Etiology and Clinical Characteristics of Influenza-Like Illness (ILI) in Outpatients in Beijing, June 2010 to May 2011

Background Since May 2009, exposure of the population of Beijing, China to pH1N1 has resulted in an increase in respiratory illnesses. Limited information is available on the etiology and clinical characteristics of the influenza-like illness (ILI) that ensued in adults following the pH1N1 pandemic. Methods Clinical and epidemiological data of ILI in adults was collected. A total of 279 throat swabs were tested for twelve respiratory viruses using multiplex RT-PCR. Clinical characteristics of influenza A in outpatients versus test-negative patients were compared using Pearson's χ2 and the Mann-Whitney U test. 190 swabs were tested for pH1N1 by virus isolation. Consultation rates for ILI were compared between 2009 and 2010. Results One or two virus were detected in 29% of the samples. Influenza A virus (FLU-A) accounted for 22.9% (64/279). Other viruses were present at a frequency less than 3.0%. Cough was significantly associated with Influenza A virus infection (χ2, p<0.001). The positive rate of FLU-A was consistent with changes in the ILI rate during the same period and there was a significant reduction in the incidence of ILI in 2010 when compared to 2009. During the 2010–2011 influenza season, the incidence peaked in January 2011 in Beijing and north China. Conclusions Exposure to pH1N1 had no impact on typical influenza seasonal peaks, although FLU-A remained the predominant virus for 2010 in Beijing. Symptomatically, cough was associated with FLU-A infection. The positive rate of influenza virus was consistent with changes in the ILI rate during the same period and there was a significant reduction in the incidence of ILI in 2010 when compared to that of 2009.


Introduction
Influenza virus is a seasonal acute respiratory infection and is associated with significant morbidity and mortality in adults every winter [1]. Similar to most other countries, the peak influenza season in China is associated with higher health care utilization [2]. In China, weekly cases of influenza-like illness (ILI) in acutecare settings throughout the country are reported to the Chinese National Influenza Center(CNIC) and are mainly attributed to influenza virus, predominantly influenza A virus (FLU-A) and influenza B virus (FLU-B) [3,4]. In addition to the influenza viruses, other respiratory viruses can also cause ILI symptoms in adults, such as human respiratory syncytial virus (HRSV), human parainfluenza viruses (HPIV), including parainfluenza virus type 4 [5], rhinovirus, adenovirus, human coronaviruses and human metapneumovirus [6]. Thus, multiple respiratory viruses may circulate among adults and cause influenza symptoms [5]. 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 studies have attempted to identify signs or symptoms specifically associated with influenza virus or other viruses [7,8], but no definitive conclusions have been drawn.
ILI data such as ILI rate and ILI count can to some extent reflect influenza activity, and ILI attack rates may be higher among adults during pandemics [9]. In China, ILI data from larger hospitals can deliver valuable information that could be used for monitoring the onset of an epidemic [3]. Peking University People's Hospital (PKUPH), an affiliated and teaching hospital of Peking University, is a nonprofit health care institution that provides an exceptionally caring environment for patients and families, and extends care to patients of any nationality. PKUPH receives approximately 2,000,000 outpatient cases annually and patients from areas beyond Beijing (mainly northern China) account for forty percent of this figure. The Infectious Diseases Department of PKUPH is a national influenza surveillance sentinel unit.
Some studies have focused on changes in ILI when pH1N1 is present. However, few studies have reported on whether there have been changes in viral etiology and distribution, as well as clinical and epidemiological characteristics of ILI, following a pH1N1 pandemic. In this study, we focused on the etiology and epidemiological characteristics of adults with ILI symptoms after the 2009 influenza (H1N1) pandemic in Beijing.

Ethics statement
This research study did not involve any health-related patient interventions. The study was conducted in accordance with the principles and guidelines expressed in the Declaration of Helsinki and was approved as less than minimal risk research by the Research Ethics Committee at PKUPH. Written consent forms were not required and approved information sheets were used instead of consent forms. Detailed information was given to all patients, and once informed consent had been received, swab samples were collected and the data analyzed anonymously.

Case definition and study population
A human influenza-like illness (ILI) case was clinically defined as a sudden onset of fever ($38uC), cough and/or sore throat of less than three days duration, and no laboratory-confirmed evidence of another etiology [3,10]. ILI cases were defined as FLU-A-positive ILI-cases when the RT-PCR results from throat swab samples were positive for influenza A virus. A co-infection was diagnosed when more than one virus tested positive on RT-PCR. The study population included adults ($16 years) who sought medical attention in the fever outpatient service of PKUPH, a national influenza surveillance sentinel unit, between June 2010 and May 2011.

Clinical specimens and data collection
Data was collected at the time of medical consultation using a case report form for each adult patient who met the ILI-case definition criteria. Data collected included demography, clinical observations (such as symptoms, chronic medical conditions, treatment, temperature) and laboratory test results. For RT-PCRpositive patients, we additionally used telephone interviews to determine treatment, co-morbidity, vaccination status and duration of fever. All variables and analyzed data are shown in Table 1. A single throat swab, using Sterile Rayon Swabs( 167KS01), was collected from each ILI-patient who had not received antiviral drugs. The swab was immediately placed in virus transport media tubes. Each sample was frozen within 24 hours at 270uC until analyzed. In total, 279 clinical specimens were analyzed.

Laboratory investigations
Viral RNA extraction and PCR amplification. Viral RNA was extracted from 140 ml of viral transport medium containing the swabs, using the QIAamp Viral RNA Mini Kit (Qiagen Viral RNA Mini Extraction Kit, Germany), following the manufacturer's instructions. Extracted RNA was then used as the template to perform the RT-PCR reaction, using a reversetranscriptase polymerase chain reaction (RT-PCR) kit (Fermentas, Shenzhen). cDNA was compounded under the action of reverse transcriptase and random primers from RNA. The RT-PCR amplification system was as follows: total RNA 8 ml, random primers 1 ml, 56RT buffer 4 ml, 10 mM dNTP 2 ml, RNA inhibitor 1 ml, and 1 ml reverse transcriptase added to DEPC treated water to 20 ml. Conditions used for the two-step RT-PCR were as follows: the primary RT-PCR was incubated at 80uC for 3 minutes, cooled down on ice for 2 minutes; followed by 37uC for 90 minutes and 94uC for 2 minutes, then cooled down on ice for a further 2 minutes, and stored at 220uC until collected.
Testing the virus using multiplex PCR. RT-PCR products were tested as per manufacturer's instructions for the following viruses: influenza A and B virus; respiratory syncytial virus (RSV); human coronaviruses 229E, OC43, NL63, HKU1; parainfluenza virus; enterovirus or rhinoviruses; adenoviruses; and finally metapneumovirus, using the 12 league respiratory virus multiple PCR detection kit (Neuro-Hemin Hangzhou, RV1211). The PCR amplification system was as follows: cDNA templates 3 ml, 56RV Primer 4 ml, 8-Mop Solution 3 ml, and 26Multiplex Master Mix 10 ml. PCR was carried out in a PCR Amplifier and the conditions were as follows: 94uC for 15 minutes, followed by 40 cycles of 94uC for 0.5 minutes, 60uC for 1.5 minutes and 72uC for 1.5 minutes, with extension at 72uC for 10 minutes, followed by storage at 4uC until collected. The samples were divided into two groups for testing, Group A samples were tested for the presence of adenovirus (Adv), metapneumovirus (MPV), human coronavirus 229E/NL63 and parainfluenza virus (PIV) types 1, 2, and 3. Group B samples were tested for the presence of influenza B virus, human coronavirus OC43/HKU1, rhinovirus (HRV A/B), RSV B/A, and FLU A.
All PCR products were analyzed in 2% agarose gels by electrophoresis, stained with ethidium bromide and visualized on a UV light box. PCR products were compared with markers A and B bands in the gel to determine the size of the specimens tested, and to decide whether a swab was infected with a respiratory virus.
Viral culture and further typing. Sections of throat swab specimens were randomly selected and periodically shipped in an icebox to the Beijing Center for Disease Control and Prevention (CDC) for virus isolation and confirmation of pH1N1. Further isolation and sub-typing of FLU-A was performed in 190 samples. Virus isolation was processed using MDCK cell lines, and positive results were followed by the haemagglutination inhibition test using reference antisera (as recommended by the WHO manual: http://www.who. int/csr/resources/publications/influenza/whocdscsrncs20025rev.pdf). Once a negative or decreased reaction to the reference antisera occurred with the isolated virus, RT-PCR and sequencing was used to clarify the infection.

Statistical Analysis
Data was entered into a database and analyzed using SPSS software version 16.0 (SPSS, Chicago, IL, USA). Continuously distributed variables were compared between FLU-A-positive ILIcases and FLU-A-negative ILI-cases using the Student's t-test (t) if data was normally distributed and the Mann-Whitney U test if nonnormally distributed. Categorical data or proportions were compared using Pearson's x 2 or Chi-square test. In order to evaluate the association between each variable and the FLU-A-positive cases at a univariated level, continuously distributed variables were recategorised into binary variables, and univariate analysis using x 2 allowed calculation of odds ratios with 95% confidence intervals. P values,0.05 were considered statistically significant.

Clinical features of the ILI case-patients
The clinical characteristics are summarized in Table 1

Clinical characteristics of ILI patients who tested positive for influenza A virus
Among the 279 patients tested by RT-PCR, 69 (24.7%) had influenza A virus and 13 (4.7%) had other respiratory viruses detected. Complete information concerning clinical presentation and routine examination was available for 190 patients (68.1% of the total), 61 of the FLU-A-positive cases and 129 of the FLU-Anegative cases. Data on the number of samples, clinical and univariate characteristics, P value and odds ratio (OR) (as well as the 95% confidence intervals (CIs)) in FLU-A-positive and in FLU-A-negative cases can be seen in Table 2. In comparison to FLU-A-negative ILI-case patients, cough was more likely to be reported in patients whose RT-PCR tested positive for influenza A virus ( Table 2). Patients infected with FLU-A experienced more cough than patients infected with other viruses or those viruses free (x2, p,0.001). Other clinical characteristics (such as sore throat, headache) appeared to be more frequent in those infected with the FLU-A virus than those non-infected, but the difference was not statistically significant. FLU-A-infected patients presented with a mean temperature of 38.6uC [95% CI (38.4, 38.7)] which on average, commenced 1.3 days before admission [IQR [1][2]. Approximately 30% of FLU-A-infected patients reported hyperpyrexia (temperature .39uC). The median age of patients who tested positive for influenza A was 32 years [IQR , and 44.3% were male. There was no significant difference in male and female rates. There was no difference in the neutrophil percentage between those whose RT-PCR tested positive for influenza A and those who tested negative. The clinical and univariate characteristics of influenza-A virus ILI-case patients are showed in Table 2.

Temporal distribution of viral etiology
Among the 279 swabs collected, 82(29.4%) samples tested positive for at least one respiratory virus by multiplex RT-PCT. The temporal distribution of viral agents from confirmed cases tested by RT-PCR is shown in Figure 1. Influenza viruses (FLU) were the predominant viral etiology comprising 80.5% (23.7% of the total specimens) of the confirmed cases between June 2010 and May 2011 in Beijing. Almost all of these FLU-positive specimens were FLU-A cases, which accounted for 97.0%. Conversely, two influenza B virus-positive samples were identified among the samples. All other viruses were present at a frequency less than 3%: HRV and ADV in three cases individual (1.1% of 279); HRCV in two cases (0.7%); and HRSV in one case (0.4%). HMPV, FLU-C and HPIV were not detected, even thought the primers were designed to detect these strains. In addition, seven samples (2.5% of the total number of specimens) tested positive for more than one virus or were co-infections: HRSV-A and HRV in two cases; OC43 and FLU-A in one case; HRV and FLU-A in one case; HRSV-A and FLU-A in one case; and a combination of HRSV-A, HRV and FLU-A in two cases. The identity of each virus in these co-infections was confirmed by amplification with primers specific for each virus.

A reduction in the incidence of ILI
A decrease in the number of patients seeking medical attention was observed in 2010 ( Figure 2). In 2010, the overall level of outpatient consultations was less than that of 2009, particularly between May and December. From May 2009, when pH1N1 emerged in Beijing [11], the capacity of the outpatient service of the Infectious Diseases Department of PKUPH increased rapidly. The peak in numbers attending the outpatient department occurred in November 2009. From then, the outpatient volume declined, reaching normal levels in the third month of 2010. The numbers of outpatients remained relatively stable in 2010. The volume again increased during November and December 2010, and peaked in January 2011.
In addition, a significant reduction in the percentage of outpatient consultations for ILI was detected during 2010-2011 in North China (Figure 3). During the 2009 pandemic, the highest weekly ILI rate was 12.1 cases per 100 consultations in week 44 of 2009 ( Figure 3). The ILI rate was relatively stable, based on data available from the weekly ILI surveillance system of North China in 2010. The percentage of patient visits for ILI peaked at 5.0% in late January 2011.

Discussion
RT-PCR is considered to be one of the most sensitive and specific tests for the diagnosis of influenza [12]. In addition, multiplex PCR methods can detect a panel of respiratory viruses or co-infections simultaneously [13]. In this study, pharyngeal swab specimens were collected from ILI case-patients, who strictly complied with Chinese CDC standards [3]. The general population in Beijing, China has been exposed to the novel pandemic H1N1 influenza virus since mid May 2009 [11]. According to the Chinese CDC, as of May 2011, 136869 confirmed cases and 875 deaths from pandemic influenza H1N1 2009, have been reported nationwide in mainland China [4,14].
In this study, approximately 29.4% of the samples were positive for at least one virus, which is consistent with the results of other studies, in which between 0.9-27% [4,15] and 44% [16]   A alone. The incidence of FLU-positive specimens was high. This may reflect the possibility that our sample collection was biased towards patients exhibiting ILI, which is a clinical or symptomatic definition of influenza to identify potential influenza cases, in other word, influenza-like illness case-definition make influenza viruses as the virus most commonly detected [5,18,20,21]. A study of human-to-swine transmission of pandemic influenza A virus concluded that the human ILI case definition has a high specificity and a low sensitivity for FLU-A [22]. Influenza viruses usually account for a much greater proportion of positive specimens of  influenza-like illness in adults than other respiratory viruses during the peak seasons [23].
A total of 7 samples (8.5% of the total number of RT-PCR positive cases) revealed the presence of co-infections. In five FLU-A-positive samples, viral co-infections were observed, including one co-infection with HRCV in July, one with HRV in November, one with HRSV-A, and one with HRSV-A and HRV in January, and one with HRSV-A and HRV in February. A population challenged by multiple infectious agents may result in an epidemic and restructure of various viruses [24].
Rates of ILI are an indicator of trends for influenza pandemics [25]. Beijing is located in the temperate zone of the Northern Hemisphere, where influenza typically peaks seasonally once a year, and Beijing experience one peak of influenza activity and the peak occurred during December-January next year before 2009 [3,26,27], but ahead to November in 2009 [17,26]. That the peaks in ILI and the increase in acute respiratory infections (ARI) are due to influenza is supported by the seasonal pattern of highprobability ILI, the low level of respiratory syncytial virus infections, and laboratory results in the influenza season [23,28,29]. During the 2010-11 influenza season, a seasonal pattern in ILI activity was observed and influenza activity peaked in late January 2011 in Beijing. Compared with the previous pandemic year (2009), lower outpatient numbers were observed during 2010-11( Figure 2). Overall, the rates of influenza-like illnesses in outpatients were lower during the 2010-2011 season, than during the 2009-10 pandemic influenza season ( Figure 3).
As a result of the requirement for fever in our definition of ILI, our calculated incidence may underestimate the true incidence of ILI in the cohort. Our study shows that the positive rate of influenza virus was consistent with changes in the ILI rate during the same period.
More than 70% of ILI-case patients were infected with p-H1N1 between June 2009 and January 2010 [4]. Compared to the previous year, the age of FLU-A patients ranged from 18 to 61 years with a mean age of 36.7 years, which is older than the age of 2009 ILI-case patients (mean age was 23.4 years) [11].
In 197 (70.6%) specimens, no viral etiology was identified or the virus was not detected. This may reflect the fact that only viruses known to cause respiratory symptoms were tested, and therefore the remainder may have been caused by other respiratory viruses or by other micro-organisms, which could also have been additional pathogens in the positive specimens [5]. It is also possible that some viruses could not be detected due to low levels of shedding.
In this study, we attempted to identify symptoms associated with influenza A infections. However, it is widely known that a clinical diagnosis of influenza is not straightforward, and it is difficult to find symptoms or combination of symptoms specific for influenza [20,30]. We found that cough was significantly associated with influenza A (x2, p,0.001). This is consistent with the results of Boivin, of Ohmit and of Monto, who reported that cough and a fever.38uC were associated with a positive PCR test in the influenza population, when influenza was prevalent within the community [7,31,32]. Our specimens were collected during the influenza season in Beijing and the cough appears to be specific for influenza in this cohort. This is also consistent with the view of Call et al, who believe that both fever and cough are more specific for influenza among elderly individuals when influenza virus is circulating in an area [30]. We did not find any statistically significant difference in the occurrence of fever (temperature .39uC) between the FLU-A-positive group and the FLU-Anegative group (x 2 , p.0.05). A study that involved all age groups demonstrated that muscle and joint pain and headache were associated with influenza [20]. However, we did not find any statistically significant differences in the occurrence of muscle and joint pain, and headache between the FLU-A-positive group and the FLU-A-negative group (x 2 , p.0.05).
This study has limitations. No testing for other etiologies of acute respiratory illness was performed. As is generally known, respiratory viruses, bacteria and other micro-organisms can cause respiratory illness with influenza-like symptoms [5]. Without doubt, other micro-organisms could have been additional pathogens in the positive specimens.
In conclusion, pH1N1 did not affect typical influenza seasonal peaks, although FLU-A remained the predominant virus in Beijing in 2010. Symptomatically, cough was associated with FLU-A infection. The positive rate of influenza virus was consistent with changes in the ILI rate during the same period and there was a significant reduction in the incidence of ILI in 2010 compared to 2009. The findings of this study may facilitate the clinical discrimination of influenza A virus infection, as well as providing data and distribution information for virologic surveillance of influenza.