Virological Surveillance of Influenza Viruses during the 2008–09, 2009–10 and 2010–11 Seasons in Tunisia

Background The data contribute to a better understanding of the circulation of influenza viruses especially in North-Africa. Objective The objective of this surveillance was to detect severe influenza cases, identify their epidemiological and virological characteristics and assess their impact on the healthcare system. Method We describe in this report the findings of laboratory-based surveillance of human cases of influenza virus and other respiratory viruses' infection during three seasons in Tunisia. Results The 2008–09 winter influenza season is underway in Tunisia, with co-circulation of influenza A/H3N2 (56.25%), influenza A(H1N1) (32.5%), and a few sporadic influenza B viruses (11.25%). In 2010–11 season the circulating strains are predominantly the 2009 pandemic influenza A(H1N1)pdm09 (70%) and influenza B viruses (22%). And sporadic viruses were sub-typed as A/H3N2 and unsubtyped influenza A, 5% and 3%, respectively. Unlike other countries, highest prevalence of influenza B virus Yamagata-like lineage has been reported in Tunisia (76%) localised into the clade B/Bangladesh/3333/2007. In the pandemic year, influenza A(H1N1)pdm09 predominated over other influenza viruses (95%). Amino acid changes D222G and D222E were detected in the HA gene of A(H1N1)pdm09 virus in two severe cases, one fatal case and one mild case out of 50 influenza A(H1N1)pdm09 viruses studied. The most frequently reported respiratory virus other than influenza in three seasons was RSV (45.29%). Conclusion This article summarises the surveillance and epidemiology of influenza viruses and other respiratory viruses, showing how rapid improvements in influenza surveillance were feasible by connecting the existing structure in the health care system for patient records to electronic surveillance system for reporting ILI cases.


Introduction
Identification and characterization of circulating influenza viruses is essential to detect the emergence of antigenic drift variants causing influenza epidemics and novel A subtypes with the potential to cause an influenza pandemic. Thus, virological surveillance of influenza provides a basis for selection of the virus strains to be included in the annual formulation of influenza vaccines [1]. Surveillance data from the African continent has increased substantially in the past five years [2][3][4][5], but they are still insufficient to allow for a thorough understanding of influenza virus circulation patterns on the continent and their associated morbidity and mortality, or to inform influenza control strategies [6,7]. The primary objective of this study was to develop or strengthen influenza sentinel surveillance systems in line with WHO standards in selected North African countries. In the past decade, information of epidemic strains from Tunisia was largely unknown due to lack of any systemic study. In this present study we are reporting the activity and circulation of influenza viruses during three seasons (2008-2009, 2009-2010 and 2010-2011) as a part of global influenza surveillance network, which was expanded to Tunisia since 1980.
A subset of sentinel primary care physicians participating in virological surveillance schemes in the community submits respiratory samples for virological testing from patients presenting in primary health care with an ILI, as well as all regional emergency centres and hospitals that take on surveillance of influenza from community, hospitalized and fatal cases. The surveillance of influenza and other respiratory viruses is undertaken by 268 primary care centres for adult and pediatric patients ( Fig. 1) distributed in 24 governorates covering 2.7% of general Tunisian population (Table 1). Sentinel physicians report weekly the total number of patient visits to their facilities for influenza-like illness (ILI) and acute respiratory infection (ARI) within four age categories (0-4 years, 5-14 years, 15-64 years, and 65+ years). Sentinel physicians are asked to collect respiratory specimens from patients with symptoms of ILI or ARI. ILI was defined as an outpatient with fever (38uC) and cough or sore throat with onset less than five days prior to presentation in the absence of a specific diagnosis. ARI was defined as an outpatient with sudden onset of respiratory signs including cough, difficulty breathing, rhinitis and general symptoms such as fever, headache, fatigue, and myalgia less than five days prior to presentation. The physicians sent specimens for influenza testing and basic demographic data from a subset of patients with ILI each year during October through May to the National Influenza Centre (NIC) situated at the Charles Nicolle's Hospital Tunis, provided that the number of consultants is 10% above the total number of consultants in the sentinel centre. In addition, according to World Health Organization (WHO) recommendations, severe acute respiratory infection (SARI) related to influenza viruses has been added to existing outpatient surveillance systems to fully describe the spectrum of disease related to influenza and to identify individuals at highest risk for severe disease. As a result, various existing routine influenza surveillance systems in Tunisia were enhanced or supplemented to gain a rapid understanding of this novel virus, to monitor its spread and impact, and to evaluate the uptake, impact and effectiveness of the various countermeasures that were implemented.
During each season, oro-pharyngeal and naso-pharyngeal swabs were collected from ILI and SARI patients enrolled under the virological surveillance system and placed in viral transport medium (Vircell, SpainH). Oro-pharyngeal and nasopharyngeal swabs collected from the same patient were placed in one cryovial, stored at 4uC at the sentinel site, and transported daily to the NIC. The samples taken from severe cases were from children and/or adult hospitalised in the care unit or patients hospitalized with severe acute respiratory infection were additionally tested for other respiratory viruses. For all participants, respiratory specimens collection were performed after informed consent, under the supervision of local sanitary authorities. It is just an informal statement of the study and members of the ethic committee have previously always approved the work of the NIC. In fact, patients included in this study were of all ages and consulted the sentinel clinics or were hospitalized in state or private hospitals for influenza like symptoms infection. The consent was verbal because the patients, or parents in the case of minors, accept the test for Influenza viruses since it is free and safe. Until now, written consent is judged not necessary by the ethics committee.
Samples were analysed for diagnostic purposes using the real time PCR CDC protocol [8] for detection of influenza viruses and using xTAGH Respiratory Viral Panel Fast (Abbott Molecular, Germany) and the LuminexH technology for other respiratory viruses. All influenza positive samples were sub-typed using specific real-time PCRs for influenza A(H1N1)pdm09, A/H3N2 and for influenza B viruses using ''Influenza Virus B Real Time RT-PCR Kit'', ''Subtype H1 of Influenza virus A Real Time RT-PCR Kit'' and ''Subtype H3 of Influenza virus A Real Time RT-PCR Kit''(Shanghai ZJ Bio-Tech Co., Ltd). Every year, we realized the exchange of influenza strains with WHO Collaborating Centre for Influenza in London. In the pandemic year, we succeeded to cultivate some Tunisian strains of influenza A(H1N1)pdm09 and influenza B in National Influenza Centre Madrid Spain. In fact, cell culture virus isolation will be implemented in the National Influenza Centre in Tunisia in the next few years.
A representative number of influenza viruses were genotypically characterized by analysis of the nucleotide sequence of partial haemagglutinin HA1 chain (931 nucleotide residues) and partial neuraminidase (836 nucleotide residues) genes in order to know if circulating viruses were well-matched with vaccine viruses and check for the most frequent amino acid key changes related to neuraminidase inhibitors resistance respectively. All viruses analysed were amplified and sequenced according to the protocol of National Influenza Centre Madrid [9].
Tunisian sequences were aligned with other sequences from reference influenza A viruses available at the NCBI Influenza Virus Resource (http://WWW.ncbi.nih.gov/genomes/FLU/ SwineFlu.html) and Global Initiative on Sharing Avian Influenza Data database (http://WWW.platform.gisiad.org) using Clustal W program implemented in MEGA version 4 under default conditions [10]. The nucleotide sequence data reported in this work were deposited in the GenBank nucleotide sequence database with accession numbers JN037697 to JN037779.  were associated with influenza A/H3N2 infection. Despite the importance of these preliminary results, our surveillance system had limitations in season 2008-2009, and the rate of influenza virus detection remained low (,10%). In fact, the specimen collection and storage techniques may not always have been optimal. In addition, the identification of influenza viruses was performed primarily using immunofluorescence assays which are less sensitive for the detection of influenza viruses than viral culture The curve of cases of influenza viruses spread out over 11 weeks (from Wk 1 to Wk 11). There is a decrease of this curve during two weeks (Wk 5 and Wk 6) (in January 24-30 th and January 31 st -February 6 th ). This descent of curve is probably due to the disturbances in the schooling, the transport and the work in this period. Two peaks have been reported in the Wk 4 and Wk 9 (58.3% and 69.2%).We observe the decrease of influenza A (H1N1)pdm09 in the sentinel samples since week 11. Nine influenza-associated deaths were confirmed (ranging 19 to 57 years old) in an ante mortem or post mortem specimen. Five of these fatal cases were pregnant women with an underlying  clinical risk factor (mean 32 year old). It is important to note, that all deaths were associated with influenza A(H1N1)pdm09 infection. In fact, pregnancy was identified as a particular risk group for adverse influenza outcome during previous pandemics, and also during seasonal influenza [11,12].

Epidemiological findings
Two respiratory disease outbreaks in closed settings were reported during the 2010-2011 season; one in the intensive care unit of Charles Nicolle's hospital concerning eight members of the medical staff and the second in the intensive care unit of Rabta's hospital of Tunis when a patient 67 year old diabetic died after infection. Ten deaths were virologically confirmed in two separate outbreaks with influenza A(H1N1)pdm09 detected. It was expected that the virus would behave as a seasonal virus and continue to circulate in the population. Its behaviour, however, could not be reliably predicted [13], although it was considered likely that the virus would continue to cause serious disease in a minority of those infected in younger age groups and people in high-risk groups [14]. Severe cases were defined as any condition or clinical presentation requiring hospital admission for clinical management according to WHO guidance criteria [15]. Overall, influenza activity in Tunisia in 2010-2011 reached a level higher than that seen in the winter of the 2008-2009 season, but lower than during the first wave of the pandemic in the summer of 2009. In fact, in season 2010-2011 over 50% of sentinel specimens were tested positive for influenza. This intensity of influenza activity was similar to that observed during the peak of the 2009-2010 'pandemic' season. This may reflect a greater intensity of influenza circulation resulting from the introduction of a novel virus into naïve human populations, and/or improvements in the sensitivity of laboratory diagnostic methods to detect influenza in use in Tunisia. However the percentage of consultants for ILI or ARI in the sentinel centres in Tunisia in 2010-2011 season was lower (8.3%) than the percentage of consultants in the pandemic year (30.3%) (Fig. 3). Clinical consultation rates for ILI or/and ARI were declining overall, whereas the percentage of consultants in the next year of the pandemic was not correlated to virological analysis of influenza viruses in Tunisia. Despite the increased number of samples obtained in season 2010-2011 under the enhanced surveillance system, there were limitations to our ILI surveillance system. This is probably due to the participating clinicians may have not correctly identified all ILI or SARI cases. Also it is possible due to a sub-declaration of the clinical services in the sentinel centres in comparison with the last year in which the medical crow was more motivated by the Ministry of Health because of the pandemic. The same situation has been reported in England by WHO European Influenza Network [16]. In fact, both the burden of severe respiratory infection and the proportion due to viral etiologies including influenza are largely undocumented in Africa highlighting the need for continued development of respiratory illness surveillance in this continent [17].

Virological findings
During the 2008-2009 season, 420 specimens were collected from patients at sentinel sites. Of those, 80 (19%) were positive for influenza viruses. 45 influenza A/H3N2 (Perth lineage) (56.25%), RSVs (26.8%) and 32 hMPVs (21%). In season 2010-2011, a total of 160 specimens were tested positive for respiratory virus and the most frequent respiratory viruses were: RSVs (97/160; 60.6%) and Rhinovirus/enterovirus (37/60; 61.6%). The most common nonflu pathogen circulating in three seasons causing the lower respiratory tract infections leading to hospitalisation especially in children was RSV (207/457; 45.29%). In 2010 in Chile there have been more cases of acute ARI in children but this is attributable to epidemics of respiratory syncytial virus infections (RSV) rather than influenza [18]. This emphasises the importance of countries being able to test for a suite of respiratory pathogens, not just influenza. This data contribute to a better understanding of the circulation of influenza viruses and other respiratory viruses especially in North-Africa.
Phylogenetic analysis of the HA1 nucleotid sequence of 23 influenza A(H1N1)pdm09 viruses from mild, severe (patients hospitalized with severe pneumonia and severe acute respiratory syndrome) and fatal cases, shows that all viruses characterised in Tunisia during season 2009-2010 were outside the seven genetic groups described in the European Centre for Disease Prevention and Control (ECDC) report [19]. A total of 27 HA genes of influenza A(H1N1)pdm09 viruses regardless of whether they were from 2011 from mild, severe cases and from influenza cases admitted to intensive care units were also sequenced and analysed  [20]. This genetic diversity of Tunisian strains compared to A/California/7/2009 was consistent with expected patterns of virus evolution. Additional substitutions in the position 222 of HA gene were found in 4 Tunisian sequences out of 50 viruses (8%). D222G substitution observed in a total of three viruses analysed (6%). Most of viruses with this mutation were found in severe cases (2/3), one of them from a fatal case [21]. Although most of studies have demonstrated the presence of D222G substitution in severe cases, it was also reported in mild cases [22][23][24][25]. D222E substitution was found in one out of 50 viruses studied (2%). This sample was taken from a patient with severe clinical syndrome. In fact, D222G mutation has been considered relevant for the acquisition of a hypervirulent phenotype during the 1918 influenza pandemic [26], while the role of D222E in virulence has been ruled out [27]. The clinical significance of D222E mutation is still unclear. Analysis of sequences of neuraminidase gene of influenza A(H1N1)pdm09 from 18 severe cases, not received any antiviral treatment, did not show presence of any of the known mutations associated to neuraminidase inhibitors resistance.
Influenza B viruses were grouped as Victoria lineage or Yamagata lineage on the basis of the HA gene sequence (Fig. 6 [28]. Most of Yamagata strains were detected in severe cases (9/13; 69%). Inspite of lower mutation rate and lower pathogenicity than influenza A, influenza B infection contributes to significant proportion of acute respiratory infection among Tunisian people.

Conclusions
Notably, in season 2008-2009 influenza A/H3N2 viruses, followed by influenza B, have been the predominant influenza viruses circulating in Tunisia. In season 2009-2010, pandemic influenza A(H1N1)pdm09 viruses were the predominant circulating viruses but, in contrast to the 2010-2011 season, there is a higher rate of co-circulation with influenza B viruses. The circulation of other winter viruses such as human respiratory syncytial virus and the particularly cold weather were also identified during three seasons in Tunisia. Unlike the vast majority of influenza B viruses circulating in the world during season 2010-2011, which were from the B/Victoria lineage, most of influenza B Tunisian strains belong to the B-Yamagata lineage, not included in the 2010-2011 vaccine. Therefore a virus strain belonging to the B-Yamagata lineage was indeed recommended for the vaccine composition of the Northern Hemisphere for the season 2012-2013. Appearance of D222G substitution in HA of A(H1N1)pdm09 viruses circulating in Tunisia might be related with severe respiratory disease. Mutations associated with resistance to neuraminidase inhibitors oseltamivir and zanamivir were not detected in the neuraminidase gene. To this end proper surveillance systems should be set up in already existing and wellestablished national influenza centres to understand the epidemiology of influenza and other respiratory viruses in North Africa, which in turn may help the processes of decision making regarding influenza vaccination on the continent, which may have a high impact on health in Africa.