Northern Hemisphere [4][5][6][7].However, much less is known about the etiology of ILI in China, especially in Southeast China, which is located in a temperate region.
Pudong New Area is the largest and most developed district in Shanghai, with an extremely dense population of 5.5 million in 2013, representing one-fifth of the population of Shanghai. Shanghai Dongfang Hospital and Shanghai Zhoupu Hospital have been national influenza surveillance sentinel units since 2007. As in other regions of China, ILI data from hospitals can provide valuable information that could be used for monitoring the onset of an epidemic, especially the circulation of influenza virus. However, the epidemiology in Pudong of viral etiology for ILI is poorly understood. With support from the Chinese Key SciTech Program for Infection Control(under The Twelfth Five-year Framework), a laboratory-based influenza-like illness surveillance system was established in 2011 in Pudong New Area. We adapted multiplex RT-PCR assays for the detection of influenza virus and a number of other respiratory viruses that have recently been introduced, and these assays are also more sensitive than culture [8][9][10]. The objective of this study was to examine the clinical and etiological characteristics of influenza-like illness (ILI) in outpatients in this area from January 2011 through December 2013.

Ethics statement
This study did not involve any health-related patient interventions. The surveillance and sampling protocols were approved by Shanghai Pudong New Area Center for Disease Control and Prevention Review Board. A verbal informed consent was obtained from each subject; however, the patient identities have not been disclosed at any stage. The institutional board was informed of the specific needs with reference to the study setting and approved this mode of consent. A check box was included in the data form to document the consent-taking procedure.

Case definition and study population
An ILI case was defined as a sudden onset of fever(38°C), cough and/or sore throat, with no laboratory-confirmed evidence of etiology [11,12].A co-infection was diagnosed when a patient tested positive for two viruses on RT-PCR. The study population, including children (<15 years) and adults (15 years),consisted of the ILI cases who sought medical attention in the outpatient fever clinics of the above mentioned two hospitals from January 2011 to December 2013.

Specimens and data collection
Five to ten ILI cases per hospital who had not received antiviral drugs were invited to participate in this study each week. Nasopharyngeal swabs were collected by well-trained nurses. The swabs were sent in 3 ml of viral transport medium to Pudong New Area Center for Disease Control and Prevention within 48 hrs after collection and were stored at -70°C. Data were collected simultaneously using a standardized case report form, which included both demography and clinical information (such as symptoms, treatment, and temperature).

Nucleic acid extraction
Nucleic acid was extracted by the NucliSens easy MAG platform with the NucliSens magnetic extraction reagents (bio Me´rieux, The Netherlands) according to the manufacturer's instructions using a sample volume of 250μl. The elution volume of nucleic acid was 50μl.
Real-time PCR for human influenza C (IFVC) and human bocavirus(HBoV) was performed using an Influenza Virus C Real Time RT-PCR kit (Shanghai ZJ Bio-Tech Co., Ltd., People's Republic of China)and a Diagnostic Kit for Human Bocavirus DNA (GuangzhouDAAN Gene Co., Ltd., People's Republic of China). In accordance with the manufacturer's instructions, PCR for ICV was performed in a volume of 25μl using ABI 7500 fast (ABI, USA). Each reaction consisted of 19μl IFVC Super Mix, 1μl RT-PCR Enzyme Mix and 5μl Extraction RNA template for a final volume of 25μl. The PCR reaction was performed in the instrument at 45°C for 10 minutes (min), and 95°C for 15 min, followed by amplification consisting of 40 cycles of 15 seconds (s) at 95°C and 1 min at 60°C. Positive and no-template controls were included in each run.PCR for HBoV was performed in a volume of 20μl using ABI 7500 fast. Each reaction consisted of 14μl HBoV PCR Reaction Mix, 0.7μl Enzyme Mix, 3.3μl Primer and Probe Mix and2μl Extraction DNA template for a final volume of 20μl. The PCR reaction was performed in the instrument at50°C for 8 min, 93°C for 2 min, 10 cycles of 45 s at 93°C and 1 min at 55°C, followed by amplification consisting of 30 cycles of 30 s at 93°C and 45 se at 55°C. Positive and no-template controls were included in each run.

Electrophoresis
PCR products were sized using QIAxcel DNA Screening gel cartridge (QIAGEN) on the QIAxcel system (QIAGEN, Switzerland), which enabled high-resolution capillary electrophoresis. A 50-800 bp QX DNA Size Marker (Qiagen) was included on QIAxcel runs, and the size of the products was determined using the QIAxcel ScreenGel software (Qiagen). The QIAxcel system produces a digital gel image and an electropherogram for fragment analysis.

Statistical Analysis
Data were collected by the staff of the sentinel hospitals and laboratories on a standardized case report form and entered into an online data management system established by China CDC. The data were analyzed with SPSS software. The mean and SD or median and IQR were calculated for continuous variables, and percentages were calculated for categorical variables. A Pearson chi-square test and one-way ANOVA were used for categorical data or continuous data, respectively, wherever appropriate. A multivariate unconditional logistic regression analysis was used to determine factors associated with various clinical signs of ILI in outpatients. All statistical tests were 2-sided. P values<0.05 were considered statistically significant.

Results and Discussion
Clinical characteristics of the ILI case-outpatients A total of 1970 ILL outpatients who met the ILI-case definition criteria were enrolled in this study. Of these outpatients, 51.8% were male. The mean age was 19.7±16.9 years. A temperature reading of >39°C was documented in 262(13.3%) of the subjects. A total of 67.9% of the subjects reported a cough, 51% pharyngalgia, 40.5% running nose, 32.4% fatigue, 28.2% headache, and 7.6% myalgia. Patients infected with influenza virus experienced higher temperatures and more coughs, running noses, headaches and fatigue than patients infected with other viruses or patients who were virus-free(p <0.001) ( Table 1).

Viral etiology of ILI cases
All the subjects' samples were analyzed ( Table 2). All of the 12 respiratory viruses were detected. At least one agent was detected in 638 specimens (32.4%) by multiplex RT-PCR. The viruses most commonly detected were influenza A virus (IAV) and influenza B virus (IBV) at 230 (11.7%) and 162 (8.2%), respectively. All of the others were present at a frequency less than 4%: HRV in 61(3.1%) patients, HADV in 53(2.7%) patients, HCoV in 51(2.6%) patients, HPIV in 48(2.4%) patients, RSV in 47(2.3%) patients, HMPV in 5(0.3%) patients, and HBoV in 2 (0.1%) patients (Table 2)A single infection was identified in 616 (31.3%) patients, and co-infection was observed in 22 (1.1%). There were 14 cases with co-infection in children less than 15 years of age, not significantly different (p >0.05) from the corresponding result for adults (15) (Table 3).Of the 22 co-infection specimens, the most frequent combination was IAV and HRV (n = 6), followed by IAV and HPIV, IBV and HCoV, HPIV and HCoV, RSV and HPIV, RSV and HCoV. Each of these combinations was observed in two cases. The combinations IBV and HRV, HPIV and HRV, RSV and HRV, RSV and HADV, IBV and HPIV, and HRV and HCoV were each observed in one case.

Seasonal distribution of respiratory viruses
The characteristics of the seasonal patterns of respiratory viral pathogens were identified in our study. The monthly distribution of viruses is shown in Fig. 1A, 1B, 1C and 1D. We observed a winter peak in the detection rate in ILI patients during 3 years of surveillance and a summer peak in 2012. Influenza virus was the most commonly detected agent in the three winter seasons (i.     (Fig. 1C).HADV and HCoV were other major viral pathogens of ILI. HADV infection presented predominantly in 2011 and 2013, while HCoV presented predominantly in 2011 and 2012, and one peak was observed in September (Fig. 1C). HPIV, RSV, and HMPV appeared sporadically during the study period but presented an epidemic peak in September-October of 2011 (Fig. 1D). The results suggested that multiple respiratory viruses may circulate concurrently in the population and cause a large proportion (approximately 40%) of ILI.

Gender and age distribution of respiratory viral etiology
The distribution of respiratory viral pathogens from confirmed cases tested by RT-PCR by age and gender is shown in Table 3. Respiratory viruses were detected in 314(49.2%) of the male patients and 324(50.8%) of the female patients. There was no gender difference in susceptibility to ILI in Shanghai (Table 3).
IAV and IBV were most frequently identified in all age groups. The proportion of samples positive for IAV was higher in subjects 15-24 and 25-59 years of age (44.0%, and 41.8%, respectively) than in other age groups. This difference was statistically significant (p = 0.001). HCoV-229Eand HCoV-OC43 were detected frequently (13.3%, and 7.6%, respectively) in children less than 5 years of age. These values differed significantly from the corresponding values for other age groups (p = 0.002 and p = 0.01, respectively).HPMV and HBoV were only detected among subjects less than 15 years of age. HPIV was detected more frequently in sub-jects60 years of age compared with other age groups, but the difference between these frequencies was not statistically significant (Table 3).  Multiplex RT-PCR is considered to be one of the most sensitive and specific methods for the diagnosis of respiratory virus infections [13,14]. The Seeplex Respiratory Virus Detection assay (Seegene Inc., Seoul, Korea) was introduced to the respiratory research community in 2007 and can efficiently differentiate 12 respiratory viruses [10]. The application of multiplex design RT-PCR allows the simultaneous detection of multiple infections. In our study, multiplex viral agents, including IV (A and B), RSV (A and B), HPIV1-3, HRV, HCoVs (229E and OC43), HADV and HMPV were analyzed in a large population with ILI. To our knowledge, this is the first study of respiratory virus etiology with ILI in both children and adults in Shanghai. The overall detection rate of respiratory virus was 32.4%in the 3 years of laboratory-based  surveillance. All of the 12 respiratory viruses were detected. Influenza virus was the most commonly detected agent in both children and adults; other viruses were present at a frequency less than 4.0%.

Factors Associated with Clinical Signs of ILI
We observed a winter peak in the detection rate in ILI patients during 3 years of surveillance and a summer peak in 2012.Similar results have been reported in other studies in South China [15]. However, the ILI incidence only peaked during winter in Beijing and North China [6]. HCoV, HADV, and HRV were more frequently detected in children than in adults; HMPV and HBoV were only detected in children less than 15 years of age. Coughs, sore throats, running noses, fatigue, headache and temperatures >39°C were higher in influenza virus-infected outpatients than other respiratory virus-infected patients.
In our study, the investigation of the clinical characteristics of ILI showed that patients with IV experienced more coughs and temperatures >39°C than patients with other respiratory virus or those who were virus free. This result is consistent with the findings of previous research [6].
Children less than 5 years of age experienced more coughs, running noses and temperatures >39°C than adults. In the study by Friedman and Attia [16], the symptom triad of cough, headache, and pharyngitis was found to be a predictor of isolation of influenza virus in febrile children who were selected from emergency department patients at a tertiary pediatric center according to predetermined criteria indicating influenza infection. Respiratory viruses can cause influenza symptoms [17]. There is inconsistency about whether certain symptoms can be used to distinguish specific infections, and it is generally accepted that there are no symptoms specific to any viral infection. Some other studies have attempted to identify signs or symptoms specifically associated with influenza virus or other viruses [18,19], but no definitive conclusions have been drawn.
During the study period, 32.4% of the specimens were positive for at least one virus. This result is consistent with the results of several previous studies that found percentages between 27-34.6% [6,[20][21][22], but it is lower than those reported by several investigators who found percentages between 48.7-59.5% [23,24]. There are multiple explanations for these differences. The differences could be due to true differences in the overall burden, to differences among study populations, or to detection methods that differed from among studies. It is difficult to compare the results, as the reported data were obtained with different detection methods or PCR primers [21,25]. Future comparative studies to evaluate the sensitivity and specificity of these detection methods should clarify this issue. Second, the infection rates may vary with geographical location and with the particular period chosen for testing [25,26]. Third, the patient population and its environment may influence the results. In our study, up to 1332 specimens were negative in RT-PCR although all of them matched well with the inclusion criteria for ILI. Negative findings could have resulted from the low load of viral material in samples, or to infection with bacteria or other viruses such as enterovirus.
The seasonal pattern of ILI cases in Shanghai was quite unique according to our study.
Results differ markedly among studies in terms of the respiratory viruses most commonly found among studies. As in many previous studies [4,6,7,20,[27][28][29][30], influenza A and B virus were the most commonly detected viral pathogens in outpatients with ILI in our region. An epidemic of influenza virus usually occurs in the winter. However, an epidemic peak of IAV infection was observed in the summer of 2012. More patients with influenza-like illness were observed during this period. This result was consistent with the Chinese National Influenza Center (CNIC) (2012 2013) China flu Weekly Report [31]. In addition, HRV, RSV, HADV, HCoV, HPIV and HMPV were identified from outpatient specimens and contributed collectively to 38.4% of all ILI laboratory-confirmed cases in our study. The detection rate of influenza virus was significantly higher among adult patients, whereas other viruses, such as HADV, HCoV, and HMPV, were prevalent among children less than 15 years of age. These findings are consistent with reports from other studies [5,[32][33][34].
HRV was not present in 2011, but it was the second most frequently detected viral agent in both children and adults in 2012 and 2013, representing approximately 10% of all positive cases. The positive rate (3.1%) was lower than that found by previous studies, which detected HRV in 10% to >40% of respiratory infections [29,35]. HRV may circulate year-round in our region, a result found by previous studies [20,25,28]. RSV-positive samples occurred at a low level in our study. This finding contrasts with other studies [27,36], where the predominant virus detected in children less than 5 years of age was RSV. The reason for this difference is that we did not include hospitalized children, as RSV has been shown to be a common cause of lower respiratory tract infection in children admitted to the hospital [5,37,38]. Previous studies on RSV infections have primarily been conducted in hospitalized elderly adults with medical conditions such as cardiopulmonary diseases [39] that may predispose them to viral infections that are relatively uncommon in the general population.
In our study, a total of 22 samples (3.5% of the total number of RT-PCR positive cases) revealed the presence of co-infections. In co-infection cases, influenza A virus, PIV, and RSV were found together most frequently, at rates of 36.4% (8/22). Previous studies have reported that co-infections were associated with more severe signs than mono-infections [40,41]. Children were found to be more likely to be co-infected than adults in the present study. However, the subsequent clinical conditions of ILI illness patients were not obtained, and thus the association between co-infections and severe signs cannot be analyzed. We will address such research themes in the future.
This study has limitations. First, no testing for other etiologies (such as enterovirus or bacteria) of acute respiratory illness was performed. As is generally known, respiratory viruses, bacteria and other microorganisms can cause respiratory illness with influenza-like symptoms. Without doubt, other microorganisms could have been additional pathogens in the negative specimens, and our results may underestimate the role of virus infection. Second, the histories of influenza vaccine in ILI outpatients were not obtained. Thus, the analysis of clinical characteristics in ILI patients may be biased. Third, we did not collect all ILI cases presenting at the above two sentinel sites from Monday through Sunday. Facility staffs were involved in the project on a voluntary basis, with frequent shifts of personnel to other facilities.
In conclusion, the spectrum, seasonality, age distribution and clinical associations of respiratory virus infections in children and adults with influenza-like illness were analyzed in Shanghai for the first time in this study. To a certain extent, the findings can provide baseline data for evaluating the burden of respiratory virus infection in children and adults in Shanghai. It also can provide clinicians with helpful information about the etiological patterns of outpatients presenting with complaints of acute respiratory symptoms, but further studies should be conducted, and longer-term laboratory-based surveillance would give a better picture of the etiological characteristics of ILI.
Supporting Information S1 Dataset. Detailed information for all the case involved in our study. (XLS) S1 File. Editorial certificate of this manuscript. (PDF)