1.      Introduction

In Central- and North-America, six months after its emergence, the novel pandemic swine-origin influenza virus A(H1N1)2009 had spread worldwide, requiring implementation of national pandemic control plans in many countries [1,2].

In Europe, a combination of containment procedures and atmospheric conditions, such as dry and warm weather might have been responsible for the delayed and sporadic circulation and transmission of the virus. In the Southern hemisphere the pandemic occurred mainly in regions with a more temperate climate [3,4]. The proportion of hospitalized A(H1N1)2009 cases observed in EU countries started to increase since mid-June.

One month later, there were already 8,936 confirmed cases of A(H1N1)2009 reported by 28 EU countries, 67% of which by the United Kingdom [5].

In Belgium, the first case was identified on the 12th of May, followed by sporadic cases clearly linked to international traveling.FrommidJuly 2009 on,an increasing number of indigenous cases were detected all over the country after several big international music festivals. Between week 40 and 49the number of flu cases surpassed the epidemic threshold, with a peak in week 44. An adjuvant pandemic vaccine (Pandemrix^®) was available after the peak incidence in October and was administered in priority to risk groups(defined as health care workers, pregnant women, obese patients, children below 6 months of age and chronically ill patients)[6]. The Belgian surveillance system estimated that 214,531 people were infected, 733,000 could have benefited from a vaccine and 19 patients died in conditions attributable to A(H1N1)2009 infection [7]. They clearly saw a higher number of ILI consultations in Brussels compared to Flanders and Wallonia, without obvious reason.

Although A(H1N1)2009 infection had several unique features which included rapid transmissibility, fast growth and high morbidity in patients of risk groups, the clinical characteristics did not significantly differ from previous influenza seasons and mainly consisted of influenza like illness (ILI). In addition to influenza viruses A/B (IA & IB) responsible for 5 to 15% of the total upper respiratory tract infections worldwide during annual epidemics [8], ILI can be attributed to a broad range of other respiratory viruses, such as rhinoviruses (hRV), Parainfluenza viruses (PIV1,PIV2,PIV3,PIV4), human metapneumoviruses (hMPV) and respiratory syncytial viruses (RSVA, RSVB). Even though the exact viral etiology of ILI has been extensively investigated, limited and conflicting information is available on the possible role of viral co-infections on clinical severity, especially during the 2009 pandemic [9-13]. However, some recent data suggest that some viruses alone or in co-infection (e.g. A(H1N1)2009 and RSV), could be independently correlated with severity of infection (i.e. length of hospitalization, ICU admission, respiratory failure requiring mechanical ventilation etc...) [12,14]. This highlights the importance of diagnosingallcausal pathogens during an ILI episode, in addition to therapeutic purposes when antiviral drugs are available.

Facing the lack of clinical presentation specificity, determination of the viral agent(s) involved requires an extensive work-up.Over the past decade, several nucleic acid amplification tests, including multiplex Real-Time Polymerase Chain Reaction (RT-PCR) and micro-array assays have shown high reliability in detecting the presence of one or more viruses in respiratory tract samples [10,15]. These assays have demonstrated superior sensitivities and specificities compared to traditional methods such as direct fluorescent-antibody assays, shell-vial culture, and Rapid Antigen Tests (RAT) [16].

In this study, we describe viral pathogens yielded from respiratory samples of patients suffering from ILI and attending the emergency room, outpatient clinics and all types of hospitalization wards (including ICU) in seven hospitals during 2009 pandemics in Brussels. Furthermore, we were able to evaluate the frequency of viral co-infection over 5 influenza seasons (between October 1^st and March 31^st) with observation of a substantial reduction of viral co-infections in general and IA-associated specifically in the winter 2009-2010, as compared with other influenza seasons.

2. Methods

2.1.  Study Design

Between 15 July and 31 December 2009, respiratory specimens from 4,101 patients with ILI, attending seven public hospitals -emergency rooms or policlinics (4 university and 3 general hospitals)-were prospectively examined by RAT, viral culture and/or PCR (see Table 1 for stratification) in a licensed laboratory for medical microbiology. For each successive ILI episode (separated by at least 8 weeks) one sample was included, so patients could be included more than once. In total, 4,895 samples were included in the study period. Detection of a virus was recorded only once for each patient. ILI was defined as a patient suffering from at least one general symptom (fever >38.0°C, asthenia, myalgia, chills or headache) and one respiratory symptom (cough, dyspnoea/tachypnea, wheezing, rhinitis or pharyngitis). A pre-existing comorbidity was defined as a chronic condition requiring long term medication or medical follow-up. Comorbidities were listed based upon CDCA(H1N1)2009 guidelines [17]. The type of viral diagnostics was defined in advance according to patients’ clinical condition (Table 1).

For patients at risk of complications [1] or presenting with a severe illness who required >48h hospitalization, a

Specific Real-Time RT-PCR assay for the detection of IA, and, if positive, additional sub typing for A(H1N1)2009 was performed on a daily basis (except Sunday). In case the patient was admitted to ICU (n=620), additional multiparameter PCRs for the detection of 16 respiratory viruses were applied.

Criteria for severe clinical presentation were a temperature <35°C or >39°C, a heart rate ≥120/min, a respiratory rate ≥30/min, respiratory distress with oxygen need, a systolic arterial pressure <90 mmHg or an altered consciousness.

Beside age and sex, predisposing factors for complicated diseasesuch asobesity, pregnancy, chronic respiratory disease, diabetes mellitus, immunosuppression and chronic cardiac disease were registered [1]. To comparethe circulation of the pandemic strain in 2009 with the circulation of influenza strains in other seasons, we considered 2 preceding (2007-08, 2008-09) and 2 following influenza seasons (2012-13, 2013-14), with the season 2009-10, each time from October 1^st till March 31^st to objectively compare the same seasonal time periods. Critically-ill patients were examined using multiparameter molecular tests.

2.2.  Respiratory Specimen Acquisition and Handling.

Nasopharyngeal aspirates (NPA) or flocked swabs (Micro Reologics, Copan, Brescia, Italy) from throat or nasopharynx were used, according to the instructions of the manufacturer. The swabs were placed in 1 mL of homemade viral transport medium (VTM, for details see AnnexI) and samples were transported at room temperature, submitted at the Virology Laboratory of Saint-Peter University Hospital and held at 4°C prior to processing.

2.3.  Viral Culture 

Within 24h after sampling, three classical cell lines (LLC-MK2, Vero, MRC5) were inoculated (Annex I). Viral culture was performed on 4,275 samples. 

3. Molecular Assays

3.1. Target Sequence Selection and Primer and Probe Design 

Specific primers and probes were selected and designed by using the Primer Express software, version 2.0 (Applied Biosystems (AB), Foster City, CA) according to sequences available from public databases (National Centre for Biotechnology Information; NCBI). The oligonucleotide concentration which gives lowest threshold cycle (C_t) value and maximum amplification efficiency was selected (Supplement 1). 

3.2.  DNA Extraction 

Genomic DNA and RNA were isolated using the MagNA Pure LC (Roche Diagnostics) instrument by using the Total Nucleic Acid Isolation Kit - Large Volume.25µL of Phocid Distemper virus was spiked into each sample prior to extraction as an inhibition and amplification control. 

3.3.  Real-Time PCRs 

An in house Real-Time (rt) Reverse Transcriptase (RT) PCR (rtRT-PCR InflA) targeting the matrixprotein-coding gene, was used for the detection of IA, able to detect all subtypes including A(H1N1)2009. For specific detection of the circulating variant, two monoplexrtRT-PCR assays according to the Centres of Disease Control (CDC) protocol were initially used (from 15/7/2009-31/8/2009): SW InfA PCR (swInfA) and SW H1 PCR (swH1), further called as RT-PCR A/H1N1 [18]. From September on, A(H1N1)2009 subtyping was based only on swInfA detection, due to numerous observed discrepancies in laboratories worldwide and the suspicion of mismatches in the swH1 primer, probe binding regions and of existing variants of the pandemic strain [19]. 

Five reactions were set up as duplex rtPCRs: human bocavirus (hBoV)-human adenovirus (hAdV), IA-RSVA, IB-PIV2, RSVB-hMPV (hMPV-1A,1B,2A,2B) and hRV-PIV3 duplexes.The sixremaining rtRT-PCR reactions (human corona viruses: hCoV-229E, -NL63, -OC43, PIV1 and PIV4, hEV[19]. were set up as singleplex PCRs. For concentrations of primers and probes see Supplement 1. 

In total, on 2,367 samples a separate IA rtRT-PCR with subtyping of A(H1N1)2009 was performed. On 158 IA-positive ICU samples (in multiparameter rtPCR) a subtyping PCR was directly performed supplementary. And also, on IA-positive cell cultures (without parallel rtRT-PCR-result for A(H1N1)2009 or with negative IA RT-PCR result) this typing step was performed in batch retrospectively, once weekly, to stratify the samples.

During seasons 2012-13 and 2013-14, all samples of critical-ill patients were prospectively tested once for respiratory pathogens with an in house customized TAC (Taqman™ Array micro-array Card) respiratory panel which includes rtPCR testing for the same viral pathogensby rtPCR(Annex I).

4.  Data Analysis

Statistical analyses were performed using Graph Pad Prism Software (Inc, 2003, San Diego, USA). Chi square or Fischer’s exact test were used to compare non-continuous variables and Mann Whitney test to compare continuous variables. A two-tailed p-value less than 0.05 were considered as statistically significant. 

5. Results

5.1. Sample Distribution 

We received 4,895 unique respiratory samples (after exclusion of 645 replicated samples) between July and December 2009, corresponding to 4,101 patients with medically-attended ILI episodes, including 954 (19.5%) throat flocked swabs, 2,176 nasopharyngeal aspirates (44.5%), and 1765 nasopharyngeal flocked swabs (36.1%). Distribution of the samples over the study period is shown in (Annex II and Figure 1). Delivery time before processing of specimens ranged from 45 minutes to 36h (median time 8h45). Overall, 2,371 samples yielded a respiratory virus (viral prevalence of 48.4% in our cohort).

For the comparison of viral co-infection between different seasons all unique samples from medically-attended ILI-patients were included from the 1^st of October until the 31^st of March to cover the classic influenza season. 

5.2. Description of the Study Population 

The samples came from 2,102 males and 1,999 females. Out of the whole cohort, 22.1% of patients were infected with A(H1N1)2009 equally distributed according to gender (p-value =0.73). 

As presented in (Figure 2), a age distribution of patients harbouring respiratory viruses ranged from 1 day to 94.9 years old with a median age of 1.4 years [interquartile range (IQR) of 0-8.25 years] in the global cohort and 0.8 years [IQR 0-43 years; range 0-86 years] in the ICU group. Overall, 57%, 12%, 24% and 7% of all included respiratory samples were taken from children <5y of age, children between 5-10y, patients from 11-60 years of age and patients above 60 years, respectively.

Children <5 years accounted for the major part (72.4%) of the 2,371 virus-positive samples; whereas only 2.7% of positive samples came from patients older than 60 years (11.0% and 13.9% of positives was taken from people between 5-10y and 11-60y respectively). 

Overall, pandemic H1N1 strain explained 21.9% (361/1,645), 77.2% (193/250), 80.1% (331/413) and 31.7% (20/63) of the total proven viral episodes in the <5y, 5-10y, 11-60y and >60 y age categories, respectively.As illustrated in Figure 2, the prevalence of other respiratory viruses also significantly varied according to age groups.

In children <4 years, the main isolated viruses were RSV and PIV; together with IA they were responsible for 50.4% of infections. While pandemic H1N1 infections predominated in age groups from 2-3y to 71-80y, RSV exceeded the prevalence of IA in age group < 1 y (32% of children RSV+) and 1-2 y, and of the patients >80y none was infected with A(H1N1)2009. Between 61-94y, 21% (63/299) was virally infected, 2/3 of them due to non-IA viruses (mainly hRV, HSV, PIV, hEV& RSV). 

Overall, age-specific notification rate was highest in the age subcategory “11-20y” where 47.4% of patients was infected with A(H1N1)2009, followed by age subcategory “6-7y” where 45.7% of patients showed infection with the pandemic strain. 

Out of our global cohort, 3,065 (74.7%) patients belonged to an officially declared WHO risk group for pandemic influenza. A total of 2,605 patients (63.5%; corresponding to 2987/4895 infectious episodes) required hospitalization out of which 2,474 (94.97%) harbored 1 or more comorbidities/risk factors listed above. This proportion even increased when focusing on ICU admitted cases, where 608/620 (98.1%) patients belonged to one or more not mutually exclusive risk groups. Pregnant women in particular represented 27% and 13% of 16-42 years old female patients admitted to medical wards or ICU, respectively. 

When analysing patients with a confirmed A(H1N1)2009 infection, 53.7% belonged to an officially declared WHO risk group. 

There was no clinically relevant difference between the median ages of the described tested population (0.8y in ICU & 1.4y in whole study population) and the median ages of examined patient groups in the 4 other influenza seasons (Table 2), namely 0.6y; 0.7y; 0.5y and 1.7y for 2007-2008; 2008-2009; 2012-2013 and 2013-2014 respectively [IQR were 0-1.5y; 0-1.8y; 0-2.7y and 0-11.5y respectively]. 

5.3. Severe Cases and Outcome 

In 337/620 (54.4%) ICU patients, a respiratory viral pathogen could be identified, which was A(H1N1)2009 in 47% of them. Overall 25.5% of all ICU patients were infected with the pandemic strain (Figure 3). 

During the study period, 51 A(H1N1)2009 infected patients died leading to an in-hospital fatality rate possibly related to influenza A of 5.6%. One or more underlying co morbidities were present in most of these patients (49/51), including five children (4 with a chronic neuromuscular disorder, 1 with severe immunosuppression). The main comorbidities in adults were immunosuppressant (66%), chronic cardiac disease (49%), chronic respiratory disease (44%), and morbid obesity (32%). Two previously healthy patients died: a 43 years old man and a 28 years old pregnant woman in whom no other attributable cause of death could be found apart from A(H1N1)2009 infection. Only 4 co-infections were identified among the IA-related fatal cases, involving hAdV twice, hRV once and PIV3+hRV once. These patientshad no statistical different risk on fatality compared with patients suffering from IA-monoinfection (P=0.7475; OR: 1,216; 95% CI: 0.3391 - 4.359), but the numbers are too low for conclusions. 

5.4. Respiratory Virus Circulation Over the Pandemic Period 

Figure 1 shows the temporal evolution of isolation of IA and other viruses. From the 2,371 positive samples, the four most detected viral species were A(H1N1)2009 (40.1%) with a peak in October, RSV (24.3%) with a peak in December, PIV1-4 (13.8%) with a peak in September and hRV (13.5%) with a peak in September-October. 

5.5. Viral Co Infection Rate                                                            

5.5.1. Viral culture

Overall, 85 of the 1988 positive samples (15/7-31/12/2009) from non-critical ill patients (4.30%) were infected by multiple viruses, as assessed by viral culture. But for a fair comparison with other seasons we looked specifically at the viral winter season (1/10/2009-31/3/2010), where 108 of 2487 positive samples (4.34%) were virally co-infected.This number of viral co-infections is significantly fewer compared to the 2007-08 and 2008-09 seasons harbouring 6.95% and 6.70% of co-infection rate in this population, respectively (P=0.0004, X²=15.48, 2). Children <5 years were affected by the vast majority of co-infections (82.5 %), especially among those aged <3 years (76.5% of all co-infections in culture). As for co-infections involving specifically A(H1N1)2009, they were only detected in 2.9% of all described ILI patients, and in 3.99% in the whole winter season 2009-10, which was significantly lower than observed with IA strains from previous seasons 2007-08 and 2008-09, 12.7% and 6.4%, respectively (P<0.0001, X²=37.24, 2).

5.5.2. Molecular assays

A similar trend was observed when analysing samples from 620 ICU patients, tested by multiparameter PCRs. Out of the 337 virus containing samples, 44 (13.1%) showed mixed viral infection compared to global viral co-infection prevalence’s of 27.23%, 21.83%, 29.1% and 26.4% in 2007-08, 2008-09, 2012-13 and 2013-14 seasons, respectively (P<0.0001, X²=33.84, 4). During the whole 2009-10 winter period, there were 43 co-infected samples out of the 322 positives (13.35%) within 550 ICU patients (Table 2). The majority of them suffering from co-infection were again <3 years old (88.2%).Within the ICU population, only 7% of A(H1N1)2009-positive samples contained another virus, whereas in the A(H1N1)2009-negative group, the co-infection rate was significantly higher and reached 18.4% (P=0.0018, X²=9.733,1; OR 0.3311; 95%CI: 0.1612 - 0.6801) (Figure 3).

Looking at the seasons 2007-08, 2012-13 and 2013-14, no statistical difference was seen between viral co-infection numbers in IA-positiveand IA-negative samples from ICU-patients (2007-08: P=0.27, X²=1.217, 1; 2012-13: P=0.8889, X²=0.01951, 1; 2013-2014: P= 0,8896, x²=0.01925, 1). In 2008-09, in contrast to the following (pandemic) season,significantly more IA-positive samples were virally co-infected than IA-negative samples (P=0.038, X²=4.307, 1; OR: 2,316; 95%CI: 1.030 - 5.206).

The percentages of viral co-infections in critically-ill IA-infected patients in 4 surrounding seasons differ significantly from what was observed during the pandemic season (P<0.0001, X²=32.88, 4), (Table 3).

·         Testing Performances

5.5.3.  Throat versus nasopharyngeal flocked swabs

In 272 patients, throat and nasopharyngeal swab obtained simultaneously could be both analysed separately. Of them, 65 throats and 91 nasopharyngeal samples were positive in viral culture. A number of samples was additionally positive by multiparameter PCR examination (12 and 13 in throat and nasopharyngeal groups, respectively). Overall, the nasopharyngeal swabsshowed significantly greater sensitivityin detecting any respiratory virus compared to the throat swabs (38.2% versus 28.3%, p=0.018). Considering viral distribution (Supplement 2), hRV, A(H1N1)2009, and PIV predominated in the nasopharynx, whereas AdV, hEV and RSV similarly occurred in both sites. Only the nasopharyngeal swabs of these 272 patients were included in final epidemiologic analysis.(Supplement 2).

5.5.4. Culture versus real-time RT-PCR for the detection of influenza A(H1N1)v2009

The pandemic strain A(H1N1)2009 largely predominated among circulating influenza viruses (94.04% and 91.6% in culture and PCR assays, respectively). Out of 1993 samples for which both PCR and culture were simultaneously performed, A(H1N1)2009 was detected by RT-PCR in 641 (32.2%) samples but only in 532 (26.7%) by culture. The sensitivity of viral culture compared to PCR was 83.0% and increased to 87.8% if only NPA were considered.

The mean time to positivity of culture for A(H1N1)2009 on LLC-MK2 cells was 3.1 days (CI 95%: 2.1 - 7.3 d), and the mean time to result for RT-PCR IA was 1.3 day (CI 95%: 6 hours-3.6 d). In previous seasons with a circulating seasonal IA strain, growth was seen after a mean time of 6.1 days and 4.4 days in 2007-08 and 2008-09 respectively.

6. Discussion

In this study, we provide an extensive description of viral epidemiology during the first pandemic wave of influenza A(H1N1)2009 by examining 4,895 samples of 4,101 ILI patients of all age groups, attending the Emergency department or polyclinics from seven hospitals in Brussels area. We were thereby able to demonstrate that co-infection rate during medically-attended ILI was clearly decreased when patients were infected by the pandemic strain, compared to other viruses or other influenza seasons. We also observed that the pandemic influenza variant circulated quite differently its first season compared to previous influenza strains, not only considering the targeted population (adolescents/young adults without risk factors), but also considering its impact on circulation of other respiratory viruses and its specific and rapid growth pattern in vitro.

Viral respiratory infections often show a yearly or biennial appearance during the winter months, such as observed with IA/IB, RSV, PIV and hMPV. Whereas these latter viruses are quite easily detected with rapid antigen assays, others are more challenging to diagnose in routine but could potentially be involved in co-infections and contribute to morbidity. So far, there have been several publications discussing the relative importance of mixed viral infections among ILI, particularly in combination with IA [14,20,21,22,]. However, findings and conclusions remain divergent. While some authors observed an increased severity of respiratory illness in children when infected with two or more viruses [13,23], others supported exactly the opposite [22] or even didn’t find out any association between co-infections and severity of respiratory illness [24]. Discrepancy of these findings may be partly explained by a large heterogeneity in study design, included populations (age range, comorbidity, or illness severity), geographical location, seasonal timing (influencing virus circulating) and method of viral detection (traditional culture and direct immunofluorescence assays versus molecular assays, variable performance characteristics of commercial kit versus in house molecular tests etc…). Among others, Esposito et al reviewed community-acquired pneumoniae etiologies in children and highlighted that the association of hRV with either RSV or hMPV could increase severity [25]. Furthermore, Goka et al. concluded from a large study including all age groups in North West England that co-infection with RSV and hAdV was associated with increased risk of admission to ICU, even though their results did not reach statistical significance. They further reported that co-infection with seasonal IA and IB was associated with a higher risk of ICU admission or death [14]. During the first pandemic wave in Argentina, Torres et al also observed that viral co-infection with RSV was associated to increased mortality in pediatric ICU, illustrating the critical scenario of two virulent pathogens circulating in parallel in the Southern hemisphere [26]. Luckily RSV peaked only at the end of the first pandemic wave in Belgium and Europe, so that we did not encounter much morbidity of RSV-A(H1N1)2009 co-infections. Unfortunately, our study could not assess whether co-infection resulted or not in a higher severity of illness, for two main reasons: firstly, rate of co-infections involving the pandemic strain was very low among ICU patients, irrespective of the outcome (11/620 IA-associated co-infections in total, fatality rate 4/11 vs. 47/147 among persons with multiple or single H1N1 infections, respectively, P=0.7475) and the patients could not be matched for underlying comorbidities. Any conclusions on the pathogenicity of each respiratory virus are therefore difficult to draw. Secondly, the use of multiparameter PCR was restricted to critically-ill patients (due to the lack of reimbursement in Belgium), rendering any comparison between general wards and ICU patients poorly reliable in terms of co-infection numbers.

Although the exact influence of viral co-infection on pathogenicity remains unclear, some viruses seem more prone to arise in co-infections than others. hRV is regardless of age groups in general the most prevalent virus in mixed viral infections, followed by hBoV, hCoV and hMPV. But during annual flu seasons influenza viruses are not infrequently encountered in viral co-infection neither. In this study,it was clear that in addition to almost complete exclusion of seasonal influenza strains [27], the pandemic variant, during its first circulation, was associated with a remarkably low rate of viral co-infection compared to other viruses overall and to other flu seasons. Indeed, only 7% and 2.9% of IA-positive samples harbored multiple pathogens compared to 18.4% and 5.5% in IA-negative specimens, as obtained by multiparameter PCR and viral culture respectively. This observation remained true irrespective of patient’s status (critically ill or not, age, risk factors etc…) or included period. In absolute numbers, more A(H1N1)2009 positive samples were seen in the group below 5y (highest number in toddlers <1y) than in age group 11-60y. In this age group one should expect a huge number of viral co-infection, but even there most of the IA+samples showed mono-infection in contrast to other seasons.

IA-related co-infections most frequently included PIV (PIV-2 epidemics observed during the study period), hAdV and hRV. Looking at other published data, our observation was similar to the study of Lees EA [20], reporting a co-infection rate of only 7.4% when involving the pandemic strain. Furthermore, authors from Israel also corroborated our findings of lower viral co-infection rates during the first pandemic season [9]. However, our data contrast with the 13.1% co-infections described by Esper et al. [21] as well with data from France reporting higher co-infection rates (up to 19%) in A(H1N1)2009-positive samples [28]. Surprising was therefore the conclusion of the French authors claiming that the presence of hRV in the nasopharynx, though being normally not seldom associated with IA, reduced the likelihood of co-detection with A(H1N1)2009 and that hRV circulation delayed the spread of the pandemic flu [29]. In our study, hRV infections in Belgium showed a parallel increase in time with A(H1N1)2009 infections from August to September 2009 (Figure 1) and seemed not to have had an important impact on A(H1N1)2009 circulation in Brussels and surroundings. Once the pandemic peaked in October 2009, hRV stayed at a steady level as did hAdV and hCoV, while hEV arose and PIV decreased significantly with more than 50%. Probably slight geographical and atmospheric differences between the two countries could be responsible for this delay of appearance of A(H1N1)2009 in France. As for hRV, the pattern of RSV circulation was not much influenced by the early occurrence of flu pandemics in Belgium neither; with a peak around week 49-50 [this study; 30], in contrast to the delayed RSV epidemics observed in some countries [9].Moreover,RSV remained by far more prevalent than A(H1N1)2009 in children <2years similarly to other winter seasons.

Further, we were intriguedby the difference we observed in the same epidemiological context (region, population, sample types) during previous and following flu seasons. During the pandemic 2009, the co-infection rate within the group of IA-positive samples was significantly decreased in comparison with rates measured when seasonal A/H1N1 and A/H3N2 were predominating in 2007-08 and 2008-09, respectively (P <0.0001). Furthermore, the total viral co-infection rate we measured in the pandemic season (including all respiratory viruses with/without IA) was also significantly lower than during previous influenza seasons, even when calculation was extended till the end of March 2010 to include the whole winter season (P=0.0001). Mixed viral infections in ICU globally predominated among children below 5y of age, which could be expected, because early in life children acquire 3-8 viral upper respiratory tract infections each year [30].

As found in the literature, an advanced age did not appear to be a major risk factor associated with influenza A(H1N1)2009 infection with only 7.2% positive patients above 60 years, and none above 80y old. This low incidence could be due to cross-reactive antibodies from previous exposure to IA. It was shown that IA antibodies cross-react with A(H1N1)2009, and they were detected in up to one third of healthy adults >60 years [31]. In the Southern hemisphere, pandemic IA showed two incidence peaks, one in children ≤5y old and the second in individuals between 20-29 years [4,32]. However, here only one large incidence peak was seen: between 6 and 60 years old >70% of positive patients were infected by A(H1N1)2009, this feature being the single major epidemiologic signature of a pandemic. IA prevalence was highest in the age category “11-20” where 47.4% of patients were infected with the pandemic variant. Another Belgian study showed that IA mostly affected the 6-15 years old age group, comparable with our findings [33].

In our study as in others [21,28], hRV was the virus most frequently involved in co-infection overall, regardless of the presence of the pandemic strain. Debate is ongoing about the exact role played by hRV in virus-virus co-infections. Whereas HRV is often co detected with other viral pathogens as shown above [31], other authors have suggested that this virus possesses a competitive relationship with other viruses. In a study of 1,742 specimens Brunstein et al. reported a number of instances of suspected pathogen cosuppression between specific viral combinations, particularly between single-stranded RNA-viruses [32]. It was described that hRV may render the host less likely to be infected by other viruses, for a certain time. On another hand, Esper et al. [21] found that hRV co-infection had little impact on severity of influenza disease; in fact, such patients had a lower median clinical severity score, while the opposite was true for non-hRV co-infections.

Focusing on ILI etiology, more than 40% of our samples were positive for at least one virus and A(H1N1)2009 was the most common virus causing ILI overall with a global prevalence of 22.1%. These results were in line with previous reports on viral etiologies of ILI during the A(H1N1)2009 pandemic, with incidences of viral infection ranging from 37% to 89%, depending on the study design [10,11,34]. In the ICU group 54.4 % of the samples were positive for at least one virus and 25.5% for A(H1N1)2009, similar to reports from the literature [35].

In conclusion, our study confirmed that during the 2009 flu pandemics, the novel A(H1N1)2009 strain was the most prevalent virus responsible for medically-attended ILI, except inside extreme age groups (infants and the elderly). Thanks to extensive epidemiological data collected, it was demonstrated that the pandemic strain during its first season was remarkably associated with a reduced likelihood of coincidental viral infection as compared to other respiratory viruses or other flu seasons. Much less co-infections were indeed detected in ILI episodes at that time, regardless of age and patient's status. This unique feature could probably partly be due to the early circulation of the virus, but secondly due to changes in viral interferences related to the first circulation of A(H1N1)2009, since influenza A related co-infections increased again during consecutive flu seasons. The reasons why some viral strains areprone to arise in mono-infection as well as whether the presence of multiple viral pathogens together influence ILI pathogenicity and severity remain to be clearly elucidated.

7.  Acknowledgments

We would like to thank all the medical laboratory technicians of the Virology Lab and the Molecular Microbiology service for working efficiently on a daily basis during the 2009 pandemics and the surrounding winters.

8.  Conflict of interest

The authors declare that they have no conflict of interest..

Annex I. Technical Details of Samples Collection, Treatment and Analysis

Respiratory Specimen acquisition and handling

VTM was prepared as follows: 12.5 g of Difco Veal Infusion Broth (VIB, Becton & Dickinson, Erembodegem, Belgium) was dissolved in 500 mL of distilled water and the solution was autoclaved at 121°C for 15 minutes.

After cooling, 2.5 g of Bovine Serum Albumin-Fraction V (Sigma-Aldrich, Diegem, Belgium) was dissolved in Sterile VIB and filtrated through a 0.45 μm filter and added to the rest of the VIB to a final concentration of 0.5%. VTM was supplemented with gentamicin, vancomycin and amphotericin B to a final concentration of respectively 333.33μg/mL, 66.66μg/mL and 33.33μg/mL. After reception, 3 mL of VTM was added to the respiratory sample and vortexed for 30 seconds. The samples were subsequently divided into two aliquots, 0.5 mL, to which an additional 2 mL of VTM was added, for virus culture, and 1 mL for molecular testing.

Viral culture

Daily, three classical cell lines (LLC-MK2, Veroand MRC5) were inoculated, with respectively 300 μL, 150 μL

And 150 μL of the diluted sample and subsequently incubated in a 10% CO2 atmosphere at 36°C. 

The presence of a virus was determined by routine microscopic examination after 2, 3, 5, 7, 10, 14, and 21 days of incubation.

The culture medium was replaced weekly.

Molecular assays

Target Sequence Selection and Primer and Probe Design

Specific primers and probes were selected and designed by using the Primer Express software, version 2.0 (Applied Bio systems (AB), Foster City, CA) according to sequences available from public databases (National Center for Biotechnology Information; NCBI). The oligonucleotide sequences, PCR products lengths, locations,

And GenBank accession numbers of the corresponding target genes are displayed in Supplement 1.

As indicated in this table, some of the primers reveal a degenerated code. This was a requirement for the detection of viral subspecies differing from each other by single nucleotides. All primers and probes included in this study were synthesized by Eurogentec (Seraing, Belgium) and Applied Bio systems. Prior to experimental testing, the primers and probe sequences were assessed for specificity by comparing them to sequences of other prokaryotic and eukaryotic organisms by using standard nucleotide-nucleotide BLAST (NCBI) alignment software. None of The selected oligonucleotide displayed significant homologies to any other sequences. Then, the optimal concentration of oligonucleotides used in real-time PCR was assessed. The oligonucleotide concentration which gives the lowest threshold cycle (Ct) value and maximum amplification efficiency was selected (Supplement 1).

DNA Extraction

Genomic DNA and RNA were isolated using the MagNA Pure LC (Roche Diagnostics) instrument by using the Total Nucleic Acid Isolation Kit - Large Volume. 25 μL of a nonhuman control virus (Phocid Distemper virus, (Kindly provided by dr. G. van Doornum, University of Rotterdam, The Netherlands), was spiked into each sample prior to DNA and RNA extraction. In total, for nucleic acid extraction, the input volume was 500 μL and the elution volume was 100 μL. The nucleic acid recovered was stored at -70°C until further testing.

Real-time PCRs

An in house reversed transcriptase real-time PCR (rtRT-PCR InflA), targeting the matrix protein-coding gene,was used for the detection of IA virus (Supplement 1). This one-step RT-PCR is able to detect all subtypes of IA Viruses including A (H1N1) pdm2009. For specific detection of the circulating pandemic variant, two monoplexrt RT-PCR assays according to the Centers of Disease Control (CDC) protocol were initially used (from the 15^th of July till the end of August 2009): the SW InfA PCR (swInfA) and the SW H1 PCR (swH1), further called as RT-PCR A/H1N1 [16]. From September on, A (H1N1) pdm2009 sub typing was based only on swInfA detection, due to the numerous observed discrepancies in laboratories worldwide and the suspicion of mismatches in the swH1 primer, probe binding regions and of existing variants of the pandemic strain [17]. Five reactions were set up as duplex PCRs. The reaction was composed of 12.5 μL of reaction mix, 50 to 900 nM concentrations of primers of each virus, 100 to 300 nMTaqMan probe of each virus (Supplement 1), and 10μL of nucleic acid extract tot a final volume of 25 μL. For the human Boca virus (hBoV) human Adenovirus (hAdV) duplex, the reaction mix consists of the Light Cycler 480 Probes Master (Roche Diagnostics). For IARSVA, IB-PIV2, RSVB-hMPV (hMPV 1A,1B,2A,2B) and hRV-PIV3 duplexes, the reaction mix consists of 1x TaqMan EZ buffer (AB), 3 mM manganese acetate solution, 300 nM each dNTP, 2.5 U rTth DNA polymerase and 0.25 U Amperage UNG (AB). The six remaining RT-PCR reactions (229E, NL63, OC43, hEV, PIV1 and PIV4) were set up as single plex PCRs in a total volume of 25 μL containing 1X TaqMan A buffer (AB), 5.5 mM magnesium chloride, 300 nM each desoxyribonucleotide triphosphate (dNTP), 0.625 U AmpliTaq Gold DNA polymerase (AB), 6.25 U Multi Scribe reverse transcriptase (AB), 10 U RNase inhibitor (AB), 50 to 900 nM concentrations of primers, 100 to 200 nMTaqMan probe, and 10 μL of nucleic acid extract. The concentrations of primers and probes are specified in Supplement 1. The mixtures were then prepared in 96-well optical micro titer plates (Roche, Vilvoorde, Belgium), centrifuged for 1 min at 250 X g and amplified by using the following cycling parameters: for duplex PCRs performed on the Light Cycler 480 (Roche Diagnostics, Vilvoorde, Belgium): 2 min at 50°C, 30 min at 60°C, 5 min at 95°C, and 50 cycles of 20 s at 95°C and 60 s at 62°C. For single plex PCRs performed on iCycler (Bio-rad, Diegem, Belgium): 30 min at 48°C, 5 min at 95°C, and 50 cycles of 15 s at 95°C and 60 s at 60°C. During the weekends, PCRs were not performed, and samples that arrived on Friday after 1 p.m. were examined on Monday morning. During seasons 2012-2013 and 2013-2014, samples of critical-ill patients in ICU were prospectively tested for respiratory pathogens with an in house customized TAC (Taqman™ Array micro-array Card) respiratory panel which includes testing for the following pathogens: IA (H1, H3, H5, H7), IB, RSV A, RSV B, PIV 1 to 4, hAdV, hRV, hEV, hMPV, coronavirus (hCoV) (229E, HKU1, OC43, NL63, SARS), hBoV, cytomegalovirus (CMV), paraechovirus, mumps virus, measles virus,  and nine bacterial respiratory pathogens. From each Respiratory patient sample, 78 μL of nucleic acid extract and 26 μL of Taqman Fast Virus 1-step mastermix (Life Technologies, Carlsbad, CA) were mixed and added to the TAC sample port. A reversed transcriptase RTPCR was performed on the Viia 7 (Thermofisher, Carlsbad, CA) using following amplification protocol: 50°C for 5 min, 95°C for 20s, and 40 cycles of 95°C for 1s followed by 60°C for 20s. Multiple genetic targets per pathogen are being detected; sample adequacy and extraction/amplification inhibition is assessed. The system reports a cycle threshold for each positive PCR assay so that the load of the micro-organisms present in the clinical sample can be estimated. Based on the internal QC (Phocid Distemper Virus) data a %CV of 9.5% was registered, indicating a highly reproducible method. 

Figure 3: ICU patients and the distribution of their viral infections.

 Supplement 2: Evaluation of diagnostic sensitivity for respiratory viral detection of two different anatomical sitesby comparing 272 parallel nasopharyngeal and throat samples.

1. Novel Swine-Origin Influenza A (H1N1) Virus Investigation Team, Dawood FS, Jain S, Finelli L, Shaw MW, et al.(2009) Emergence of a novel swine-origin influenza A(H1N1) virus in humans. N Engl J Med 360:2605-2615.
2. Khan K, Arino J, Hu W, Raposo P, Sears J, et al. (2009) Spread of a novel influenza A(H1N1) virus via global airline transportation. N Engl J Med 361:212-214.
3. Lowen AC, Mubareka S, Steel J, Palese P (2007) Influenza virus transmission is dependent on relative humidity and temperature. PLoSPathog 3:1470-1476.
4. Oliviera WK, Carmo EH, Penna GO, Kuchenbecker RS, Santos HB, et al. (2009) on behalf of the Surveillance Team for the pandemic influenza A(H1N1) 2009 in the Ministry of Health.Pandemic H1N1 influenza in Brazil: analysis of the first 34.506 notified cases of influenza-like illness with severe acute respiratory infection (SARI). Euro Surveill14: 19362.
5. ECDC Surveillance Report. Pandemic (H1N1)2009: Analysis of individual case reports in EU and EEA countries.
6. Commissariat interministériel Influenza/Interministrieel commissariaat Influenza Belgium.
7. Belgian National Influenza center.
8. Williams BG, Gouws E, Boschi-Pinto C, Bryce J, Dye C (2002) Estimates of world-wide distribution of child deaths from acute respiratory infections. Lancet Infect Dis 1:25-32.
9. Meningher T, Hindiyeh M, Regev L, Sherbany H, Mendelson E, et al. (2014) Relationships between A(H1N1)pdm09 influenza infection and infections with other respiratory viruses. Influenza and Other Respir Viruses 8:422-430.
10. Renois F, Talmud D, Huguenin A, Moutte L, Strady C, et al. (2010) Rapid Detection of Respiratory Tract Viral Infections and Coinfections in Patients with Influenza-Like Illnesses by Reverse Transcription-PCR DNA Microarray Systems. J Clin Microbiol 48:3836-3842.
11. Nisii C, Meschi S, Selleri M, Bordi L, Castilletti C, et al. (2010) Frequency of detection of upper respiratory tract viruses in patients tested for pandemic H1N1/09 viral infection. J Clin Microbiol 48:3383-3385.
12. Debiaggi M, Canducci F, Ceresola ER, Clementi M (2012) the role of infections and coinfection with newly identified and emerging respiratory viruses in children. VirolJ 9:247.
13. Scotta MC, Mattiello R, Marostica PJ, Jones MH, Martins LG, et al. (2013) Risk factors for need of mechanical ventilation in children with influenza A(H1N1)PDM09. J Pediatr (Rio J) 89:444-449.
14. Goka E, Vallely P, Mutton K, KlapperP (2013) Influenza A viruses dual and multiple infections with other respiratory viruses and risk of hospitalisation and mortality. Influenza and OtherRespirViruses 7:1079–1087.
15. Pretorius MA, Madhi SA, Cohen C, Naidoo D, Groome M, et al. (2012) Respiratory Viral Coinfections Identified by a 10-Plex Real-Time Revers-Transcription Polymerase Chain reaction Assay in Patients Hospitalized with Severe Acute Respiratory Illness, South Africa, 2009-2010. J infect Dis 1: 159-165.
16. Reynders M, De Foor M, Maaroufi Y, Thomas I, Vergison A, et al. (2012) Prospective evaluation of CorisInflu-A&B Respi-Strip and of BinaxNOW Influenza A&B assay against viral culture and real-time PCR assay for detection of 2009 pandemic influenza A/H1N1v in Belgian patients.Acta ClinBel 67:94-98.
17. http://www.cdc.gov/flu/.
18. World Health Organization (WHO). CDC protocol of realtime RTPCR for swine influenza A (H1N1).
19. Klungthong C, Chinnawirotpisan P, Hussem K, Phonpakobsin T, Manasatienkij W, et al. (2010) The impact of primer and probe-template mismatches on the sensitivity of pandemic influenza A/H1N1/2009 virus detection by real-time RT-PCR. J ClinVirol 48:91-95.
20. Lees EA, Carrol ED, Gerrard C, Hardiman F, Howel G, et al. (2014) Characterisation of acute respiratory infections at a United Kingdom paediatric teaching hospital: observational study assessing the impact of influenza A(2009 pdmH1N1) on predominant viral pathogens. BMC Infect Diseases 14:343.
21. Esper FP, Spahlinger T, Zhou L (2011) Rate and influence of respiratory virus co-infection on pandemic (H1N1) influenza disease.J Infect 63:260-266.
22. Martin ET, Kuypers J, Wald A, Englund JA (2012) Multiple versus single virus respiratory infections: viral load and clinical disease severity in hospitalized children. Influenza and Other Respir Viruses 6:71-77.
23. Drews AL, Atmar RL, Glezen WP, Baxter BD, Piedra PA, et al. (1997) Dual respiratory virus infections. Clin Infect Dis 25:1421-1429.
24. Peng D, Zhao D, Liu J, Wang X, Yang K, et al. (2009) Multipathogen infections in hospitalized children with acute respiratory infections. Virol J 6:155.
25. Esposito S, Daleno C, Scala A, Castellazzi L, Terranova L, et al. (2014) Impact of rhinovirus nasopharyngeal viral load and viremia on severity of respiratory infections in children.Eur J ClinMicrobiol Infect Dis 33:41-48.
26. Torres SF, Iolster T, Schnitzler EJ, Farias JA, Bordogna AC, et al. (2012) High mortality in patients with influenza A pH1N1 2009 admitted to a pediatric intensive care unit: a predictive model of mortality. PediatrCrit Care Med 13: 78-83.
27. Gröndahl B, Ankermann T, von Bismarck P, Rockahr S, Kowalzik F, et al. (2014) The 2009 pandemic influenza A(H1N1) coincides with changes in the epidemiology of other viral pathogens causing acute respiratory tract infections in children. Infection 42:303-308.
28. Schnepf N, Resche-Rigon M, Chaillon A, Scemla A, Gras G, et al. (2011) High burden of Non-influenza Viruses in influenza-like illness in the early weeks of H1N1v Epidemic in France. PLoS One 6: 23514.
29. Casalegno JS, Ottmann M, Bouscambert Duchamp M, Escuret V, Billaud G, et al. (2010) Rhinoviruses delayed the circulation of the pandemic influenza A(H1N1)2009 virus in France. ClinMicrobiol Infect 16:326-329.
30. Chonmaitree T, Revai K, Grady JJ, Clos A, Patel JA, et al. (2008) Viral upper respiratory tract infection and otitis media complication in young children. Clin Infect Dis 46:815-823.
31. Franz A, Adams O, Willems R, Bonzel L, Neuhausen N, et al. (2010) Correlation of viral load of respiratory pathogens and co-infections with disease severity in children hospitalized for lower respiratory tract infection. J ClinVirol 48: 239-245.
32. Marchand-Austin A, Farrell DJ, Jamieson FB, Lombardi N, Lombos E, et al. (2009) Respiratory infection in institutions during early stages of pandemic (H1N1) 2009, Canada. Emerg Infect Dis 15: 2001-2003.
33. Hombrouck A, Sabbe M, Van Casteren V, Wuillaume F, Hue D, et al. (2012) Viral aetiology of influenza-like illness in Belgium during the influenza A(H1N1)2009 pandemic. Eur J ClinMicrobiol Infect Dis. 31:999-1007.
34. Echevarría-Zuno S, Mejía-Aranguré JM, Mar-Obeso AJ, Grajales-Muñiz C, Robles-Pérez E, et al. (2009) Infection and death from influenza A H1N1 virus in Mexico: a retrospective analysis. Lancet 374:2072-2079.
35. Libster R, Bugna J, Coviello S, Hijano DR, Dunaiewsky M, et al. (2010) Pediatric hospitalizations associated with 2009 pandemic influenza A (H1N1) in Argentina. N Engl J Med 362:45-55.